550/9-73-008

     PROCEEDINGS of the
INTERNATIONAL CONGRESS
         on NOISE as a
  PUBLIC HEALTH PROBLEM
       DUBROVNIK, YUGOSLAVIA
           May 13-18, 1973
 U.S. ENVIRONMENTAL PROTECTION AGENCY
         Washington, D.C. 20460

-------
                 PROCEEDINGS of the
            INTERNATIONAL CONGRESS
                     on NOISE as a
             PUBLIC HEALTH PROBLEM

                DUBROVNIK, YUGOSLAVIA
                     May 13-18, 1973
               SPONSORS of the CONGRESS
          Union of Medical Societies of Yugoslavia
        Environmental Protection Agency, U.S. Gov't
          American Speech and Hearing Association
                 World Health Organization
                       Prepared by
THE U.S. ENVIRONMENTAL PROTECTION AGENCY
        Office of Noise Abatement and Control
             For sale by the Superintendent o! Documents, U.S. Government PrintinK Office
                      Washington, IXC. 20402 - Price $U.«

-------
                                     Foreword

     In  1968, a Conference on Noise as a Public  Health Hazard was organized by the
American  Speech and Hearing Association. At this  conference, an attempt was made to
bring together a group of speakers who could present summaries of the current state of
knowledge on all aspects of the "noise problem", ranging all the way from fairly technical
treatises to completely non-technical statements of personal opinion.  Such a wide-ranging
representation was judged to be necessary for the purpose of that conference, which was to
present a broad overview of what "noise pollution" was all about, to government personnel
and other intelligent laymen who saw that it was probably going to become a hot issue, and
give at least a few examples of the scientific evidence underlying arguments about just what
effects noise does have.
     At that time  it was  realized that  as  the  environmentalist movement gathered
momentum, a rapid development of public concern could be expected, and so a permanent
Committee of ASHA was established, one of whose charges was to plan another conference
when it was judged appropriate.
     The  burgeoning of interest in noise in the intervening 5 years has clearly met, if not
surpassed, our expectations  at that time. In the developed areas of the world, millions of
dollars or their equivalent are being spent  on surveys of noise levels and  exposures, and
increasingly stringent noise regulations are being imposed by all levels of government. And,
although  the measurement  of  the effects  of noise  is nowhere near as simple  as the
measurement of the noises themselves, many laboratories, mostly with federal support, are
engaged in full-time research on the hearing losses, sleep disturbance, speech interference,
alteration of physiological state, and annoyance caused by noise.
     Accordingly, in 1971 we began looking for a sponsor for a second conference—one
who would  agree, we hoped, to fund attendance  by a substantial number of researchers
from abroad, so that certain areas of knowledge less intensively studied in the USA could be
included in the subject matter. Fortunately, the head of the newly-created Office of Noise
Abatement  and Control (ONAC) of the Environmental Protection Agency, Dr. Alvin F.
Meyer, had need of just such a conference, as a source material for a document summarizing
all known criteria that might be used to establish national standards for noise control—that
is, provided that the Congress passed the bill, then being duly debated and amended,  that
would make such a document necessary. Furthermore, certain PL 480 funds (money  that
must be  spent in  other  countries)  were  available, which  meant  that  the  degree of
participation by foreign scientists might  be even greater than we had hoped. Not only that,
but the particular PL 480 funds in this case were in Jugoslavia, the country that includes
one of the garden spots of the world, Dubrovnik.
     On the assumption that our Congress would pass some  form of the bill in question
(which it did on October 27,  1972), we forged ahead with plans for  our meeting, now
upgraded  to an International Congress. With the help of Dr. Grujica iarkovic, the energetic
President  of the Jugoslavian Medical Association, and Dr. Mario Levi of the University of
Sarajevo, a planning meeting was held to which we invited a representative from most of the
countries  in which noise research was being done (I say "most" because we could not quite
afford to  pay for attendees from Japan, Australia, and South Africa because of the distance
involved, even though considerable research is being done there). At this meeting the formal
agenda was decided  on, and the list of invited participants prepared. It was agreed that we
would try to limit the Congress content strictly to the effects of noise on health, thereby

-------
excluding discussions of engineering aspects of noise reduction and control, descriptions of
methods for legal control, and presentation of viewpoints of special-interest groups. There
was some debate about how much time  to allot to public opinion surveys of annoyance,
some of us contending that annoyance, as measured in that manner, is not a health hazard at
all in the ordinary sense of the term.  However, proponents  of the WHO definition of
"health", in which any deviation from "optimum well-being"  is regarded  as undesirable,
carried  the  field,  and the  final day of the Congress was therefore given  over to the
sociologists.
    Despite a series of crises precipitated by governmental red  tape originating both in
Washington and Belgrade, the Congress was held on May 13-18, 1973 at the Libertas Hotel
in Dubrovnik. We had two major disappointments; one was  the failure of our Russian
invitees to appear due to  the fact that  our official invitations  had not been sent early
enough. The other was that the Xerox machine at  the Libertas  was out of  commission.
However, the general success of the Congress can be gauged by the fact that the audience
was as large on the final afternoon as at any other time.
    A side benefit of the Congress (or so we hope)  was the formation of an international
organization consisting of 5 "teams" who will try to accumulate and coordinate knowledge
about the effects of noise on (1) temporary and permanent hearing loss; (2) extra-auditory
function; (3) speech; (4) sleep; and (5) community reaction. The  parent group, or "basic"
team, will attempt to consolidate this knowledge for use by governmental  agencies, and will
make plans for the next Congress. Although the organization is now alive,  its name is still in
question. At the moment  it is  still the "International Scientific Noise  Teams", but the
resulting acronym has a negative connotation that pleases  few of us. Other names are being
considered.
    I regret that the length of the invited papers made it impracticable to publish at this
time any of the short contributed papers that were presented  at the Congress, many of
which were excellent, or the often-lively discussions that followed each session. It is hoped
that these can be included if another printing of the Proceedings is to be made.
    An enterprise of this scope cannot be a success without hard work on the part of many
people. Without doubt the most effort of all was  put forth by Dr. Levi, who managed all the
mechanical  details  of the  Congress, with the help  of his  and Dr.  3Larkovic's staff,
particularly, Felih Vesna.
    Official thanks are extended to our sponsoring organizations: The Jugoslavian Medical
Association, The American Speech and Hearing Association, the World Health Organization,
and of course most of all the Office of Noise Abatement and Control.
                                         u

-------
                         SPONSORS OF THE CONGRESS

    Union of Medical Societies of Yugoslavia
    World Health Organization
    Environmental Protection Agency, U.S. Government
    American Speech and Hearing Association

                      CO-PRESIDENTS OF THE CONGRESS

Grujica Zarkovic, M.D.      President  of the  Union of Medical Societies of Yugoslavia

Alvin F. Meyer, Jr.          Deputy Assistant Administrator of the U.S. Environmental
                              Protection Agency

Robert Goldstein, Ph.D.      President  of the American Speech and Hearing Association

                            PROGRAM COMMITTEE

    Angelova, M., Bulgaria                Kylin, B., Sweden
    Borsky, P. USA                      Levi, M., Yugoslavia—Standing Secretary
    Curlee, R., USA                      Nixon, C., USA
    Cerkez, F., Yugoslavia                Pearsons, K., USA
    Dieroff, H., DDR                    Pinter, L, Hungary
    Dzumhur, M., Yugoslavia              Raber, A., Austria
    Gulian, E., Rumania                  Spoor, A., Netherlands
    van Hattum, R., USA                 Sulkowski, W., Poland
    Hinchcliffe, R., UK                   Thiessen, J.G., Canada
    Jansen, G., Germany                  Tobias, J., USA
    Jokic, J., Yugoslavia                  Ward, W. D., USA-President
    Konig, E., Switzerland                Webster, J., USA
    Kryter, K., USA                      Whitcomb, M. USA


             SESSION ARRANGER AND PROCEEDINGS EDITOR

                              W. Dixon Ward

-------
                             TABLE OF CONTENTS

                                    Session 1

Introduction and Masking Effects of Noise
Chairman: Henning von Gierke (USA)                                           Page

     A Preview of the Congress Content, W. Dixon Ward (USA)  	    3
     Systems of Noise Measurement, Karl S. Pearsons (USA)	    7
     The Effects of Noise on the Hearing of Speech, John C. Webster (USA)	   25
     Reception of Distorted Speech, Jerry V. Tobias, F. Michael Irons (USA)  	   43
     Hearing Loss and Speech Intelligibility in Noise, Jerzy J. Kuzniarz (Poland)  ...   57
     The  Long-Term  Planning of a  Noise Control Program,  Michael  J.  Suess
         (Denmark)  	   73

                                    Session 2

Noise-Induced Hearing Loss (NIHL)—Empirical Data
Chairman: D. Robinson (UK)

     Basis for Percent Risk Table, Aram Glorig, William L. Baughn (USA)  	   79
     A Critique of Some Procedures for Evaluating Damage Risk from Exposure to
         Noise, Karl D. Kryter (USA)	103
     The Incidence  of Impaired  Hearing in Relation to Years of Exposure and
         Continuous Sound Level, (Preliminary  Analysis of 26,179 Cases), A. Raber
         (Austria)  	115
     Some Epidemiological  Data  on  Noise-Induced  Hearing Loss in  Poland,  Its
         Prophylaxis and Diagnosis, Wieslaw Sulkowski (Poland)	139
     On  the Problem  of Industrial  Noise and Some Hearing Losses in Certain
         Professional Groups Exposed to Noise, J. Moskov (Bulgaria)	  157
     Noise-Induced Hearing  Loss from Exposure to Intermitant  and Varying Noise,
         W. Passchier-Vermeer (Netherlands) 	169
     Evaluation of the Hearing Damage Risk from Intermittent  Noise According to
         the ISO Recommendations, B. Johansson, B. Kylin, S. Reopstorff (Sweden)  201

     Noise-Induced Hearing  Loss from Impulse Noise: Present Status, R.R.A. Coles,
         CG.  Rice, A.M. Martin (UK)   	211
     Hearing Loss Due to Impulse Noise.  A Field Study, Tadeusz Ceypek, Jerzy J.
         Kuzniarz, Adam Lipowczan (Poland)  	219
     Hearing Damage  Caused by  Very Short,  High-Intensity Impulse  Noise,  H.G.
         Dieroff (DDR) 	229

-------
                                    Session 3

Noise-Induced Hearing Loss—Mechanism
Chairmen: H.G. Dieroff (DDR), R. Hinchcliffe (UK)                             page

     Behavioral, Physiological  and  Anatomical  Studies of  Threshold  Shifts  in
         Animals, Donald H. Eldredge, James D. Miller, John H. Mills, Barbara A.
         Bohne (USA)	237
     Presbyacusis in Relation to Noise-Induced Hearing Loss, A. Spoor (Netherlands)  257
     Noise Exposure, Atherosclerosis and Accelerated Presbyacusis, Z. Bochenek, W.
         Bochenek (Poland) 	267
     High-Frequency Hearing and Noise Exposure, John L. Fletcher, (USA)  	271
     Susceptibility to TTS and PTS, W. Dixon Ward (USA)  	281
     Growth  of TTS and Course of Recovery for Different Noises; Implications for
         Growth of PTS, Wolfgang Kraak (DDR)	293
     Experiments  on  Animals Subject to Acute Acoustic Trauma, Wiktor Jankowski
         (Poland) 	301

                                   Session 4 A

Interaction of Noise with Other Noxious Agents in Production of Hearing Loss
Chairman: E. Lehnhardt (BRD)

     Influences of Chemical Agents on Hearing Loss, M. Haider (Austria)  	307
     Hearing Loss of Forest Workers and of Tractor Operators, (Interaction of Noise
         with Vibration), Istvan Pinter (Hungary)	315
     Infrasound and Hearing, Charles W. Nixon, Daniel L. Johnson (USA)   	329
     The Effects of Airborne Ultrasound and Near Ultrasound, W.I. Acton (UK)  . . . 349

                                   Session 4 B

Performance and Behavior
Chairman: D. E. Broadbent (UK)

     Psychological Consequences of Exposure  to  Noise, Facts, and Explanations,
         Edith Gulian (Romania)	363
     Similar and Opposing Effects of Noise on Performance, L. Hartley (UK)  	379
     The Effects of Different Types of  Acoustic Stimulation  on Performance, C.
         Stanley Harris (USA)	389
     Behavioral  Effects and Aftereffects of Noise,  David C. Glass, Jerome E.  Singer
         (USA)	409
     Effects  of Noise on a Serial  Short-Term Memory  Process, G. Wittersheim, P.
         Salame (France)	417
     The Effect of Annoying Noise on Some Psychological Functions During Work,
         Irena Franszczuk (Poland)  	425

                                       vi

-------
                                    Session 5

Non-Auditory Physiological and Pathological Reactions
Chairmen: E. Grandjean (Switzerland), S. Kubik (Czechoslovakia)                   Page

     Non-Auditory Effects of Noise, Physiological and Psychological Reactions in
         Man, Gerd Jansen (Germany)  	431
     Industrial Noise and Medical, Absence, and Accident Record Data on Exposed
         Workers, Alexander Cohen (USA)  	441
     Factors Increasing and Decreasing the Effects of Noise, D.E. Broadbent (UK)  . .  455
     Examples of Noise-Induced Reactions of Autonomic Nervous System during
         Normal Ovarian Cycle, Barbara Griefahn (DDR)	.^	459
     The Influence of Noise on Auditory Evoked Potentials, J. Gruberova, S. Kubik,
         J. Zal5ik (Czechoslovakia)  	469
     Some Data  on the Influence of Noise on Neurohumoral Substances in Tissues
         and Body Fluids, Lech Markiewicz (Poland)   	473
     Stress  and  Disease  in  Response  to Exposure  to  Noise-A Review, Gosta
         Carlestam, Claes-Goran Karlsson, LennarJ Levi (Sweden)   	479
     Some  Laboratory  Tests of Heart Rate  and Blood Volume  in Noise, Karl  D.
         Kryter (USA)	487
                                    Session 6

Sleep and Its Disturbance by Noise
Chairmen: B. Metz (France), M. Levi (Yugoslavia)

     Effects of Noise on Sleep-A Review, Harold Williams (USA)	501
     Predicting the Response to Noise During Sleep, Jerome S. Lukas (USA)	513
     The Effects of Noise-Disturbed Sleep on Subsequent Performance, M. Herbert,
         R.T. Wilkinson (UK)  	527
     Effects on Sleep of Hourly Presentations of Simulated Sonic Booms (50 N/m2),
         William E. Collins, P.P. lampietro (USA)	541
     Prolonged Exposure to Noise As a Sleep  Problem, Laverne C. Johnson, Richard
         E. Townsend, Paul Naitoh, (USA), Alain G. Muzet (France)  	559
     Relationship   between  Subjective   and   Physiological   Assessments  of
         Noise-Disturbed  Sleep,  A.  Muzet,  J.P.  Schieber, N.  Olivier-Martin,  J.
         Ehrhart, B. Metz (France)   	575
     The Effects of Aircraft Noise on  Sleep  Electrophysiology as Recorded in the
         Home,  Gordon Globus, Joyce Friedmann, Harry Cohen, Karl S. Pearsons,
         Sanford Fidell (USA)  	587
     Noise and Mental Health—An Overview, W. Hausman (USA)  	593
     Observations of the Effects of Aircraft Noise Near Heathrow Airport on Mental
         Health, C.F. Herridge, L. Low-Beer (UK)   	599
                                       Vll

-------
                                    Session 7

Community Response I
Chairmen: G. Thiessen (Canada), P.N. Borsky (USA)                              Page

     Methodological Aspects of Studies of Community Response to Noise,  EHand
         Jonsson, Ola Arvidsson, Kenneth Berglund, Anders Kajland (Sweden)  ...  611
     Decision Criteria Based on Spatio-Temporal Comparisons of Surveys on Aircraft
         Noise, Ariel Alexandre (OECD)  	619
     Psycho-Social Factors in Aircraft Noise Annoyance, Aubrey McKennell (UK)  . .  627
     A Survey on  Aircraft Noise in Switzerland,  Etienne Grandjean, Peter Graf,
         Anselm Lauber, Hans Peter Meier, Richard Muller, (Switzerland)  	645
     Aircraft  Noise  Determinants for the Extent of Annoyance Reactions, Ragnar
         Rylander,  Stefan Sorensen (Sweden)  	661
     Reaction Patterns  in Annoyance Response to Aircraft Noise, Stefan Sorensen,
         Kenneth Berglund, Ragnar Rylander (Sweden)	669
     The Reduction of Aircraft Noise  Impact  Through  a  Dynamic Preferencial
         Runway System, Martin Gach (USA) 	679
     A  Causal  Model  for  Relating Noise Exposure, Psychosocial  Variables and
         Aircraft Noise Annoyance,  Skipton Leonard, Paul N. Borsky (USA)	691
     Community Responses to Aircraft Noise in Large and Small Cities in  the USA,
         Harrold P. Patterson, William K. Connor (USA)   	707
                                     Session 8

Community Response II
Chairman:  R. Rylander (Sweden)

     Measurements of Street Noise in Warsaw and Evaluation of Its Effect on the
         Acoustic Climate of  Dwellings,  Schools, Offices, Hospitals,  Hotels and
         Parks;  the Degree  of Offensiveness to Inhabitants in the  Light of  a
         Questionnaire, Aleksander Brodniewicz (Poland)  	721
     A New Field Survey-Laboratory Methodology for Studying Human Response to
         Noise, Paul N. Borsky, H. Skipton Leonard (USA)   	743
     An  Interdisciplinary  Study  on the  Effects of Aircraft Noise on Man,  B.
         Rohrmann,  R.  Schiimer, A. Schiimer-Kohrs,  R. Guski, H.-O.  Finke
         (Germany)  	765
     Rating the  Total Noise Environment. Ideal or Pragmatic Approach? D. W.
         Robinson (UK)  	777
     Motor Vehicle Noise: Identification and Analysis of Situations Contributing to
         Annoyance, William J. Galloway, Glenn Jones (USA)  	785
                                       vui

-------
                                   Session 9

Summary and Integration
Chairmen: G. Zarkovic (Yugoslavia), W.D. Ward (USA)                            Page

    Summary, I.J. Hiish (USA) 	807
                                       IX

-------
            SESSION 1





INTRODUCTION AND MASKING EFFECTS




     Chairman: H.E. von Gierke, USA

-------
                     A PREVIEW OF THE CONGRESS CONTENT

                                   W. Dixon Ward

                             Hearing Research Laboratory
                               University of Minnesota
                            Minneapolis, Minnesota 55455
     Although the story is too long to recount in detail, I think that most of my fellow
participants here know that we owe our presence in this historic  and delightful city of
Dubrovnik not only to the hard work of Dr. Zarkovic and Dr. Levi, but also to a series of
lucky coincidences, culminating in the Noise Control Act of 1972, enacted by the Congress
of the USA on October 27, 1972. Two  provisions of this Act are as  follows: (1) The
Environmental Protection Agency shall "within nine months of the  date of the enactment
of this Act, develop and publish criteria with respect to noise. Such  criteria shall reflect the
scientific  knowledge most useful in indicating the kind and extent of all effects on the
public health  or welfare which may be expected from differing quantities and qualities of
noise." (2) It shall also, 12 months after enactment, "publish information on the levels of
environmental noise the attainment and maintenance of which in defined areas under var-
ious conditions  are  requisite to protect the public health and welfare with an adequate
margin of safety."
     Because of the obvious urgency of the charges (and in spite of their vagueness), EPA's
Office of Noise Abatement and Control was willing to subsidize our efforts to get together a
truly international meeting devoted exclusively to the effects of noise on human health and
welfare. It seems only fair, therefore, to look a little more closely at the task they have been
assigned.
     Now the term criteria, as used by Congress in the first provision above, consists of a
specified effect  or set  of effects that are set up as some sort of target—generally, a set of
conditions not to be exceeded. These criteria, in  general,  can be of two different types,
depending on whether they reflect concomitant  effects or after-effects of noise. If the
former, they may properly  be termed "noise criteria"; however,  the latter are more accu-
rately called "noise exposure criteria" because after-effects depend  not only on the charac-
teristics of the noise but also on the duration of exposure of a person to it. I believe that
this distinction is of paramount importance, though legislators do  not always understand it.
     For example, noise criteria could be (a) a certain degree of masking of ordinary speech,
or of radio or television perception; (b) a specified degree of vasoconstriction; (c) a definite
degree of probability of shifting the sleep stage  from a deeper to a lighter level; (d) an
average "comfortable loudness" as judged by some fraction of the population; or, conceiv-
ably, (e) the point at which aural pain is felt.
     On  the other hand,  noise exposure criteria could be  based  on a specified degree of
temporary or permanent threshold shift, or on a certain amount of hair-cell damage, or on a
change in circulatory problems in a specified fraction of the population, or a similar definite
change in any aspect of health.

-------
     There are, of course, certain effects that may be  both concomitant and residual. A
decrement in task performance—that most elusive of the many effects of noise that seem so
obviously  real but which generally vanish into thin air when one  looks for them in the
laboratory—might serve as a criterion for either noise or noise exposure. The same holds for
annoyance; we could generate criteria based on the arousal of the feeling of annoyance in a
specified fraction of the individuals concerned by  a specific noise, or we could use an
expression of integrated annoyance, as displayed by complaints and legal action.
     Even in the latter cases, however, it remains important to keep the two types of criteria
separate, and to try to educate the public—and particularly lawmakers—to the distinction.
The  latter, naturally, want to regulate noises, because they are relatively easy to measure.
However, the most important effects on human health come not from noise but from noise
exposure.  Noises per se are not hazardous, sleep-disturbing, or annoying-only noise expo-
sures can be. Therefore in this symposium our attention will be focussed primarily on noise
exposures: a duration of exposure must usually be involved, in addition to the intensity and
spectrum of the noise, when one calculates a noise "dose".
     Ideally,  in order to specify the relation between noises or noise exposures  and their
effects,  what is needed is a scale for measuring noise that  would result in each noise being
assigned a specific number whose magnitude  would reflect the relative noxiousness of that
noise. Let us assume for the  moment that such a scale could be found; since in science the
units of scales are  often named for famous men in the  field  concerned (for  example,
newtons of force, watts of power, amperes of current),  the unit of  this scale might well be
the peyser, in honor of one of the early pioneers in the noise field, Alfred Peyser. A noise
whose rating was 50 peysers would be twice as noxious in all respects  as one rated at 25
peysers, five times as noxious as a 10-peyser noise, half as noxious as a 100-peyser noise, etc.
     Furthermore, if noise exposure consisted of the instantaneous  value of the noise inte-
grated over time, noise exposure could then be expressed in peyser-hours. For example, a
man who worked 8 hours in a noise of 10 peysers, went home and  cut his lawn for half an
hour in  a 40-peyser noise, and then listened to his son's music group practicing at 50 peysers
for 2 hours before retiring for the night would have had a total noise exposure that day of
8x  10 + ^x40 + 2x50 = 200 peyser-hours. His effective exposure on that day would be
the same as that of another person whose noise exposure at work consisted of 10 hours at 20
peysers  and negligible the rest of the day, or that of a third man whose work environment
was  quiet, but who spent an hour at a rifle range in a 200-peyser noise without wearing any
ear protection. If the  size of  the "peyser"  had been defined  in such a  way  that  100
peyser-hours  were the maximum tolerable daily noise-exposure "dose",  then each of these
three individuals would have experienced  twice as much noise as he  should have, and if this
were continued day after day for many years, then he would be expected to show twice as
much hearing loss as the person exposed to only 100 peyser-hours each day.
     Unfortunately, nature has not been so obliging as to furnish us with such a scale, nor
indeed has she provided the uniformity of degree of effect that would make such a scale
even possible. That is, a noise mat is twice as likely as another to cause a person to awaken
is not twice  as annoying nor twice as  hazardous to hearing, nor does it produce twice as
much of a change in the circulatory system nor interfere with twice as much speech. In fact,
many noises that  are highly irritating and  hence should have a high  peyser index may

-------
produce no effect on hearing whatsoever, and so should be rated on that basis as being at
near zero peysers.
     Furthermore, it is known that a noise exposure of say 1 hour at 10 peysers will have
quite a different effect on temporary threshold shift measured immediately afterward than a
cumulative one-hour exposure again at 10 peysers, but in which 5-min periods of noise are
separated  by say  30 min  of quiet.  Therefore  if TTS  were  our criterion, the index  of
noxiousness would have to include some factor  that takes such intermittency into account.
     I fear, therefore, that the search for a single index of noise exposure as an indicator
that a given criterion effect has been  reached is foredoomed to failure. Perhaps I am
unnecessarily pessimistic—or perhaps some first-order approximation such as the concept of
"equal  A-weighted energy"  with appropriate  correction  factors will prove  to be close
enough to reality to justify its use in the absence of a true unifying principle. 1 know that
we will hear something more of this concept during the rest of our symposium.
     Most of us,  however,  find  it difficult enough  to cope  with the complex relations
between noise or noise exposure and our own favorite effect—in my case, for example, TTS.
Thus a host of specific questions will come under scrutiny in the some 90 papers that follow
this one,  questions whose answers are mostly still debated rather hotly. The following are
some that can be expected to appear:
     (  1)  Is hearing above 3000  Hz important to the  perception of speech? If so, under
what conditions?
     ( 2)  What frequency  weighting scheme, such as A-weighting and D-weighting, gives
the closest prediction of the speech-masking ability of a noise?
     ( 3)  Can there be damage to hearing without a change in sensitivity?
     ( 4)  What single exposure (8 hr or less) will just  produce a "significant" permanent
threshold shift (PTS)?
     ( 5)  What relatively steady-state  exposure, 8 hr/day, for many years, will just produce
PTS that exceeds that ascribable to presbyacusis plus sociacusis?
     ( 6)  Is there any  way  to correct audiometric  data for  presbyacusis-plus-sociacusis
other than simple (and probably incorrect) subtraction?
     ( 7)  Under what conditions does the  equal-energy hypothesis hold for steady expo-
sures?
     ( 8)  Can individual differences in susceptibility to PTS be predicted?
     ( 9)  Can this susceptibility be changed by drugs or diet?
     (10)  What is  the evidence for and against  the microtrauma theory as opposed to the
critical-incident hypothesis in the production of PTS?
     (11)  To what extent does it make any sense to speak of a "critical intensity" or even a
"critical exposure" for a given ear?
     (12)  Is a damaged ear more susceptible to further damage than a nondamaged one?
     (13)  In such case, what is "equal further damage" in the first place?
     (14)  Is some aspect of the TTS produced in a group of listeners a valid index  of
average expected PTS after years of exposure to that noise?
     (15)  If so, which parameter—initial TTS, recovery time, or  what?
     (16)  To what extent is the  auditory hazard from  noise enhanced by other noxious
influences such as vibration, fumes, exertion?
                                         5

-------
     (17)  To what extent does intermittency reduce the hazard from a given (cumulative)
noise exposure?
     (18)  In recovering from ITS, what noise level constitutes "effective quiet"?
     (19)  Is 4000 Hz  the place to first look for auditory  damage, or are the very-high
frequencies more susceptible?
     (20)  Does infrasonic noise or ultrasound at commonly-found intensities pose a hazard
to health?
     (21)  What are the effects of repeated awakenings or forced changes in depth of sleep
every night (by noise or by any other agent)?
     (22)  How much does simple reaction time to visual stimuli change in noise?
     (23)  Does such a  change persist after exposure (or, under what conditions does it
persist)?
     (24)  If so, is there any evidence that a permanent change might ensue?
     (25)  Does chronic arousal of the vegetative system lead to circulatory problems?
     (26)  Does chronic noise exposure increase mental problems?
     (27)  Is there any evidence in humans for changes in the adrenals due to noise expo-
sure, as commonly found in rats? Or in any of the other stress-reaction indicators?
     (28)  Is there any way to measure the "fatigue" that many workers complain of after
noise exposure?
     (29)  Does chronic  noise exposure increase the incidence of gastro-intestinal problems?
     (30)  What task performance is adversely affected by noise?
     (31)  Do  workers in high levels of noise show a significantly higher absentee and illness
record?
     (32)  To  what extent do individuals "adapt" to noise that does not pose a hazard to
hearing?
     (33)  To  what extent  can  such  "adaptation" be   manipulated by  propaganda
techniques?
     (34)  What constitutes a significant increase in complaints about neighborhood noise,
Le., how  much greater than the baseline of chronic  complainers must a  complaint  level
attain before a practical problem exists?
     (35)  Is utter silence seriously advocated at being the "best" acoustic environment?
     I  hope that  by our careful consideration of the evidence, answers to at least some of
these questions can be reached and accepted  by the majority of us here. I also hope that we,
at least, can keep from confusing noise and noise exposure.

-------
                        SYSTEMS OF NOISE MEASUREMENT

                                   Karl S. Pearsons
                           Bolt, Beranek and Newman, Inc.
                                Los Angeles, California
     Probably the most universally used system of noise measurement is something we all
carry with us all the time, our ears. It's an extremely versatile device that normally measures
sounds over a range of 120 dB and a frequency range from about 15 Hz to 20,000 Hz.
However, my talk today  is not  about our ears, nor is it a history of noise ratings,  but rather
a summary of the noise ratings which are currently in use today. Much of the material which
I will speak on is contained in a Handbook of Noise Ratings. This Handbook was prepared
for  the  National  Aeronautics  and  Space  Administration, Langley  Research  Center  in
Hampton, Virginia. I had hoped to have  copies of this available at  this time; however,
preparation delays and printing delays probably will prevent the Handbook from  being
available for 6 to 8 months.
     Hopefully my talk will prepare you for the kinds of noise ratings and measurements
which will be discussed in relating certain effects of noise to some noise measures.  Details of
the calculation procedures for  determining the various noise ratings will not be  presented
here since there is not enough time. Rather  the approach is to summarize the various classes
of noise ratings and provide some indication of the type of jobs that rating is supposed to
do. Some comparisons will also be made among the noise ratings but remember that each
noise rating is individualistic and cannot be translated directly to another noise rating except
for perhaps a particular sound which is being measured. To facilitate the discussion of the
noise measures  let us consider them  in five groups:  1) direct measures, 2) calculated
measures, 3) calculated measures for long term exposure (community response measures), 4)
graphical measures, and 5) measures specifically related to hearing level.
     The last category is not contained in  the Handbook of Noise Ratings which I men-
tioned earlier. However, because of the nature of this meeting it seemed important to briefly
touch on the nomenclature to facilitate the understanding of those sessions concerned with
hearing damage.
     Before we discuss the various measures, let me first mention some of the terms which
will  be employed in describing the various measures even  though most of you are familiar
with them. Although there are special units for measuring certain aspects of noise,  in general
noise is measured in decibels. This is a logarithmic quantity chosen because of the  very large
range of sounds  which people perceive. The decibel, usually abbreviated (dB), is  a measure
of a magnitude of a particular quantity such as sound pressure, sound power intensity with
respect to a standard  reference value. This standard reference value is usually 20 micro-
newtons per square meter. This  is about the  threshold of hearing for young ears at 1000 Hz.
     The other major aspect of  sound is its  frequency content. This is measured in terms of
hertz (Hz), formerly called cycles per second.  As I mentioned earlier, the range of hearing is
about 15 Hz to 20,000 Hz. This is really the number of times which something oscillates or
vibrates per second. For an example, the musical pitch "A" is an oscillation of 440 times per
second. A truck passing  by may have energy in the vicinity of 200 Hz. The high pitched

-------
whine of a jet engine would be about 3000 Hz. To further describe a sound in terms of both
its level and  frequency content the latter is sometimes divided into various bands. Such
bands as octave bands are sometimes employed. An octave band is a frequency band whose
upper and lower  cutoff frequencies have a ratio of 2. It is characterized by its upper and
lower frequency bounds or its center frequency which is the geometric mean of the upper
and lower bounds. Noise or sounds may be measured in terms of octave band sound pressure
level as shown  in Figure 1.  This is the  sound pressure level which is contained within an
octave band.  Finer resolutions may be made by employing third-octave bands or one-tenth-
octave bands  or even narrower bands.

DIRECT MEASURES

     The first measure is the overall sound pressure  level or sometimes simply the sound
pressure level with  various abbreviations such  as OASPL or SPL  or L or Lp. The overall
measure, which is approximated by the C-weighted network of a sound-level meter, provides
equal weight  for  all frequency components of the  noise. It is primarily used by engineers
who need a  measure which  indicates the total noise energy. Weighted sound levels are
measured on  sound-level meters in  terms of fast or  slow response.  These terms refer to the
speed with which an indicating meter follows the fluctuating sound. The approximate time
constants of this sampling procedure are about 1/10 of a second and 1 second respectively.
     The most  common and  widely used  sound measure in the world  is the A-weighted
sound pressure level or more simply the  A-level. This measure is also quantified in units of
dB although a shorthand technique has been employed to eleminate the necessity for saying
A-level each time a  measurement is quoted. This shorthand is to consider the unit a dB but
with an (A) following the dB and is usually read dB (A). It should be emphasized that dB (A)
is not an  actual  unit but rather a shorthand method to tell the  reader which weighting
network was employed to make the measurements. Figure 2 shows a diagram of A weighting
along with other weightings which we will  discuss shortly. Notice that the low frequencies
are attenuated.  The reason for this is to  more closely approximate the way people perceive
sounds. Originally it was designed for sounds of less than 55 dB in level; however, currently
it is used for all level sounds.
     The B-level sound does not discriminate  as much  against  the low frequencies. It is
shown  also on this figure, but currently is not widely used. The C-level as mentioned  earlier
provides an indication of the  flat response. Its frequency range was somewhat dictated by
the  frequency range of the ear but originally was influenced by available instrumentation.
Essentially the C-weighting limits the high and low  frequency response, but in spite of this
limitation  it  still  provides a reasonable  measure  of overall sound  pressure level for most
common noises.
     A relatively new addition to the weighting levels is the D-level. The weighting network
shown  in Figure 2 used for this measure is more complicated than the earlier ones and tries
to incorporate more accurately the frequency response of the ear. Actually it was originally
developed  as  a simple approximation of  perceived noise level PNL which I will discuss later
under calculated measures. Originally the  D-level was described as N-level with the  differ-
ence between the D and the N being 7 dB. In other words if 7 dB is added to the D-level,
one should obtain approximately the perceived noise  level of a given sound. Modifications

                                         8

-------
    no
    100

-------
 to D-level have been suggested by Kryter to account for new data. Mainly they affect the
 low frequency portion of the weighting.
     E-level has also been suggested and is fairly similar to the D-level. It was suggested by
 Stevens as an approximation to what he termed perceived level. Differences between the
 E-weighting and D-weighting are relatively small. At this time, the E-level is not standardized
 nor is it available on any sound level meter.
      -o
      c

      •o
      c
      o
      to
     JJ
      o
 20

 10


  0


-10


-20


-30


-40


-50
          20     50   100  200    500  1000
                            Frequency  In  Hz
                                                               5000     20,000
                  Figure 2. Frequency response of various weighting functions.
GRAPHICAL MEASURES

     The two graphical measures which will be described today are the noise criterion curves
and the preferred noise criterion  curves. Other measures such as Zwickjer's calculation of
loudness and some of the community response measures also employ graphical techniques,
however they will be discussed later under the calculated measures. The noise criteria curves
shown in Figure 3 were developed to provide a single number rating for octave band spectra.
They are, mainly employed by architects and engineers to specify the maximum noise levels
                                         10

-------
permitted in each octave band. In using them, the octave-band spectrum is plotted and an NC
value assigned to the noise, a value that corresponds to the highest NC curve to which the
spectrum is anywhere tangent. Thus the NC rating is almost always determined by the sound
pressure level at a single octave band frequency. For example, in Figure 3, an NC rating of
59 characterizes  the noise spectrum.  These NC curves were originally developed for office
spaces; however, they have been used in other environments such as auditoriums, sound
studios, restaurants, etc.
    Recently the  noise  criteria curves have been modified both to accommodate more
precisely the new octave band center frequencies  and also to answer the many objections
made  about the adequacy of the NC curves. Mainly changes are made in the higher and
lower frequencies since spaces designed in accordance with the previous NC  curves were in
some  cases  too  hissy or too rumbly. The new curves are shown in Figure 4. Here the
previously-mentioned noise spectrum would  have  a value of PNC 61  or 2 higher than the
rating given by the NC curves.

CALCULATED MEASURES FOR INDIVIDUAL EVENTS

     Table I provides a list of various calculated measures. For the most part, these measures
utilize octave- and  third-octave-band levels  of  noise which  are  employed  in  various
calculation schemes to come up with a single number rating of that noise. LLS stands for
loudness level, "s" refers to its originator, S. S. Stevens from the United States, who devoted
a good deal of his life refining the techniques for predicting the loudness level of sounds.
The scheme is intended to provide a level of the sound which is numerically equal in level to
that of a 1000-Hz tone which is judged  equal  in loudness to the sound being rated.  The
technique now is a calculation procedure which essentially transforms octave band levels to
a loudness quantity called sones that are added up  in a particular way and transformed back
to a decibel-like quantity known as phons.
    Another  scheme for calculating loudness  level  identified  as LLZ, "z" for Zwicker,
employs graphs and also allows for the upward  spread of masking, (the masking of higher
frequencies by low frequencies). This technique uses  one-third-octave band data and the
result is  intended to represent the level of a one-third-octave band centered at 1000 Hz
judged equally loud to the sound being rated.
    PNL or perceived noise level is similar  to  the loudness level by Stevens except  that
noisiness is employed instead of loudness. The units for this measure are PNdB or perceived
noise decibels. The numerical value was intended to represent the sound pressure level of an
octave band of noise at  1000 Hz which would be  judged equally noisy to the sound to be
rated. Equally noisy means that in a comparison of sounds one would just as soon have one
noise as the other at his home during the day or night.
    Stevens continued improving on his loudness  level calculation and came up with a new
rating technique  called perceived level which was similar in concept to the loudness level but
utilized more  information and included  noisiness  as  well as loudness judgment tests.  The
main difference between perceived level and the loudness level and perceived noise level is in
the numerical value. This time the levels were lower by approximately 8 dB than the earlier
loudness level calculation scheme. Also the  units are PLdB for'perceived level decibels

                                        11

-------
                                            1200
                                            2400
2400
4800
4800
9600
                                              NOISE CRITERIA

                                               NC  CURVES
                                                       IMC-40
                                                       NC-30
   APPROXIMATE
   THRESHOLD OF HEARING
  UFOR CONTINUOUS NOISE
   REF: ACUSTICA 14 (1964)
   PAGE 33, FIG. 14
10
 31.5     63     125      250      500     1000     2000    4000
         OCTAVE BAND CENTER  FREQUENCIES IN Hz (Cps)
            Figure 3. Noise criteria curves with noise spectrum example.
          8000
                              12

-------
                     1971 PREFERRED NOISE CRITERIA
                               PNC-CURVES
 Approximate threshold of
 hearing for continuous noise
 Ret:  Acustica 14 (1964)
      Page 33, Fig.14
H.5
63     125    250    500    1000   2000   4000
OCTAVE-BAND  CENTER FREQUENCIES IN  Hz

     Figure 4. PNC curves with noise spectrum example.

                   13
8000

-------
                                                 TABLE I
                    SUMMARY OF CALCULATED MEASURES FOR INDIVIDUAL EVENTS
      MEASURE
ABBREVIATION
UNITS
                   PROCEDURE
Loudness Level - Stevens
Loudness Level - Zwickei
Perceived Noise Level
Perceived Level
Tone Corrected Perceived
Noise Level

Effective Perceived Noise
Level
Single Noise Exposure
Level
Articulation Index
Speech Interference Level
    LL«
    LL,
    PNL
    PL
    PNLT
    EPNL
    SENEL
    Al
    SIL
Phons
Phons
PNdB
PLdB
PNdB
EPNdB
dB
(none)
(none)
Using 1/3 octave* band SPL's, loudness index (S) values
are determined from tables**. Total loudness S, is then
determined by S, = Smax + F (£S - Smax), F = .15

then LLS = 40 + 10 Iog2 S,

Graphical procedure using 1/3 octave band SPL's; includes
upward spread of masking effects

Using 1/3 octave* band SPL's, noy values (n) are deter-
mined from tables**. The total noy value (Ntot) is then
determined by Ntot = nmax + F (En - nmax), F = .15

then PNL = 40 +33.22 log ,0{N,ot)

Using 1/3 octave or octave band SPL's, sone values (S)
are determined from tables**. The total sone values (St)
is then determined by S, = Smax + F (ES - Smax) where
F is a function of level and bandwidth

then PL = 32 + 9 Iog2 St
PNLT = PNL + Tone correction
                        PNLT
                                          2d
EPNL = 10 log £   antilog (
             r=0         10
                             ) - 1 3
where PNLTj is value of ith .5 second sample
                n        AL
SENEL = 10 log (E  antilog - ) fit
                i=l        10

where AL is level of ith 1 second sample
      At is time interval between samples in seconds
      n  is the number of events

AI can be calculated from one-third octave band or octave
band differences in speech and background noise levels
where a weighting correction is applied to each band to
account for the relative contribution of each band to
speech intelligibility

SIL is the arithmetic average of the sound pressure levels
of the noise in the four octave bands with center frequencies
lying between 5 00 and 4000 Hz
                                                        *Octave band data may be used by employing !•' = .3
                                                       **Tablesare available in "Handbook of Noise Ratings".
                                                         Sec acknowledgement.
                                                     14

-------
instead of phons. The summation procedure for combining the octave or third-octave contri-
butions is more sophisticated and accounts for masking as a function of level.
     With the advent of discrete frequency components or tones in aircraft flyover noise it
seemed advisable to add a correction for the presence of these tones. Perceived noise level
does not  attempt to include these and therefore additional  calculations were employed to
account  for  the increased noisiness caused by  these tones.  The tone-corrected perceived
noise level, PNLT, is the current method for applying the corrections.  This method essen-
tially adds a correction to  the final value of the perceived noise level depending on the
amount that  the third  octave band containing the tone exceeds its adjacent bands. A fairly
complicated  procedure is actually employed utilizing computer techniques to determine
whether  or not a tone exists in a spectrum. This technique  essentially determines what the
noise floor is by reiterative  averaging process. The technique can be accomplished by hand
but is fairly cumbersome and a reasonable approximation can be made by averaging the two
adjacent bands and  subtracting this average from the band  level containing the tone. This
difference then is divided by three for frequencies between 500 and 5000 or divided by six
for other frequencies  and the result added to  the numerical value of the perceived noise
level. The limit of the tone correction is 6.7 for third octave bands between 500 and 5000
and 3.3 for all other third octave bands in the frequency range of 100 Hz to 10000 Hz.
     Since long-duration flyovers appear to be more annoying or noisier than short-duration
flyovers, a duration  correction was applied to the tone-corrected perceived noise level and a
new quantity called effective perceived noise level (EPNL) came into being. The method for
applying the duration  correction is essentially one which integrates or sums the PNLT levels
in half-second periods. This is equivalent to adding 3 dB for every doubling of duration of
the sounds. Currently  this measure is employed  in the aircraft noise certification procedures
in the United States. The units for the measure are EPNdB.
      SENEL  or single  event noise equivalent level is another measure of single events-in
particular, individual aircraft flyovers.  In this sense it is similar to effective perceived noise
level, but has no tone correction and employs the A-level weighting instead of perceived
noise level.  Presently its main use is in conjunction  with  the determination of community
noise equivalent level which will be described later.
     Moving  into the area of speech related noise measurements, we find two main calcu-
lated measures. One which will be mentioned in the following paper is articulation  index,
AI. Essentially it is a measure between 0 and 1 which purports to indicate speech intelligibil-
ity. It is based on the proportion of the normal speech signal that is available to the listener.
An articulation index  of  .6 or greater indicates reasonably  good intelligibility while levels
less  than .2  indicate  poor intelligibility. The technique  used in  determining articulation
index is  to divide the speech and sound into 20 bands from 200 Hz to 6100 Hz. The bands
are specially  selected such that for speech signals each contributes equally to intelligibility.
The value of articulation index is then  determined by the  sum in dB of the  differences
between the  peak speech levels and the noise spectra in each of the 20 corresponding bands
relative to an ideal speech to noise ratio of 30 in each band. Approximations are available
for determining articulation index from third-octave and octave band data using appropriate
weighting factors to account for the relative contribution of each band to speech intelligibil-
ity.  Many tests have been conducted using steady state noise to determine the percent of

                                          15

-------
various  types of speech material correctly understood for various levels of articulation
index.
     A simplified method, useful for engineers to determine approximate effects of noise on
speech,  is the speech interference level (SIL). Currently this is an arithmetic average of 4
bands centered at 500, 1000, 2000 and 4000 Hz, although other bands have been suggested.
No measures are made of the speech but in utilizing the speech interference level various
levels of voice are  assummed such as "normal", "raised", etc. Graphs can then be made
indicating the  distances over which speech is reasonably understood. Articulation  index
values for these various speech levels characterized by normal voice, raised voice, etc. have
been about 0.5. More about this matter will be described in the paper by John Webster.

CALCULATED MEASURES FOR MULTIPLE EVENTS
(COMMUNITY RESPONSE MEASURES)

     Probably the  biggest proliferation  of noise measures exists in the various schemes to
rate community noise. Several of these measures are in use in the United States and other
measures are employed in various countries around the world. The measures discussed here
do not include all those in use today but rather  should provide an indication of the various
ways in which community noise is assessed. All of the measures attempt to relate in some
fashion  to the noise impact on the community. As such, they eventually come to some type
of descriptive meaning for the various levels. Table II provides a summary of these measures.
In most cases the units for these  measures or ratings are in "dB-like" units. This means that
they are not actually in terms of decibels in the  normal sense of the word but they are in
logarithmic quantities which relate somewhat to decibels. Thus for example the measures
would increase by 10 units if the level of the contributing signals went up by 10 dB.
     CNR stands for Composite  Noise Rating or sometimes Community Noise Rating. It
was one of the  earlier attempts to evaluate community  reaction to noise in 1952. The
technique assumes that sounds were measured  in octave bands and that the values are
obtained by averaging over a reasonable time interval for critical locations in the commun-
ity. It utilizes a  family of curves that ranks the noise level on  a scale from A through M,
Thus a noise level rank is determined by plotting the octave band spectra on a set of level
rank curves in a manner similar to that used in determining NC levels described earlier. The
level rank thus determined is corrected for:
     1)  Discrete frequency components
     2)  Impulsive nature of the sound
     3)  Repetitiveness of the sound
     4)  Background noise level in the community
     5)  Effect of the time of day
     6)  Previous community exposure to the noise
     After all, corrections were applied, the final CNR, is determined as a new letter. The
letters are then  converted to  various  community  reaction  such as no annoyance,  mild
annoyance, mild complaints, strong complaints,  threats of legal action, and vigorous legal
action.  These categories have  been changed slightly throughout the years but remained
essentially the same. Later the CNR was employed for rating aircraft noise and at this time

                                        16

-------
the letters were replaced by dB-Iike numbers. The numbers were still derived from the same
level rank. The scheme also tended to shift from one which evaluates community noise to
one which predicted it on the basis of aircraft flyovers. Special contours were developed for
landings, takeoffs, and ground run-ups of various  types  of aircraft.  The end result still
remained the same which was to categorize the CNR into various community responses.
     As methods for rating aircraft noise improved, it was decided to create a new rating for
impact of aircraft noise on communities rather than continually change or update the CNR
rating method.  Thus the  noise exposure forecast NEF was born, which employed E, the
effective perceived noise level,  as its basis rather than the perceived noise level as mentioned
earlier. This measure accounted for both the additional noisiness  of discrete frequencies or
tones and the effect of duration.  A correction for nighttime operations of 10 dB was still
employed. In other words, for the same average number of aircraft operations per hour, the
NEF value for the nighttime operations would be 10 units higher than for daytime opera-
tions. The NEFs around  an  airport were lowered  in absolute value  by  subtraction  of  a
constant of 88  to avoid confusion with the previously developed CNR. An example of NEF
contours are shown in Figure 5. The final NEF values were converted to 3 levels of com-
munity response as follows:

     NEF          Description of Community Response

     Less          Essentially no complaints would be expected. The noise may, however,
     than 30       interfere occasionally with certain activities of the residents.

     30 - 40        Residents  in the  community may complain, perhaps vigorously. Con-
                   certed group action is possible.

     Greater       Individual  reactions would likely include repeated, vigorous complaints
     than          and recourse  to  legal action. Concerted  group action would  be ex-
     40            pected.
                                                             NEF 25,
                            Figure 5.  Example of NEF contours
                                         37

-------
     Before going on to the next type of community noise rating let us discuss briefly a very
general measure. This is called the equivalent sound level, or Lgq. Very simply, this is the
energy average of the noise level (usually in A-level) for some specified amount of time. This
measure is used to allow quantification of fluctuating sounds over a long period of time. In
essence it is that numerical value of the fluctuating sound which is equivalent in level to a
steady state sound with the same amount of total energy. If, for example, one had a sound
which was 80 dB and it was on for half an hour and then went down to a level of 40 dB for
half an hour, then  the Lgq would be the energy average, or 3 dB less than the maximum
value.  Thus the energy Lgq for this sound  sample would be 77 dB and not 60 dB as we
might  expect if we just averaged the dB levels. Leq is usually approximated by taking several
samples in time of A-level, then averaging the samples by first dividing by 10 and taking the
antilog and actually averaging those quantities, then reconverting back to a dB level by
talcing the log and  multiplying by 10. Since this measure is employed in several measures
which  will follow, I  felt it was important to briefly discuss it at this time.
     One measure which does indeed use this Leq is CNEL or Community Noise Equivalent
Level. This rating represents the average noise level on an  energy basis determined for a
24-hour period with different weighting factors for noise levels occuring  during the day,
evening, and nighttime periods. Essentially, then, this is an L eq for the day for the 24 hour
period  but with  special weightings of 5 dB  and  10 dB respectively to account for the
increased disturbance caused by noise events during the evening (1900-2200) and nighttime
(2200-0700) hours. To facilitate these calculations, an hourly noise level or Leq for an hour
is employed and weighting factors are applied directly to  this measure.
     The next group of measures, which includes  the Isopsophic index designated as (N)
which  is used in France, the mean annoyance level (Q) which is used in Germany, the noise
and number index (NNI) used in England, the noisiness index (NI) used in South Africa, the
total noise load (B) which is used in the Netherlands, and the weighted equivalent contin-
uous perceived noise level,  WECPNL which was suggested  by the International Civil Aviation
Organization are  all somewhat similar to the  measures of the CNR or CNEL measures
already discussed. They do differ in some detail.
     First the isopsophic index: This is a noise rating which  takes into account the energy
average maximum perceived noise level of aircraft noise and the number of events. Another
French measure, the classification index (R), is identical in all aspects to the isopsophic
index.  The big difference  between this and other measures is in the  handling of nighttime
events. In the first place, the nighttime events are broken into early night (2200-0200) and
late  night (0200-0600) time periods.  The early night period is viewed as three times  as
significant as the second or late night time period. Also, the effect of doubling the number of
operations at night  is not as great as during the day, since doubling the number of opera-
tions increases the nighttime portion of this measure by less than 2  units as opposed to 3
units for the normal daytime operations. The measure is used to determine zones for various
types of buildings. The zones include areas where all buildings are prohibited down to levels
for which no building restrictions apply.
     The mean annoyance level Q is another noise rating for aircraft noise impact on a
community. It uses sampled A-level to  provide an average noise level  for a specified tune
period-for example, day,  night  or 24 hours. Again, it is similar to CNEL except there is no
nighttime  weighting and a doubling of the number of events increases the measure by 4

                                         18

-------
units as opposed to 3 units for CNEL. The Q is also employed to designate 4 zones as did
the previously discussed isopsophic index.
    NNI uses the average perceived noise level (averaged on an energy basis) in combination
with the number of aircraft heard within a specific period. Unlike the previously discussed
measures,  this rating increases by 4Vz units for a doubling  of the number of events. No
distinction is made between daytime and nighttime in the calculation procedure; however,
different levels of NNI are employed in determining reasonable levels for daytime or night-
time operations. Thus a level of 50 to 60 NNI is assumed  to be unreasonable during the
daytime whereas an NNI of 30 to 45 is intolerable during the nighttime.
    The total noise load, B, employs maximum A-level  and  number of aircraft with ap-
propriate  weightings for time of day. B was developed to be  numerically equal to the
percentage of mean relative nuisance. The Dutch authorities have chosen a B rating of 45
which  is  equivalent to  45%  mean relative nuisance score. The main  difference  in  this
measure from the previous ones is the fact that a doubling of the number of events increases
the rating by 6 units rather than 3 for normal energy summations. The number of time
periods for which weightings are given is increased to include  10 different periods during the
24-hour day, with nighttime hours weighted  much more heavily  than other measures. The
noisiness index (NI) used in South Africa is the energy average noise level  based  on a tone
corrected  A-level for a 24-hour period. Appropriate weightings are applied for time of day
and also season of the year.  The tone corrections for A-level are determined from third-
octave-band levels before summing to obtain  a corrected A-level. The actual tone-correction
procedure is taken from the techniques employed for EPNL or PNLT tone corrections. Two
sets of weightings for day and nighttime activities are provided for two different  groups of
periods; for example, if the day is divided into two periods, there is  a 10-db weighting for
nighttime  events occuring during the hours of 2200 and 0700, while if the day is divided
into three periods, then  a weighting of 5 dB for evening hours of  1900 to 2200 is employed
and a 10 dB weighting for nighttime events occuring between the hours of 2200  and 700.
Seasonal corrections are based on the number of hours in a  month which the temperature
falls in the range of 20 degrees centigrade to 25 degrees centigrade. This is done to provide
more weighting for the situations when the windows are open  in the summer months.
    The weighted equivalent continuous perceived noise level, WECPNL is an attempt to
provide a standardized measure for the impact of aircraft noise on the community. This is
quite similar to the CNEL described earlier but uses tone corrected perceived noise level as
its base for energy averaging rather than A-level. Also, weightings are included for  season of
the year, and time-of-day corrections  for only two periods rather than three three periods
used in the  CNEL calculation procedure are  employed. To provide an indication  of the
approximate values of the various measures we have discussed in terms of number of aircraft
operations per day, see Figure 6, which shows the levels for the various rating techniques.
The figure assumes a flyover noise of 110 PNdB with an effective duration of 10 seconds.
Approximations had to be made in certain cases in order to make this comparison possible.
For example  in the WECPNL and the NEF a flyover of 110 EPNdB  instead of 110 PNdB
was employed. Notice that the lines on the graph are not all  parallel to one another. This is
because the number of operations per day is not always summed in an energy fashion for all
of the  measures as discussed earlier. If we are to pick an average number for CNR  of 110

                                        19

-------
I J
c
      !      o    <3
      
-------
this would provide an NEF of approximately 35 and a NNI of 50, an isopsophic index of
92, a Q of 75, total noise load B of 50, a noisiness index of 70, and a weighted equivalent
continuous perceived noise level of 83. It has been assumed  that the number of aircraft
operations per day fall in the range of about 10 to 20.
    The rating sound level or Lr  is similar in  many  respects to the original CNR rating
method, except that L equivalent or A-level is used instead of level rank curves. Corrections
are provided for such things as the impulsive nature of the sound, the duration of the sound,
and whether  or not a whine or tone is present. In addition, although corrections are not
applied directly to the Lj., other corrections  are employed to the basic criterion associated
with Lj. to include such effects such as the time of day, the background noise and previous
exposure to the background noise.  If octave-band levels are employed in this procedure, a
set of noise rating (NR) curves are available which are somewhat similar to the original level
rank curves. This measure is included in the ISO recommendation on noise assessment with
respect to community response.
     Because of the feeling that the variation of noise levels was not adequately accounted
for in  the measures described above, in particular for the normal variations observed in
traffic  noise, another measure was  developed called the traffic noise index, or TNI, which
uses LIQ  and Lgg. These indices represent the levels which on the average were exceeded
10% and 90% of the time. The difference between the values provides an indication of the
variability  of the  sound. Actually,  the TNI  as shown in  Table II is  then equal to 4 times
this difference plus the background noise level  which is represented by Lp0. In using this
measure as a limit the problem exists that for sounds with very small variation a fairly high
permitted background level  would result.
     An improvement over the TNI developed by Robinson of England is the noise pollu-
tion level,  or NPL. The noise pollution level is a little more sophisticated than the traffic
noise index but tries  to accomplish the same sort of thing. In this case it uses the energy
mean of Leq of the sound and to  this is added the standard deviation of the noise (noise
level, not  the noise energy) multiplied by some constant. Typically the formula is as shown
in Table II. An approximation of this is provided by formulas utilizing LJQ and Lgg; thus,
the noise  pollution level is equal to  Leq plus the  quantity  L\Q  - L^Q.  Still  another
approximation is shown in  Table II.  The latter approximations  are  only  valid if the
distribution of noise levels is reasonably normal.

MEASURES RELATED TO HEARING LEVEL

     Table III shows some abbreviations that are employed in  research concerning hearing.
The first is hearing level (HL) in dB, sometimes referred to as "hearing loss". Essentially it is
the level of an individual's hearing relative to a standardized hearing level determined for
young  adult ears. It is the measure of hearing threshold. Thus a hearing level of 40 dB would
mean that the person's sensitivity to sound is 40 dB less than the standard or average level.
A hearing level of - 5 dB would mean that the person had hearing of 5  dB better than the
average young adult ear. Hearing levels are established at various frequencies usually,  starting
at 125  Hz or 250 Hz and proceeding in octaves and half-octaves up to 8000 Hz. Hearing level
sometimes  refers to the  average  of the levels at various frequencies. For example,  hearing

                                         21

-------
                                             TABLE II
                               SUMMARY OF COMMUNITY MEASURES
         MEASURE
ABBREVIATION
                   PROCEDURE
Community Noise Rating
    CNR
Noise Exposure Forecast
    NEF
Community Noise Exposure Level
    CNEL
Equivalent Sound Level



Isopsophic Index
    N
Mean Annoyance Level
Originally determined from level rank curves plus corrections.

Presently,

CNR = PNLmax + C

where:   PNLmax is maximum perceived noise level
        C       is sum of corrections for time of day,
                frequency of flights, and season of
                the year

                     EPNL
                         n
NEF = 10 log [£ antilog (	) + 16.67 £ antilog
                       10
                     Daytime
                     Events


                                    EPNL
                                          n
                                    (	)]-88
                                       10
                                   Nighttime
                                    Events

            where:  n is the event number
                          HNL

              £ w • antilog (-TT—)
CNEL =10 log {      24           ]

                         SENEL
                                n
              £ w • antilog (—rr—)
CNEL =10 log [   864000      U    ]

where:


w is the tone of day weighting factor (1,3,10)
h  is the number of hours (0-2 3)
n  is the number of events
                         10 log   Q/  antilog AL(t)dt
                   "night =
                                                                  6 Iog10 (3ni + n2 - 1) - 30
                                                   24 hours - 10 1Q8 t«tiog ef) + antilog
                   Q> 13.3 log [
                                                             n       AL_
                                                             £ antilog 13.3

                                                             M
                                                                   T
                                                 22

-------
                                            TABLE II
                          SUMMARY OF COMMUNITY MEASURES (Continued)
         MEASURE
ABBREVIATION
                   PROCEDURE
Noise and Number Index
Noisiness Index
Total Noise Load
Weighted Equivalent
Continuous Perceived
Noise Level
 Rating Sound Level
 Traffic Noise Index
 Noise Pollution Level
    NNI
    NI
                                 WECPNL
    TNI
                                  NPL
NNI = PNLmax+ 15 log N - 80
_
NI = 10 log 2 antilog
                                                                        101og(t/t0)-f-C
                                                                           10
                                        AL
                   B = 20 log [S w • antilog ^-] - C
                                                                    EPNLi  .„.  To
                   TNEL = 10 log £ antilog ——+ 10 log -—
                                          10        °
                                                 ECPNL = TNEL-10 log f-
                                                                5       ECPNL (D)2
                                                 WECPNL = 10 log [5 antilog
                                                                O
                                              10
                                                 •,       ECPNL (N), + 10
                                                 j- antilog	ft	]  + S
 Lj = L A* Impulse Noise Correction

 L,. = Leq + Fluctuating Noise Conection


 TNI = 4(L10-L90) + L90-30


 NPL = Leq + 2.56 a
                                                                         (L  0-L90)2  Assuming Normal
                                                     • L50 + 
-------
levels associated with speech are sometimes the average of hearing levels at 500,  1000 and
2000 Hz.
     After exposure to some levels of noise, hearing levels change. This change is referred to
as threshold shift. It is the difference in the hearing level before and after exposure to some
stimulus. The abbreviation TTS or temporary threshold  shift is the amount of shift that
occurs at a given time after exposure. For example, typically TTS2 refers  to the threshold
shift occuring 2  minutes after cessation of the noise. If no recovery exists after  exposure,
the  threshold shift is considered permanent and is dubbed permanent threshold shift PTS.
Sometimes this threshold shift is further described as NIPTS or noise induced permanent
threshold shift.
                                      Table III

              SUMMARY OF MEASURES RELATED TO HEARING EVALUATION

 Hearing Level                                                               HL

 Hearing Threshold Level                                                      HTL

 Temporary Threshold Shift                                                   ITSf*

 Permanent Threshold Shift                                                    PTS

 Noise Induced Permanent Threshold Shift                                      NIPTS


 *Subscript "t" refers to time after exposure threshold was determined.


                              ACKNOWLEDGEMENT

    The material contained in this paper is primarily taken from the "Handbook of Noise
Ratings" by  K.S. Pearsons and R. L. Bennett. This Handbook was prepared under contract
for the National Aeronautics and Space Administration, Langley Research Center.
                                        24

-------
               THE EFFECTS OF NOISE ON THE HEARING OF SPEECH

                                   John C. Webster
                          Naval Electronics Laboratory Center
                             San Diego, California  92152


     To cover this subject matter, I will talk about  the Articulation Index or AI, and the
Speech Interference Level, or SIL, and in particular the relation between them. I will show
that  the best octaves to choose in calculating the SIL depend on what Articulation Index
(AI) you want  to work at or design for. And to make this meaningful, I will have to show
you  what scores you can expect to get on syllable, word, or sentence tests at various
Articulation Indices.  Beyond this, I will discuss what sort of  tests  can be  used to  test
systems or listeners  operating at high-AIs, that is, in relatively quiet environments. If this
seems off the subject, I will relate these types of tests to methods of evaluating hearing aids
and/or different kinds of hearing losses including noise-induced hearing losses.
     To talk of these things intelligently, I will have to spend a little bit of time discussing
the pros and cons of efficient intelligibility tests.  At the first of these conferences (Webster,
1969) I traced the early history of intelligibility testing. I will not repeat it here, but I would
like  to stress a  single distinction made by the early Bell Telephone Laboratory investigators,
namely, articulation  testing as opposed to intelligibility testing. Articulation testing involves
the use of nonsense syllables to determine what single speech sounds, phonemes, distinctive
features, or consonants are misheard.  Once any aspect of redundancy or language enters
the  testing it is no longer articulation  but  intelligibility  that is being tested. Articulation
testing centers  on speech sounds per se. Intelligibility testing involves  both the ear and the
brain or involves both speech sounds and language.
     To summarize very briefly the problems associated with speech testing, I must mention
that the construction of speech intelligibility tests varies along two dimensions—the redun-
dancy and/or vocabulary size of the input stimulus (language) and the constraints or number
of possible  choices in  the output or response.  Within vocabularies of the same size the
relative familiarity  of the word and the number of syllables in  a  word and the context
within which it is imbedded influence  its intelligibility. The constraints on the response,
open vs. closed sets,  also affect intelligibility scores. Closed set  or Modified Rhyme Tests
(MRT) (House  et al., 1965; Clarke, 1965; Kreul et al.,  1968) and pseudo-closed set rhyme
word tests (Fairbanks, 1958) are largely replacing the open-set Phonetically Balanced (PB)
(Egan  1948) and Spondee tests and other  multiple choice tests  at the present time.  The
major reason is the time and effort required to train both talkers and  particularly listeners in
the open-set PB-type word test.
     This  is not the document to trace out in any more detail the history, the rationale, the
strengths  and weaknesses, nor  the actual listings of syllables, words, phrases, or sentences
used this  century to evaluate the effects of noise on  speakers, listeners, communication
components and systems, etc. Recently, however, Webster (1972) has  compiled 24 lists of
word,  phrase,  and  sentence  tests in English. A  very  good reference  for more details on
intelligibility tests is Clarke, Nixon, and Stuntz (1965)  because it has abstracts of over 160
earlier references.

                                          25

-------
     Table 1  shows the relationship between intelligibility scores and Articulation Index (a
special form of speech-to-noise ratio) of many standard speech tests. The generalizations to
be made from Table 1 are that the smaller the stimulus vocabulary and/or the size of the
response set,  or the more redundant in terms of context, the higher the score for a given AI
or speech-to-noise ratio.


                                       Table 1

 Expected Word  or Sentence  Scores  for  Various  Articulation Indices (AI)
A!
0.2
0.3
0.35
0.40
0.50
0.60
0.80
PB*
22
41
50
62
77
85
92
MRT**
54
72
78
86
91
94
98
SENT*
77
92
95
96
98
98
99
 *From Kryter and Whitman (1963)

 **From Webster  and Allen (1972)
     So far I have mentioned only the printed stimulus and response variables that affect
intelligibility testing. The talkers, listeners, and the noise environment around them  have
very large effects on test validity and reliability. For example, Dreher and O'Neill (1957)
had  15 naive speakers read in 5 different noise levels. When the words and sentences  were
played to listeners at a constant speech-to-noise differential the speech originally recorded in
noise was the more intelligible. Pickett (1956) shows, however, that if vocal effort measured
one meter in front of the lips exceeds 78 dB, intelligibility drops.
     It should  be apparent by now that intelligibility  test results require some interpre-
tation. It is neither  simple nor straightforward to assess the affects of noise on speech using
word testing methods. It  would be advantageous to specify the effects of noise on speech in
terms of the spectra and level of the noise and of the speech. Two  such physical schemes

                                        26

-------
exist-the Articulation Index, AI, and the Speech Interference Level, SIL. The Articulation
Index or AI assumes that there are 20 bands in the speech spectra between 200 and 6100 Hz
that differ in bandwidth such that each band contributes 1/20 of the total articulation. Each
band  contributes linearly  to the extent that the speech peak level exceeds the  RMS noise
level by from 0 to 30 dB. The AI is a specialized method of specifying the speech-to-noise
ratio.  It is a non-dimensional numeric that varies from zero to one, but it can be considered
to be a decibel scale ranging from zero to 30 such that for example an AI of 0.5 corresponds
to a complex signal-to-noise ratio of 15 dB, 0.8 to 24 dB, etc.
     The AI was introduced by French  and Steinberg (1947), generalized and simplified by
Beranek (1947a), and refined and validated by Kryter (1962a, 1962b). The AI was discussed
at the first Congress of Noise as a Public Health Hazard by Webster (1969) and by Flanagan
and Levitt (1969) in sufficient detail that it will not be belabored further here.
     Almost simultaneously with the introduction of the AI, Beranek (1947b)  proposed a
simplified substitute for it, the Speech Interference Level (SIL) of noise. The definition of
SIL is the arithmetic  average of the decibel levels in  three or four selected octaves. The
choice of octaves will be discussed later. The SIL is only a measure of noise, and to interpret
it in terms of permissible distances between talker(s) and listener(s), reference must be made
to a  table (Beranek (1947b)) or a graph (Botsford (1969), Webster  (1969)).  An updating,
Mark II, of the Webster (1969) graph is shown as Figure 1, It differs from Mark I unveiled at
the first of these  conferences by (1) adding  two new physical measures,  the four octave
PSIL (.5/1/2/4) and  the proposed  SI-60  weighting which will be discussed in more detail
later; (2) appending an AI scale to help orient people in the real meaning of the figure; and
(3) a droopoff in the communicating voice level curve to reflect the fact that at voice levels
above 78 dB intelligibility does  not increase as fast with vocal effort as at lesser levels. The
gist of the figure  is that  for an AI of 0.5 using "normal" vocal effort (65 dB  at 1 meter)
conversation at  16 feet or 5 meters can take place in noises as high as 50 dB as measured on
the A-weighting network of a sound level meter.
     The one aspect of AI that has been alluded to by many (see Webster (1965)) but not
fully  appreciated is that as  the AI and its correlate, word intelligibility, increase, the most
important speech  frequencies  and/or the  frequency range of noise  that masks the speech
most  effectively increases from between 800 and  1000 Hz to between 1700 and 1900 Hz.
This  of course  should be reflected in  the octaves chosen to calculate the SIL, and this
relationship will  be developed in  the next four figures.
     Figure 2  shows a method  of calculating the AI by  counting the proportion of dots
between the noise spectra and the upper limit of the conversational level speech spectrum.
The example shows how it can  be used to specify the AI for a -6 dB per octave (-3 when
measured in octaves) noise.
     This figure was developed from the Cavanaugh et al., (1962) procedure of deriving AI's
from dot patterns spaced in a 30-dB range in the shape of the normal male speech spectrum.
The concentration of dots reflects the relative importance of different frequency bands to
the intelligibility of speech heard in noise. Figures 3, 4, and 5 show the results of calculating
AI's at 0.2, 0.5, and 0.8 for 5 theoretical noises, and show why and how the octaves chosen
for SILs should vary accordingly.
     Note from  Figure 3 that the spectra lines cross each other (with about a 2 dB spread)
at  1000 Hz. Since these  are all well-behaved, theoretical noises with constant slopes, the

                                         27

-------
        SI-60
     SIL(LLB)
    32


    16
LU
u_
5   8
LU
:/
       CO
            2

            1
                  35
                  30
                       45
F       X   X'X
      APPROXAI FOR NORMAL VOICE
       \      \  X \
 0.1    0.8     0.6  X 0,4
     A  LEVEL 30
     SIL  (4) 21
        PSIL 23
                                                                                    NOISE-DISTANCE AREA
                                                                                       WHERE UNAIDED
                                                                                        FACE-TO-FACE
                                                                                      COMMUNICATIONS
                                                                                      AREA IMPOSSIBLE
                NOISE-DISTANCE AREA
                WHERE FACE-TO-FACE
                 COMMUNICATIONS IN
                  NORMAL VOICE IS
                   POSSIBLE AI >0.5
                   31
                   33
                        50
                        41
                        43
60
51
53
70
61
63
80
71
73
90
81
83
100
 91
 93
110
101
103
             Figure 1. Necessary voice levels as limited by ambient noise for selected distances between talker and
             listener for satisfactory face-to-face communication. Along the abscissa are various measures of noise, along
             the ordinate distance, and the parameters are voice level. At levels above 50 dB(A) people raise
             level as shown by the "expected" line if communications are not vital or by the "communicating"
             communications are vital. Below and to the left of the "normal" voice  line communications are at  an Al
             level of 0.5. 98% sentence intelligibility. At a shout, communications are possible except above and to the
             right of the "impossible" area line.

-------
                          OCTAVE  PASS BANDS  IN  CYCLES PER SECOND
                    90  —   ISO  —   55S  —   TIC  —  1400  —  2tOO  —  S600  —  11200
    20
                     100                         1000
                       FREQUENCY  IN CYCLES  PER  SECOND
10000
Figure 2. Al speech region for "conversational level" speech. The number of dots in each band signifies the
relative contribution of speech in that band to the Al. A series of idealized thermal noises with -6 dB/oct
spectra are drawn in 5 dB steps. The number of dots above each noise contour is proportional to the Al of
conversational level speech in that level of noise. (After Cavanaugh  et al., 1962.
                                           29

-------
     II 0
                          OCTAVE PASS BANDS IN CYCLES PER SECOND
                     90  —   IBO  —   355  —  TlO   —  HOO —  Z«00 —  S«00 —  llfcOO
                                 250      500      1000     2000     4000
     40
                      100                        1000
                        FREQUENCY IN  CYCLES  PER SECOND
10000
Figure 3. Allowable octave band sound pressure levels of steady state noises with spectrum slopes of -12, -9,
-6, flat, and +6 dB per octave for an Al of 0.2 and conversational level speech. The superimposed St-70
contour is a proposed frequency weighting network for evaluating the speech interfering aspects of noise at
Al = 0.2.
                                         30

-------
crossing point at 1000 Hz is also the S1L for the octaves centered at 500, 1000, and 2000
Hz (.5/1/2 SIL). Note also that the spectra cross, a hypothetical line at 1414 Hz (.5/1/2/4
SIL) with a spread of about 9 dB and that the spread at 2000 Hz (1/2/4 SIL) is about 18
dB. Obviously, the  .5/1/2 SIL is the measure with the least variability for specifying the
level of diverse-spectrum noises at an AI  of 0.2,  which corresponds roughly to Fairbanks
(1958) Rhyme Test (FRT) and Modified Rhyme Test (MRT) score of just over 50%, a 1000
word phonetically Balanced  (PB) word score of just under 25%, and a sentence score of just
over 75%.
     Interpreting Figure 4 in the same way, it is evident that (1) at 1000 Hz (.5/1/2 SIL) the
spread  is about 10 dB; (2) at 1414 Hz (.5/1/2/4 SIL) the spread is minimal, about 2 dB; and
(3) at 2000 Hz (1/2/4 SIL)  the spread is about 8  dB. It is equally apparent therefore that
the .5/1/2/4  SIL shows the least variability in specifying an AI of 0.5 which corresponds to
a PB score just over 75%,  an MRT (and FRT) score of about 90%, and  a  near perfect
sentence score.
     Figure 5 shows the  1/2/4 SIL to be the  least variable in specifying an AI of 0.8 which
results in near-perfect socres  on all word and sentence testing materials.
     It should now be apparent that the  choice of octaves in calculating SIL is directly
related to the intelligibility required of the system to be evaluated or to the AI expected of
the system. But just to summarize it once more let us look at Figure 6.
     Note for example that  the slope of AI versus SIL decreases with decreasing SIL levels
as the spectral slope changes from -12 to -6, to 0, to +6 dB per octave. It therefore follows
and it is evident from Figure 6 that when  these 4 theoretical noises are equated in level to
give an approximately equal AI of 0.2, as measured by the .5/1/2 SIL, they are not equal at
AIs of 0.5 or 0.8. When equated by the .5/1/2/4 SIL, the noises are generally equivalent in
level at an AI of 0.5 but  not at 0.2 nor 0.8. Finally, if equated by the 1/2/4 SIL they are
generally equivalent at an AI of 0.8 and not at 0.5 nor 0.2.
     Interpolation  shows that a 50% PB  score (AI = 0.35)  could  be about  equally well
specified, over a large diverse  sample of noises by  an .5/1/2 or a .5/1/2/4 SIL. An AI of
0.35, MRT (FRT) score  of  80 and sentence score  of 95% has been recommended as the
minimum acceptable specification  for  certain military  communication equipments (see
Webster and  Allen, 1972) operating in highly adverse environments. Even lower levels for
acceptance have been suggested for use  in  the past (see Webster, 1965) and  thus lend
credence to using the .5/1/2 SIL for measuring the effects of Navy noises. Architects and
others  working in  quieter  environments  and requiring higher levels of communication
efficiency naturally  prefer AIs of 0.5 for which the .5/1/2/4  SIL  is the least variable
measure. Only the perfectionist would need to design or operate at  AI levels of 0.8 and so
there is probably no serious reason for considering the 1/2/4 SIL for practical engineers.
     Probably the best validating data concerning  the  change of SIL frequency with AI are
those of Cluff (1969). Guff equated the spectra and levels of 112 industrial noises to give
one-third octave AIs of 0.1, 0.2—0.9, and then  determined the bandwidth that gave the
best prediction (least standard diviatibn) over all noises for (1) an average level in one third
octave  bands—similar to an SIL—(2) an overall or band level—similar to a Gweighted
(but band-limited) sound level meter reading—as well as broadband measures of (3) the
A-weighting,  and (4) the  proposed SI-70 weighting. He found as the AI increased from 0.1
to 0.9  the center frequency  of the optimum bandwidths  increased  from 848 to 2264

                                         31

-------
                        OCTAVE  PASS BANDS IN CYCLES PER SECOND
                   *0  —   ISO  —  355  —   710  —  1400  —   ?§00 —  5»00 —   11200
   40
                    100                        1000
                      FREQUENCY IN CYCLES PER  SECOND
10000
Figure 4. Allowable octave band sound pressure levels of steady state noises with spectrum slopes of -12, -9,
-6, flat, and +6 dB per octave for an Al of 0.5 and conversational level speech. The superimposed SI-60
contour is a proposed frequency weighting network for evaluating the speech interfering apsects of noise at
Al = 0.5.
                                        32

-------
CD
O
cr
o
CM
O
o
o
UJ
or

CD
T3
UJ
>
UJ
DO

UJ

5
r-
O
O
60
 50
40
       30
       20
 10
                            OCTAVE PASS  BANDS  IN CYCLES PER SECOND

                  —   90   —   ISO  —  555  —   TlO  —   1400 —  2800 — 5600 —  IIJOO
                                                                   \
                                                                       \
                                                                \
                                                                   \
                                                                             \
-9


-12
                   -.'
                           l?5      250
                                           500
                                                    1000     2000     4000    8000
                1

                5
                        IOO                        IOOO

                          FREQUENCY  IN CYCLES PER SECOND
                                                                        IOOOO
 Figure 5. Allowable octave band sound pressure levels of steady state noises with spectrum slopes of -12, -9,
 -6, flat,  and +6 dB per octave for an Al of 0.8 and conversational level speech. The superimposed SI-50
 contour is a proposed frequency weighting network for evaluating the speech interfering aspects of noise at
 Al = 0.8.
                                          33

-------
5
3
        70
                   RELATION BETWEEN  ARTICULATION  INDEX  (Al) AND
                            SPEECH  INTERFERENCE LEVEL (SID
               60         50         40          50         40         50
                            40
1,0

0,8

0,6

0.4

0,2

0.0
       I
re Al ALA
CAVANAUGH ETAL
T
 -12
                                             -!*•
                                           (PSIL)
4 BAMHSIL)
 ,5/1/2/4
                  T
        70
               60         70         60          70         60         50
                PREFERRED FREQUENCY SPEECH INTERFERENCE LEVEL IN oB
Figure 6. Relation between Articulation Index (Al) and Speech Interference Level (SIL) for 4 noises with
spectrum level slopes of -12, -6, flat, and +6 dB/octave. Three different sets of octaves are shown for
calculating  SIL; from  left to right: 500, 1000, and 2000 Hz (.5/1/2); 500,  1000, 2000, and 4000 Hz
(.5/1/2/4);  and 1000, 2000, and 4000 Hz (1/2/4). The overall level of each noise is adjusted to obtain Al
levels of 0.2, 0.5, and 0.8, and then the SIL is calculated for each of 3 sets of octaves. The data points at an
Al of 0.18 are actual experimental points (50% Fairbanks Rhyme  Scores) from Klumpp and Webster
(1963), i.e., these are the SILs for noises No. 1, 4,10, and 15 in their study.
Hz—average—or 709 to 2530—overall. For the average measure (SIL-type) the center
frequencies and bandwidths were 1135 Hz (3.33  octaves) for an Al of 0.2,  1421 Hz (433
octaves) at 0.5 Al, and  1797 Hz (3.33 octaves)  for  0.8. These values compare very well
indeed to those proposed in this paper of 1000, averaged over the three octaves 500, 1000,
and 2000  Hz; 1428, averaged over the four octaves 500,  1000, 2000, and 40000 Hz; and
2000, averaged over the three  octaves  1000,  2000, and 4000 Hz. Cluff also found  the
SIL-type measure gave standard diviations varying from 0.3  to 0,9 (ave.-0.54)  while  the
standard deviation of the A-weighted levels varied from 1.6 to 3.8 with an average of 2.20.*
     *CIuff,  G. L. (1969), "A  comparison of selected methods of determining speech interference
calculated by the Articulation Index," J. Auditory Res. 9,81-88.
                                           34

-------
     The previous analysis has shown that the octaves chosen to calculate an S1L vary
according to what AI the SIL is trying to estimate. Webster (1964a,  1964b) constructed a
set of contours (see Figure  7)  for predicting AIs or SILs that also showed the increasing
importance of the high speed frequencies for increasing levels of intelligibility (and AI). It is
suggested that weighting networks for sound level meters could be built to predict AI levels
of 0.2 (SI = 70 dB); 0.5 (SI = 60 dB); and 0.8 (SI = 50 dB). A good set of noises on which to
test these hypotheses are the 16 noises of Klumpp and Webster (1963).
     Calculations made on Klumpp  and Webster's 16 noises equated  in level at AIs of 0.2,
0.5, and 0.8, comparing 4 sound level weighting networks A, SI-70, Sl-60, and SI-50, and 3
ways of calculating SIL, namely using the 3 octaves 500/1000/2000, the 4 octaves from 500
to 4000 and the 3 octaves from 1000 to 4000 are shown in Table 2. The results generally
confirm everything  that has just been stated, namely, that at a level of intelligibility corre-
sponding to (1) 0.2, the Sl-70 and the 500 to 2000 SIL are the best (lowest a and R) (2) 0.5
and 0.8, the SI-60 and the 500 to 4000 SIL are the best, and (3) 0.8, the SI-50 and the 1000
to 4000  SIL are good. A-weighting  appears slightly inferior to the  proposed SI-60 and any
SIL that included 500 Hz.
     If  the  manufacturers  of  sound level meters are  seriously considering weighting net-
works other than A, B, and C, an SI-60 should be considered. It is appreciably better than A
for predicting speech intelligibility at all AI levels.
     I have shown how the choice of frequencies for SILs or weighting networks is depend-
ent on the level of intelligibility to  be specified. Now we get back to intelligibility testing.
What tests should be used for various levels of AI?
     Efficiency factors in test  design dictate that the functional relationship between the
dependent and independent variable should be steep and linear in the critical testing region.
Therefore,  consideration should  be given  to  using  different  language tests for different
communication effectiveness areas.  For example, for marginal  conditions, AI = 0.2, closed
set rhyme words  (Fairbanks, 1958;  House et al., 1965; Kreul et al.,  1968; Griffiths, 1967;
Clarke,  1965), which  yield scores  of about 50%, would  make very  efficient tests. If a
listening situation—room or communications equipment—required adequate  intelligibility,
i.e., an AI of 0.35,  then open-set, 1000-word PB tests would yield scores close to 50% and
therefore be efficient  in test  design, although inefficient  in  terms  of crew  training, test
scoring, etc. The use of closed response-set rhyme words would be on the border line of
acceptability since the expected scores would be around 75%.
     At AI levels around 0.8, no intelligibility  test is inherently difficult enough to be an
efficient  test. Even  1,000 nonsense syllables have an intelligibility of greater than 90% at AI
levels of 0.8. To discriminate between listening conditions—communication systems, com-
ponents, etc.—at AI levels of 0.8 requires something more than a simple  intelligibility test.
Reaction times, quality judgments, scores on secondary  tests, or interference tasks, such as
competing messages, have been used or suggested. We will have time to discuss only one of
these promising approaches, namely the competing message paradigm. Tillman, Carhart, and
Olsen (1970) show  the decrement in performance on a competing message task due merely
to adding the equivalent of a hearing aid between the sound field and the listener's ears. The
listener's task was to recognize  in turn one  of 50 phonetically  balanced (PB) words from a
loudspeaker in one corner of a  room while  competing sentences at levels 6  or 18 dB down
were coming from  a loudspeaker in the  other corner  ahead of  the  listener,  i.e., the 2

                                         35

-------
     no
                           OCTAVE PASS  BANDS (N CYCLES PER SECOND
             4ft   —   tO  —   ICO  —   996  —  710   —  1400 —  MOO — MOO  —  lltOO
     40
                      100                        1000
                        FREQUENCY IN CYCLES PER SECOND
10000
Figure 7. Noise rating contours for estimating SILs based on different averaging octaves. Use SI-70 for
estimating the .5/1/2 SIL. = Al of 0.2; the SI-60 for the .5/1/2/4 SIL, = Al of 0.5; and the SI-50 for the
1/2/4 SIL, = Al of 0.8. The inverse of these contours could be used as frequency weighting networks in
sound  level meters to measure the speech interfering aspects of noises. The SI-60 is the best compromise
contour.
                                         36

-------
                                                                   Table 2
        Cells  on the diagonal show  the  Mean (M), Standaid Deviation (a), and Range (R) for  the designated paiameteis for AI levels of
        (from left to right) 0.2, 0.5, and 0.8. Cells above the diagonal show the physical statistics for the  16 Klumpp and Webster (1963) noises
        (AI is not a factor). The numbers in the cells (above the diagonal) represent the M, a, and R of the difference distribution between the
        parameters listed at the left and across the top.
                                     Weighting Networks
                A-Wtng.
SI-70
SI-60
SI-50
   3L
.5/1/2
  SILs

    4
.5/1/2/4
A-Wtng
SI-70
SI-60
Sl-50

3L
.5/1/2
.5/1/2/4
3H
1/2/4
 3H
1/2/4
83.1 72.3 62.6
3.7 2.8 3.2
11.3 8.3 9.7

5.6
2.6
10.0
77.4 66.6 56.9
2.4 3.1 4.0
9.0 9.0 11.5

6.6
2.1
10.0
1.0
2.8
9.5
76.5 65.7 55.8
3.0 1.6 2.3
11.0 7.6 10.6

3.2
4.4
14.7
2.4
5.8
18.9
3.3
3.2
9.7
79.7 68.9 59.2
5.6 3.1 2.6 x
16.5 9.8 7.8

9.8
3.3
9.3
4.2
1.9
8.2
3.2
1.9
6.8
6.6
4.6
14.6
73.0 62.3 52.5
1.5 1.6 2.7
6.3 6.4 10.4

10-6
3.2
11.5
5.0
3.5
12.7
4.1
1.5
6.2
7.4
2.7
8.2
0.8
1.7
6.9
72.2 61.9 51.6
3.0 0.7 1.4
9.4 4.5 6.6

11.6
4.8
16.4
6.0
5.8
19.7
5.0
3.2
11.5
8.3
0.9
5.2
1.8
4.2
13.0
1.0
2.2
8.1
71.2 60.4 50.6
5.4 3.2 2.7
17.4 10.9 9.4
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           tr
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           £7
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R
                                                                                                           M
                                                                                                           a
                                                                                                           R

-------
loudspeakers were 45° to the left and to the right of the listener's nose. Unaided listening
was (1) binaural; (2) monaural direct, in which the speech was on the side of the listening
ear, and the other ear was occluded with a muff; and monaural indirect, in which the speech
was on the side of the occluded ear. Aided listening used an artificial head in the sound field
with two  hearing aids and connections via amplifiers and calibrated attenuators to insert
earphones in the ears of the remote listener. Again three conditions were tested—binaural,
monaural direct, and monaural indirect.
     Four groups  of 12 subjects each were tested  including (1) those classified audio-
logically as normal (average age 22); (2) with moderate hearing losses diagnosed as conduc-
tive (average age 42); (3) sensorineural (average age  51); and (4) presbyacusic (average age
70). The groups will be indicated by N, C, S and P, respectively.
     All listening  was at  a  level 30  dB above  the  threshold for spondee words, 30 dB
Sensation  Level (SL), under each of the 6 conditions. Figure 8 shows the results, which can
be summarized as follows: Compared  to an earlier reference group of 20 normal hearing
subjects, on  the  PB word/sentence competition task (Northwestern University Auditory
Test  2, Carhart et al., 1963) the N and C groups sitting in the sound field (unaided) heard
essentially at reference level; the  S and  P groups required, on average, a  14 dB better
word-to-sentence differential than the N and C groups in the sound  field; the N group
required about the same increase in word-to-sentence  differential when a hearing aid was
interposed between them and the sound field; the C group required an even greater increase
for the aided conditions, about 18 dB more; and the S and P groups, who required a 14 dB
word-to-sentence (W/S) improvement  in the unaided case, required further improvements
which increased as the basic word-to-sentence (W/S) differential increased. Restated, the S
and P groups are worse off than the N and C groups in listening to competing speech signals
30 dB above their speech threshold, whether listening with or without hearing aids.
     These results show both a hearing deficiency penalty and an equipment-imposed
penalty  when listeners are placed in competing message listening conditions. This is bad
news for people  incurring noise-induced hearing losses which are generally sensorineural in
nature. No only do  they have more difficulty than their normal-hearing or conductively-
deafened friends in cocktail  party environments, but they cannot look forward to a hearing
aid to help equalize their relative disadvantage.
     The last point  I want  to make  concerns listening to  speech in noise while wearing
earplugs or muffs. It  has long been established that in noise levels greater than 90 dB, speech
is heard better when wearing hearing protection. This early work of Kryter (1946) was for
young normal hearing subjects. However, there  is at least one study by Frohlich (1970)
which shows that unlike young normal hearing males,  senior aviators with high-frequency
sensorineural losses do not  discriminate digits better in noise levels above 100 dB when
wearing good noise-attenuating ear muffs. He shows that this could be expected by plotting
hearing-level  and hearing-level-under-muff for senior aviators on the speech area and noise
masking area. This procedure shows that the muff cuts out a region of speech  frequencies
where the speech is well above the masking noise. It seems safe to say that acoustic-trauma
listeners have more difficulty than  normals in discriminating speech in quiet, in noise, and
particularly  in competing  message  situations. They do  not get the full benefit  enjoyed by
normal listeners of increased intelligibility in high noises by wearing hearing protectors, and
they cannot expect a hearing aid to help them untangle competing messages.

                                         38

-------
          -12   -60     6     12    18    24
QUIET
      EFFECTIVE  WORD-TO-SENTENCE  RATIO  IN  dB
Figure 8. Percentage of 50PB words correct in the presence of competing sentences at various word-to-
sentence differentials. The parameters are: REF-20 reference normal hearing listeners; IM/C = normal or
conductive pathology experimental listeners; P/S = presbyacusic or sensorineural pathology experimental
listeners. Aided refers to listening via hearing-aid circuitry. Unaided refers to listening normally in a sound
treated room. From Tillman. Carhart, and Olsen (1970).
    In summary, I have tried to tell you in this presentation that the octaves chosen to
calculate the SIL and/or the weighting networks that could be built into a sound level meter
to measure the interference of noise with speech vary as a function of what level of speech
communication you desire to design for. Correspondingly, the tests you use to evaluate a
listener or a system vary in the same manner, sentence intelligibility tests being best for a

                                       39

-------
basically  bad system, word or nonsense syllable tests for a good system, and competing
message tests or judgment tests for an excellent system. Persons with noise-induced hearing
loss cannot hear as well as normals when wearing plugs or muffs in moderate to high levels
of noise nor can they by wearing a hearing aid unscramble competing messages ^at a cocktail
party) as well as normals.

                                    References

 1. House,  A.S., C.E. Williams, M.H.L. Hecker, and K.D. Kryter (1965),  "Articulation
    testing methods:  Consonantal differentiation with  a closed-response set,"  J. Acous.
    Soc. Amer. 37, 158-166, also USAF EST-TDR 63^03.

 2.  Clarke, F.R. (1965), "Technique for evaluation of speech systems," Final Report of
    Stanford Research Institute Project 5090 on U. S. Army Electronics Laboratory Con-
    tract DA 28-043 AMC-00227(E), August 1965.

 3.  Kreul, E.J. Nixon, J.C., Kryter, K.D. Bell, D.W. Lang, J.S. and Shubert, E.D. (1968),
    "A Proposed clinical Test of Speech Discrimination," J. Speech and Hearing Res. 11,
    536-552.

 4.  Fairbanks, G. (1958), "Test of phonemic differentiation: The rhyme test," J. Acoust.
    Soc. Amer. 30, 596-600.

 5.  Egan, J.P. (1948), "Articulation testing methods," Laryngoscope 58, 995-991.

 6.  Webster, J.C.  (1972), "Compendium  of speech testing  material and  typical noise
    spectra to be used in evaluating communication equipment," NELC Tech Doc. 191 of
    13 Sept.

 7.  Clarke, F.R., J.C. Nixon,  and S.E. Stuntz, (1965), "Technique for evaluation of speech
    systems," Stanford  Research Institute  Semi-annual Report of SRI Project 5090 for
    U. S.  Army Electronics Laboratory, Contract DA 28-043 AMC-00227(E) AD 462836.

 8.  Dreher, J.J. and J.J. O'Neill (1957), "Effects of ambient noise on speaker intelligibility
    for words and phrases," J. Acoust. Soc. Amer., 29, 1320-1327.

 9.  Pickett, J.M. (1956), "Effects of vocal force on the intelligibility of speech sounds," J.
    Acoust. Soc. Amer., 28, 902-905.

10.  French,  N.E.,  and  J.C.  Steinberg (1947), "Factors governing  the  intelligibility of
    speech sounds," J. Acoust. Soc. Amer.  19, 90-119.

11.  Beranek, L.L.  (1947a), "The design of speech communication systems." Proc. Inst.
    Radio Engrs., 35, 880-890.

                                        40

-------
12.  Kryter, K.D., Methods for the calculation and use of the articulation index. J. Acoust.
    Soc. Amer., 34, 1689-1697 (1962a).

13.  Kryter,  K.D.,  Validation of  the articulation  index.  J. Acoust. Soc. Amer., 34
    1698-1702 (1962b).

14.  Webster, J.C. (1969), "Effects of noise on speech intelligibility," P. 49-73 in National
    Conference on  Noise as  a Public Hazard. Proceedings, 13-14 June 1968, The American
    Speech and Hearing Association (ASHA) Reports 4.

15.  Flanagan, J., and H. Levitt  (1969), "Speech  interference from community noise", p
    167-174 in  National Conference  on  Noise as a Public Health Hazard, Proceedings,
    13-14 June 1968, The American Speech and Hearing Association, ASHA reports 4.

16.  Beranek, L.L. (1947b),  "Airplane quieting II. Specification of acceptable noise levels."
    Trans Amer. Soc. Mech.  Engrs., 69, 97-100.

17.  Botsford, J.H,  (1969), "Using sound  levels to gauge human response to noise," Snd.
    and Vib. 3(10) 16-28.

18.  Webster,  J.C.  (1965),  "Speech communications as limited  by ambient  noise," J.
    Acoust. Soc. Amer., 37, 692-699.

19.  Cavanaugh, W.J., W.R. Farrell, P.W. Hirtle, and  B.C. Walters (1962), "Speech Privacy
    in Buildings", J. Acoust. Soc. Amer. 34,475-492.

20.  Webster,  J.C.  (1946a), "Generalizes  speech  interference contours," J.  Speech and
    Hearing Research, 7, 133-140.

21.  Webster, J.C. (1964b). "Relations between speech-interference contours and idealized
     articulation index contours," J. Acoust. Soc. Amer., 36, 1662-1669.

22.  Klumpp, R.G.  and J.C. Webster (1963), "Physical Measurements of equally speech
     interfering Navy noise," J. Acous. Soc. Amer. 35, 1328-1338.

23.  Griffiths, J.D.  (1967),  "Rhyming minimal contrasts: A  simplified diagnostic articu-
     lation test", J. Acoust. Soc Amer. 42,  236-241.

24.  Tillman, T.W.,  R. Carhart and W.O.  Olson (1970), "Hearing  aid efficiency in a com-
     peting speech situation", J. Speech and Hearing Research 13, 789-811.

25.  Carhart,  R., T.W. Tillman  and L.  Wilber (1963), "A test  for speech discrimination
     composed of CNC monosyllabic words," Perceptual and Motor Skills 16:680.
                                        41

-------
26.  Kryter, K.D. (1946), "Effects of ear protective devices on the intelligibility of speech
    in noise," J. Acoust. Soc. Amer. 18, 413-417.

27.  Frohlich, B. (1970), "The effects of ear defenders on speech perception in military
    transport aircraft," North Atlantic Treaty Organization (NATO) Advisory Group for
    Aerospace Research and Development (AGARD) Advisory Report 19.

28.  Webster, J.C. and C.R. Allen (1972), "Speech intelligibility in Naval Aircraft radios,"
    NELCTR 1830 of 2 Aug.
                                         42

-------
                        RECEPTION OF DISTORTED SPEECH

                         Jerry V. Tobias and F. Michael Irons
                           Aviation Psychology Laboratory
                           FAA Civil Aeromedical Institute
                           Oklahoma City, Oklahoma 73125
     Noise has direct physiological, psychological, and social consequences. It also has indirect
consequences that are associated with public health and that certainly are not limited to
damage to the auditory physiology, to the psyche, or to the community's acceptance of loud
sounds. Consider the effect  of a bit too much noise on an airline pilot's reception  of an
air-traffic-control message: the physical  well-being of hundreds of passengers and of un-
known numbers of people on the ground  can be changed by the inaccurate understanding of
an instruction. A missed warning in a  steel mill can produce frightening-even  deadly-
effects on personnel; the physiological results are not confined to the temporal bone.
     We  know ways to  measure speech  interference; and we know something about the
acoustic  factors that determine how well a  listener will  be able to understand masked
speech. Dr. Webster covered  those things in detail in the previous paper. However, there is
another kind of influence on speech intelligibility that all of us have had experience  with,
but that no one has measured before: people can leam  to manipulate signal-to-noise ratios
mentally as well as acoustically. The procedure involves no poltergeists, no telekinesis, no
meditative  or metaphysical  manipulations. It only requires that the  listener's brain be
adequately  exposed to  the masked signal. This  exposure allows the signal-selection
mechanisms to search out the best methods for processing the speech-plus-noise, and, after a
time, produces  greatly  improved intelligibility.  One of the questions  that has not been
answered before is how much time it takes to learn that new analyzing process. Now there
are experiments that suggest that it takes  less time than you might have believed.
     Here is a practical illustration of what  the phenomenon is. People often whistle or hum
or sing while they work. Many nod their heads or tap their fingers to keep time with their
music. In offices, you can sometimes see three or four people, each tapping out a different
rhythm, oblivious of the tempo being strummed on the next desk. Sometimes, though, in a
noisy work environment, a bizarre variation of this behavior appears: a group of employees
who could  not possibly hear each other's  humming because of nearby loud machinery all
move in time to the same invisible drummer. The first time we saw such a thing, we asked
one of the  workers what he  was waving  at. He said, "It's the music," and we pretended to
understand. Of course, when  we went into a storeroom a little distance from the machinery,
there was music. Whether for morale or for entertainment or for setting a working pace, the
company pumped recorded  music into  the factory. The workers heard it even though a
visitor could not make it out  above the din of the equipment.
     Similar stories can  be picked up from anyone who measures noise. All the anecdotes
lead  to the same conclusion:  the ability to  hear masked signals that are inaudible or unintel-
ligible to the untrained or inexperienced observer can be improved by listening practice. The

                                        43

-------
anecdotal evidence has been overwhelming. The laboratory evidence has been nonexistent.
Now there are  a few experiments that were designed to quantify this  process that we all
know to exist.

Signals.

     Masking and distortion are similar kinds of operations; each  covers up part of the
otherwise-available information with extraneous matter to make the signal "noisy." They
both decrease the intelligibility of speech signals. The effect of an intelligibility-decreasing
distortion can be nearly indistinguishable from that of masking.  For example, in a masked-
speech experiment, Kryter (1946) showed that the measured intelligibility of highly reverb-
erant speech  that'is masked  by enough noise to raise it to a level 60 dB above threshold is
comparable to the intelligibility of non-reverberant speech raised 80 dB. In that study, the
reverberation had a masking effect similar to the effect of an extra 20 dB of noise.
     A learning process permits man to overcome the change that the  noisiness produces.
Practicing listening to the speech without the noise (or without the distortion), however,
seems not  to help intelligibility much. Exposure to the noise alone or to the distortion of
nonsemantic  signals seems not  to help. But practice  listening to the combination of speech
and its intelligibility-destroying noise leads to rapid improvement.  The available data cover
studies of both masked and distorted speech; the results from experiments with one kind of
signal are similar to the results from experiments with the other..
     Subjects were all taught to shadow (Cherry, 1953) while listening  to recorded speech.
In shadowing, the listener immediately repeats every word he hears, even as he is hearing
new  material. Although the  idea may sound difficult, subjects are quite  adept at learning it,
and  intelligibility-test scores measured by  shadowing are similar to scores earned in other
kinds of tests. Indeed, if anything,  shadowing is a particularly  sensitive measuring  tool
(Pierce and Silbiger,  1972). Most subjects reach 95-100% intelligibility  scores on clear,
continuous speech within a  few minutes. Our subjects were trained in shadowing until they
had scored higher than 95% in five successive one-minute intervals.
     The speech used for the speech-learning experiments was not the same as that used to
teach shadowing. The experimental speech is a series of easy-to-understand  120-word pas-
sages, read by a male talker who monitored himself during the recording session in order to
insure a constant speaking level. Later, slight variations in level were made from passage to
passage in order to  insure that  all would be equally  intelligible in  a  simple masking experi-
ment.  Each  passage  is  approximately 50 seconds long, with a 10-second pause  between
passages. A total of  54 such one-minute segments  was available, and  the segments were
spliced together in many randomized orders.
     For some subjects, the passages  were masked  with a wide-band Gaussian noise; for
others, the speech was infinitely peak clipped; for still others, the signal became a pulse train
whose  spacing was determined  by the  line-crossings of the speech  wave; and finally, in one
series of tests,  the speech became a  carrier  that was  amplitude-modulated by a band of
noise.  Subjects selected their own signal levels; for a 1000-Hz  tone adjusted to the same
peak level as the speech, the  sound-pressure  level  was 75  ± 4 dB, which was near the
optimum  choice  according to  preliminary tests of the relation between level and intelligi-
bility.  Figure 1  illustrates the kinds  of distorted signals that were used.  In the  masked

                                         44

-------
                                     SPEECH SIGNALS
                           LJ
                           Q
                           CL
                           2
                                      Clear-Undistorted
                           UJ
                           Q
                           ID
                                    Infinite peak-clipping
                          UJ
                          Q
                          D
                           Q.
                           S
                           <
                           i
                          T
                          u
                          Q
                                       Pulse modulated
                                       Noise modulated
                                         TIME
                           Figure 1.  Waveforms of test signals used.
condition,  the speech wave is  simply added to the  noise. In the modulated condition,
though, a multiplication transform is used, with the effect that each partial in the original
instantaneous speech spectrum is replaced by a steep-skirted band of noise, 1200 Hz wide,
centered on the partial. In the pulse-modulated procedure, all that is retained of the original
waveform is the time and polarity of axis crossings; infinitely peak-clipped signals look to
have only that same information (Licklider and Pollack, 1948), but they are generally much
more intelligible (Ainsworth.  1967), even when experienced listeners adjust  the levels of
                                          45

-------
 both types for maximum intelligibility. Clipped speech sounds harsh; pulsed speech sounds
 harsher.  Noise-modulated  speech sounds very noisy, but is generally reported to be much
 clearer than one would expect with "that much noise" present.
     The masking level and the modulator bandwidth were  selected  to produce approxi-
 mately .the same maxjmum intelligibility score (80% correct)  for highly trained listeners as
 unmasked pulsed speech does. Clipped speech is a bit easier to understand, and maxima near
 90% are common.

Subjects.

     Each segment of these studies used six university students as listeners. All subjects had
normal hearing, and none had any previous experience  with this kind of task. A total of 13
series of experiments used 78 subjects. Everyone was  trained in shadowing before being
exposed  to  the distorted or masked signals. Most subjects were then  simply instructed to
shadow whatever they could hear. Several groups, though, received special treatment; some
shadowed for a total of only eight minutes in a 54-passage session; another few shadowed
everything, but were informed that they'would be given  a monetary incentive to do well.

Speech Learning.

     The basic outcome of all these experiments is perfectly predictable: intelligibility starts
at a low level and improves with listening practice up to a plateau value. Figure 2 shows a
learning curve for each of the four kinds of signal. The rates of change are fairly similar from
one condition to another, although ;the  plateau values vary  somewhat. The  immediately
apparent pojnt to note about all of these data is that learning seems to be complete within
 15 or 20 minutes. The auditory  system  makes its  analysis of the signal-plus-noise, deter-
mines how to extract  the maximum information,  and makes whatever modifications are
necessary in order to perform the extraction—and it does all that  in less than half an hour.
The listeners are  probably not especially conscious  of what they are doing in order to get
this analytical processing under way; most of them report no special effort to get better, and
generally they have little recollection of how well they performed.
     Although the curves are  similar in shape, it is inappropriate to try to draw inferences
and  conclusions  about the speech-learning mechanism from that fact. Learning curves
simply look  alike. That does not necessarily demonstrate anything about similarities or
differences in the  analysis of modulated and pulsed speech signals.
     The curves  that represent  what happens in  this learning mechanism do have one
particularly  fascinating .aspect, though (Figure 3). They show that subjects returning after
one or two weeks away from the task start the first couple of passages with scores slightly
lower than  their  previous maxima, but  then, almost  immediately, they rise  to a higher
plateau than the one they attained during their first test session. The change from the first
to the second plateau is statistically significant at better than the  .05 level; final scores on
session two are  12 to  15 percentage  units  above those  on session  one, making a  total
improvement that is equivalent to about an 8-dB shift in signal-to-noise ratio. During their
time away  from the laboratory, the subjects had no opportunity  to listen to  the kinds of

                                         46

-------
 LJ
 
-------
oo    iz
IUU
90
80
70
*~* 60
o^ 50
*~* 40
S 30
or 20
R 10
o
CO
1 st Session
-
MASKED
^^
- f
I
2nd Session





1 1

-
i i I I I


i i

i \j\j
90
80
70
60
50
40
30
20
10

1st Session
-
PULSED
2nd Session
^*
f
- x^ If
/
J
'-
-
_
1 1 1 1 1





I i

0 10 20 30 40 50 0 10 20 30 0 10 20 30 40 50 0 10 20 3
t
-J 100
QQ 90
O 8O
^•^ w \J
-J 70
UJ 60
H 50
— 40
30
20
10



1 st Session
— M/*\f\iiiiv"^rf\
MODULATED
Y
i
-
-
-
i i i i i

2nd Session
^^^•••••••••••v
~
/"





i i
100
90
no
ou
70
60
50
40
30
20
10



1st Session
CLIPPED
Y
-
-
-
-
i i i i i

2nd Session
X^





i i
            0    10  20  30  40  50   0    10   20   30        0   10  20  30   40  50   0   10  20  30
                                               TIME  IN   MINUTES
        Figure 3. Mean learning curves for second experimental session. Note the improvement following a week or more without practice.

-------
     Whatever the solution to that mystery, though, one thing is clear. The brain, once it
has organized itself to determine the transformations  necessary  for  analyzing difficult
speech messages, continues to refine the analyzing process. These changes and refinements
are the kinds that allow the listener to generalize or transfer the techniques of decoding one
sort of distortion to the interpretation of other sorts.

Transfer.

     In tests of the transfer of speech learning, subjects were trained during the entire first
session with material that had been treated with one variety of distortion. For the first half
hour of the second session, two weeks later, they continued with that same distortion; by
the end of the 30 minutes, it was certain that they had reached their new higher asymptotic
intelligibility score. Then,  for the last half hour of that session, they were given material that
had been subjected to a  different distortion.  For example, subjects who spent all of their
hour and a half of practice time on modulated speech were tested on pulsed speech (Figure
4). Within three  to five minutes,  they reached plateaus that were at least as high as those
reached  by similar subjects  who  had  listened to nothing but pulsed speech during both
sessions. Transfer is equally good in the opposite direction.
     Although pulsed  speech lacks most of the spectral information of the original wave-
form, and modulated speech lacks most of the temporal information, the transfer of ability
from one to the other seems complete. The suggestion is strong that human observers, once
they have learned how to listen to difficult speech, can successfully understand almost any
form of it (for another example, see Beadle,  1970). Perhaps that idea helps to account for
the fact that  some people are able to  understand  English spoken with many kinds of
dialects,  but that others  cannot get much intelligibility out  of what is said to them by
talkers with just moderately variant speech.
     A practical result of  this finding is that a person probably does not need to be trained
to listen to distorted or masked signals that are  of the precise form that will occur in his
work.  Once he has mastered some types of speech learning, he will be able to assimilate
others rapidly. Should we decide  to make a set of recordings to train aviators to listen to
radio transmissions, those records need not  contain precisely the same  kinds of signal
degeneration that actually arise in the  cockpit; anything similar—or maybe even dissimilar-
ought to do as well.

Passive Listening.

     Must such training  involve  continuous  speech-related activity on the part  of the
listener?  To find out, we ran an experiment  on  the learning that accrues to subjects who
hear the  distorted signals,  but who do not have to shadow them (Figure 5). If the shadowing
activity is contributing to the  learning, then levels of performance ought to be higher for
those who are thus employed during their listening. Twelve subjects were used to test this
idea. Before their first exposure to the  distorted speech, they were trained in shadowing
clear speech, just as all the'subjects had been. Then they were asked to shadow the first pah",
the last pair, and two equally spaced intermediate pairs of distorted passages. For the rest of
the time—46 passages—they were  silent. When these passive listeners' responses are  com-

                                         49

-------
    lOOr-
LU
    20
     I 0
                      I st  Session
                       MODULATED
                                    2nd  Session
                           MODULATED
                                                                    PULSED
                                          r
               i
i
_L
I
I
I
I
I
I
        0    10   20   30   40   50  0     10   20   30   40   50

                              TIME    IN    MINUTES

 Figure 4.  Mean teaming curves for subjects who received all of their listening experience on modulated
 speech. When tested on pulsed speech, their scores were almost immediately at the maximum expected for
 subjects experienced with pulsed-speech listening.
pared with the responses of active, continuously shadowing subjects, two differences are
apparent  from the  data and from the  subjects' reports.  First, the passive group's final
first-session scores are consistently higher than the active group's (second-session curves are
alike). Second, active shadowers have almost no retention of the material that they listen to,
but passive listeners remember  a great deal of what they hear. One interpretation of this
finding is that our active listeners are giving us scores that do not represent improvements in
                                      50

-------
   100

o^  90

to  so
LJ
g  70
u
CO  60
    50

    40

    30

    20

    10
OQ
_J
LJ
f-
2
                              1st Session
                        PASSIVE  PARTICIPATION
                        GROUP A AND  B
        2nd  Session
— — GROUP A  ACTIVE

— GROUP B  PASSIVE
                         20       30      40       50        0
                                         TIME  IN   MINUTES
            20
                                                                                        30
40
            Figure 5.  Mean learning curves for passively listening subjects. In the second session, half were asked to
            shadow.

-------
shadowing technique, but rather real improvements in the ability to understand difficult
speech  signals. The higher  passive scores might be interpreted to mean  that shadowing
somehow interferes with speech learning, and we cannot refute that possibility. However, it
is possible too that the  listener who can understand and retain what he listens to is better
motivated to learn. That possibility can be tested.
Motivation.

     Two groups of subjects were tested for motivation effects-one group on masked and
one  on pulsed speech—but, unlike previous listeners, these were given monetary incentives
to do well. After a subject had worked through the first three passages, the three scores were
averaged, and he was told that for each 1% by which the 54th passage was better than that
beginning average, he would be paid a bonus of $.05. Also, in order to keep him working at
a high level during the entire test  session, he  got an  additional $.10 for each passage on
which the score was above 90%.
     The results (Figure 6) are similar to those for the passive listeners: curves continue to
rise for a longer period of time during the first session, and, within the first hour, they reach
values that are comparable to second-session plateaus. This relation between passive-listening
results and motivated4istening results certainly suggests that the passive subject continues to
improve because he is more interested in the  task than the active subject is. He is able to
relax a bit, and he can actually attend to what the talker is saying (remember that his
retention is better than the active subject's).
     Second-session scores for these subjects are indistinguishable from those of any other
subjects. Changes in ultimate peak scores, if they occur as a result of monetary reward, are
not large enough for us to measure with these techniques.


Masked Speech.

     Experiments with masked speech at a -3  dB signal-to-noise ratio show one kind of
quantitative difference from the other experiments: first-session subjects reach  two quite
different kinds of asymptotes, apparently  as a function  of their  earliest scores. Listeners
who do well in the first few minutes are like most listeners; they improve rapidly to plateaus
of 80% or so. But those  who  start with intelligibility scores  of approximately 10% reach
peaks in the neighborhood of only 50 or 55% (Figure 7). The intention  had been to set a
signal-to^ioise ratio that would give a first-minute intelligibility score of 20 or 30%, but the
selection was not right for these subjects; their initial scores actually ranged from 5 to 29%.
The  low-plateau subjects show greater variability  than might be  accounted for by  the rela-
tively unrestricted range in which  they were working. Their learning curves rise  compara-
tively slowly, sometimes taking 35 or 40 minutes to get to the asymptotic value. The curves
are unlike any others we have seen.
     An explanation may lie in an  evaluation  of the learning experience that each group
receives: the people who start off well get exposed to  large numbers of correctly heard
words;  those  who start poorly  receive relatively little  information that helps  them in

                                         52

-------
   100 r
     90  -

-------
     100

o^    90

cn    80
LU
ct:    70
o
   o
   en

   H
   .j
   DO
   C)
   _J
   _J
   LU
      60

      50

      40

      30

      20

      10
                                       I st Session
                       10         20        30         40
                            TIME    IN   MINUTES
                                                                 50
             Figure 7.  Individual teaming curves for subjects who listened to masked speech.
magnify that success into higher scores. The listener who starts low may spend most of his
training time struggling to hear anything at all. His decoder never gets enough samples of the
difficult sounds to permit the formation of useful hypotheses about how to listen.

Retention.

    Most subjects were retested in a third session one month following the second (six
weeks after the first). Third-session curves  and scores are similar to second-session curves
and scores. One group was retested after a year had passed with no known intervening
practice. Their latest performances are similar to their second sessions, too.
                                        54

-------
Non-Normal Hearing.

     Normal-hearing subjects learn to understand badly mangled speech after a short period
of practice. There is no evidence that people with pathological hearing do as well, but it is
certainly possible—even reasonable-to conclude that they do.
     The plateaus of subjects  in their second sessions, no matter what the conditions of
their  training  (except  those who are  able to receive only small  amounts of information
during the first session), and no matter what kinds of signals they were trained with, are all
improved, and all look similar. That fact may  partially explain why hearing aid users are
reported to do much better at speech discrimination after they have used an aid for a week
than they did  when the instrument was first tried. The early listening presents them with a
kind of sound.that is somehow different to them; they have to learn about it before they
can get maximum sense from it. If this learning process works with some kinds of pathologi-
cal ears as well as it does with normals, we might also expect to find that, for these ears,
experience with any hearing aid will transfer to any other.
     The hard-of-hearing person may have a greater problem learning to understand dis-
torted speech  than  the normal-hearing listener, though, for the very reason that he cannot
hear enough of the signal to work out an appropriate analysis strategy. He could be like the
low-plateau  subjects in the masked-speech experiments. However, usually, the overall sound-
pressure level of his  work environment will be high enough to overcome much of the
problem caused by an elevated threshold, so he can learn as well as  his colleagues.  This
likelihood leads to  the interesting possibility  that  the results of some audiometric tests of
the ability to understand  speech that  is immersed in noise may be more a function of
learning than of hearing.


Training.

     Even two minutes of listening can improve the ability to understand a talker (Peters,
1955). Six to  eight  hours may be needed to teach people to understand speech sounds  that
are transformed by a spectral  inversion (Beadle,  1970), and even then, it takes longer to
learn from an  unfamiliar talker. But for optimum training for the reception of non-inverted
speech, about an hour is needed.
     How should the  time  be  spent?  If you want to improve your reception of distorted
speech, it is not enough to  listen to the right  kind of interfering noise.  It probably will not
help to be exposed  to nonspeech sounds that are subjected to the same sorts of distortions
that affect the speech; the analyzing activities of the brain are quite different for speech and
for non-speech signals (Stevens and House, 1972). Student pilots take far longer than half an
hour  of flight time  to learn to  understand air-traffic-control communications; factory
workers do not begin to understand what is said to them in noise until days of listening have
passed, not minutes. Both groups commonly hear speech-plus-noise for only short moments
at a time, and then return  to listening to noise alone. The requirement for rapid learning,
though, is that the listener be able to hear the combination of signal and noise at a signal-to-
noise ratio that is high enough to permit him some success in interpreting the messages  that
are being transmitted.  If his motivation  to learn is high (or heightened), he can reach his

                                         55

-------
maximum  capacity in an hour. Hell probably learn still faster if his training is done by a
talker whose voice is familiar to him.* And when he has found the knack of how to listen,
he will keep it for a long, long time.
     Once, after a static-filled, phase-distorted, narrow-band,  whistling, short-wave broad-
cast of a concert, Sibelius is reputed to have pointed at the radio receiver and said, "I can't
understand how anyone but a musician could enjoy listening to that thing." In the  same
way, one who is not  trained to listen to speech will not enjoy it. Indeed, in many circum-
stances, he may  not even to able to hear it.

*Schubert and Parker (1955) reported in a paper on a speech intelligibility study, "A puzzling phenomenon
occurs with three of the subjects, who were wives of the talker. In each  case the wife exhibits only  a very
slight dip in intelligibility, if any,	when her husband is the talker, but shows about the average dip for
either of the other two speakers. This obviously falls beyond what has previously been considered the
boundary of auditory theory and the authors, who were two of the talkers, are relieved of the risk of discussing
it further." So are Tobias and Irons.
                                      References

Ainsworth,  W.  A.,  Relative intelligibility of different transforms of clipped  speech. /.
     acoust.  Soc.  Amer.,  41 1272-76 (1967).
Beadle,  K.  R., The  effects  of  spectral  inversion  on  the  perception  of  place of
     articulation.  Doctoral dissertation,  Stanford University (1970).
Buxton, C.E.,  The  status  of research  in  reminiscence. Psychol.  Bull,  40, 313-350
     (1943).
Cherry,  E.  C.,  Some  experiments  on the  recognition  of  speech,  with one and  with
     two ears. J.  acoust.  Soc.  Amer.,  25, 975-979 (1953).
Kryter,  K.   D.,  Effects  of  ear  protective  devices on the intelligibility of  speech in
     noise. J. acoust. Soc. Amer.,  18, 413-417 (1946).
Licklider, J. C. R.,  and Pollack,  I., Effects of differentiation,  integration,  and infinite
     peak  clipping   upon  the  intelligibility  of  speech.  /.  acoust.  Soc.  Amer.,  20,
     42-51  (1948).
Peters,  R. W., The effect of length of exposure to  speaker's voice upon listener reception.
     Joint Project Report No. 44, U.S. Naval School of Aviation Medicine (1955).
Pierce,  L.,   and Silbiger, H.  R.,  Use  of  shadowing in  speech  quality  evaluation. J.
     acoust.  Soc.  Amer.,  51, 121  (1972).
Schubert,  E.  D., and  Parker,  C.  D.,  Addition  to  Cherry's findings on  switching
     speech  between the  two  ears. /.  acoust.  Soc.  Amer.,  27,  792-794 (1955).
Stevens,  K.  N.,  and  House, A.  S.,   Speech  perception.  In  Tobias,  J.  V.  (Ed.),
     Foundations of Modem  Auditory  Theory.   Volume II.  New York:  Academic
     Press  (1972).

                                          56

-------
                           HEARING LOSS AND SPEECH
                            INTELLIGIBILITY IN NOISE

                                  Jerzy J. Kuzniarz
                             Oto laryngology Department
                              Silesian Medical Academy
                                   Francuska 20
                              40-027 Katowice, Poland

     While testing patients with sensorineural hearing loss, I have frequently noticed their
complaints about difficulties with understanding speech in the presence of some background
noise. Similar observations were reported  by many authors during the last 10 years (Harris,
1960, 1963; Kryter, 1963; Watson, 1965; Robin, 1967; Tonkin,  1967; Niemeyer, 1967;
Groen, 1969), recently by Carhart et al. (1970), Tillman et al. (1970) and Lipscomb (1972),
so it has become a fairly well established  clinical fact. For simplicity, that symptom will be
called in the following paragraphs a "noise  distractability (ND) phenomenon" in sensori-
neural deafness. The purpose of my presentation is to  show some results of my studies on
that phenomenon in noise-induced hearing loss (NIHL), especially concerning the frequency
area which influences speech intelligibility in noise.
Tests procedure and results.

Experiment I.

     An  NIHL was approximated by  low-pass filtering  (Fig. 1) of speech tests:  mono-
syllables, PB words and sentences (each filtered test recorded separately).
     The intelligibility of each filtered test was examined on 30 normal listeners. The tests
were presented binaurally at an intensity of 65 dB SPL via Pedersen earphones linked with
"Y"-type connection  (no stereophonic effect) with a tape recorder and an audiometer
(Peters SPD 2). The intelligibility was tested both in quiet and in presence of two kinds of
noise: white noise (Fig. 2) and a low-frequency one (Fig. 3).
     The noises were generated by the audiometer and mixed with the speech materials at
different S/N ratios.

Results.

     In quiet, sentences were fully understood when the speech materials consisted of fre-
quencies up to 1000 Hz; monosyllables were understood at a 90% level when the upper
cutoff frequency was 2000 Hz. It is concluded that frequencies up to 2000 Hz are entirely
sufficient for understanding Polish in quiet.
     In background noise, however, the results were quite different. The intelligibility of
filtered speech was markedly reduced in the presence of noise, even at S/N ratios which did
not impair the intelligibility of non-filtered speech (Fig. 4-7). That effect was particularly
evident in the low-frequency noise (Fig. 6 and 7).

                                        57

-------
oo
              0    W5       OJ50       0,50  0.75  W   15   2    3     4     6    6    ^ KHl
           10
           30
          50
                                   Figure 1. Characteristics of low-pass filters used in this study.

-------
MS
                                                                                                 I—1—I	1—
                  0.1
5  6    9   fO  KHZ
                                    Figure 2. Spectrum of white noise, as measured in Pedersen earphones.

-------
Ch
O
                  0.15    a*
0/0
                                    Figure 3. Spectrum of low-frequency noise, measured in Pedersen earphone.

-------
     The difference between the intelligibility of non-filtered speech and that filtered at
3000 Hz was statistically significant for monosyllables in white noise and for both mono-
syllables and sentences in low-frequency noise.
     It can be concluded that frequency bands lying above  2000 Hz and 3000 Hz carry
information which becomes important for speech understanding when lower speech fre-
quencies become masked by low-frequency noise.

Experiment II.

     Thirty selected subjects of the ages 25 - 40 years with NIHL (mean audiograms are
shown  in Fig. 8,10 persons in each group) were tested binaurally with speech audiometry in
a semireverberant room (reverberation time 0.2  sec) in  quiet and in presence of low-fre-
quency noise (Fig. 9).
     Another 30 people, 20-30 years, with normal hearing, were used as a control.
     Arrangement of the apparatus is shown in Fig. 10. Note that, as in Exp. 1, there is no
stereophonic effect. The speech intensity was held constant at  70 dB, monitored with a
VU-meter and controlled with a Bruel and Kjaer Impulse  Sound Level Meter  (mean peak
level).  The noise level was changed in  5-dB steps to obtain S/N ratios from + 15 dB to -15
dB.
Results.

     All persons with NIHL were evidently handicapped even in presence of a mild intensity
of noise (S/N = +10 and +5 dB) that did not disturb normal listeners (Fig. 11). This effect
was seen even in group III, whose members have normal hearing thresholds at 2000 Hz and
good speech intelligibility in quiet. The curves closely resemble those obtained with filtered
speech.
     The results of both experiments may be summarized in three points:
     1.  Persons with NIHL have difficulties in understanding speech in presence of back-
         ground noise (ND phenomenon) even at S/N ratios close to those of everyday
         conditions (S/N = +10 or +5 dB), which do not disturb normal listeners.
     2.  Similar results may be obtained on normal listeners when tested with low-pass
         filtered speech in presence of background noise ("everyday noise").
     3.  The loss of speech intelligibility was  seen  in both  situations even when speech
         frequencies up to 2000 Hz were perceived.
Comment and conclusions.

     The nature of ND phenomenon is not clear. It seems that three factors may play a role
here:
1. Masking  effect: low-frequency  speech spectrum components,  on which patients with
NIHL  mainly depend, are effectively masked  by background noise, especially the  most
common low-frequency one (Niemeyer, 1967).

                                        61

-------
                                           Mynosulob/es PB
               Quiet     +20     +10     +5      0
                                      S/N   (IB
-5     -10
Figure 4. Intelligibility of phonetically-balanced monosyllables, non-filtered and low-pass-filtered (cut-off
frequency is the parameter) in white noise.

                                     62

-------
duiet
                                             Sentences
                                             Unfilled
                                              3000
                                       0     -5     -10
                                5/N     ctB
Figure 5. Intelligibility of sentences, non-filtered and row-pass-filtered (cut-off frequency is the parameter)
in white noise.
                                   63

-------
                                          Monosylables  PB
                                                                -20     -?5
                                    S/V    dB
Figure 6. Intelligibility of monosyllables, non-filtered and  low-pass-filtered (cut-off frequency as the
parameter) in low-frequency noise.
                                      64

-------
                  Sentences
                    Unfiltered
         S/H   dB
Figure 7. As figure 6 for sentences.
          65

-------
•-
              BOA
                                 250
500
1000
2000
4000
6000   Hi
                   Figure 8. Idealized average audiograms of subjects with noise-induced hearing loss, used  in Exp. 2 (10
                   persons in each group).

-------
figure 9. Spectrum of low-frequency industrial noise used in Exp. 2.

-------
0\
ex
              Y/////////. //,
                 Figure  10. Scheme of apparatus:  1-tape recorder with monosyllable test, 2-audiometer Kamplex DA2,

                 3—examiner, 4—control phones, 5—tape recorder with recorded noise, 6—amplifier, 7—loudspeaker unit,

                 8—subject, 9—control microphone, 10—microphone amplifier.

-------
                                         Monosulab/es  PB
                                                     Normal hearing
                                                 Hearing loss:
                                                     > WOO Hz (ff)
                                                     > 500  Hz (I
  §    50.
             Quiet
                                   S/N   dB
Figure 11. Intelligibility of PB monosyllables, tested in quiet and in noise, on subjects with noise-induced
hearing loss and on a group of normal listeners (control).

                                    69

-------
2. Cumulative effect of speech distortion:  reported by Harris (1960): a single distortion,
which by itself does not produce any adverse effect on the intelligibility of speech, may
induce strong degradation of speech when combined with another similar deformation.
3. Inefficiency of organ ofCorti: due to damage induced by noise. It may not be evidenced
by pure tone audiometry, but becomes evident while testing with a complex stimulus such
as speech test, or better, speech-in-noise test (Lipscomb,  1970).
     The most important clinical implication of these findings is that speech tests performed
in quiet do not provide information on ability to understand speech in everyday conditions
as far as patients with sensorineural hearing loss are concerned. This applies also to NIHL, of
course.
     In consequence, the present concept of so-called "most important speech frequencies":
500 - 2000 Hz, based on speech tests performed in quiet  (Report, 1959), should be changed,
as it does not apply to everyday conditions. That has also serious influence on compensation
for NIHL. As we all know very well, under "everyday" conditions, especially at work, there
is nowadays almost no quiet at all. That is why our Congress takes place.
     According to present rules, supported  by the  latest ISO recommendation, (1971) we
are expected to believe that a subject having total hearing loss at all the frequencies over
2000 Hz will have normal ability to understand speech  at work. Is this true? Certainly not.
     I would therefore  propose  the inclusion of the frequencies 3000 Hz and 4000 Hz into
the list of those that are most important for understanding speech in everyday conditions.
There is a lot  of evidence that  these and higher frequencies participate in carrying speech
information (Mullins and Bangs, 1957; Kryter,  1962, 1963; Harris, 1965; Huizing,  1963;
Palva, 1965; Ceypek and Kuzniarz, 1970). Moreover, frequencies  up to 4000 Hz have been
already used in the AMA method for computing hearing loss for speech from the pure-tone
audiogram (after Harris, 1956),  and  frequencies up to 6000 Hz are still used for computing
the Articulation Index (Kryter, 1962).
     That point  of view was accepted by the Ministry of Health in Poland, and so since
1968 the extent of NIHL has been tentatively estimated  in my country on the basis of mean
hearing loss at frequencies  of 1000, 2000 and 4000 Hz, as being the most important for
speech intelligibility in everyday conditions.
     It seems also that for reliable  estimation of a sensorineural listener's performance in
everyday conditions, speech tests in a noise background, similar to those proposed by Kreul
et al. (1968), Groen (1969) or Carhart and Tillman (1979), should  be applied.
Final conclusions.

1. Persons with high-frequency sensorineural hearing loss suffer from disability to under-
stand speech  in  everyday noise, although the noise may not  be  disturbing to normal
listeners.
2. This disability appears even when the hearing threshold up to 2000 Hz is not changed, so
a new  concept of basic speech frequencies for everyday conditions should be developed,
with special attention to the frequency area of up to 4000 Hz.

                                        70

-------
                                    References

Carhart, R., Tillman, T.W.,  Interaction  of competing speech signals with hearing losses.
     Arch. Otolaryngol., 91, 273-289 (1970).
Ceypek, T., Kuz"niarz, J., Znaczenie ograniczonych  pasm  czestotliwosci dla rozumienia
     mowy polskiej. Otolar. Pol., 24,429-433 (1970).
Huizing H.C., Kruisinga, R.J., Taselaar, M., Triplet audiometry: an analysis of band discrimi-
     nation in speech reception. Acta Otolar. (Stockh.), 51, 256-259, § (1960).
Groen, J.J. Social hearing handicap:  its  measurement by speech audiometry in noise. Int.
     Audio., 8,  182, (1969).
Harris, J.D., Combinations of distortions  in speech. Arch. Otolaryngol., 72, 227-232 (I960).
Harris, J.D., Haines,  H.L., Myers, C.K., A new formula for using the audiogram to predict
     speech hearing loss. Arch.  Otolaryngol, 63, 158-176(1956).
ISO - Recommendation R  1999:  Assessment of occupational noise exposure for hearing
     conservation purpose, May, 1971.
Kreul, E.J., Nixon, J.C., Kryter, K.D., Bell, D.W., Lang, J.S., Schubert, E.D., A proposed
     clinical test of speech discrimination. /. Speech Hear. Res., 11, 536-552 (1968).
Kryter, K.D., Methods for the calculations and use of the Articulation  Index. /. Acoust.
     Sac. Am.,  34, 1689-1697  (1962).
Kryter, K.D., Williams, C. Green, D.M. - Auditory acuity and the  perception  of speech. J.
     Acoust. Soc. Amer., 34, 1217-1223  (1962).
Kryter K.D., Hearing impairment for speech. Evaluation  from pure tone audiometry. Arch.
     Otolaryngol., 77, 598-602 (1963).
Lipscomb, D.M., Noise exposure and its effects. Oticongress 2, 1972.
Mullins, C.J.,  Bangs, J.L., Relationships between speech discrimination and other audio-
     metric data. Acta Otolaryngologica (Stockh.), 47, 149-157, (1957).
Niemeyer,  W., Speech discrimination in  noise-induced deafness. International Audiology, 6,
     42^7 (1967).
Palva, A., Filtered speech audiometry. Acta Otolaryngologica (Stockh.), Suppl. 210, (1965).
Report of the  Committee on Conserv. of Hearing: Guide  for the evaluation of hearing
     impairment. Trans. Amer. Acad.  Ophthalm. Otolar., 63, 236-238 (1959).
Robin, J.G., The handicap of deafness./  Laryngol. Otol., 81, 1239-1252, (1967).
Tillman, T.W.,  Carhart,  R., Olsen, W.O., Hearing aid efficiency in a competing speech
     situation. /. Speech Hear.  Res., 13, 789-811, (1970).
Tonkin, J., The diagnosis and  assessment of peripheral sensori-neural deafness. /. Laryngol.
     Otol.,&\,  1187-1239, (1967).
Watson, T.J.,  Speech audiometry  in varied  acoustic conditions. Int. Audiol., 4, 102-104
     (1965).
                                         71

-------
           THE LONG-TERM PLANNING OF A NOISE CONTROL PROGRAM

                                   Michael J. Suess
                               World Health Organization
                                 Copenhagen, Denmark

     Noise, that is to say, an annoying and unwanted sound, has been recognized as a public
health hazard, and endangers both the mental and  physical state of man. Consequently,
major activities have  already been undertaken by  WHO in the mid-sixties to study the
implications of Noise on human health.
     As part of its concern for the improvement of the human environment, the WHO
Regional Office for Europe published a two-year  study on The Environmental  Health
Aspects of Noise  Research and Noise Control.** However, it was the favorable acceptance
of the Office's over-all Long-term Program on Environmental Pollution Control and its
approval by the Regional Committee of the European Region at its 19th session in Budapest
in 1969  that  led to  the detailed planning  and implementation  of a program for Noise
Control (see figure  1). The first  activity within this program was the convention of a
Working Group in The Hague in October 1971.*** The members of that Working Group
reviewed and assessed the noise situation prevailing in Europe and its control, studied future
trends  and  developments, discussed  needed activities of special  importance, and recom-
mended actions and projects to be undertaken. The Working Group stated that

     "Noise must be recognized as a major threat to human well-being."

     "Available knowledge on  the effects of noise and on methods of noise control is not
     being adequately utilized."

     "Progress in noise reduction can be made by setting specific noise limits. While such
     limits  must  necessarily take  technical  and  financial  constraints into account, most
     existing limits cannot be considered as reflecting the prerequisites for well-being, which
     must be the ultimate goal."

     The various  activities shown in Figure  1  have been developed to  support two major
objectives:

(a)  the implementation of investigations for the study of health effects  from noise in order
     to complement existing data and fill research gaps.

(b)  the preparation of a Manual on Noise Control in order to provide the decision maker in
     national and local government with the necessary information for the development of a
     local noise control program.

 **Lang, J. & Jansen, G. (1970), Copenhagen, (WHO document EURO 2631)
***Development of the Noise Control Programme" report on a Working Group, WHO, Copenhagen (document EURO
   3901)

                                         73

-------
    1211
                            1973
                                        1974
                                                    1975
1977
                                                                                        197B
                        1979
Development
of
programme
we-








Survey and
directory of
Institution* •
CS


Study of
nolle limit!
and reiearch ^
priorltlei
CS


Survey of
legislation
*
CS
(continue)
4
CS
Prtpar
of train
manual
CS
1
Preparation
of *tudy
protocol
^
CS
it ion
Ing





T raining
cour*«
(under
EURO 3101)

Field invettlgatlon* of health effect* from
nolle CS
'
: :.
Ditcuiiion
of *tudy
• protocol
WG
T
(continue)
CS
Co-
ordination
of
activities
C

Development
of criteria
and guide*
• r 1
lor nos*e
control
CS

Preparation
of model
chapter on
noi*e control •
(or building
code*
CS
Di*cu*ilon
of model
chapter
r
WG

Ma.nua
T
(continue)
CS


Di*cu*iion
of chapter
*
WG
j 	 '
Study of
enforcement
of nol*e
control
CS

Co-
ordination
of
activities
C


Evaluation
~* of itudie*
WG

4
Disc in lion
of chapter
WG
Di«cua«ion
of nolle
control in
SY
I 	 1
on Noi»« Control
T
Preparation
of chapter*
on Regional
planning, sur-
veillance of
nolle lour-
cei, econo-
mic aipecti.
etc.
CS
Dl*cu**ion
of
*• chapter*
WG


•*• Preparation
of manual for
CS
Type of




Evaluation
of long-
term prog
ramme
WG
i

activities;
C a Consultant
CS = Contractual Service
SY = Symposium
WG =s Working Group
This programme is periodically  reviewed and as such subjected to changes  as appropriate
                Figure 1. The long-term programme on noise control of the WHO Regional Office for Europe.

-------
     The various activities include a survey and the preparation thereafter of a Directory of
European institutions active in the study of health effects from noise and noise control. This
activity will lead to the  identification of appropriate collaborating institutions for the
investigations mentioned above. A study of noise limits and research priorities has just been
completed in draft and will serve both for the preparation of the mentioned investigations as
well as for the preparation of criteria and  guides. The criteria and guides will, in turn, serve
as a chapter for the Manual. Another chapter will serve as a "Model Chapter on Noise
Control for Building Codes", and additional ones will review and discuss various subjects
related to noise control such as regional planning, surveillance of noise sources, economic
aspects, the setting up of standards, manpower needs, etc. A survey of existing legislation
and administrative regulations related to  noise control is under way and will assist in the
identification of supplementary ones needed. A study  on law enforcement will follow and
eventually serve also as a chapter of the  Manual. The urgent subject of noise control in
Europe will be brought to the attention of the general public and the responsible European
authorities through  the convention of a European symposium, probably in  1978.  The
Office's long-term programme will then be evaluated and re-examined for activities needed
in the future.
                                          75

-------
                    SESSION 2




NOISE-INDUCED HEARING LOSS (NIHL)-EMPIRICAL DATA



            Chairman: D.W. Robinson, UK
                       77

-------
                        BASIS FOR PERCENT RISK TABLE

                                  Aram Glorig
                         Callier Hearing and Speech Center
                                  Dallas, Texas

                                William L. Baughn
                                Anderson, Indiana
    The fact that hearing loss produces an impairment is indisputable. The question is how
to evaluate the degree of impairment with presently available measurement techniques.
    The highly complex sound world we live in forces us to make value judgments on what
elements of our sound world are more important than others. The developmental history of
man has passed through stages which have emphasized  different aspects of the auditory
system.  The  development of speech and language as a means of communication has shifted
the emphasis from a simple  warning system for protection from danger to a highly complex
and unique system for storing and dispensing information. Our present  civilization has all
but eliminated the need for the auditory warning system of early man and the importance
of the auditory function now rests mainly with language acquisition and speech. Obviously
hearing  is involved in many  other listening experiences, such as music, etc.,  but I believe I
can say  without fear of contradiction that none are as important as communication  through
speech.
     If  this  concept is acceptable,  it  would  seem reasonable to  evolve  a method of
determining hearing impairment which correlates with "hearing everyday speech". "Hearing"
is used in its broad sense which includes an appropriate and correct response indicating that
what was said can be repeated by the listener.
    On  the assumption  that hearing speech is a common denominator for determining
"impairment"  or "handicap" due  to hearing loss, let us review the history of the present
American Medical Association-American Academy  of Ophthalmology and Otolaryngology
method  for evaluating hearing impairment from pure tone hearing levels.
     Because of the state of confusion that existed prior to the published recommendation
the Subcommittee on Noise of the Committee on Conservation of Hearing of the American
Academy of Ophthalmology and Otolaryngology arranged a conference on Determination
of Handicaps Resulting from Hearing Loss. The Conference was jointly sponsored by the
National Advisory Neurological Diseases  and  Blindness Council (USPHS) and the  Sub-
committee on  Noise. This conference was held on February 12-14, 1958. The following is a
summary of the proceedings of that conference:
     Although much is known about the measurement of hearing with various stimuli, the
use of these measurements to determine  the amount of handicap produced by hearing loss is
in a state of confusion. The conference afforded the first real  opportunity for individuals
representing various disciplines related to hearing to come together to discuss this important

                                       79

-------
problem.  It was the purpose of the conference to pool information and opinions upon
which policy  makers of the American Academy of Ophthalmology and Otolaryngology
might base recommendations for calculating handicap resulting from hearing loss.
     The  group included men  and women  who  represent the fields of Acoustics, Bio-
acoustics, Bio-communications,  Linguistics, Otology, Physics, Psycho-acoustics, Psychology
and Speech (including Speech Analysis  and Speech Synthesis). The following organizations
were  represented: The American Academy of Ophthalmology and Otolaryngology, Ameri-
can Medical Association, Bell Laboratories, Central Institute for the Deaf, Haskins Labora-
tories, Massachusetts Institute of Technology, United States Naval Research Laboratory,
Northwestern  University, University of Illinois, Purdue University, American Speech  and
Hearing Association and  the United States Naval Electronics Laboratory.

                        THEORETICAL CONSIDERATIONS

     Following presentation of and some  brief discussion of the background material, the
conference was thrown open for free  discussion which  centered mainly around (1) the
problems involved in the transfer of information from person to person through the medium
.of  speech and (2) the kind of investigation necessary to allow us to evaluate handicap for
communication from auditory measures.
     The discussion was based on three questions:
      1.   Which  kind of auditory communication efficiency should  be used to estimate
          handicap?
     2.   Which existing auditory  test is the best predictor of mis efficiency?
     3.   What test or type of test would you like to see used to estimate handicap?
Discussion of the first question included the following comments: (a) The normal speaker of
English has to perform certain tasks. If he is consistently  unable to do this he certainly has
"trouble", but  it is  very difficult  to  estimate the handicap  caused by this  "trouble".
Language  impairment, which is relatively easily measured, it is not necessarily the same as
handicap, (b) The question might be rephrased to say first that there is a "normal" listener
and then  to ask  "by what degree does the subject fall  short of meeting minimum normal
standards  for listening?" There are two dimensions of adequacy to be considered here: (1)
the signal level required  to be heard, and (2) discrimination of fine elements of speech, (c)
An approach similar to  that  used by industrial psychologists in areas other than audition
might be adopted here, namely, to specify that there are two ways of defining the criterion
for handicap: (1) as a job sample which would require a full replica of speech conditions and
(2) as an item analysis which would assess critical features of speech communication.
     Discussed as critical  auditory  efficiencies that might be used to estimate auditory
handicap were phonemic differentiation, auditory communication, performance of auditory
communication  tasks, and deficiency in the  reception of speech signals.  It was also noted
that speech and hearing are not necessarily the only factors responsible for a breakdown in
communication, which, after all, is the ultimate measure of handicap due to hearing loss.
     After some discussion, the conference advanced to a consideration of questions 2 and 3
about test materials. A  test of phonemic differentiation  (the Rhyme Test) was discussed.
The test measures word recognition, but confines the basis therefor to the initial consonant

                                        80

-------
and consonant  vowel transition  and yields a score that is heavily weighted with auditory-
phonemic factors, non-auditory factors being attenuated.
     It was suggested that different auditory abilities are  used in different ways depending
on the amount  of handicap and  that ultimately an intelligibility  test might be used to make
grosd distinctions and then batteries of tests applied within the discrete steps of the gross
scale.
     Also discussed were such questions as: Can everyday speech  be represented by specially
constructed  sentences or will carefully  selected words,  phonemes, or nonsense syllables be
more  useful  for determining the change in  information received  from speech when some
part  of the  auditory system is malfunctioning? What consideration should be given to the
effect of education, intelligence and language background of the subjects under test? Should
hearing be measured at levels above threshold?  What  part is played by  the environment
surrounding  both the speaker and the listener? What, for example, are the effects of noise
level  and noise spectrum? The  consensus was that no single functional  test could  apply
under all conditions. Throughout  the discussion, it was  evident that much more research
must  be done before a completely satisfactory method of determining handicap could be
formulated.

                          PRACTICAL CONSIDERATIONS

     Having  accepted the necessity  of further research, the group turned  to the immediate
practical problem of assessing handicap from results  of existing auditory tests and to a
discussion of the possibility of agreeing on an interim  method of handicap determination.
The  need for an interim method for determining handicap resulting from hearing loss was
great. This need was attested to  in part by the confusion that existed  in  the various states
where compensation for hearing loss is provided; at that  time, no two states used the same
method  of  rating disability due to hearing loss.  Further, there was no agreement on a
method for  rating improvement following surgical procedures used to correct conductive
hearing losses. There was a practical need to provide surgeons, legislators and others with an
interim method, even though that  method might not be completely satisfactory and would
have to be changed several years later when more  information became available.
     The conference members agreed  (1) that  they could not,  as scientists,  designate a
completely satisfactory  method  at that time; (2) that the need for a method to evaluate
handicap was urgent and (3) that it would be better for an authoritative group  to recom-
mend the use of an interim method rather than to condone by default  the continued use of
methods formulated, in  some  instances, by groups with little or no  knowledge of the
subject, (4)  that it was reasonable to propose a method  which would be useful, provided the
limitations  of  the  method were understood. It was  eventually agreed that  if sentence
intelligibility is accepted  as a representative measure  of everyday speech,  there was enough
information  to recommend an interim method  of determining handicap. An objection to
using sentence intelligibility as a measure was that currently available sentence tests do not
take into account the effects of background noises, talker identification, localization, etc.
     It  was  agreed  that sentence intelligibility  depends more  on  hearing level than  on
discrimination;  Acknowledging the lack  of  sufficient quantitative information about tests

                                         81

-------
for discrimination, the group agreed that it would be impossible at that time to recommend
a simple method  that includes an evaluation of the changes in discrimination that accom-
pany hearing loss.

        CONCLUSIONS AND RECOMMENDATIONS OF THE CONFERENCE

I.  Although there was no general agreement about how to define or assess handicap due to
hearing loss, the overwhelming majority of the conferees agreed that from a practical point
of view an adequate assessment of handicap could be made from pure tone measurements. It
was agreed by majority vote of 10 to 6 that (a) if the ability to hear  and repeat sentences
correctly in a quiet environment is accepted as  currently the best representation of hearing
for everyday speech and (b) if the measures used  to calculate hearing loss for this everyday
speech are to  be  weighted equally, then: an average of the hearing  level in decibels at the
three frequencies  500, 1000 and 2000 Hz is an acceptable interim method for determining
hearing  loss for  everyday  speech as an  estimate  of handicap. The minority voted to
recommend the use of the average of hearing levels at 500, 1000, 2000 and 3000 Hz to
determine hearing loss for everyday speech. The  chairman and three conferees abstained
from voting.

II. Not  being completely satisfied with the foregoing methods,  the  conference members
unanimously recommended that research,  such as investigation of the effects of different
talkers, of various types of hearing loss, of variations in environmental noises, of various
types of stimuli  both  at threshold and at suprathreshold levels, etc., be undertaken to
discover the factors which should be employed  in a more valid predictive test of the ability
to receive everyday speech.

III. It was unanimously accepted that the principle of the  high and low fence be incor-
porated in the design of any method for determining handicap due to hearing loss.
     The principle of the high and low fence is essentially the principle that the range of
handicap is smaller than the range of auditory sensitivity measured  by pure tones. The low
fence is the point on the audiometric hearing level  scale at which significant handicap begins
and the high fence is the point on the audiometric hearing level scale at which the handicap
is total. Audiometric zero, which is presumably  the average normal hearing threshold, is not
an acceptable low fence because it is not the point at which significant  handicap begins. The
low fence is definitely higher than audiometric zero. Handicap is total at hearing levels lower
than the maximum  output of the standard audiometer; therefore, the high fence is neces-
sarily  lower than the maximum hearing level. It is  evident that  the American Medical
Association-American Academy of Ophthalmology and Otolaryngology method was evolved
with much consideration and a large amount of background information.
    The following is a direct quotation from the American Medical Association Guide to
the Evaluation of Permanent Impairment: (1)
    "Ideally, hearing impairment  should be evaluated in terms of ability to hear everyday
speech under  everyday conditions. The ability  to hear sentences  and to  repeat  them
correctly in a quiet environment  is taken as satisfactory  evidence for correct  hearing of

                                        82

-------
everyday speech. Because of present limitations of speech audiometry, the hearing loss for
speech is  estimated  from  measurements made with  a pure tone audiometer. For this
estimate, the simple average of the hearing levels at the three frequencies, 500, 1000 and
2000 Hz is recommended.
    "In order to evaluate the hearing impairment it must be recognized that the range of
impairment is not nearly as wide as the audiometric  range of human hearing. Audiometric
zero, which  is presumably the average  normal threshold level, is not the point at which
impairment begins. If the average hearing level at 500,  1000 and 2000 Hz is 25 dB (15 dB
ASA - 1951) or less, usually no impairment exists in  the ability to hear everyday speech
under everyday conditions. At the other extremes, however, if the average hearing level at
500, 1000 and  2000 Hz is over 92 dB (82 dB, ASA-1951) the impairment  for hearing
everyday speech should be considered  total. For every decibel that the estimated hearing
level for speech exceeds 25 dB (15 dB, ASA-1951) 1.5% of monaural impairment is allowed
up to a maximum of 100%. This maximum is reached at 92 dB."
    Impairment  in each ear is determined and binaural impairment  is calculated on a 5
(better ear) to one (poorer ear) basis.
    There are several points in the above statement that need discussing.
     1.   "Hear everyday speech"
         Hearing in this context is used in the broad sense indicated by correct repetition
of what was said. Criticisms  of the formula have included statements that to hear is not
necessarily to understand. Understanding  may be used (1)  to indicate discrimination of
word parts or (2) to indicate correct interpretation of the message content. One may hear all
the component parts of a foreign language and not understand the message.
     2.   "Everyday speech under everyday conditions"
         It would  be difficult to define everyday  speech to everyone's satisfaction, par-
ticularly to  define what is included under  "everyday conditions". We have chosen to con-
sider  speech in sentence form to  be  "everyday speech" and correct repetition of the
sentences in quiet as satisfactory evidence of having heard the speech. After consideration of
the varied ambient  conditions that  prevail from day to day during periods of communi-
cation "quiet" was considered to be the condition which prevails the majority of the time,
and therefore best represents prevailing everyday conditions.
     Attempts were made to estimate an average of the noise levels and noise spectra which
accompany speech communication on infinitely varied occasions. Since we could not solve
the question of a representative ambient noise "everyday conditions" are considered to be
listening in quiet
     Obviously it can be  argued that "quiet" is not representative  of everyday conditions
but in our judgment it was the only condition that could be repeated infinitely without
significant variation and therefore subject to standardization.
     3.    "Low fence" and "high fence"
          Twenty-five dB average hearing level (AHL)  was chosen as the level above which
impairment begins. This level was chosen for several reasons among which are:

          1.   The  output  level of "everyday speech"  is usually  between  60-70  dB.
              Listening at threshold is  generally not the case.

                                        83

-------
          2.   Audiometric zero is the central statistic of a range between plus or minus 15
              dB implying that many of the subjects included as normal "hearers" were 15
              dB worse than zero hearing level.
     The "low fence" principle  is supported by at least two studies, one reported by Glorig
et al. (2) and another by Baughn (3). Both studies included pure-tone audiograms and a
"self evaluation of hearing" based on a response to the question. "Is your hearing good, fair
or poor?" in the Glorig study and "How well do you think you hear?" in the Baughn study.
Responses were restricted to the words "good", "fair" or "poor". Both studies indicate a
gradual change in the  direction of higher AHL (average of 500, 1000 and 2000 Hz) with
increase in age in each of the categories, "good", "fair" or "poor". The  twenty-year-olds
stated  their hearing was "good" when the average hearing level was 10 dB. (Figure 1 from
Baughn).  Figure 2 shows a comparison of self-evaluation and speech reception threshold.
Notice the 10-dB change  from 10 dB (ASA) to 20 dB (ASA) for "good" as age increases to
70. The clinical  assumption that noticeable impairment does  not begin until the AHL
exceeds 15 dB ASA or 25 dB  ANSI is  well supported by these data.  In fact, both "low
fence" and "high fence" concepts are seen to be reasonable. Figure 1, from Baughn.
     A review of the literature indicates that the use of 500, 1000, and 2000 Hz does indeed
prove to be quite adequate to represent "hearing of everyday speech" even in noisy situa-
tions.
     Ward et al. (4), in a study related to characteristics of hearing loss found in  subjects
exposed  to  gunfire and  steady noise, determined differences in discrimination for PB's,
paired  consonant  weighted words and paired phonemes between these two groups.  Subjects
with severe  losses above  2000  Hz showed only small differences in discrimination from
those with little or no loss above 2000 Hz. These tests were done in quiet but it is reason-
able to assume that the additional information provided by speech in sentence form would
easily compensate for this small difference when "listening" in noisy environments.
     Ward says, "These results  constitute additional justification for the practice of con-
sidering hearing losses  only at 2000  Hz and below in estimating "handicap" in everyday
speech perception." (4)
     Myers and Angermeier (5)  and Murry and Lacroix (6) tested subjects  with losses above
2000 Hz in noise  and found 5 to 10% difference between normal subjects and subjects with
losses above 2000 Hz. Thus, at least from these data, it appears that the  difference between
subjects with losses  above 2000  Hz and normal subjects is small and it could be expected to
be compensated for by the redundancy of information in speech in sentence form. Further-
more, the 5-10% lower discrimination scores in the hearing loss group appear to be the same
in "quiet" as in "noise". In my opinion, a recommendation to change a well-established and
much-used method of calculating impairment due to hearing loss should be  based on more
significant advantages than those shown in the present data before a change is made in the
method of calculating impairment due to hearing loss.
     Because the  concept of the "percent risk"  table is based on the determination of
impairment  due  to hearing loss I have included  this discussion of the presently recom-
mended evaluation formula.
                                        84

-------
     70i
     60
     5O
     4O
X
• J,
3O
   LLJ

   o
     20
      10
                                 FAIR
                                   FP
                               J.
                                            POOR
                                                           VP
DEAF
               10     20     30      40      50      6O     7O     8O


                      AUDIOMETRIC HEARING  LEVEL  IN DECIBELS


                       Figure 1.  A graphic representation of the results of a self evaluation study.
                                                                          90
             1OO

-------
•/
       -5

   G  10
   w
   Q

       20
       30
       40
    50
CO
CO


3

o
                                          Good
                                             Fair
                                             Poor
       10-19  20-29   30-39  40-49   50-59  60-69  70-79


                              AGE
                 Figure 2. Self evaluation compared with speech reception threshold scores.

-------
                                "PERCENT RISK"

    The concept of "percent risk" was first suggested by Glorig in 1962 as a possible means
of separating the effects of noise on an exposed population from the non-noise effects on a
non-exposed population as a function of equivalent sound level (7). In order to implement
this suggestion, Baughn organized a program to gather the essential data.
    More than 6000 audiograms  and matching noise-exposure data were  subjected to
regression analysis based on three parameters as follows:
     1.   Hearing level  should be defined by the simple arithmetic average of audiometric
         hearing levels  at 500, 1000 and 2000 Hz.
    2.   Exposure intensity should  be defined  by the  reading displayed by a standard
         sound level meter  set at "A" frequency weighting and "slow" inertial dynamics
         with the microphone in a position to intercept the characteristic sound prevailing
         at the subject's ear in his work environment.
    3.   Exposure duration should  be defined  as  years on  the job, approximated for
         groups by subtracting 18 years  from the center year of the  chronological age
         group.
    The noise studies were taken  from over 15,000 detailed sound analyses of work loca-
tions, covering a period of 14 years. Studies of work records, interviews with employees,
and comparative testing of older and more  recent equipment and processes made it possible
to extend exposures back for nearly 40 years with reasonable confidence. Actually the noise
analyses included octave bands, A, B, and C weightings, SIL's and other computed indices in
both slow and faster meter dynamics and in some  areas  repetitions with a General Radio
impact meter. Only the "A" weightings and slow meter readings were used for this study.
All measurements were made  and/or checked by  competent engineers and  checked for
consistency. Three specific levels were used: 78, 86  and 92 dB(A), because the majority of
the sample  clustered around these  points. The group assigned 78 dB(A) spent 90% of their
time in no more than 81  and no less  than 66 dB(A). None exceeded  82  dB(A). The group
assigned 86 dB(A) spent 80% of their time between 86 ± 4  dB(A). The group assigned 92
dB(A) spent 87%  of their time in 92 ± 5 dB (A). (Detailed distributions of time vs. level are
shown in Table 1.)
     All audiograms were made  by two individuals who had done more than 25,000 audio-
grams over  a period of  eight years, all of which had been  submitted to scrutiny by the staff
of the Subcommittee on Noise Research Center. Samples were subjected to consistency tests
and mathematical analysis by a senior research member of this staff. Similar tests applied to
the data used in this study  confirmed its  self consistency.  Approximately 20,000 audio-
grams were culled for consistency of exposure history leaving 6,835 audiograms.
     The audiometric test  environment conformed fully with  the  specifications of the
American  Standards  Association.  The audiometers were Maico  H-l  models and  were
checked against normal experienced ears before each day's use, and were calibrated in the
laboratory  of the Maico Company periodically. They were never found to be out of the
acceptable calibration range.
     The population under study was a very  stable  one since the work force came from a
relatively small community. The age ranged from  18 to 68 years and many were in the same
job for more than 40 years. Age was used  as  a uniform measure  of exposure in years since
                                        87

-------
TABLE 1. MEAN PERCENT TIME SUBJECTS SPENT IN EACH ASSIGNED NOISE LEVEL.


    dBA               78            86            92
65 - 66
66 - 67
67 - 68
68 - 69
69 - 70
70 - 71
71 - 72
72 - 73
73 - 74
74 - 75
75 - 76
76 - 77
77 - 78
78 - 79
79 - 80
80 - 81
81 - 82
82 - 83
83 - 84
84 - 85
85 - 86
86 - 87
87 - 88
88 - 89
89 - 90
-90 - 91
91 - 92
92 - 93
93 - 94
94 - 95
95-96
96 - 97
97 - 98
98 - 99
99 - 100
100 - 101
.75
1.
1.
1.
1.
1.
2.
2.
3.
3.
4.
10.
12.
16.
20.
12.
5.
2.
1.
.5
















                                    .5
                                   2.
                                   2.
                                   3.
                                   4.
                                   6.
                                   8.              .5
                                  10.            2.
                                  13.            2.
                                  14.            2.5
                                  14.            3.5
                                  10.            4.5
                                   8.            6.0
                                   5.           10.0
                                   3.           12.0
                                                14.0
                                                12.0
                                                10.0
                                                 6.0
                                                 5.0
                                                 3.5
                                                 2.0
                                                 1.0
                              88

-------
attempts to  allow for various interruptions in work life provided no advantages. Subjects
with seriously mixed exposures,  or unknown exposures, were excluded in an attempt to
keep the data as nearly ideal as  possible.  Selections for history of otological disease, etc,
were not made. Studies  of large populations  have shown that exclusions for non-noise
effects (exclusive of age) show only very small changes in statistical distributions. (7)
     Following the exclusions noted  above we were left with 6,835 audiograms matched
with exposure history in terms of three exposure groups identified as 78 dB(A), 86 dB(A)
and 92 dB(A).
     The audiograms were divided into two groups; those with AHL's at 15 dB or less and
those with AHL's above 15 dB (15 dB ASA-1951) (25 dB ISO-1964).
     All audiograms were  done prior  to the end of 1965  and all  were done to American
Standards Association 1959 standard  audiometric zero calibrations. (All data were  entered
on punched cards and all sorting and calculations were done by  electronic data processing
equipment).  There were 852 subjects in the group assigned 78 dB(A), 5,150 subjects in the
group assigned 86 dB(A)  and 833 subjects in the group assigned 92 dB(A). Table 2 shows
the distribution of the sample as a function of five year age groupings and exposure assign-
ments.
     These data were used to construct Table 3 and Table  4. It is  obvious that when these
tables are compared there are discrepancies between them. Careful examination will  reveal a
difference related to beginning age; 18 in Table  3 and 20 in Table 4. When the numbers are
picked from  a curve constructed from extrapolations of the three sound levels the 2 year
difference in  starting age accounts for the  differences seen  in the various published Tables.
There is not time to include a complete discussion of the methods used to  extrapolate the
numbers given in Tables 3 and 4.  Figure 3 shows the extrapolated functions by merely using
the three noise-exposure assignments. Figure 4 shows the slight upward curves produced by
extrapolation and curve fitting procedures. A thorough discussion of  the extrapolation is
given in an Aerospace Medical Research Laboratory Technical Report. (8)
     Since  "percent  risk"  is based  on the percentage difference between  a non-noise-
exposed group and a noise-exposed group, the non-noise-exposed group is a critical factor.
The non-noise-exposed data used in these tables were derived from a study by Glorig et al.
(7) A comparison of these data with a U.S. Public Health Service study  is shown in Table 5.
There are differences between these  studies which are probably sample differences since
both studies were designed and supervised by Glorig.
     Table  5  shows  differences  between  two samples whose derivations  are completely
different. The  Glorig sample consisted of professional men  whose history included  no
occupational noise-exposure and very little non-occupational  noise-exposure. The USPHS
sample was  picked to represent the general population of the United States. Theoretically
the general  population should have  more occupational noise-exposure influence than a
population chosen to exclude this. The logic of this difference is supported by the fact that
the Glorig  total sample  shows a lower percentage of "better ears"  that exceed 15 dB
(ASA-1951) AHL (average of hearing levels at 500, 1000 and 2000 HZ)  than is foundin the
USPHS sample. The age differences noted are quite consistent until the Glorig sub-group N's
become small. Further evidence  of consistency is provided by a comparison of non-noise
exposed samples and the 78 dB(A) group studied by Baughn. (Figure 5).

                                         89

-------
          TABLE 2. NUMBER OF SUBJECTS AS A FUNCTION OF AGE GROUPS AND SOUND LEVEL A.
Age Group                 Exposure I   Exposure II    Exposure  III
 Number      Age Span       78 dBA       86  dBA         92 dBA	Total
1
2
3
4
5
6
7
8
13 -
24 -
30 -
36 -
42 -
48 -
54 -
60 -
23
29
35
41
47
53
59
65
N = 10
66
144
148
183
159
95
45
N - 107
476
544
860
1041
1070
723
329
N « 4
39
76
124
189
197
127
77
121
583
764
1132
1413
1426
145
451
                             852          5150             833           6835

-------
TABLE 3. PERCENT RISK TABLE USING AGE 18 AS YEAR EXPOSURE STARTED.
AGE
EXP. YEARS
EXP. LEVEL
80 dBA
EXP. LEVEL
85 dBA
EXP. LEVEL
90 dBA
EXP. LEVEL
95 dBA
EXP. LEVEL
100 dBA
EXP. LEVEL
105 dBA
EXP. LEVEL
110 dBA
EXP. LEVEL
115 dBA
(AGE - 18)
TOTAL '/. EXPECTED
7. DUE TO NOISE
7. DUE" TO OTHER
TOTAL '/,
7, NOISE
7. OTHER
TOTAL 7.
7. NOISE
7. OTHER
TOTAL 7.
7, 'NOISE
7. OTHER
TOTAL %
7o NOISE
7» OTHER
TOTAL %
7. NOISE
7o OTHER
TOTAL 7.,
7. NOISE
7. OTHER
TOTAL 7.
7» NOISE
7» OTHER
18
0
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
.5
0
.5
23
5
1.7
0
1.7
2.5
.8
1.7
6
4.3
1.7
9.0
7.3
1.7
14
12.3
1.7
20
18.3
1.7
28
26.3
1.7
38
36.3
1.7
28
10
3
0
3
6
3
3
13
10
3
20
17
3
32
29
3
45
42
3
58
55
3-
74
71
3
33
15
4.5
0
4.5
9
4.5
4.5
18
13.5
4.5
28
23.5
4,5
42
36.5
4.5
57
52,5
4.5
75
70.5
4.5
87
83.5
4.5
38
20
6.5
0
6.5
12.5
6
6.5
22
15.5
6.5
34
27.5
6.5
48
41.5
6.5
64
57.5
6.5
84
77.5
6.5
93
86.5
6.5
43
25
9.7
0
9.7
16.5
6.8
9.7
26
16.3
9.7
39
29.3
9.7
53
43.3
9.7
70
60.3
9.7
88
78.3
9.7
94
84.3
9.7
48
30
14
0
14
22
8
14
32
18
14
45
31
14
58
44
14
76
62
14
91
77
14
95
81
14
53
35
21
0
21
30
9
21
41
20
21
53
32
21
65
44
21
82
61
21
93
72
21
96
75
21
58
40
33
0
33
43
10
33
,54
21
33
62
29
33
74
41
23
87
54
33
95
62
33
97
64
33
63
45
50
0
50
57
7
50
65
15
50
73
23
50
83
33
50
91
41
50
95
45
50
97
47
50

-------
            TABLE 4. PERCENT RISK TABLE USING AGE 20 AS YEAR EXPOSURE STARTED.
Age
Exposure
(age = 20)
Total
80 Due to
Noise
Total
85 Due to
Noise
Total
, 90 Due lo
PQ Noise
20
0
0.7
0.7
0.0
0.7
0.0
26
5
1.0
No
2.0
1.0
4.0
3.0
30
10
35
15
1.3 2.0
increase in
3.9
2.6
7.9
6.6
6.0
4.0
12.0
10.0
40
20
46
25
3.1 4.9
risk at this
8.1
5.0
15.0
11.9
11.0
6.1
18.3
13.4
60
30
7.7
level
14.2
6.5
23.3
15.6
55
35
60
40
13.5 24.0
of exposure
21.5
8.0
31.0
17.5
32.0
8.0
42.0
18.0
65
Years
46
40.0
46.5
6.5
54.5
14.5

B
— — « .
 g     Total              0.7    6.7  13.6  20.2  24.5  29.0  34.4   41.8   52.0  64.0
•3  95 1)^  jo
H     Noise   	0.0    5.7  12.3  18.2  21.4  24.1  26.7   28.3   28.0  24.0
       Total              0.7   10.0  22.0  32.0  39.0  43.0  46.5   55.0   64.0  75.0
   100 Due  to                                                                              •$
       Noise	0.0    9.0  20.7  30.0  35.9  38.1  40.8   41.5   40.0  35.0        «
       Total              0.7   14.2  33.0  46.0  53.0  59.0  65.5   71.0   78.0  84.5        f
   105 Due  to                                                                               |
       Noise	0.0   13.2  31.7  44.0  49.9  54.1  57.8   57.5   54.0  44.5        fc>
       Total              0.7   20.0  47.5  63.0  71.5  78.0  81.5   85.0   88,0  91.5       £
   110 Due  to
       Noise	0.0   19.0  46.2  61.0  68.4  73.1  73.8   71.5   64.0  51.5
       Total              0.7   27.0  62.5  81.0  87.0  91.0  92.0   93.0   94.0  95.0
   115 Due  to
       Noise              0.0   26.0  61.2  79.0  83.9  86.1  84.3   79.5   70.0  55.0

-------
c
Ul
I

• I


CD
o
                                 BO
                                                                                    1 15
                                                                                            120
                                                                                                    1 25
                                                                                                            130
                                     90     95      1OO    1O5     110



                                        EXPOSURE   dB(A)


          Figure 3. Extrapolated curves representing percent risk due to noise-exposure using the dB(A) sound level

          assignments 78, 86, and 92. The parameter is age.
                                                                                                                   135
                                                                                                                          140

-------

DO
a
        TO
                     80
                            85
90     95     100    1O5     10


    EXPOSURE     dB(A)
                                                                       1 1 5
                                                                               20
                                                                                     1 25
                                                                                            1 3O
       Figure 4. Same data as shown in figure 3 except for use of further extrapolation procedures. The parameter

       is age.

-------
TABLE 5. COMPARISON BETWEEN GLORIG NON-NOISE SAMPLE AND THE U.S.P.H.S. GENERAL
POPULATION SAMPLE. BOTH STUDIES WERE BASED ON AMERICAN STANDARD ASSOCIATION
{ASA} 1951 AUDIOMETER CALIBRATION LEVELS.
         PERCENTAGE OF MEN  OVER  15 dB-AHL.   BETTER EAR.
All
Years
N =
Glorig
N *
US PHS
1323
2.9
52744
7.6
18-24
Years
120
0.8
7139
1.2
25-34
Years
693
1.7
10281
1.4
35-44
Years
438
3.
11373
3.7
45-54
Years
111
7.2
10034
4.1
55-64
Years
30
23-3
7517
10.6
65-74
Years
1
0
4972
30.5

-------
             40
C
3
         X   30
         CD
         T3
         m
A   20
IT
O
         •I
         _J
         r^
         O
         a
10
                                                                              NON-NOISE
                                                                                         /
                                                                     \
                                                                              80 dB (A)
                              10
                               20
                                                    30             40
                                                         AGE
50
60
70
                     Figure 5. Comparison of percent risk curves from Glorig's non-noise data and Baughn's 78 dB(A) data.

-------
    Table 6 shows a comparison of six studies of median hearing levels as a function of age.
All studies were corrected to zero hearing level at age 20 to rule out sample differences on
the  assumption that  the  majority of  20-year-olds have normal hearing. The Table clearly
shows differences in  the  numbers which must be  related to the population samples. In
general there is reasonably good agreement, at least when medians are compared. Distribu-
tions were not available.
    These comparisons confirm the validity of the Glorig sample for use as a non-noise-
exposed group to determine percent risk. Obviously one should expect differences in studies
of different groups. Perhaps the most important comparison in this  discussion relates to
Table 5. It is difficult to disregard so  large a sample but it  is also difficult to classify these
data as non-noise-exposed or disease-free since no attempt was made to exclude either.
    A careful examination  of Table  6 shows that the median hearing levels of sample B
(Glorig) and sample  F (USPHS) at 1000 and 2000 HZ are very close together. When the
hearing levels at 3000 and 4000 HZ are examined the USPHS data show significantly higher
levels  from age 50 up. The most likely explanation of these differences is the effect of noise
exposure in  the  general population sample. What  to  use  as a non-noise-exposed  baseline
remains a dilemma. It is this baseline that determines the percent risk as a function of noise
exposure.
    An examination of Tables 3 and 4 reveals a reversal in direction of the percentages as
exposure years increase. This reversal  at first glance appears to be paradoxical since longer
exposure duration should increase hearing loss. The explanation lies in the fact that noise-
induced hearing loss  assumes  a decelerating function after 15-20 years exposure while
hearing loss due to age assumes an accelerating function.  Figure 6 is a diagramatic repre-
sentation of the change in percentages of population with AHL's above 25 dB due to age,
(OG)  noise-exposure  plus age (OP) and noise-exposure minus age (OK). It is obvious that as
the effects of age on the baseline (non-noise-exposed) population accelerate with increase in
age and the effects of noise exposure on the subject population decelerate the difference
between these two functions  decreases. This decrease is reflected in  Tables 3 and  4 by a
decrease in "percent risk". Figure 7  shows an idealized set of curves taken from Table 3
showing the effects of noise-exposure versus presbyacusis.

                               GENERAL REMARKS

     The previous discussion  presents the background and data-gathering history  of the
numbers used  to establish the percent risk  due to noise as a function of increasing noise
level.  The data used  were carefully  obtained from a representative industrial  population
under conditions which prevail in above-average industrial hearing conservation programs.
Even  if it were practical to obtain this amount of data under more stringent controls it is
doubtful if the  stringently  controlled data  would provide as useful information as the
present data. Hearing conservation programs and compensation are related to data obtained
under field conditions. Highly refined data may provide false impressions.
     We must agree  that further study should  be attempted to confirm or disaffirm the
non-noise  baseline but repeating such a study is not a simple matter. Finding larger N's at
the higher ages is rather difficult.

                                         97

-------
TABLE 6. COMPARES MEDIAN HEARING LEVELS AT 1000, 2000, 3000, 4000 AND 6000 Hz AS A
FUNCTION OF AGE. DATA BASED ON ASA-1951 AUDIOMETER CALIBRATION LEVELS WERE
TAKEN  FROM 6 STUDIES DONE IN VARIOUS PARTS OF THE WORLD. ALL DATA ARE COR-
RECTED TO ZERO AT AGE 20.

AGE
A
B
C
D
E
f

AGE
A
B
C
D
E
f
AGE
A
B
C
D
E
F
AGE
A
B
r
V.
D
E
F

AGE
A
B
C
D
E
f
A- CORSO
B- GLORIG

20
0
0
0
0
0
0

20
0
0
0
0
0
0
20
0
0
0

0
0
20
0
0
0
0
0
0

20
0
0
0

0
0



30
1
0
2
5
0
0

30
0
1
1
5
0
1
30
0
3
3

4
3
30
8
6
4
14
4
5

30
7
7
3

2
5


1000
40
2
0
2
8
0
1
2000
~4U~
0
4
3
7
1
4
3000
1
6
6
NO DATA
5
8
4000
40
8
10
9
18
5
14
6000
40
5
12
9
NO DATA
7
12
C- HINCHCLIFFE
D- JOHANSEN

50
5
1
4
13
0
5

50
7
7
7
14
5
8
50
11
12
16

9
15
50
20
17
18
27
10
23

50
19
20
21

7
19



60

4
5
19
1
6

60

14
12
24
5
12
60

21
29

10
29
60

27
34
38
14
36

60

33
46

8
35
E - ROSEN
F- USPHS

70

14
12
29
4
13

70

27
25
34
5
24
70

34
37

10
44
70

37
42
46
10
48

70

46
47

13
51


                                     98

-------
t
a

E
o


CO
 a
 o
a.
 0)
 o
 k.
 a>
o.
              A                B                     C

              Exposure  Duration   (Age-18)
      Figure 6. Diagramatic representation of accumulating effects of noise exposure and age.
                                  99

-------
                                                           FIGURE 7

   100
 I
••

 •
 •
 :
 •
 •
Q
'
D
                                LIMIT   f ^,

                                FIGURES     4
           Figure  7. Idealized curves showing accumulated effects of noise exposure and age as a function of sound
           level A.

-------
    We are convinced that attempts to improve on these data will prove difficult and are
some years away. Furthermore, new data will probably not change the interval increases in
"percent risk" significantly. It must be remembered that the  risk numbers can never be
applied  to individuals, only to populations of adequate size and  that they never refer to
"amount" of hearing impairment, only to the  percentage of a population having trans-
gressed a certain criterion.

                                    SUMMARY

1.  Impairment due to hearing loss must be based on hearing and understanding everyday
    speech.
2.  The use of the three-frequency average at 500, 1000 and  2000 Hz has proved reason-
    ably accurate to determine impairment.
3.  The low fence of 25 dB (ISO-1964) or 15 dB (ASA-1951) and a high fence of 92 dB
    (ISO-1964) or 82 dB (ASA-1951) are realistic.
4.  The "percent" risk tables are based on carefully done studies. Any differences among
    tables are due to manipulating such factors as age, allowance for TTS or attempts to
    avoid effects of noise-exposure and/or disease.
5.  Attempts to refine the data should be encouraged but no changes should be made until
    further specific studies unquestionably show valid reasons for change.
                                    References

 1.   Guides to  the Evaluation of Permanent Impairment; Committee on Rating of Mental
     and Physical Impairment, Amer. Med. Assoc., 1971. Page 103.
 2.   Glorig, A., Wheeler, D., Quiggle, R., Grings, W., Summerfield, A.; 1954 Wisconsin State
     Fair Hearing Survey, Published by the American Academy of Ophthalmology and
     Otolaryngology, 1957.
 3.   Baughn, William  L.; How Well Do  You Think You Hear? Proceedings of 13th Inter-
     national Congress on Occupational Health, Book Craftsmen Assoc., N.Y. 1960.
 4.   Ward, W.D., Fleer, Robert E., and  Glorig,  A.; Characteristics of Hearing Losses Pro-
     duced by Gunfire and by Steady Noise. J. Audit. Res., Vol. 21, 1962. Pp. 325-356.
 5.   Myers, C.K., and  Angermeier, C.; Hearing Loss at 3 Kilohertz and the Chaba "Proposed
     Clinical Test of Speech Discrimination in Noise." Naval Submarine Medical  Research
     Lab., Report No. 720.
 6.   Murry, T., and Lacroix, Paul G.;  Speech Discrimination in Noise and Hearing Loss at
     3000 Hertz. Naval Submarine Med. Research Lab. Report No. 719.
 7.   Glorig, A., Summerfield, A.,  and Nixon, J.;  Distribution of Hearing Levels in Non-
     Noise-Exposed Population. Proceedings 3rd  International Congress  on Acoustics.
     Elsevier Publishing Co., Amsterdam.
 8.   Relation Between Daily Noise-Exposure and Hearing  Loss as Based on the Evaluation
     of 6835 Industrial Noise-Exposed Cases. Aerospace Medical Research Lab. - Technical
     Report-73-53-1973.

                                        101

-------
References for source of data used in Table 6:

Corso, John K: Age and Sex Difference in Pure Tone Thresholds. J. Acous. Soc.
Amer. 31: 498-507, Apr. 1959.

Glorig, A., Nixon,  J.: Hearing  Loss as  a Function  of Age. Larngoscope, Vol.
LXXII No. 11, pp 1596-1610, Nov. 1962.

Hinchcliffe, R.: The Threshold  of Hearing as a Function of Age.  Acoustica 9:
303-308, 1959.
Johansen, H.: Loss of Hearing Due to Age, Munksgaard, Copenhagen, pp 165.


Rosen, S., Bergman, M., Plester, D., El-Mofty, A., Satti, M.H., Presbycusis Study
of a Relatively Noise-Free Population in the  Sudan, Ann. Otol. Rhin. & Larng.,
Vol. 71, pp 727-743, 1962.

National  Health Survey. United  States - 1960-1962. Hearing Levels  of Adults by
Sex and Age. Series 11, No. 11.
                               102

-------
      A CRITIQUE OF SOME PROCEDURES FOR EVALUATING DAMAGE RISK
                            FROM EXPOSURE TO NOISE

                                   Karl D. Kryter
                             Stanford Research Institute
                             Menlo Park, California 94025

                                   Introduction

    There  are fundamental  definitions and  measurements that must be made before a
proper evaluation of the damaging effects of noise on the inner ear can be performed and
before valid standards  relevant to  the  protection of man's health in this regard can be
promulgated. First, one must define a criterion of hearing, along with a quantitative means
of specifying degrees of impairment to that hearing. Second, the hearing ability, as defined,
of the population of persons who  have not  suffered noise-induced hearing loss must be
known. Thirdly, of course, hearing-ability data of populations  of persons exposed to noise
of various intensities, durations, spectra and years of exposure are  required.

                          Hearing and Hearing Impairment

    It should be obvious that the nature of so-called damage risk tables (i.e. ISO 1999) are
very much  dependent  upon the specification of what constitutes hearing and the start of
hearing impairment, and a critique of present day procedures  for evaluating noise-induced
damage to hearing must begin with that question. For purposes  of hearing conservation with
respect to noise-induced hearing loss, "hearing" is to be known, according to the rather
long-standing recommendations of the American Academy of Opthalmology and Otolaryn-
gology (AAOO)2, as the  average  of pure-tone hearing levels measured at 500, 1,000 and
2,000 HZ,  this average being taken  as an index or indicator  of an ability of a person to
understand  speech.  Speech is defined by  the AAOO as "everyday" speech in the  quiet heard
with the distance between talker and listener presumably approximately one meter or so.
The AAOO Committee further proposed that:  (1) the threshold of impairment (or lower
fence  as it was called) to the understanding or perception of everyday speech, be taken as an
average HL at 500 1,000  and 2,000 Hz of 25 dB, and (2) that there is an increase of 1-1/2%
in hearing impairment with each 1 dB increase above 25 dB until  an average HL  of 92 dB at
the three test frequencies is reached, at which point impairment for hearing everyday speech
reaches  100%. The AAOO  also has specified ranges  of hearing impairment that  are
supposedly  useful in relating the impairment to a description of "handicap" suffered by the
person with that  impairment.3 Whether the handicap is for general social communication or
in one's occupation is not made clear.

                        Hearing Level and Speech Impairment

    The general relations between hearing levels averaged for 500,  1,000 and 2,000 Hz,
percent impairment and percent of speech material correctly perceived, are shown in Fig. 1.
Also drawn on Fig. 1 are the  relations  between hearing level, average for 500, 1,000 and

                                       103

-------
  100
   60
40
   20
                               I    I
 I    I   I    I    I   I
1000 PB WORDS AT WEAK-
CONVERSATIONAL LEVEL
(50 dB) IN QUIET—Al

SENTENCES AT NORMAL
CONVERSATIONAL LEVEL
(55 dB)  IN QUIET—Al
        SENTENCES AT
        "EVERYDAY"
        LEVEL (65dB)—Al
                                                                 I    I   I
SLIGHT
X^X
^<
ZMILD

1 1
MARKED
DEGREES OF
AAOO
1 1
SEVERE EX
: HANDICAP
GUIDE
1 1 1
TREI
—

                                                                                60
                                                                                80
                                                                                    o
                                                                                    o
                                                                                    c
                                                                                    8
                                                                                    I
                                                                                    IU
                                                                            111
                                                                            OT
                                                                            DC
                                                                            o
                                                                            CO
                                                                            o
                                                                            oc
                                                                            O

                                                                            CQ
                                                                            0.
                                                                                100
    -5   0   5  10  15  20  25  30  35  40  45  50   55  60  65  70  75  80  85  90  95
          AVERAGE OF HEARING LEVELS AT 500,  1000. AND 2000 Hz (Earphones re ISO)

     I   I   I   I    I   I   I   I   I    I    I   I   I    I   I   I   I    I   I   I    I
     5  10   15  20  25  30  35  40  45  50  55  60   65  70  75  80  85  90  95  100  105

         AVERAGE  OF HEARING LEVELS AT  1000,  2000, AND 3000  Hz (Earphones  re ISO)

Figure 1.  Relations between HL (average at 500, 1,000, and 2,000 Hz and 1,000, 2,000 and 3,000 Hz) and
measures of speech understanding as calculated according to the Articulation Index (Al), impairment, and
handicap. Speech level measured in field 1 meter from talker. (After Kryter4**)
2,000 Hz, and percent hearing impairment and hearing handicap proposed by AAOO. It has
been shown by  most laboratory and clinical tests that the average of the hearing levels at
1,000, 2,000 and 3,000 Hz provides a better index to hearing impairment for speech than
does the average of 500 1,000 and 2,000 Hz. Fortunately, at least for statistical purposes, an
average difference of about 10 dB can be assumed between the hearing levels averaged for
500, 1,000 and  2,000  Hz and the average for 1,000 2,000 and 3,000 Hz; this is shown by
the lower abscissa in  Fig.  1.
     Attention is invited  to the fact that at  the lower fence  or threshold  of impairment
proposed by AAOO a person would be unable to correctly perceive individual speech sounds
and even some  sentences or  words whenever the speech level  was less than that normally
present (i.e. if the distance between the talker and listener was greater than about one meter
or if the source  of the speech was of a lower intensity than normal).  In addition, it is seen
that the person whose hearing level averaged 65 dB or more would be unable to perceive any
sentences of everyday speech, that specified by AAOO as the  speech signal for evaluating
impairment to hearing. According to AAOO, however, 100% impairment, the "upper fence"
of AAOO, does not occur until about 92 dB average hearing level.
    A particularly relevant study was reported by  Kell, Pearson, Acton and Taylor6 who
found that some 96  subjects, female weavers, whose HLs average 39 dB at 500, 1,000 and
                                         104

-------
2.000 Hz experienced some difficulty in speech communications under  many  conditions
(approximately  809f  of the weavers  reporting difficulty with  communications whereas
controls  persons of the same age but with normal hearing  for  that age about only  109?
reported  difficulty).  However,  according to  AAOO descriptions and  procedures,  these
persons would  have an impairment  of around 20''.  a  "slight handicap", and should have
"difficulty only with  faint speech". Figure 2 illustrates some relations between hearing level
and percent of sentences and other test materials heard as a function of the hearing level of
the subjects; it  is seen in Fig. 2 that when  the hearing level starts to exceed about 0 dB there
is some degradation in the ability of the subjects to perceive the speech material.7
   C
   u
   f-
   t
   -
   I
100
00
BO
70
50
X
30
10
0
-1
*o-»


o*.




o—
- X— —

S

C?

*



"""-x-.
*- --^,


HEE3



— o
	 X



\



^^V



>Y>
\>™—

AVERAGE HL AT
BOO. 1000. AND 2000 H
AVERAGE HL AT
1000, 20OO. AND 3000
I
ri
3=
~-u,


^
	 ,

I
•HI
^ SEN!
NO C
~ » 5 c
" — — x.
' 	 ^.

tf





ENCES,
JISTOR1
B S/N
***-^
riON

^-0
ALL TESTS
IPB Wordl §ne
AT DIFFEREf
AND WITH S(

— —




_*"7^~-1






SPEECH LEVEL
AT 99 dB
*^

S*nttnc«l)
JT S/N RATIOS
3ME FILTERIN*

1 	 >
















0-5 0 5 10 15 20 25 30 35 40 45 S
HEARING LEVEL — dB TA_, „,
    Figure 2. Intelligibility test scores as a function of the average hearing level, re ISO. at 500, 1,000,
    and 2,000 Hz and at 1,000, 2,000, and 3,000 Hz. Speech level at 95 dB. From Kryter7
      All in all, it would appear that the definitions of hearing and the upper and lower fence
 impairment to the hearing given by the AAOO are not consistent with present-day knowl-
 edge of the relation between puretone hearing levels and the ability to understand speech.
 particularly everyday speech.

                                          105

-------
     It must be emphasized that taking the ability to perceive everyday speech as the proper
criterion of what constitutes the sense of hearing is itself debatable, particularly when the
index to that ability is based on but several audiometric test frequencies. For example, the
ability to perceive sounds above 3,000 Hz, let alone 2,000 Hz, is a capacity that is most
helpful  in  perception of many meaningful sounds besides that of speech. It is also an
important frequency region for the maintenance of abilities to localize objects or sound
sources in space and as a safety guide to persons during locomotion. Perhaps most important
is that it is arbitrary to say that the person who loses sensitivity to weak sounds (that is,
losses up to the threshold of 25  dB of the "lower fence"), has suffered no impairment or
handicap inasmuch as in real life speech can  often be very faint or very weak due  to the
distance  between source and listener or because of weak signals coming from the source, or
due  to certain acoustic conditions. Therefore, the  person with normal hearing ability can
enjoy the perception of many weak sounds,  including speech, that are lost to the person
who falls below the "fence" that has been proposed for defining impairment to hearing.

                       Proposed New Upper and Lower Fences

     In spite of all the above factors it is perhaps of practical importance to use the ability
to perceive everyday speech as a criterion  for evaluating hearing impairment due to exposure
to noise. In any event, for present purposes, and keeping the above aforementioned reserva-
tions and conditions in mind, it is proposed that: (1) noise-induced hearing loss be evaluated
with respect to the criterion of impairment to  hearing for speech sentences heard one meter
from talker using normal conversational effort in the quiet, and as predicted by pure-tone
hearing levels averaged at 500,  1,000 and 2,000 Hz, or, preferably,  1,000 2,000 and 3,000
Hz; (2) the threshold of impairment be  taken as 15 dB for the average at 500, 1,000 and
2,000  Hz, or  25  dB at  1,000, 2,000 and 3,000 Hz; and (3) the  upper fence, or  100%
impairment, be taken as being reached at average HLs of 65 dB at 500, 1,000 and 2,000 Hz,
or 75 dB at 1,000, 2,000, and 3,000 Hz. We have taken the liberty of indicating on Fig. 1 a
linear relation between percent hearing  impairment for speech and the average of hearing
levels for three frequencies.

                          Hearing Levels as a Function of Age

     Before plotting data  that relates  the hearing level  of men who have been exposed to
various amounts of noise, and have therefore  suffered hearing impairment that exceeds the
thresholds shown in  Fig.  1, it is appropriate to determine what the population hearing-level
statistics are for people who have not  been exposed to appreciable amount of noise during
their careers. A logical and useful way  to express the effects of  noise on hearing is to
compare: (1) the percentage of the population exposed to noise whose hearing exceed the
lower fence, or start of hearing impairment as defined (or of some other hearing level value
that one may wish to choose) with (2) the percentage of people having hearing impairments
exceeding the specified hearing levels in the non-exposed population.
     Figure 3  depicts the results' of some studies and estimates of the prevalence  in the
population of HLs exceeding 25  dB, averaged at 500, 1,000 and 2,000 Hz, as a function of

                                         106

-------
            UJ
            tr
            "0 o
            So
                 90
                 80
                 70
                 60
   ~   50
i ;=
z 8
< i-
i <
3 S
CL C?
o <
in tr
a. nj
u. >
O <
            (J
            oc
                 40
                  30
20
                  10
         "NON-NOISE EXPOSED" ISO
                              R1999

         "NON-NOISE  EXPOSED" BAUGHN.
                              AFTER GLORIG
                              AND NIXON     /
          ISO  DRAFT  RECOMMENDATION
          TC431/SC-57E. OTOLOGICALLY
          DISEASED AND UNDUE-NOISE
          EXPOSED EARS EXCLUDED.

                       USPHS '60--62
                       BETTER EAR
                        "NON-NOISE
                        EXPOSED"
                                                                    /  „
                                               NON-NOISE
                                              EXPOSED" —\
                     NIOSH '72
                     "NON-NOISE"
                     EXPOSED
                                          AGE — years
                                                                  SA-5825-9R
             Figure 3. Percent of population with HLs of 26 dB or greater (aver, at 500,
             1,000 and 2,000 Hz) as a function of age and approximate years of
             exposure. Dashed  lines  are extrapolations.  Based on  Refs. Glorig and
             Nixon8; Corso9; Glorig and Roberts1 °; NIOSH'l; ISO/TC12.
age.  The 500,  1,000  and 2,000 Hz average is here used because it is the only measure
presented in some of the studies reported. It is seen in Fig. 3 that there appear to be two
sets of curves,  the upper .set being that of ISO R1999 and the non-noise-exposed data of
Glorig and Nixon as reported by Baughn.  Although the basis of the curve presented is not
given, it  would appear that  ISO  1999 was largely drawn from the Glorig and Ni^on  data
base.
                                          107

-------
     The data plotted by Baughn and labeled "non-noise" exposed men came from data
collected by Glorig and Nixon in various studies of hearing of industrial workers. Some
2,518 industrial workers, including  office workers, were asked questions regarding their
previous exposure to noise. Three hundred and twenty nine of the 2,518 men fell into what
Glorig and Nixon labeled the "non-noise"-exposed category. It is not clear from the pub-
lished literature how Baughn derived the  "non-noise"-exposed statistical data, nor whether
they represent all 2,518 men or the screened population of 329 men. Glorig and Nixon
comment that the two populations differed only at the extreme of the distributions. In any
event, Glorig and  Nixon's data must  be interpreted as  being  non-representative  of the
hearing levels of the general population not exposed to noise in the light of the other more
general survey data shown on Fig. 3, in my opinion.
     Some reservations, however, can also be expressed about each of the other functions
shown on Fig. 3: Corso's study of the hearing of a random screened  sample of men in a
small non-industrial town involves but 237 men; the draft ISO document  does not give
statistical data beyond the 25th percentile and median; the NIOSH study involved only 380
men and, further, considered men  working in less than 80-dBA noise to be in the "quiet";
the U.S. Public Health Survey, while  meeting the requirements of a large, randomly selected
population, did not screen the data for men who had been exposed to intense noise in their
work or who had suffered some otological disease.  In an attempt to at least partially remove
from the USPHS data such diseased ears were suggest that the data for the "better  ear" of
the men be used to represent "non-noise" exposed to hearing for a large population. We will
show later the data, which is not greatly  different, for the average of both ears in the U.S.
Public Health Survey.

                        Hearing Levels for Noise-exposed Men

     Baughn studied the hearing of  some 6,835 men who had been exposed for  various
numbers of years to various levels of continuous 8-hour-per-day industrial noise. Some of his
results are shown in Figs. 4  and 5, along with the data from the U.S. Public  Health Survey
for the "better-ear" and average-of-both-ears, as well as some findings from a  recent NIOSH
study.
     Several comments are in order. First, it is clear that the HLs of Baughn's men appear to
shown significant adverse effects as the result from working in noise at levels as low as 78
dBA after a number of years of exposure, with  the increase in incidence in men having
worse hearing dramatically greater for the 15 dB than for the 25 dB fence. Secondly, while
the data shown for the  NIOSH study are somewhat different from the Baughn data for the
lower noise levels—compare  80 dBA  of NIOSH with 78 dBA fqr Baughn-the data  for the
92-95 dBA noise conditions are quite comparable.  NIOSH comes to the conclusion, because
of their data for the 80 dBA noise as  compared with the hearing of "non-noise" exposed
men (defined by them as people working in less than 80 dBA noise) that 80 dBA noise, 8
hours per day will  not increase the incidence of people with HLs greater than either the 15
or 25 dB fence at 500, 1,000 and 2,000 Hz.
     This is so contrary to Baughn's findings, and to other NIOSH data, (Cohen,1 s) as shown
in Fig. 6 that, I think, the conclusions of NIOSH for the lower levels  of noise must be
seriously questioned. In that regard,  it  is interesting to note that the hearing  of the NIOSH

                                        108

-------
       DC
       HI

       <
       LU
       
-------
                          90
                         60
                       §

                      II"
0-50
                      if 0
                      o <
                      ui K
                      * 5
                         40
                      DC
                      s?
                         20
                             HL'i 16 dB
                             OR GREATER
                             (Average at 500
                            . 1000 and 200O H
                                         I
                                        STEADY
                                        NOISE
                                        8 HR/OAY
                                        (RigM Earl
                                        BAUGHN
USPHS
1960-62
BETTER EAR
(Includes
Nonr Ex-
posed and
DtKated
Ears)
                                             I
                         ALSO.
                         PROPOSED AS
                         GENERAL POPULATION
                         (Non-NotM Exposed)

                         	I	I
                           20
                                30
                                      40


                                       I
                       SO     60
                       AGE — v«en
                       I	I
    80

    I
                                                                    Ml
                                       20     30     40    SO
                                      APPROXIMATE YEARS OF EXPOSURE
                      Figure 5. Percent with HLs of 16 dB or greater (aver, at
                      500, 1000, and 2000 Hz) as a function of age and years
                      of exposure. Dashed lines are extrapolations. Based on
                      Baughn'sdata14.

data, were more substantial,-there being 314 men in the 90-94 dBA noise condition, for
example.
     Figure 7  shows  Baughn's and some NIOSH  data, plotted against noise level, in dBA,
with age or years of exposure as the parameter. Also, we have taken the liberty of extrapo-
lating the  data to lower noise levels, using as the non-noise exposed general population the
U.S. Public Health  Survey, better-ear data. Fig. 8 shows a quite  different picture of essen-
tially the same noise-exposed  data versus dBA noise exposure, the source of the difference
between Figs.  7  and 8 being due to the choice of the  "non-noise exposed" population data
chosen. Botsford1* uses  the  data of Glorig and  Nixon, as discussed above  in relation to
Baughn's  use of the same data, and data from a public health survey made by Glorig in a
highly industrialized  community  with  volunteer  subjects—survey  results which  Glorig
described  as being  significantly biased towards the inclusion of people with noise-induced
deafness.
                                            110

-------

0
V)
2
CD
^
z
UJ
W
_J
O
•I
Ul
T
-.
LJ
2

10
20

30

40

SO

60


70

PO


T)

.

-

.

.
O ttFfiPT mm ir* ^'* 4BA
w wvribt wnOUP
( M«]ion Eiperlcoct— 3>ri )
• - NON-EXPCSEO
i I i 1 1 1
     500  1000 2000 SOOO 4000 6000
                                 40

                                 50

                                 60

                                 7O

                                 80
                    O- m-PLANT GROUP
                                    85 DBA
                      ( Median Eiperienc* — 13 yn.)
                       N = 54
                    • - NON-EXPOSED
                    J	I	i	I	I	I
                                    500  lOOO 2000 3000 40OO SOOO
                                      TEST  FREOUEMCY  (Hz)
                     0

                    10

                    20

                    30


                    40

                    SO

                    60

                    70

                    60
                                                          O- IN-PLANT GROUP  66J "aA
                                                             ( Metfion E»p*fii">ee — '2/rt.l

                                                          • —NOM-EXPOSED
                                                           I	I	i    I	i	i
                                                   500  IOOO 2000  3000 4000 600O
          O
          ul
          S  20
             30
             40
O

E
LJ
i

<
Ul
             50

             60

             70

             80
O-IN-PLANTGHOUP   a86dBA
  I Mra.an E«j>ernnc« — I7jr». )
   N« 6
• -NON-EXPOSED
1	J	I  ..  I	1	
                500 IOOO iOOO 3000 4000 60M

                                      TEST
    10

    20

    30

    40

    50

    60

    70

    80
O-IN-PLANT GROUP  _!-%JJ!iS
  ( Median Eipentnc* —  3in.)
   Nr 20
• -NON-EXPOSED
           l',00

FREQUENCY (Hi)
                                                3000 4COO  «iOOO
                                                                                      TA-M2S-13
Figure 6. Mean hearing levels (re ISO) for paper bag workers in different job locations compared with
non-noise-exposed groups equated in  number, age, and sex composition. From Cohen, Anticaglia and
Jones.15


                                         Conclusion

     Standards, guide lines, etc. are, of course, no better than the validity of the data and
definitions  on  which they are based. It would  appear  that the standards  and guidelines
proposed on these  matters by AAOO,  ISO  and  NIOSH are not predicated on  the more
significant and relevant  sources  of presently available data, but rather have consistently
followed the results of rather unreliable, small population studies and rather arbitrary and
unrealistic definitions of hearing that have all  tended to lead to a significant underestimation
of the damaging effects of noise on hearing.
                                             Ill

-------
                CC
                HI
                     80
                UI
                cc
                  •= 70
                OC
                0
^8


^5
  o
  UJ
  o
                LU
                     60
                     50
                     40
                     30
                     20
                  ui
                  >
                I- -  10
                01
                U
                
-------
                   70
                   60
               O  50
               U
               g
                h  30
                Z
                u
                U
                Q<  20
                U
                CL

                    10












^^^^^>





A AGE 50- 59
A AGE 40-49
0 AGE 2.0-39
• AGE 20-29




i
'

i
A






4.
\

I
•




1

1







^

^^



/
x

X7
>-
, ^t_-«
	 ;
r






.
/
^

A y






V
/

V
^
A
, ,f
*0


•


* V
.
r
A


/
^
o/
Xo

^»


/
^


/


/

s


A  A  A
    *   " z
     w    5
                                           65
                               105
                                             SOUND  LEVEL
                                             AT WORK, DBA
                           i
                          z
                          0
                          Z
I    UJ

O   UJ
Z   O
                Figure 8. Percent of population with HLs of 26 dB or greater
                (aver, at 500. 1.000 and 2,000 Hz) of workers in steady industrial
                noise, and of  "non-noise and general  population groups" (from
                Botsford1").


5.  Kryter. K. D.. "Impairment to Hearing from I-.xposure to Noise," J. Acoust. Soc. Am.
   53. May 1973.
6.  Kell.  R. L. J. ('. (i. Pearson. \V. I. Acton, and \V. I;i\ lor in Occupational I/carinx /Lo55
   D. W. Robinson  (1 d. I. (Academic Press. London. 1971) pp. 179-191.
7.  Kryter. K. D. ( 1963). "Hearing Impairment tor Speech." Arch. Otolaryn. 77. 598-602.
8.  Glorig.  A., and  J. Nixon ( |9(iQ).  "Distribution of  Hearing Loss in Various Popula
   tions."  Anals of Otology and Laryngology d9. 2. 497-516.
9.  Corso. J. F. (1963). "Age and Sex Differences in Pure-Tone Thresholds." Arch. Otolar
   yngol. 77. 385-405.
                                          13

-------
                                          Table 1.


Table 1. Results of questionnaire survey of 96 female weavers with average HL's at 500,1000, and 2000Hz of 39dB and a
control group of same age with average HL's of 16 dB. From Kell, Pearson, Acton and Taylor.*

'The social consequences of this impaired hearing ability were:
     (a)   difficulty at public meetings (weavers 72%, controls 5%)
     (b)   difficulty talking with strangers (weavers 80%, controls 16%)
     (c)   difficulty talking with friends (weavers 80%, controls 15%)
     (d)   difficulty understanding telephone conversation (weavers 64%, controls 5%)
     (e)   81% of all weavers considered that their hearing was impaired (5% controls)
     (0   9% of weavers and no controls owned hearing aids
     (g)   53% of weavers and no controls used a form of lip-reading."
 10.  Glorig, A., and J. Roberts (1965). "Hearing Levels of Adults by Age and Sex, United
     States 1960-1962," National Center for Health Statistics,  Series 11, Number 11. U.S.
     Department of Health, Education and Welfare, Public Health Service, Washington, D.C.
 11.  "Occupational  Exposure  to Noise," National Institute for Occupational  Safety and
     Health (NIOSH), U.S.  Dept. of Health, Education and  Welfare,  Washington, D.C.,
     1972.
 12.  Draft Proposal for Hearing Levels of Non-noise Exposed People  at  Various Ages,
     ISO/TC 43/SC-l (Netherlands) 57E.
 13.  Baughn,  W. L. (1966). "Noise Control - Percent of Population Protected," Inti. Audio.
     5,331-338.
 14.  Baughn,  W. L. (unpublished). "Percent of Population Having HLs Greater than 16 dB
     and 26 dB as Function of Years of Exposure in Steady Noise."
 15.  Cohen, A., Anticaglia, J. R. and  Jones, H. H., "Noise Induced Hearing Loss-Exposures
     to Steady-State Noise." Presented at the  American Medical Association Sixth Congress
     on Environmental Health,  Chicago, Illinois, 1969.
                                           114

-------
           THE INCIDENCE OF IMPAIRED HEARING IN RELATION TO
            YEARS OF EXPOSURE AND CONTINUOUS SOUND LEVEL
                  (PRELIMINARY ANALYSIS OF 26,179 CASES)

                                     A. Raber
                 Allgemeine Unfallversicherungsanstalt, Wien, Austria

    Impaired hearing caused by occupational noise is in  Austria a compensable disease.
Hearing conservation consequently belongs to the legal obligations of the Accident Branch
of the  Social  Security Board. Since 1962, our institution, the Allgemeine Unfallversicher-
ungsanstalt, has been performing mass-audiometric investigations in noisy factories.
    For  this  purpose, we use specially-constructed busses with audiometric booths. Our
three audiometric teams have now taken about  165,000 pure tone audiograms.
    The  main objectives of our investigations are:
    1.   To single out  persons whose noise-induced hearing loss is significantly  above
         average, considering the duration and intensity of their particular noise immission;
    2.   To identify  persons suffering from  ear conditions which make  them unfit for
         further work in a noisy environment;
    3.   To check the audiograms in order to see if they  indicate  such a degree of NIHL
         that the investigated persons might be entitled to compensation by law.
    Audiometric investigations are provided for all people who are exposed at their work-
ing places to a sound level of 80 dB(A) or higher. All apprentices are tested irrespective of
their noise-exposure situation at the moment
    The  workers come directly from their working places to our busses. The interval
between the end of noise exposure and the audiometric test is 20 minutes  on the average.
The tests are  performed by  trained and  experienced  Industrial Audiometrists. Otological
investigations  are not carried out  on this occasion. The response rate is 80 to 85%. Non-
participation is generally due to  absence from work because of vacation or sickness. Persons
refuse the test only very rarely.
    The  quality of the audiograms is checked continuously.
    We have collaborated for  many years with some otologists  to whom we send the
following groups of people:
    1.   Persons with remarkable rapidly  developing NIHL;
    2.   Persons whose  social hearing might already be damaged by NIHL;
    3.   Persons in whom the form of the audiogram calls for further otological investiga-
         tion.
    The  otologist repeats the audiometric test and  performs the necessary clinical investi-
gations. On an  average we find a good conformity between our field work and the audio-
grams taken by the otologists.
    Periodic investigations in the factories are repeated every 3-5 years. The serial audio-
grams are performed without knowledge of former results.
    One member of our team is an otologist. He classifies all audiograms- after a preselec-
tion by the computer— and decides which worker has to be sent for a clinical checkup. On
this occasion he examines also the serial audiograms for significant differences which could
have been caused for instance by TTS or imperfect  cooperation of the investigated person.

                                        115

-------
     More details about the organization of audiometric investigations and determination of
noise levels in Austrian industry has been published by Surbock (1).
     In the following report, the data of our routine mass-audiometry are used to relate the
incidence of impaired hearing to the noise immission.
     From all of our audiometric data we extracted those audiograms that met the follow-
ing criteria:
     1.   The person works (nearly) the whole shift at one place;
     2.   At his working place, a continuous noise was measured;
     3,   The person got his  noise immission mainly in the industry in which he was
         working at the time of the investigation; and
     4.   In  me audiogram, the difference  between air and bone conduction is at no fre-
         quency 15 dB or more.
     Using these criteria we got a sample consisting of the data of 18,059 men and  8,120
women.
     Each data record includes the following information:
      1.  Sound level (10 classes:  less or equal 80 dBA, 85 dBA, 90 dBA, 92.5 dBA, 97
         dBA,  100 dBA,  103.5 dBA, 107 dBA, 110 dBA, 115 dBA).
      2.  Years  of exposure (according to the person's report).
      3.  Sex
      4.  Year of investigation
      5.  Age
      6.  Audiogram of the right ear (10 frequencies)
      7.  Industry in which the person was mainly exposed
      8.  Years of exposure according to ISO-Recommendation R 1999 (Years of exposure
         = Age-18)
      9.  Ear diseases yes/no
     10.  Ear trauma yes/no
     The data were subdivided in classes and subclasses according to this hierarchy:
      1.  Sex (male  or female)
      2.  Method  of determining years of exposure (according to ISO or  according  to the
         report of the person)
      3.  Years  of exposure (10 classes)
      4.  Sound level (10 classes)
     For each of these (2x2x lOx 10=) 400 "cells" of hearing losses we computed:
      1.  Number and  percentage  of right ears whose mean Hearing Level at 0.5, 1.0 and
         2.0 kHz is 25  dB or greater.
      2.  Number and  percentage  of right ears whose mean Hearing Level at 0.5, 1.0 and
         2.0 kHz is 15  dB or greater.
      3.  The 5, 10, 25, 50, 75, 90, 95 centiles of the hearing losses at 0.125, 0.25, 0.5, 1.0,
         1.5, 2.0, 3.0,4.0, 6.0, 8.0 kHz.
      4.  Mean, standard deviation, and mean square deviation of the hearing losses.
     In Table 3 of ISO Recommendation 1999 (Assessment of Occupational Noise  Expo-
sure for  Hearing Conservation Purposes,  1st Edition May 1971), the total percentages of
people with impairment  of hearing  for conversational  speech in relation to continuous
sound level and  years of exposure are to be found (2).

                                       116

-------
    In a note to that Table 3 it is mentioned that the values of the table are based on the
limited data available at  that time and that  they are subject  to revision when results of
further research become available.
    In the above-mentioned ISO Recommendation, the definition of impairment of hearing
for conversational speech is in agreement with the AAOO regulation (3). That is, the hearing
of a subject is considered to be impaired if the arithmetic average of the permanent thresh-
old hearing levels of the subject at 500, 1,000 and 2,000 Hz is 25 dB or more compared
with the corresponding average given in ISO-Recommendation R 389, Standard reference
zero for the calibration of pure-tone audiometers.
    As the first step of the evaluation of our data, we compared the percentages of subjects
with impaired hearing in our material with those predicted by the ISO Recommendation.
    The main results are summarized in Table 1 and Table 2.
    Out of the 18,059  men and 8,120 women of our sample, 16,726 males and 7,529
females were working under noise-immission conditions which are registered in Table 3 of
ISO Recommendation  1999. Applying  this table to our data,  it would be expected that
there would be 3,244 males and 1,452 females with impaired hearing in our sample. However,
if one takes the exposure years according to the persons report instead of computing them
with the formula "Years of Exposure = Age - 18", the forecast would be 1,883 males and
743 females with impaired hearing in our sample.
    We actually observed only 851  males and 142 females with impaired hearing. In other
words, relative  to the  number of expected cases, we found in the male population only

                                     Table 1.
                                                  male
                           female
     Sample
              16.726
             7525
     Cases with impaired hearing
     predicted

     (IS01999)
Age-18
Anamn.
3.244
1883
U52
 743
           observed (AUVA)
                 851
              142
                                    117

-------
                                       Table 2

t
b

-------
    100
 I 80
  o
  Q>
     60
  o


  t 40
     20
Years of

Exposure
     ISO Rec.1999

     AUVA
                        \   1
                        35    45




725.665.761.698.695.724.644.391.468.92


           Figure 1
                     119

-------
    100




 •g 80


 I

 "g 60
 • ^^
 o
 a

 •§ 40
    20
85CIBA

ISORec.1999 85dBA

ISORec.1999 ^80dBA

AUVA 85dBA

AUVA
            r	r	r	\	\	\	\	\	r
Years of

Exposure

Coses of
 AUVA *08J555.654.640.m57640&224225.27
                 Ftyun 2
                  120

-------
             90dBA
Years of
Exposure
15    25    35    45
         165.352.462.423.465386290181.17714.
                   Figure3
                    121

-------
100
 60
 20
         954BA
              15    25    35    45
Years of

Exposure

Coses of
 AUVA  121.248.30&279.264.2I6.175M 104. 13.
               Figure 4
                122

-------
  700
Q>
   60
Q.
.i40
*20
           974BA
Years of
Exposure
                 75    25     35    45
       32. 131.194. 200.193.128.106. 62. 58. 8.
                 Figure 5
                   123

-------
   100
•f 80
 CD
   60
NO 20
«N
Years of

Exposure


Cases of

 AUVA
              100 dBA
                  15    25    35    45
           K5.147138.128.119. 82.37. 48.  7.
                  Figure 6
                   124

-------
  700
2>
•| 80
o
0*
   60
.1.
o
9.

•i 40
I 20
Years of

Exposure

Cases of

 AUVA
            105dBA
                 15    25    35    45
          - 75 104-97 9ft ^ 7a
                  Figure 7
                   125

-------
700
         WdBA
Sao-
fe
"S 60
*6
.1 40
^
JS
*»•
/ears of
Exposure
'•^^ » •«M«l^rV •
I ,
1
!
! j
'
[ , '
T Tyf— ^"^ A
- 	 d i_ i •
f- f r [ 1 1 1 1 1 i
5 15 25 35 45,
Cases of
AUVA ft 75- 36 52 ^ 55- 2ft 72 72 a
              FigureB
               126

-------
    100
 •  80
  6
  Q>

 "g 60
    20
           WdBA
Years of
Exposure
Cases of
 AUVA
      15    25    35    45
17 29. 36. 36. 36. 15. 16. 16.  2.
                    Figure 9
                     127

-------
    100
S
fe
Q>
     80
   50
^^
o
&

S 40
   20
         115CIBA
/ears of
Exposure

Cases of
 AUVA
                 15    25    35    45
                  I  4.  2.  3.  2. 2  7.


                  Figure 10
                   128

-------
Table 3
25 dB-criterion years
of exposure according
to ISO; males.

<80dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
lOOdB
105 dB
(AUVA 103)
107 dB
110 dB
115dB
ISO%
AUVA%
Nt.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
1
725
1
0
408
1
1
165
1
0
121

0
32
1
0
39
1
0
49

0
9
1
0
6
1
0
1
5
2
1
665
3
1
555
6
0
352
9
1
248

2
131
14
1
145
20
0
75

0
15
28
0
17
38
0
7
10
3
1
761
6
1
654
13
2
462
20
1
308

1
194
32
3
147
45
0
104

3
36
58
0
29
74
17
6
15
5
1
698
10
1
640
19
4
423
29
3
279

1
200
42
2
138
58
4
97

13
32
76
6
36
88
0
9
20
7
2
695
13
4
601
23
5
465
35
3
264
no en
1
193
49
7
128
65
8
98
no en
13
31
85
6
36
94
75
4
25
10
3
724
17
5
576
26
7
386
39
6
246
ties
8
128
53
11
119
70
12
83
ries
15
33
88
6
36
94
0
2
30
14
7
644
22
10
408
32
13
290
45
14
175

12
106
58
13
82
76
17
70

18
28
91
20
15
95
33
3
35
21
9
391
30
15
224
41
15
181
53
16
81

11
62
65
8
37
82
22
37

33
12
93
31
16
96
50
2
40
33
16
468
43
21
225
54
19
177
62
26
104

26
58
74
38
48
87
21
34

25
12
95
44
16
97
100
2
45
50
8
92
57
11
27
65
14
14
73
15
13

13
8
83
29
7
91
50
8

0
0
95
0
2
97
0
1
                                              5863
                                              4318
                                              2915
                                              1839
                                              1112
                                               890
                                               655
                                               208
                                               209
                                                37
                                              18046
1555  2210   2701   2552   2515   2333   1821   1043  1144    172
 129

-------
                    Table 4
25 dB-criterion years of
exposure according to
anamnesis; females.
< 80 dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
100 dB
105 dB
(AUVA 103)
107 dB
llOdB
115dB
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
. Nr.
1SO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
1SO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO*
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
2
2653
1
1
1062
1
1
622
1
1
400

0
236
1
1
227
1
1
162

0
22
1
4
24
1
0
11
5
2
3
1046
3
2
859
6
2
514
9
3
354

2
203
14
3
189
20
5
150

2
41
28
3
39
38
14
7
10
3
4
764
6
2
691
13
5
499
20
3
347

3
206
32
8
142
45
4
102

5
37
58
6
31
74
0
3
15
5
6
572
10
6
611
19
8
442
29
4
273

5
173
42
9
105
58
2
64

14
44
76
8
39
88
25
8
20
7
9
394
13
6
560
23
10
431
35
12
227
no en
12
135
49
12
109
65
12
73
no en
19
33
85
13
40
94
25
4
25
10
7
300
17
12
386
26
11
300
39
12
164
tries
9
114
53
18
82
70
25
73
ries
27
22
88
13
23
94
100
2
30
14
12
94
22
20
110
32
23
77
45
21
52

9
33
58
15
26
76
33
18

50
6
91
29
7
95
0
0
35
21
8
36
30
25
28
41
8
24
53
33
12

14
7
65
44
9
82
75
8

100
1
93
60
5
96
100
1
40
33
22
9
43
22
9
54
38
8
62
25
8

50
6
74
50
4
87
0
4

50
2
95
100
1
97
100
1
45
50
0
0
57
33
3
65
0
0
73
0
2

0
0
83
0
0
91
100
1

0
0
95
0
0
97
0
0
5419  3402   2822  2331   2006  1466   423    131
52
                                                                   5868
                                                                   4319
                                                                   2917
                                                                   1839
                                                                   1113
                                                                    893
                                                                    655
                                                                    208
                                                                    209
                                                                     37
18058
                      130

-------
                   Table 5
25 dB-criterion years of
exposure according
ISO female.
< 80 dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
100 dB
105 dB
(AUVA 103)
107 dB
llOdB
115dB
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA %
Nr.
ISO%
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
0
225
1
0
234
1
0
141
1
0
104

0
48
1
0
25
1
0
30

0
9
1
0
1
1
0
0
5
1
1
313
3
0
305
6
1
149
9
0
88

0
73
14
0
22
20
2
41

0
6
28
0
0
38
0
0
10
3
0
273
6
1
267
13
1
133
20
1
88

0
53
32
3
37
45
0
40

10
10
58
0
1
74
0
0
15
5
1
244
10
0
257
19
1
145
29
1
78

2
55
42
3
44
58
4
26

0
5
76
0
0
88
0
0
20
7
1
270
13
0
277
23
1
194
35
1
108
noei
1
67
49
2
55
65
2
46
no er
33
9
85
0
1
94
0
0
25
10
3
327
17
2
300
26
2
212
39
1
114
tries
3
80
53
0
41
70
11
54
tries
10
10
88
50
2
94
0
0
30
14
4
361
22
1
342
32
3
218
45
3
115

0
67
58
2
59
76
6
52

17
6
91
0
0
95
0
1
35
21
3
183
30
4
196
41
4
142
53
4
50

0
41
65
6
34
82
21
14

0
3
93
0
0
96
0
0
40
33
4
141
43
4
131
54
8
71
62
9
43

8
38
74
14
22
87
31
13

75
4
95
0
0
97
0
0
45
50
0
10
57
33
9
65
0
6
73
0
2

33
3
83
0
2
91
0
0

0
0
95
0
0
97
0
0
                                                                  2347
                                                                  2318
                                                                  1411
                                                                   790
                                                                   525
                                                                   341
                                                                   316
                                                                    62
817    997    902   854   1027  1140   1221    663   463
32
8116
                    131

-------
                                        Table 6
25 dB-criterion years of
exposure according
anamnesis female.
< 80 dB
85 dB
90 dB
(AUVA 92)
95 dB
97 dB
100 dB
105 dB
(AUVA 103)
107 dB
UOdB
115 dB
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO*
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO%
AUVA%
Nr.
ISO*
AUVA%
Nr.
PERCENTAGES
YEARS OF EXPOSURE
0
1
1
1344
1
0
713
1
1
306
1
0
211

0
154
1
0
77
1
0
50

0
11
1
0
2
1
0
0
5
2
1
408
3
2
529
6
1
338
9
1
189

1
123
14
0
73
20
2
89

0
15
28
0
1
38
0
0
10
3
3
222
6
1
352
13
1
265
20
2
116

3
73
32
2
59
45
2
47

10
10
58
0
0
74
0
0
15
5
5
152
10
0
266
19
3
175
29
1
91

2
57
42
2
52
58
3
31

25
8
76
0
0
88
0
0
20
7
6
144
13
3
232
23
4
163
35
4
89
no en
5
57
49
9
32
65
11
46
no en
10
10
85
50
2
94
0
1
25
10
2
53
17
4
141
26
5
110
39
5
58
tries
0
39
53
6
32
70
14
35
JKS
50
4
88
0
0
94
0
0
30
14
5
21
22
2
52
32
8
36
45
5
21

6
16
58
10
10
76
27
11

67
3
91
0
0
95
0
0
35
21
0
4
30
4
28
41
9
11
53
0
10

0
3
65
20
5
82
33
6

100
1
93
0
0
96
0
0
40
33
0
3
43
20
5
54
0
6
62
0
5

0
3
74
0
1
87
0
1

0
0
95
0
0
97
0
0
45
55
0
0
57
0
0
65
0
1
73
0
0

0
0
82
0
0
91
0
0

0
0
95
0
0
97
0
0
                                                                                  2351
                                                                                  2318
                                                                                  1411
                                                                                   790
                                                                                   525
                                                                                   341
                                                                                   316
                                                                                   62
                      2868  1765   1144   832   776    472
170
68
24
8120
     One reason for the poor correspondence between the predicted and observed number
of cases with impaired hearing could be an underrepresentation of subjects with advanced
noise-induced hearing losses in our material. An answer to this question may be given by the
distribution of hearing losses, especially at the higher frequencies, after various noise immis-
sions.
     Tables 7 to 16 show the distribution of hearing losses for the 5, 10, 25, 50, 75, 90, and
95 percentfles.
                                         132

-------
Table 7
       SPL: <80
       Years of exposure: 27-32
       Nr. of cases:  644

128 Hz
256 Hz
512 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.4
5.4
5.4
5.4
5.5
5.9
7.6
10.5
14.4
11.6
CIO
5.9
5.9
5.9
5.9
6.1
7.0
10.2
15.5
17.6
15.7
C25
7.2
7.3
7.3
7.4
7.8
10.0
16.5
25.8
26.3
24.3
C50
9.5
9.6
9.6
9.9
11.2
16.8
25.9
37.5
38.8
37.8
C75
14.5
14.4
14.0
14.3
16.7
25.3
38.5
53.7
57.1
57.7
C90
18.4
18.3
18.1
18.6
23.1
35.4
54.9
69.8
77.3
81.4
C95
19.8
19.7
19.7
21.5
28.4
43.6
63.9
79.3
89.0
95.9
Table 8
       SPL:  85
       Years of exposure: 27-32
       Nr. of cases:  408

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.4
5.4
5.4
5.5
5.5
6.1
7.9
11.5
14.0
11.8
C 10
5.9
5.9
5.9
6.0
6.1
7.3
11.0
16.3
18.2
15.9
C25
7.3
7.3
7.3
7.5
7.9
11.0
17.4
26.1
27.5
25.2
C50
9.5
9.6
9.6
10.1
11.5
17.5
26.9
39.2
40.0
38.5
C75
13.9
13.8
13.8
14.6
17.3
26.8
42.4
58.0
59.6
59.7
C90
17.9
17.9
18.1
19.1
24.2
40.1
59.6
75.0
82.1
86.2
C95
19.4
19.6
19.9
24.4
34.0
54.3
67.9
79.7
89.9
96.5
  133

-------
Tables
       SPL:  92
       Years of exposure:  27-32
       Nr. of cases: 290

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.4
5.4
5.5
5.5
5.6
6.3
8.0
13.5
15.9
12.7
CIO
5.9
5.9
5.9
6.0
6.3
7.6
11.0
17.5
19.0
17.0
C25
7.1
7.2
7.3
7.4
8.1
11.8
18.4
27.4
28.6
26.9
C50
9.3
9.3
9.6
9.9
12.3
18.5
28.8
45.3
45.5
43.4
C 75
14.8
14.5
14.4
15.0
18.5
28.8
46.9
59.8
66.1
66-9
C90
19.0
18.6
18.9
20.0
27.2
45.9
65.0
75.6
83.8
90.0
C95
22.5
20.0
22.3
24.8
31.9
55.3
71.3
87.1
105.2
97.3
Table 10
       SPL:  95
       Yean of exposure:  27-32
       Nr. of cases:  175

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
40% Hz
5792 Hz
8192 Hz
COS
5.5
5.5
5.4
5.4
S.6
6.1
11.3
17.5
18.8
20.0
CIO
6.0
6.0
5.8
5.9
6.2
7.1
13.8
21.1
23.9
23.4
C25
7.4
7.4
7.2
7.3
8.1
10.8
19.7
31.6
31.6
31.6
C50
9.8
9.7
9.4
9.7
12.7
20.9
36.9
52.5
54.4
56.3
C75
14.7
15.0
14.6
15.0
22.3
36.5
54.4
67.2
74.7
77.8
C90
18.8
18.7
19.2
20.4
29.4
49.2
67.3
77.4
83.6
96.0
C95
21.0
19.9
23.5
28.3
38.7
57.4
73.7
86.2
88.0
98.0
  134

-------
Table 11
       SPL:  97
       Years of exposure:  27-32
       Nr. of cases:  106

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792Hz
8192 Hz
COS
5.4
5.4
5.4
5.4
5.6
7.0
11.1
15.8
14.1
12.9
CIO
5.8
5.8
5.9
5.9
6.1
9.1
15.2
19.1
21.5
17.3
C25
7.1
7.1
7.1
7.2
7.8
12.9
19.6
29.8
35.6
33.5
C50
9.1
9.2
9.3
9.3
11.3
20.0
32.9
48.5
50.0
47.7
C75
13.9
13.9
13.6
13.9
18.9
29.1
49.6
63.5
66.5
69.2
C90
18.6
18.8
18.2
18.7
28.9
47.4
59.3
78.4
87.8
88.4
C95
23.5
22.8
21.8
22.8
36.8
54.5
69.5
87.8
93.4
96.7
Table 12
        SPL:  100
        Years of exposure:  27-32
        Nr. of cases: 82

128 Hz
256 Hz
512 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
40% Hz
5792 Hz
8 192 Hz
COS
5.5
5.4
5.4
5.5
5.6
6.1
10.2
15.9
16.4
17.8
CIO
6.2
5.9
5.9
6.1
6.4
7.6
15.4
19.3
25.2
22.7
C2S
8.1
7.6
7.6
8.0
8.8
12.8
19.5
31.3
35.6
29.3
C50
11.5
10.5
10.5
11.5
14.4
21.7
35.7
48.3
51.0
48.8
C75
14.8
15.2
15.4
16.2
20.3
30.4
49.6
65.4
68.1
63-8
C90
18.9
19.0
18.8
19.4
29.7
47.3
64.0
76.5
76.8
77.7
C95
21.5
21.2
20.0
24.8
38.6
54.8
68.6
79.9
89.5
84.5

-------
Table 13
       SPL: 103
       Years of exposure: 27-32
       Nr. of cases:  70

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5 792 Hz
8 192 Hz
COS
5.4
5.6
5.4
5.6
5.8
7.2
11.5
16.9
16.3
16.3
CIO
6.0
6.1
6.1
6.3
6.7
9.4
15.0
25.0
19.2
22.5
C25
7.8
7.8
7.9
8.2
9.2
15.2
24.2
36.1
38.8
36.3
C50
11.2
11.1
11.5
11.9
15.0
22.7
34.4
50.0
55.6
55.0
C75
15.6
15.9
16.3
16.6
22.5
34.6
56.1
66.1
68.8
67.9
C90
20.0
19.6
20.0
22.5
28.9
48.8
66.7
77.5
85.0
92.5
C9S
26.9
27.5
30.6
37.5
45.8
53.8
69.6
85.6
88.5
97.1
Table 14
       SPL: 107
       Years of exposure: 27-32
       Nr. of cases: 28

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5 792 Hz
8192 Hz
COS
5.5
5.5
5.6
5.8
6.4
10.7
12.0
22.0
22.0
16.0
CIO
5.9
6.0
6.2
6.6
7.8
13.0
19.0
29.0
29.0
19.5
C25
7.3
7.5
7.9
8.9
12.0
18.8
27.0
37.5
38.3
32.5
C50
9.7
10.0
11.7
13.6
16.8
26.3
41.3
55.7
47.5
47.0
C75
15.0
15.0
17.1
17.8
20.0
35.0
56.7
65.0
58.3
65.0
C90
21.0
19.2
21.0
25.5
36.0
56.0
75.5
83.0
75 J
73.0
C95
28.0
28.0
28.0
29.0
43.0
68.0
79.0
93.0
79.0
93.0
  136

-------
Table 15
       SPL: 110
       Years of exposure:  27-32
       Nr. of cases: 15

128 Hz
256 Hz
5 12 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5792 Hz
8192 Hz
COS
5.6
5.6
5.9
6.3
6.9
11.9
13.8
18.8
23.8
18.8
CIO
6.3
6.3
6.9
7.5
8.8
13.8
16.3
22.5
27.5
21.3
C25
8.1
8.1
9.7
11.3
12.9
19.4
33.8
38.8
38.8
28.8
C50
13.8
13.8
14.4
16.3
21.3
36.3
57.5
63.8
66.3
58.8
C75
17.7
19.1
19.1
19.4
28.8
39.4
68.1
77.1
77.1
78.1
C90
19.6
26.3
28.8
36.3
46.3
57.5
77.5
82.5
82.5
88.8
C95
21.3
28.1
31.3
38.1
48.1
71.3
86.3
96.3
96.3
91.3
Table 16
        SPL: 115
        Years of exposure:  27-32
        Nr. of cases: 3

128 Hz
256 Hz
512 Hz
1024 Hz
1448 Hz
2048 Hz
2896 Hz
4096 Hz
5 792 Hz
8 192 Hz
COS
5.5
5.5
5.5
5.5
5.5
30.5
. 55.5
65.5
55.5
55.5
CIO
6.0
6.0
6.0
6.0
6.0
31.0
56.0
66.0
56.0
56.0
C25
7.5
7.5
7.5
7.5
7.5
32.5
67.5
67.5
57.5
57.5
C50
10.0
10.0
10.0
10.0
10.0
35.0
70.0
70.0
60.0
60.0
C75
17.5
17.5
17.5
17.5
27.5
57.5
77.5
77.5
82.5
82.5
C90
19.0
19.0
19.0
19.0
29.0
59.0
79.0
79.0
84.0
84.0
C95
19.5
19.5
19.5
19.5
29.5
59.5
79.5
79.5
84.5
84.5
  137

-------
     Data are shown for men with 30 years of exposure (calculated according to ISO) to
various sound pressure levels.
     Table 7, for example, demonstrates the situation in 644 men with 30 years of exposure
at sound pressure levels up to 80 dBA. Examination of the 75 percentile data shows that
subjects with severe hearing losses in the higher frequencies are rather common.
     For a second example, consider the sample of  70  men who were exposed 30 (ISO)
years to 103.5 dB(A) (table 13).
     At least one quarter of these persons show pathological audiograms of the advanced-
noise-damage type. In this group 17% meet the 25-dB criterion; according to ISO it should
be 62%.
     In countries in  which compensation is provided for ear injuries caused by occupational
noise, the frequency of persons with impairment of hearing, found at mass audiometric
investigations in noisy industries, is of practical interest, because this  frequency will to a
considerable extent determine the number of cases of compensation.
     So our  experience might also be of some practical interest to legislative and administra-
tive boards.

                                    References

(1)  Surbdck, A.R., 1971. In "Occupational Hearing  Loss." D.W. Robinson, Ed. Academic
     Press, London and New York.
(2)  International Organization for Standardization (1970). "Assessment of Noise Exposure
     during  Work for Hearing Conservation Purposes"  ISO Draft Recommendation  DR
     1999. Geneva
(3)  Committee on  Conservation  of Hearing  (1964) Guide for conservation of hearing
     impairment Trans. Amer. Acad. Ophthal. Otolaryng., 63, 236.
                                        138

-------
                       SOME EPIDEMIOLOGICAL DATA ON
                   NOISE-INDUCED HEARING LOSS IN POLAND,
                       ITS PROPHYLAXIS AND DIAGNOSIS

                                Wieslaw Sulkowski,
                          Institute of Occupational Medicine,
                                    Lodz, Poland
Introduction
     Needless to say, noise-induced  hearing damage reduces professional  efficiency and
work safety, is often a cause of disability leading to a loss of profession, and above all will
influence the sphere of private life. Glorig (1961) has rightly stressed its high potential costs
which are higher in industry than the losses brought about by olher professional diseases, if
we include in an economic reckoning the valiie of disability pensions and compensation, the
costs of absenteeism, the results of the loss of ability to work before reaching pension age,
social and health consequences of hearing impairment, etc.
     Statistical data regarding the number of persons with occupational  hearing loss, being
as a rule incomplete, are usually not  very representative and are hardly comparable due to
regional differences in the obligatory  diagnostic criteria in the particular countries, different
reference standards for calibration of audiometers, and different medico-legal regulations.
They give, however,  the  general idea of the importance  of the problem. The sources of
information  consist mainly of: records of hearing losses leading to authorization of compen-
sation, the  results of epidemiological  examination  of workers exposed to noise and, in
Poland, from the data obtained from  the central, uniformly-accepted registration of occupa-
tional hearing losses.
Epidemiological data

     Out of the 33 million population of Poland, about 4 million are employed in industry
(among them,  1.5 million are women), and out of those, roughly 15-20 per cent are exposed
to noise levels causing risk of hearing impairment.
     According to the evidence of the Sanitary-Epidemiological Department at the Ministry
of Health and  Social Welfare, based on  the individual reports of occupational diseases, the
number of occupational  hearing losses  registered during the period 1970-1972, compared
with other occupational diseases, has been presented in Fig. 1. Table 1 shows the structure
of occupational hearing losses in  the years 1970-1972, according to the industrial  branch
and age of workers. It  appears from Table I that the greatest incidence of hearing damage
occurred in the group of people of 40-49 years of age, especially in the industry of transport
means, the metal  and the textile industries. Similar conclusions come from French (Saulnier
1969), Czechoslovakian (Suntych 1970), and Austrian (Surbock 1971) statistical reports.

                                        139

-------
                          60
                          50
                       Q
                       £
                       O

                       I  <0
                       §
                       g  30
                       u.
                       I
                       U»
                       i
                          10
                                  OCCUPATIONAL DISEASES TOTAL
                                  OCCUPATIONAL rtEARJNO
                                       LOSS
                                 1910
                                            1911
                                                        1912
          Figure 1. Indices of occupational hearing loss and all occupational diseases in Poland.
Damage risk criteria

     Central accumulation of the cases of occupational hearing damage in Poland on which
the above  analysis has been made, is based on uniform criterion generally accepted for
prophylactic  and  diagnostic purposes: occupational hearing damage is a bilateral loss of
hearing in relation to the audiometric zero (ISO,  I 964) amounting to  at least 30 dB in the
better ear  after subtraction of age correction, calculated as an  arithmetic mean for fre-
quencies of 1000. 2000 and 4000 Hz. for each ear separately.
     That criterion is similar to that being used in  the USSR (Ostapkovich and Ponomareva,
1971) but differs from the  AAOO  (Davis  197 I) formula which does not accept the impor-
tance of frequencies over 2000 Hz  for perception of speech and thus limits the number of
persons receiving compensation. Without entering into the details of the controversial and
difficult  problem of defining what  degree  of hearing deterioration could be considered as a
                                          140

-------
                                       Table 1
                  OCCUPATIONAL HEARING LOSS IN POLAND IN ACCORDANCE
           TO VARIOUS INDUSTRIAL BRANCHES AND THE AGE OF WORKING POPULATION
                                 IN YEARS 1970 TO 1972
NR

1

2




3


4



S



6


7


8


9

BRANCHES
OF INDUSTRY

Transport
Means Industry
Metal

Industry


Textile

Industry

Coal-Mining
Industry

Machine

Industry

lion

Foundry
Building
Material
Industry

Other Blanches
of Industry
Great
Total
foi Industry
YEAR
1970
1971
1972
1970
1971

1972
1970

1971
1972
1970
1971
1972
1970

1971
1972
1970

1971
1972
1970
1971
1972
1970
1971
1972
1970
1971
1972
AGE GROUPS
-29
15
14
11
1
3

2
3

3
2
_
_
-
5

1
1
2

2
1
-
2
-
2
5
2
28
30
19
30-39
87
71
67
17
50

33
11

22
29
6
16
18
17

18
21
19

8
21
12
5
7
16
15
14
185
205
210
40-19
109
133
141
25
96

97
21

55
60
16
43
83
29

34
45
20

23
44
11
16
26
24
54
45
255
454
541
50-54
.30
61
78
5
26

25
2

20
23
5
25
57
10

12
18
7

10
17
3
7
11
9
15
14
71
176
243
55-59
28
55
74
8
28

24
7

14
32
6
15
28
11

6
15
12

16
24
11
14
7
6
16
15
89
164
219
60-64
17
29
61
5
11

14
10

28
81
1
24
17
6

6
11
9

8
12
9
13
19
4
13
15
61
132
230
65-
2
13
16
-
1

2
4

24
184
1
3
5
3

2
4
1

2
4
3
6
7
	
1
9
14
52
231
TOTAL
288
376
448
61
215

197
58

166
411
35
126
208
81

79
115
70

69
123
49
63
77
61
119
114
703
1213
1693
DYNAMIC
INDEX
100.0
130.6
155.6
100.0
352.5

322.3
100.0

286.2
70B.6
100.0
360.0
594.3
100.0

97.5
142.0
100.0

98.6
175.7
100.0
128.6
157.1
100.0
195.1
186.9
100.0
172.5
241.0
handicap  with regard to health impairment and/or to  the  loss or limitation of earning
ability, it seems that the criterion accepted in our country is fairer for all sufferers and is
justified both from  the  hearing-protection point of view as well as from the medico-legal
evaluation. Pensions and compensation are being granted to persons classified as disabled
due to occupational diseases by the medical board for employment and disability, taking
into account the degree of hearing loss and degree of invalidism.
    Apart from the criterion  of degree  of  hearing loss  mentioned here, the  diagnosis of
occupational hearing loss does not  take  into consideration  other etiological reasons for a
similar picture of deafness of sensorineural type, as well as other forms of hearing damage
not connected with  the working environment and, above all requires an adequate documen-
                                        141

-------
 tation of long-term noise exposure inducing the risk of hearing impairment. With regard to
 hearing protection in Poland, a noise level of not more than 90 dBA, lasting 5 hrs or more of
 continuous exposure daily is permissible.
     For continuous  noise lasting less than 5  hrs and for periodically interrupted noise.
 admissible levels are established with the help of the diagrams presented in Fig. 2. In case of
 fluctuating and irregular noise, the admissible level is specified as an equivalent sound level
 of 90  dBA. For impulse and  tonal noise  exposure, the permissible level should be 5  dBA
 lower.
                                                           
-------
Results of examinations

    1.    Textile industry

    Out of 600 randomly-selected workers in cotton-weaving, spinning and hose strand
twisting  departments exposed to continuous steady-state broad band noise, 98 subjects were
excluded who had suffered from morbid changes in the middle ear, had undergone skull
trauma or had had other diseases which would jeopardize proper evaluation of the influence
of the actual  exposure to  noise  on the  state of hearing. The mean  bilateral hearing
thresholds of the remaining 511 subjects (59 males and 452 females) (average age range 37.2
yrs ± 9.8 and with the average length of employment 12.2 ± 6.6) separately for the workers
of the particular three departments are shown in  Fig.  3.  No  statistically  significant
differences have  been found between hearing thresholds of the left and the right ear (p >
0.05). Greater  statistical significant hearing losses have  been noted in men, although the
dispersion around the mean was similar for both sexes (F < F 0.05)x.
    The significant differences of hearing thresholds among the workers of the separate
departments (F = 604.06 > F 0.01) seem to be connected with the different levels of noise,
as well  as  with  the structure  of the  examined groups, i.e. then* age and length of
employment (average age and  length of  employment  for  separate  departments differ
significantly [f=  19.38; P < 0.051)
    The influence of age and length of employment of the examined population on the rise
of hearing threshold seems to be clear. There is a significant increase of hearing loss accord-
ing to age (for men  F = 17.4; P  < 0.01 and for women F = 292.97; P < 0.01) and similar
one according to length of employment (F = 319.00 > F 0.01). The above is supported by
the regression lines (Fig. 4).
    In 46 workers the magnitude of the hearing loss (after deduction of the age correction)
reached the criterion for a compensable loss. Their distribution is shown in Table II.

    2.     Metal industry

    Out of 270 working population in three metal plants, viz. cutlery manufacture (con-
tinuous  noise with  fluctuating level), car'spring manufacturing  and press  plant  (impact
noise), 32 workers were excluded due to pathological symptoms in the middle ear or other
diseases. The mean bilateral hearing thresholds were determined in the remaining 238  sub-
jects (195 men  and 43 women) separately for  each  working group. Figure 5 shows the
oscilloscope photograms of noise in the particular plants. The results of audiometric exami-
nations of exposed populations are presented in Fig. 6.
    There  appear  significant differences  in the  mean hearing  thresholds of particular
groups, which are probably due to the different characteristics of exposure, especially as the
parameters  of age and length of employment are practically identical. In the population of
 xThe method of variance analysis according to the scheme of single classification has been used. (F denotes
 the value of F - Snedecor statistics).

                                         143

-------
                             M£M  tOUL N'59
                             HMHMliiWU
                             MEAN UVUMUNT W.&
                                                  WiA VINfr
                                                5f>L *
                                                      VI IW  l»jjj_
                                                    93 « dBA
                             MEN
                             WANA££41,t<
                             MEW
                             HOSf TWJ5TIN&  5PL  W-« *A
                             VfAN A6t 51)!
                                                                       -mean hearing threshold x; —.—.—.
Figure 3. Mean  bilateral  hearing thresholds of textile workers: —
range of dispersion ±5,	95.5 % confidence limits for the average general population (x - 25X) and (x + 2
               3 26

               g 2*

               ? 22
               tU
               T 20

               £ w
               ui
                : 10
                                      J yZ9 '0.31
                                                                                  ylf.B -09.
                          30    +0    50    60

                            AGE (YSS)
                                                    4    6    8    10   f2    14    16    19
                                                        LFNOTH Of fHfLOthCHT (V«S)
Figure 4. Regression lines of the  form y = ax  + b characterizing dependence of hearing loss on age and
length of employment.
                                                   144

-------
                                       Table 2

                      OCCUPATIONAL HEARING LOSS IN TEXTILE WORKERS
INDUSTRIAL
DEPARTMENTS
Cotton
Weaving
S*
8 c
o-S
"£
Hose Stiand
Twisting

Number of
Subjects
Occupational
Hearing
Loss
Number of
Subjects
Occupational
Hearing
Loss
Number of
Subjects
Occupational
Hearing
Loss
TOTAL
200
27
172
15
139
4
AGE (yrs)
20-29
42
-
37
-
63
-
30-39
82
10
54
5
38
-
40-49
57
11
42
4
3$
4
50-59
19
6
39
6
-
-
LENGTH OF EMPLOYMENT (yrs)
1-3
IS
-
15
-
12
-
4-6
25
1
26
1
58
-
7-10
22
3
24
4
43
1
11-15
81
11
37
4
23
3
16-
54
12
70
6
3
-
the punch press plant, consisting of men and  women  of similar distribution of age  and
length of employment, a significant rise of threshold has been noted for men (p < 0.001).
The regression lines traced on the basis of mean hearing thresholds for the separate age and
length of employment indicate the relationship between the age, length of employment and
the course of the hearing threshold curves (Figure 7).
    Table HI summarizes the cases of the stated occupational hearing losses.

    3.   Forest workers*

    Out of 12,000  woodcutters using motor-saws and therefore having been exposed to
noise of fluctuating level but without an impulsive component for less than 5 hours daily,
401 workers were chosen  for examination.  58 workers were excluded due to middle-ear
'pathology and other diseases or to exposure to noise in previous employment. In 11  of the
remaining 343, compensate hearing losses were found (Tab. IV).
     Figure  8 presents mean hearing thresholds for different  age groups, separately for left
and right ears.  Statistically significant differences were found between the left  and right
ears, which was not observed in the control group consisting of 47 tree-fellers of similar age
and length of employment, but who worked only with an axe (Table V). It seems likely that
 "The presented data are a part of the research work on: "The effects of chronic vibration and noise exposure on the health
  of woodcutters" being conducted within Polish-American Scientific Collaboration - contract No 05-005 3, Project Officer
  A. Cohen, Ph.D., Principal Investigator: H. Rafalski, M JD.                                             '
                                         145

-------
Figure 5. Oscilloscope photograms of noise in particular metal plants. A: cutlery manufacture: continuous
noise with fluctuating level, mean SPL 92-100 dBA, leq 95 dBA. B:  car spring plant: Leq 108 dBA. mean
level of acoustic background 82 dBA, impulse peak pressure level 121 dBA, rise time 0.2 sec, duration 0.6
sec, time interval 0.8 sec,  repetition of impulse 1/sec. C: punch press plant:  Leq 95 dBA,  mean level of
acoustic background  85 dBA, impulse peak pressure level 99 dBA, rise time 0.05 sec, duration 0.2 sec, time
interval 0.5 sec, repetition of impulses 2/sec.
the special posture of the chain - saw user during his work causes the left ear to be especially
exposed (Fig. 9). The calculated coefficients of both total and fractional correlation confirm
the relation between age and  length of  employment and  the  course of hearing threshold
curves (Table VI). An analogous conclusion arises from the regression lines, traced according
to  the  age and length  of employment  for evaluation  of the  average changes of hearing
thresholds (Fig. 10).

                                            146

-------
                                                    N«M  '" •-
                                                    COMTlim MfAM
                                          CJJT
                                             IfRY
                                                   r-^
                         N=109 (*EN)
                        /WAN A££.3M*14*
                        MEM EMPLOYMENT
                                        CAfl SP RIHtS _
                                             PRBH
                        MEN N-31
                        WAN Aȣ Hl>.
                                                    N-19  "
                                                    CMTimS MUM
                                                                > M
                        WOMEN N-
                        «MN A6E  ,
                        MEAN EMPLOYMENT 5,1! y
Figure 6. Mean bilateral hearing thresholds of metal workers
dispersion x ± «x).
                                                       mean hearing threshold; —.—.—. range of
     4.   Discussion

     The relation  between  noise  and irreversible changes in hearing sensitivity which arise
due  to long-term  daily exposure  has been the  subject of so many field studies that there
would not be enough room to enumerate them all here.
     In spite of standardized methods of hearing measurement  based on tonal audiometry,
the great number of variables makes the comparison and  evaluation of the results which
have been  obtained difficult-for example,  the wide variation  of individual susceptibility.
There exists, then, the problem of an adequate use of these experimental data gathered for a
better explanation of  the  effects  of the  exposure to noise on hearing. Therefore,  the
problem  is  to establish possibly  optimal damage risk  criteria (DRC) more accurate than
those now  in  use. DRC worked out on a statistical basis cannot ensure preservation of
                                           147

-------
                   	O SMMEH5
                   	* PUNCH-PIIESSfRS
                    - • CUTLERS
                            ACC (T«»»
                                                            UWTN Of ((MDYHHIT
Figure 7. Regression lines of the form y = ax + b characterizing dependence of hearing loss on age and
length of employment in metal workers.
                                             Table 3
                        OCCUPATIONAL HEARING LOSS IN METAL WORKERS
INDUSTRIAL
PLANTS
III
%h
a>*.
aJS a
M a
&r-
ws-
Q, WBL<

Number of
Subjects
Occupational
Hearing Loss
Number of
Subjects
Occupational
Heating Loss
Number of
Subjects
Occupational
Hearing Loss
TOTAL
109
16
55
32
74
5
AGE (yrs)
24-29
26
-
5
3
17
-
30-39
34
4
17
10
18
-
4049
30
8
24
13
26
2
50-
19
4
9
6
13
3
LENGTH OF EMPLOYMENT (yrs)
1-5
32
1
9
5
30
-
6-10
15
2
15
7
17
-
11-15
22
2
15
8
12
3
16-
40
11
16
12
15
2
                                               148

-------
                                                 Table 4
                           OCCUPATIONAL HEARING LOSS IN CHAIN-SAW USERS


Number of
Subjects
Occupational
Hearing Loss
Tf~lT A T


343
11
AGE (yrs)
20-29

40
1
30-39

101
1
4049

147
7
50-59

55
2
LENGTH OF EMPLOYMENT (yrs)
1 -3

57
-
4-6

106
1
7-10

137
3
11-15

36
4
16 -

7
3
                                 LEFT EAR
                                 WOOK UTTERS
                                 N«332
                                                            RIGHT  MR
101,5
                         
-------
                                      Table 5

                          EVALUATION OF MEAN HEARING LOSSES
                IN WOODCUTTERS (MEAN FOR FREQUENCIES 1000,2000.4000Hz)
EXAMINED GROUPS
Chain-Saw Users
Total (N = 343)
Chain-Saw Users
with Occupational
Hearing Loss (N = 11)

Controls (N = 47)
Significance of
Differences (Chain-Saw
Users - Controls)
LEFT EAR
22.1 ± 11.6
52.1 ± 9.1
17.1 ± 9.1
t = 2.831
p < 0.01
RIGHT EAR
18.6 ± 9.7
50.2 ± 7.1
18.0 ± 9.5
t = 0.376
p>0.5
SIGNIFICANCE
OF DIFFERENCES
t = 2.871
p < 0.01
t = 0.636
p > 0.05
t = 0.469
p > 0.05

normal hearing for the whole population exposed to noise hazard but only for the majority
— those of average susceptibility.
     Studies on a few different occupational environments which were not controlled with
regard to protection of hearing indicate a great spread of hearing thresholds of the examined
invididuals around the arithmetic mean of the whole group. A similar high standard devia-
tion was found by Riley et al. (1961) and Jatho (1966), who studied HLs as a function of
age and sex in populations not exposed to noise. In our audiograms we have not subtracted
age corrections, so the  data presented  illustrate the combined influence of noise exposure
and age-induced hearing losses which, according to Spoor (1967), increase with the logarithm
of age. However, comparison of the audiograms of non-noise-exposed subjects having the same
age as the subjects in our control groups gives an idea of the magnitude of, as it is called by
Passchier-Vermeer (1971), "the noise-induced part of the median hearing level". The results
are therefore  the key  relation between permanent  threshold  shifts for pure-tone  and
physical noise parameters, effective duration and years of exposure. They show  also, simi-
larly  to Burns and Robinson's (1970)  findings, mat the curve of hearing loss has no
tendency to a sudden flattening after 10-15 years of employment.
     Significantly lower hearing thresholds of women found in  our data as well as in many
other surveys  (Jankowski,  1952; Hassmann et al. 1970; Dieroff, 1967; Ward,  1965) are
ascribed to a higher absenteeism of women and to a lesser exposure of women  to non-
occupational noise.
     The  alarming percentage  of hearing impairment among  the workers of  the metal
industry exposed to impact noise shows the urgent need of establishing D.R.C. for this type
of noise, whose evaluation is still in the area of uncertainty of knowledge (Kryter, 1970;
Coles and Rice, 1971; Sulkowski et al. 1972).
     It should be stressed that in both the metal and textile industries, we had expected at
the outset to  find confirmation of our presumption of the considerable apparent risk of
healing loss there due to 8-hour daily exposure. But finding 11 cases of compensable hearing
losses among the woodcutters exposed  only about 3 hours daily  to fluctuating noise came as
                                        150

-------

                                                                  j
                       Figure 9. Chain-saw user engaged in tree-felling.
a surprise to us. In the entire group of 343 chain-saw users, the thresholds of the left ear
proved to be worse than those of the ritjit car. which had been reported to the examiners by
the workers themselves before the audiometric tests started.
     The risk  of damage among forest workers has also been noted  by  Holmgren  et al.
(1971) and  Passchier-Vermeer (1971). However,  they  did not give any details about  the
number of stated cases of compensable hearing losses. The results achieved by those authors
as well  as ours  seem to confirm the equal-energy  hypothesis which,  for evaluation of a
fluctuating noise as well as for an irregular one, assumes that the risk of hearing loss depends
on the total dose of acoustic energy, regardless to the distribution of the energy in time.

                                         151

-------
                                         Tables
                        EVALUATION OF DEPENDENCE OF HEARING LOSS
                       ON AGE AND LENGTH OF EMPLOYMENT BY MEANS
                    OF TOTAL AND FRACTIONAL CORRELATION COEFFICIENTS
DEPENDENCES
Age (xl) -Length
of Employment (X2)
Hearing Loss (y)
-Age(xl)
Hearing Loss (y)
-Length of
Employment (X2)
COEFFICIENTS
OF CORRELATION
TV v
X,Xj
Tyxl
TyXl.x2
^
Tyx2.x,
CHAIN-SAW
USERS TOTAL
+0.343
t = 6.03
p < 0.001
+0.299
t = 5.408
p< 0.001
+0.223
t= 3.949
p < 0.01
+0.288
t= 5.192
p < 0.01
+0.207
1=3.562
p < 0.01
CHAIN-SAW
USERS WITH
OCCUPATIONAL
HEARING LOSS
+0.394
t = 4.197
p < 0.001
+0.557
+0.471
+0.414
+0.255
t = 0.791
p > 0.05
CONTROLS
+0.685
t = 6.307
p<0.01
+0.548
t = 4.395
p < 0.01
+0.443
t= 3.318
p < 0.05
+0.360
t = 2.588
p < 0.002
-0.0154
t= 0.103
p>0.9
LEGENDS:  Coefficients of Total Correlation
                yX|, yx2
Coefficients of Fractional Correlation
        1 •
        3 n
        * M
        1:
        i:
                     l£f r CM
                     I 1M>     •
                                   MM    14   !!•<>»«
                                                    OF fmornciir 
-------
Prophylaxis and diagnosis of occupational hearing loss

    Prophylaxis of hearing damage, which starts with the physical measurement of noise
and evaluation of its harmfulness, is in the scope of the industrial health service activities in
Poland. Industrial physicians have an advisory voice in noise control which is carried out by
the industrial hygiene  staff in epidemiological centers - cities, towns, voivodeships and
districts — leading to diminution of noise levels, shortening of exposure times, and use  of
individual acoustic defenders.
    The pre-employment and follow-up audiometric examinations, which aim at assurance
of controlled, repeated and comparable evaluation of hearing of workers, are performed in
stationary audiometric laboratories at factorial, multi-factorial, district, town or voivodeship
industrial outpatient clinics. Supervision over these examinations is provided by the otolar-
yngologists, who also determine their frequency. In 1972, there were 205 otolaryngologists
employed in the Polish industrial health service. The consultant departments of the voivode-
ship outpatient  clinics verify the detected hearing damage; doubtful cases or diagnostically
difficult ones are referred for decision in conditions of clinical observation to the Institute
of Occupational Medicine. During the period of 1971 to 1972, 496 cases (356 men and 140
women) of hearing impairment among the workers of different branches of industry from
the whole country were referred for checking.  Occupational hearing loss was confirmed in
330 cases. The grounds for elimination of occupational etiology in the remaining 166 cases
are listed in the  Table VII.
                                       Table 7

                              MOTIVES FOR ELIMINATION OF
                                OCCUPATIONAL ETIOLOGY
ENUMERATION OF MOTIVES
NON-ORGANIC HEARING LOSS 	

OTOSCLEROSIS 	

TYMPANOSCLEROSIS 	

UNDERGONE CONSERVATIVE AND RADICAL OPERATIONS OF THE MIDDLE EAR 	

CHRONIC PURULENT OTITIS MEDIA . ...

CONGENITAL DEAFNESS 	

MENIERE DISEASE 	

LACK OF ESSENTIAL NOISE EXPOSURE 	

TOTAL 	

NUMBER OF
CASES
55

12

14

21

39

2

15

8

166

                                         153

-------
     It is  our experience, and Coles and Priede (1971) present a  similar opinion, that
conventional  tonal audiometry alone is  not always adequate for reliable  evaluation of
hearing, especially with regard to claimants of compensation.
     Nothing  less than the application of a large set of various audiological tests allows a
well-founded diagnosis in such cases. In our Institute these are routinely used: manual tonal
audiometry,  Bekesy automatic audiometry,  suprathreshold examinations  (mainly  tests
of Langebeck, Liischer - Zwislocki, SISI), speech audiometry, and among classic malingering
tests, mostly  Stenger's test and the delayed speech feedback test. The measurements of
stapedius muscle reflex and evoked response audiometry (Bochenek et al. 1973) are also
worth mentioning as valuable methods for evaluation of real hearing thresholds.


                                   References

 1.  Bochenek,  W., Sulkowski, W., Fialkowska,  D., Tolloczko, R., and Zmigrodzka, K.,
     Impedance and electrophysiologic (ERA) audiometry  in  diagnosis of occupational
     hearing loss. Med. Pracy (in press).
 2.  Borsuk, J., Sulkowski, W., Wendt, J., Miksza, J., and Zalewski, P., An evaluation of
     hearing status of workers of some textile plants on the  basis of audiometric examina-
     tions. Biul WAM, 13, 115-120 (1970).
 3.  Burns, W., and Robinson, D.W., Hearing and noise in industry. London: Her Majesty's
     Stationery Office (1970).
 4.  Coles, R.R.A., and Rice, C.G.,  Assessment of risk of hearing loss  due to impulse noise.
     Br. Acoust. Soc. Spec.. 1, 71-77(1971).
 5.  Coles, R.R.A, and Priede V.M., Nonorganic overlay in noise-induced hearing loss. Proc.
     roy. Soc. Med., 64, 194-199 (1971).
 6.  Davis, H., A historical introduction. Br. Acoust. Soc. Spec., 1, 7-12 (1971).
 7.  Dieroff,  H.G., Horschaden durch Industrielarm. Arbeitsmedizin, Sozialmed., Arbeit-
     shy g., 7,  256-260. (1967).
 8.  Glorig, A, cited  by Bell, A, Le bruit, Geneve, Organisation Mondiale de la Sante
     (1967).
 9.  Hassmann, W., Pietruski, J., Krochmalska, E., Filipowski, M., Kom-Rydzewska, H., and
     Kisielewska, L, Hearing impairments in  young textile workers. Proceedings of XXVII
     Conference of Polish Otolaryngologists, Warszawa, PZWL, 126-129 (1970).
10.  Holmgren, G., Johnsson, L., Kylin, B., and Linde, O., Noise and hearing of a popula-
     tion of forest workers. Br. Acoust. Soc. Spec. Vol. No. L, 35-42 (1971).
11.  Jankowski, W., Impairment of hearing in the workers of the cotton industry.  Otolar-
     yng.Pol, 6, 111-123(1952).
12.  Jatho, K., Population surveys and norms. Int. Audiol., 5, 231-239 (1966).
13.  Kryter,  K.D., Evaluations of  exposures to impulse noise. Arch. env. Health, 20,
     624-635  (1970).
14.  Ostapkovich, V.E., and Ponomareva, N.I., Determination of work capacity in workers
     with occupational hearing disorders. Gig. Truda, 15, 25-28 (1971).
15.  Passchier-Vermeer, W., Steady-state and fluctuating noise: its effects on the hearing of
     people, fir. Acoust. Soc. Spec.,  1, 15-33 (1971).

                                       154

-------
16.  Riley, E.G., Sterner, J.H,, Fassett, D.W., and Sutton, W.L., Ten years' experience with
    industrial audiometry. Am. Ind. Hyg. Ass. J., 22, 151-159 (1961).
17.  Saulnier, J., Le  bruit dans 1'industrie  et  la surdite professionnelle.  These pour le
    doctorat en pharmacie, Paris (1969).
18.  Spoor, A.,  Presbycusis values in relation to noise induced hearing loss. Int. Audiol., 6,
    48-57(1967).
19.  Sulkowski, W., and Andryszek, C., Investigation of the noise induced hearing losses in
    workers of chosen plants of light industry. Med. Pracy, 23, 296-314 (1972).
20.  Sulkowski, W., Dzwonnik, Z., Andryszek, C., and Lipowczan, A., Risk of occupational
    hearing loss due  to continuous,  intermittent and impulse noise. Med. Pracy, 23,
    465479(1972).
21.  Suntych, F., Kryze, B.,  and  Parizkova, B.,  Occupational diseases and Professional
    intoxications registered  in  Czechoslovakia in  1968. Prac. Lekarstvi, 22,  284-293
    (1970).
22.  Surbock, A., Hearing conservation  and noise control in industry organized and per-
    formed by  the Accident Branch of the Austrian Social Security Board. Br. Acoust. Soc.
    Spec., 1, 121-127(1971).
23.  Ward, W.D., The  concept of susceptibility to hearing loss.  /. Occup. Med., 1, 595-607
    (1965).
24.  Wojcieszyn, M., Adaptation and fatigue of hearing in workers  exposed to subcritical
    noise level. Proceedings  of XXVII Conference of Polish Otolaryngologists, Warszawa,
    PZWL, 39-41 (1970).
                                        155

-------
                  ON THE PROBLEM OF INDUSTRIAL NOISE AND
                       SOME HEARING LOSSES IN CERTAIN
                   PROFESSIONAL GROUPS EXPOSED TO NOISE

                                     J. Moskov
                                   Sofia, Bulgaria

     The wide development of industrial processes and extensive mechanization have greatly
facilitated physical  labor. The  modern way of life has created more comfortable  living
conditions. In spite of this, however, noise has  not diminished, but is still increasing. In
recent years, noise has proved to be one of the most widespread noxious factors of the
working  environment in  our country. Being an environmental factor,  it penetrates  all
spheres of life - both at the working places and in our ordinary living and social environment.
     To organize  and direct adequately the fight against industrial noise in our country it
was necessary to make first a complex evaluation of this factor, including both the physical
characteristics of the various noises encountered in the different industrial branches and also
the influence exerted by them on the workers.
     In view of this, the Research Institute of Labour Protection and Professional Diseases
undertook the task of studying the noise  produced at  most of the working places in the
basic branches.  We  took also into consideration the fact that not only the workers at the
machines and mechanical sources producing intensive noise are being exposed  to the noise,
but also these working in proximity to them.
     Simultaneously, we also carried out serial mass prophylactic examinations on groups of
workers exposed to noise. Different specialists took part in these examinations: profpathol-
ogists, otologists,  neurologists, internists, etc. In  instances of the presence of a noise effect
combined with other factors,  we also  carried out an investigation of these factors and of
their influence. Thus, for example,  we studied  also the vibrations  usually accompanying
industrial  noise in the timber and ore-mining industries; in the chemical industry, wood-
working, textile and other industries we conducted toxicologic studies and examinations of
the dust aerosols; in metallurgy and  in the smith-pressing departments we also studied the
overheating microclimatic conditions, etc.
     Besides all these complex studies,  we carried out further a threshold-tonal audiometric
examination. This examination was  performed in a labor-hygienic installation. In case  we
discovered substantial changes, the affected workers were taken over by the special otologic-
audiologic   departments  where  additional  audiologic  examinations,  including speech
audiometry, etc.,  were carried out. Besides an examination of hearing, in certain categories
of workers—mainly  those whose activity demanded considerable psychosensorial strain—we
have begun still other examinations: on the effect  of noise on peripheral vision, on the
processing of information,  reproduction of a dosed muscular strain, reproduction of a
spatial position of the hand, etc. In some  cases  we are  also studying the vibrational sensi-
tivity, heat  sensitivity, arterial pressure, pulse rate,  etc. We studied a  total of about 900
industrial objects of  14 different industries by measuring the noise  produced at several
thousand working places. A parallel study was also made of 6400 workers.
     The results of  these studies, as well as those of the prophylactic audiometric examina-
tions, are shown in the following figures:

                                         157

-------
                               Table 1.

     NOISE LEVELS IN SOME BRANCHES OP PRODUCTION* /in dB  (A)/

METALLUHGl                           50  60  70  8   9  1QO 110 120
  Cylinders for casts  cleaning
  Crushers
  Bell mills
  Cleaning devices
  Moulding departments

ELECTRICAL INDUSTRY
  Turbofeeders
  Turbine halls
  DEO

MACHINE-BUILDING AND METAL-
WORKING
  Tinsmith's departments
  Blacksmith's departments
  Work with riveting instruments
  Work with hand emery-wheels

WOODWORKING
  Circular saws
  Knife-grinding places
  Press departments
  Abricht-mschines

TIMBER INDUSTRY

SHOE INDUSTRY
PRODUCTION OP READY-MADE
CONSTRUCTION
  Vibromasses
  Vibroriddles
--«
     -•«
   Some most noisy sources and working processes are also  shown.
                                 158

-------
               Table 1 pg. 2 (Noise levels in some branches of production).
CHEMICAL INDUSTRY
  Soda production
  Production of nitrogen
  fertilizers
  Plastics production
  Antibiotics production
  Cement production
  Carbide and gunpowder production
  Other working places
  Production of bricks and cement

COAL NINES
  Separations

ORE-MINING
CRANES
ROPE-LINES
TEXTILE INDUSTRY
  Loons
  Spinning looms
  Ring machines
  Shuttles

POOD AND TASTE INDUSTRY
  Meat-manufacturing enterprises
  Sugar industry
  Cereals-manufacturing industry
  Production of non-alcoholic
  drinks

TELEPHONE CENTRALS AND TELETYPE
  Computer centres

LEATHER INDUSTRY

OFFICES, PLANT MANAGEMENTS AND
LABORATORIES

                                 159
0 6<










* -









MM^MI


9




















i


o e























o S










• -1





"






0 10











j
^ «•»

—4





,
"4


01'



.


4
•
,


•











0 1£























>0
























-------
250
WOO
600G
                                                              6OOO
    -/Q

     0

    JJ

    20
    40

    SO

    $0
    90

    JOO

    HO
Figure 1. Average loss of hearing in workers engaged in woodworking.

Length of work: — Less than 1 year;	1-3 years; -   - 4 - 6 years;	7-9 years,	10-15
years;	 15-20 years.
                                         I 60

-------
115     250    $00     /OOO
                                                        fa?    6000   &000
•JO
0
10
20
60
40
do
60
ro
#0
30
100
no

















	
— • —


























^w • •









JSOO



^












s*^









3000



X
\










,^
*" —
\











St
X
N>
'*" —











- — — -
	



















Figure 2. Average loss of hearing in workers engaged  in the production of ready-made construction
elements.

Length of work:	Less than 1 year; — 1-3 years; - • - Over 3 years.
                                            161

-------
         125    250    £00    '000   2000   4010   £000    40OO
-JO
  10

 20
  40

  60

  SO

  70

  00

  90

  JOO

  /10

_x
                    Figure 3. Average \oa of hearing in textile workers.

                                    —Spinners
                                   	Reel workers
                                   - • -Weavers
                                  - •- -Other
                                     162

-------
                   250    $00    1000    2000    4000   6000   80OO
   -10

     c

    10
60

40

 60

 60

 To
                                                           ,  •
    30
    /Oo

    /to
Figure 4. Average loss of hearing in workers engaged in the shoe industry.

Length of work: — Less than 1 year;	1-3 years; - • — 4 - 6 years;	 7 - 9 years;	10-15 years;
	16-20 years;	Over 20 years.
                                          163

-------
25*0
                                 JOOO   2000   4000
                                             MOO
w
20
40
 50
 fO
 ro
                                             \
 90
100
 HO
  Figure 5. Average loss of hearing in miners.
  Length of work: — 1 - 3 years; - - 4 - 6 years; -   - 7 - 9 years. -   - 10 and over 10 years.
                                      164

-------
               2SO    SDO    JO00   20DO   400G   G0OO   £000
-/O
                                   /500
jooo
20
30

So
 50
 70
 60
 90
 100
 110
       Figure 6.  Average loss of hearing in workers engaged in the timber industry.
       Length of work: — 1  • 3 years; — — 4 - 9 years; — • — 10 and over 10 years.
                                    165

-------
                SOD    fOOO   2000   30 OO   4000   SOffO  #000
'10

  0

 fO

2.0
40

 $0

 60
 no
           Figure 7. Average loss of hearing in crane and rope-line workers.

                                —Crane workers
                               	Rope-line workers


-------
         /Z5     250    500    1000   2000  4090    ***  4000
                                              3000
   20
   30
   40
   60
   60
    TO
   SO
   90
   no
Figure 8. Average data of hearing in professions unexposed to noise (administrative personnel, laboratory
workers).
                                        167

-------
     The Table indicates the average results of the noise data. The audiometric results (Figs.
1-8) are hearing levels after the age correction suggested by Hinchcliffe.
     The noise measurements are all card-indexed. A separate card is made for each  work
area studied. These  separate cards contain data about  the  kind of production and the
professions by which work is performed under the effect of noise, the category and number
of workers exposed to this noxious influence. Data are also entered concerning the duration
of exposure at the various working places.
     The purpose of our studies is the making of a contribution to elucidating the  noise
status in  our country. They are intended to draw the attention of the audiologist, industrial
health pathologist, the trade union, administrative and other bureaus, to the noise problem.
     On  the  basis of these studies, we worked out a number of prophylactic programs,
instructions for the fight against noise in various branches,  hygienic norms of industrial
noise, etc. A number of important problems still stand before us, waiting to be solved. Thus,
for example, there is still no solution to the problem of the degree of hearing needed for the
various professions; no Bulgarian standard has been worked out, as yet, for age correction in
audiometric examinations. At present,  we either do not use corrections for age changes,  or
we use those taken from foreign authors, worked  out for the population of other countries,
based on different demographic and other indices. The criterion  for  determining the work-
ing capacity of a person, in an industrial hygiene context, should depend on his profession.
Speech audiometry has to be introduced (as a rule) in establishing the  social adequacy of the
examined individual.
     Pre-employment and monitoring audiometric examinations  of workers engaged in the
so-called  noisy professions have already become standard practice. We are also having good
results in the field of preliminary sanitary control. The noise factor  is being already taken
into consideration in giving a permit and approval for different devices and machines, for
factory departments, for building a plant, etc.
     The  training programs  of medical  students and  doctors  specializing in industrial
hygiene already include a section  on noise  and vibrations.  During recent years, several
symposia and national congresses were held in our country on the problems of noise, with
participation of a number of distinguished specialists from  many foreign countries. The
development of a special law for the fight against noise and vibration is in prospect and this
will be an important step in the difficult fight against these disagreeable and noxious factors.
     We  hold  to the  concept that in establishing the noise norms in the future, both for
industrial noise and for  noises in our ordinary living and social environment, we should also
take into consideration  their extraaural effects, which in many  cases are being manifested
early, at levels that do not lead to changes in hearing.
     The fight against noise is a complex problem. The achievement of favorable results will
be possible only on the  basis of the mutual efforts of specialists from many different fields
and with the cooperation of the administrative and social organs and organizations.
                                         168

-------
               NOISE-INDUCED HEARING LOSS FROM EXPOSURE
                   TO INTERMITTENT AND VARYING NOISE

                              W. Passchier-Vermeer
                Research Institute for Public Health Engineering TNO
                              Delft, The Netherlands

INTRODUCTION

    The original intent of this review  was to consider exposure to intermittent noise only,
but exposures to varying noise have been included as well.  Since varying noise includes
intermittent noise, the  latter being nothing but noise varying between a "high" and a "low"
noise level and since it is quite unclear what "low" in this respect means, it  seems advan-
tageous to start from the more general subject of exposures to varying noise.
    At the moment no field studies are  known which give relations between exposures to
varying or intermittent noise  and  noise-induced hearing loss. In the  second part of this
review, an attempt  has been made to relate both  quantities, by using data from several
papers. Exposure to  impulsive noise, such as gunfire, is not referred to.
    In the first part of this paper, those noise exposure limits will be reviewed which refer
to exposure to intermittent and varying noises. All these  limits  are based on temporary
threshold shift by assuming that there exists a certain relation between temporary threshold
shift and permanent threshold shift.
    This review does not discuss criteria, the basic limits of safe noise exposure, concerning
the percentage of people  to be protected from so and so many decibels hearing loss at these
and those frequencies after  so and so many years of exposure, related to this or that group
of non-noise exposed people. This subject seems to be sufficiently covered by other papers
presented at the Congress (Glorig, Kry ter).

I.   TEMPORARY  THRESHOLD  SHIFT FROM  EXPOSURE TO VARYING AND IN-
    TERMITTENT NOISE

    1.  Damage risk contours prepared by NAS-NRC CHABA Working Group 46 (Kryter,
        Ward, Miller and Eldredge 1966)
        Before discussing  these contours, it should be pointed out that a tremendous
        number of  TTS  experiments  form the  basis  of these contours.  Criticism on
        extrapolations used in the derivation of these contours should be seen in the light
        of this remark. The contours are based on three postulates, (1) TTS2 (Temporary
        Threshold Shift measured 2 minutes after the  end of a noise exposure) is a
        consistent measure of a single day's exposure. This  is supported by evidence that
        TTSs maintain  their rank order  during recovery (Ward et al. 1958, 1959a) and
        that recovery does not depend on how TTS was produced (Ward et al. 1959b). (2)
        All exposures that produce a given TTS2 are equally hazardous as far as NIPTS
        (Noise-Induced Permanent Threshold Shift) is concerned (3) TTS2 is about equal
        to NIPTS  after ten years  of exposure. Noise exposures with parameters lying on

                                       169

-------
         the damage risk contours should restrict the average TTS2 at 1000 Hz and below
         to  10 dB,  at 2000 Hz to  15 dB and at  3000 Hz and above to 20 dB. These
         TTS2-values are called in the following criteria TTS2S.
             Three sets  of contours were constructed. They  were derived by using equa-
         tions, expressing the increase of TTS2 at several frequencies with exposure time,
         for constant octave band noise with  octave band  sound pressure level as the
         parameter, and  by using general frequency-independent curves for the recovery of
         TTS after exposure. These three sets are:
         a.   Damage risk contours for a single exposure daily. These contours are shown
             in  figure 1. In preparing these contours, extrapolation to the longest dura-
             tions and to the shorter ones was involved.
         b.   Damage risk contours  for  short  burst intermittent  noise (noise  bursts  2
             minutes or less in duration, alternating with effective quiet). These contours
             are  mainly  based on  the on-fraction rule (Ward 1962), which states that
             when noise alternates with quiet and is on for \% of the time, but no longer
OCTAVE  BAND  SOUND
PRESSURE  LEVEL IN  dB
         uo
          DURATION
           r!N,
          MINUTES
                           34561
34568.
           100                         '1000                          10000

                        OCTAVE  BAND  CENTER  FREQUENCY  IN  Hz
    FROM  K.D.KRYTER  ET AL

Figure 1. CHABA damage risk contours for one exposure per day to octave bands of noise. This graph can
be applied to the individual band levels present in broad-band noise (From Kryter et al, 1966).
                                      170

-------
    than 2 minutes per exposure cycle, the resulting TTS2 is x% of the TTS2
    produced by a continuous  exposure  for  the  same time. For frequencies
    below 2000 Hz,  this rule needed modification, due to  the action of the
    middle ear muscles. Again, extrapolation was involved in  applying this rule
    to higher octave band sound  pressure levels and lower on-fractions.
c.   Damage  risk contours for longer burst intermittent noise. In deriving these
    contours the "equivalent exposure rule" was used.  According to this rule,
    the residual TTS present from one noise, exposure at the start of the second
    exposure is converted to the time it would take this second noise to generate
    an amount of TTS equal to the residual TTS; this equivalent exposure time is
    added to the exposure time  of the second noise to calculate the TTS2 at the
    end of the second exposure. This rule was only shown to hold for TTS  at
    4000 Hz. Also, it was shown in the relevant paper (Ward  et al, 1959c) that
    recovery during the quiet intervals did not proceed in the expected way, but
    that recovery from higher  TTS-values was slower  than  from lower TTS-
    values. Nevertheless, the equivalent exposure rule was expanded for other
    frequencies  as well and  in deriving the damage risk contours for longer burst
    duration intermittent noise, the general recovery curves were used.
An important shortcoming  of the CHABA report concerns  the definition  of
"effective quiet": According to the report, and ignoring Ward's objections (Ward,
1966), effective  quiet is assumed to exist whenever the  noise level drops below
the octave band  sound pressure levels allowable for 8 hours a day (89 dB SPL for
the octave band with midfrequency 500 Hz, 86 dB at  1000  Hz, and  85  dB  at
2000 Hz and  400 Hz).

However,  the recovery curves used in the report were established for quiet and
they  may not be valid for 85 to 89 dB octave band SPL's. Figure 2 shows the
effect of certain  noise levels on  recovery as found  by Schwetz et al. (1970), by
Lennhardt et al. (1968) and by Ward et al. (1960). Schwetz found that recovery
from  TTS at  1000, 2000, 3000 and!4000 Hz is statistically significantly retarded
in white noise with an overall sound pressure level of 75 dB (which might be equal
to at most 70 dB SPL per octave band) and still more in white noise of 85 dB.
Results from Lehnhardt  are even more pronounced than those  from  Schwetz.
Although white  noise of 70 dB  overall  SPL (probably equal  to  at most 65 dB
octave band  SPL) allowed the same recovery of TTS at 2000, 3000,4000, 6000
and 8000 Hz as  in quiet, white noise of 80 dB SPL allowed recovery only during
the first 15 minutes or so after the noise exposure; after that time TTS increased
again! A similar effect was shown by Ward. He considered the course of TTS after
an exposure  to 105 dB SPL octave band noise, immediately followed by exposure
to 95 dB SPL octave band  noise. Ward's interpretation, which clearly emerged
from his results, is that the "excess" TTS (i.e. the difference between the TTS
produced by  the 105 dB noise level and the TTS that would have been produced
by an exposure for the same time to the 95 dB noise level) recovers in the 95 dB
noise level as in quiet, independently from  the simultaneous growth Jof TTS

                                171

-------
 TTS
                                                 Ward
                                                         Lehnhardt
                                                      75 dB
      20
      10
                               Schwetz
                                        Quiet
65 dB
   Lehnhardt
  Quiet
                  5             20       50           200
                     MINUTES  AFTER  EXPOSURE
Figure 2. Average temporary threshold shift as a function of the time after exposure, according to —
Schwetz (1970), TTS averaged over 1000, 2000, 3000 and 4000 Hz - Lehnhardt (1968), TTS averaged
over 2000, 3000. 4000. 6000 and 8000 Hz - Ward (1960), TTS averaged over 3000 and 4000 Hz Octave
band sound pressure level is parameter.
                                   172

-------
attributable to the 95 dB noise level. (By the way, by accepting the concept of
"excess" TTS, it should also be accepted that the course of ITS at a later stage
does depend on how it was produced). However, Ward's interpretation does not
hold for Lehnhardt's results, since 75 dB SPL octave band noise alone does not all
cause that much TTS during the time considered.

Klosterkotter (1971) stated that recovery from TTS in a sound level of 70 dB(A)
was slower  than recovery in 35 dB(A), with a difference in recovery of 7 dB.
Details cannot be given here, since the original paper was not at hand at the time
this paper was written.

All in all, it seems that only octave band sound pressure levels of at most 65 dB
should be considered to be "quiet", in view of recovery from TTS. This now is of
great  importance,  when  considering industrial situations.  Although CHABA
damage risk contours  for intermittent noise might be applicable, with effective
quiet at 65  dB SPL or lower, it seems that their use is limited to a few industrial
situations only.

CHABA rules are  also given for varying noise levels, at least when they do not
remain at any level for  more than two minutes. The effective level of such a
varying noise is equal  to the time-average sound pressure level of the noise over
the  exposure period.  Again,  however this is a  broad generalisation of the  one
relevant research (Ward et al, 1959a), in which it was only shown to be correct for
TTS at 4000 Hz due to exposure to noise of alternate 30 sec.-periods of 106 and
96 dB SPL. The way in which varying noises with levels remaining for more than
2 minutes at different values should be treated, is in fact not known at all. Only in
one publication, just cited (Ward 1960) was the subject touched on.

In 1970, Ward conducted  new TTS-experiments to determine,  as he states, the
degree to which the CHABA damage risk contours are in error (Ward 1970). In
general, it turned out from all experiments that the TTS2-values due to exposure,
chosen from the CHABA damage risk contours, were in the range of about 70%
up to 115% of the criteria TTS2's. Looking more closely at the results concerning
TTS2, it was shown that:
(1)   Single uninterrupted exposures up to 8 hours meet the criteria TTS2 within
     10%.
(2)   Short-burst intermittent noise does not meet the criteria TTS2- In two out
     of three experiments TTS2 after exposure was only about 70% of the criteria
     TTS2.  However,  the  other  experiment was terminated after 6 hours, al-
     though according to the CHABA contours 8 hours exposure was permitted,
     because of large values of TTS in some ears. Anyhow, even 6 hours exposure
     resulted in an average TTS2 above the criterion TTS2-
(3)   For long burst  intermittent  noise,  the criterion TTS2-values are exceeded.
     Again,  in these experiments it was demonstrated that recovery during quiet

                               173

-------
     intervals does not proceed according to the general recovery curves, but that
     recovery during successive intervals is retarded, thus not permitting sufficient
     recovery from TTS during quiet intervals.
Thus Ward's  latest published results on TTS  indicate that recovery does  not
necessarily  follow the general  recovery curves used  in  the  derivation  of  the
CHABA damage  risk  contours.  Although these curves seem to be all right for
single uninterrupted noise exposures, intermittent exposures to high frequency
(above 1500 Hz or so) high level noise, either long or short  bursts, often produced
a delayed recovery. This delayed recovery was in earlier experiments (Ward et ah,
1958,  1960)  shown to occur, after single uninterrupted  exposures, only when
TTSo was more than about 40 dB. In these experiments, however, for high level
noise it already  occurred when TTS2  was  only about  20 dB. Most alarming,
however, is the  fact  that TTS2 is not a consistent measure  of a single day's
exposure,  since TTSs do not maintain their rank order during recovery,  and
recovery from TTS does depend  on how TTS was produced. Ward, then, looking
for a practical solution suggested that TTS3Q (TTS measured  30 minutes after the
exposure) might be a useful index, since after  30 minutes the  rank order of TTS is
more or  less constant. Although this  may  be right,  it is quite unclear which
relation exists between TTS30 and NIPTS for  exposures to  intermittent noise.

Contrary to the equal-energy principle (which applied to TTS2 states that TTS2S
resulting  from exposures to noises with the  same total sound energy are equal,
irrespective of the distribution of  the energy over the exposure period), it has
been shown throughout all TTS-experiments that the distribution of the sound
energy does make a difference in the TTS2 produced.

Analogously  it has been shown  that the same TTS2  is caused by exposure to
noises with quite  different total sound energies. In figure 3 this is again shown for
Ward's latest published  results (Ward,  1970) for TTS2 at 3000 Hz, caused by
exposures to  quite different noise patterns (single exposure daily, long and short
noise bursts). In this figure, TTi>2 has been plotted against the energy-equivalent
sound  pressure level (Leq). Since the definition of Leq is given later, it is suffi-
cient to state here that Leq is nothing but a measure of the total sound energy for
an 8 hour exposure, converted into a sound pressure level.  Although Lgq in figure
3 has a range of 16 dB (from 83 to 99 dB), which corresponds to a factor of 40 in
sound energy, TTS2 is about  20 dB for all exposures. However, plotting TTS200
(200 minutes being the longest recovery period examined in all tests) against Lgq,
results in a appreciable increase of TTS2QO w^n Leq-

From this  it  is clear  that recovery from TTS is dependent on the sound energy
which  created the TTS. Since  it  is unknown at the moment which processes
underly the realization of permanent threshold shifts and  how TTS is involved in
these processes, it may be possible that recovery, and  hence, total sound energy
over a workday, plays  a more important role than TTS2 alone.

                               174

-------
         TTS  AT
         3000 Hz
          30i-
          dB
          20
           10
                            A TTS2

                            O TTS200
0
80
                                     I
                                                  I
                                    90          dB        100

                      EQUIVALENT SOUND  PRESSURE  LEVEL
Figure 3. Temporary threshold shift (TTS) at 3000 Hz, measured 2 and 200 minutes after exposure to
noise, as a function of the equivalent sound pressure level of the noise.
                                   175

-------
2.    In  1967 Botsford simplified the CHABA damage risk contours.  He recognized
     that all octave-band SPLs on the same contour for one single exposure daily (see
     figure  1), were  always  assigned practically  the  same exposure time limits for
     exposures to intermittent noise (long bursts, as well as short bursts). For instance,
     the CHABA damage risk contours for exposures to intermittent noise show that
     95  dB SPL at the octave band with midfrequency 1000 Hz and 90 dB SPL at the
     octave band with midfrequency 400 Hz (lying on the same contour of figure 1)
     both require, after an on-time of 55 minutes, an  off-time of 60 minutes; and for
     short noise  bursts, in both instances an on-fraction of  0.6 is allowed  for an
     exposure  time of 480  minutes  a  day. This fact,  that all octave  band  sound
     pressure levels on the same contour  have  the same exposure time limits for
     intermittent noise, irrespective of the exposure pattern assumed, indicates that
     the contours are general curves of equinoxious noise. The  next step of Botsford
     was to assign to each  contour an A-weighted sound level, by comparing from 580
     manufacturing noises their height  of penetration of their octave  band  sound
     pressure levels into the contours with their A-weighted sound level.
         The  results of the analysis by Botsford are shown in figure 4. The three
     curves with the highest sound levels should not be  relied on, since they come from
     an extrapolated CHABA-curve.
3.    In the "Guidelines for noise exposure control" (1970) of the Inter-society Com-
     mittee, the Botsford curves have been used; these guidelines have been included in
     the Department of Labor Standards-Walsh Healy Act-(Federal Register 34,1969)
     and are legally applicable  in the USA to industries performing work under the
     Governments Public Contracts Act.
         In the  Guidelines,  the curves given by Botsford have been modified some-
     what by shifting the 90 dB(A) curve to 480 minutes and by shifting the curves for
     the  lowest numbers of exposure cycles per day (up to 3 exp. cycles per day) to
     higher total on-time values per day. Apart  from  this figure in tabular form, the
     Guidelines also present the simple rule that exposure to 90 dB(A) is allowable for
     a full 8 hours, with an increase  of 5 dB(A)  for each halving of exposure time. As
     the  document states:  here an allowance is made for the number  of occurrences
     ordinarily found in high level noise. Referring to figure 4, this rule is overpro-
     tective when the noise  comes in short bursts,  but is highly underprotective for
     single uninterrupted exposures. E.g. calculating the TTS2,  due to an exposure for
     30 minutes to 110 dB(A)  (which is allowable  according to the rule mentioned)
     results in a TTS2 at 2000, 3000 and 4000 Hz of 25, 36 and 33 dB resp. which is
     on the average more than  10 dB above the TTS2-values from an 8 hour exposure
     to a constant sound level of 90 dB(A).
         A rule for varying noises is also given.  The ratio of the time spent at  a given
     sound level  to the allowable time at that level is calculated  and these fractions for
     all  occurring sound levels are added. If the resulting number is less than 1.0, the
     exposure is safe, and if it is more than 1.0, the exposure is unsafe.
         Here, sound levels below 90 dB(A)  do not enter into the calculations, al-
     though exposure to sound levels just below  90 dB(A) is hardly safe for 8 hours a

                                   176

-------
TOTAL  ON-TIME  PER
DAY  IN  MINUTES
          400

          200

          100
           60
           40

           20

           10
            6
             2
             1
                                        A-WE:IGHI   D
                                                   LEVEL
                          FOR  MANUFACTURING
                                   NOISES
                         4 6  10   20   40 60  100  200  UP
                        NUMBER  OF   EXPOSURE
                        CYCLES   PER  DAY
Figure 4. Total duration of a noise allowable during an 8-hour day as a function of the number of exposure
cycles per day. An exposure cycle is completed each time the sound level decreases to or below 89 dB(A).
The interruptions of potentially harmful noise are assumed to be of equal length and spacing so that a
number of identical exposure cycles are distributed uniformly throughout the day. The A-weighted sound
levels assigned to the curves were determined from manufacturing noises and may not apply to noises from
sources of other types.
                                177

-------
         day. A more realistic approach should have been to take into account sound levels
         below 90 dB(A) as well.
    4.   The American Conference of Governmental Industrial Hygienists (1968) proposed
         threshold limit values for noise, allowing 92 dB(A) for 4 to 8 hours a day, 97
         dB(A) for 2 to 4 hours, 102 dB(A) for 1 to 2 hours and 107 dB(A) for less than 1
         hour. In evaluating exposure to varying noise, the same line as in the guidelines
         was followed, exposures to sound levels of less than 92 dB(A) do not enter into
         the calculations. However,  one single  exposure for one hour  to  107 dB(A)
         (assuming a spectrum according to the CHABA equinoxious curves) results in an
         average TTS2 at 1000 and 2000 Hz of 25  dB and at 3000,4000 and 6000 Hz of
         40 dB. It seems difficult to understand why a conference of hygienists proposed a
         limit for exposure to constant noise greater than 90 dB(A)—a level which is now
         more or less generally agreed to be the maximal allowable limit—and also pro-
         posed limits for varying noise that are even less stringent than those for constant
         noise.

II.  NOISE-INDUCED HEARING LOSS FROM  EXPOSURE TO INTERMITTENT AND
    VARYING NOISE.
    Selection data
    A thorough study of the relevant literature  has been undertaken to select papers in
    which data were given that  could  enable us to relate noise exposure to noise-induced
    hearing loss for exposures  to varying  and  intermittent noise over  the workday.  In
    selecting papers the following considerations were taken into account:
    (1)  HLs  of the subjects measured a considerable time (mostly more than 12 hours)
         after their last exposure to job noise,  to permit significant recovery from tem-
         porary  threshold shift  from such noise. From the group of miners reported by
         Sataloff et al. (1969) audiograms were taken right after coming up from the
         mines, but in a pilot study it was shown that no significant TTS was at that time
         included in the hearing levels.
    (2)  Subjects selected with no previous exposure to noise at other jobs nor any prior
         ear damage or clinical abnormality. Although Sataloff reports exposure to gunfire,
         his group of miners was nevertheless included. Reasons will be given below.
    (3)  Number of subjects at least 25 per group, unless data were taken  from  a sub-group
         of a larger group. Selection  of a particular subgroup was based on number of
         subjects and exposure time.
    (4)  Total exposure time preferably more  than 10 years. When data were given for
         shorter exposures only, but for more than 4 years, these data have been included.
    (5)  Occasional wearing of ear protection at the time of the survey, reported in two of
         the papers, is included in our analysis. Both, however, report  a long service
         without ear protection and no differences were found between  the men wearing
         ear protection and those without ear protection.
    (6)  Noise exposures reported to be to several sound levels, the difference between the
         highest and lowest sound level  at  least 25 dB(A) or so. Only those surveys were
         included, for which it was sure that the noise environment did not change over

                                       178

-------
    the years. Papers dealing with exposure to noise fluctuating on a short time scale
    (impulse, impact noise) have not been included in the analysis.
(7)  Sufficient data on noise exposure to allow the calculation of a characteristic noise
    parameter of such  exposures.  Data on  the  noise  exposures were given  in the
    following ways:
    —    Overall time-distribution of  sound levels over the workday (e.g. x%  of the
         time the sound level lies between a and  a+5 dB(A), y% of the time between
         a+5 and a+10  dB(A), z%  etc.) This includes also the exposure to one par-
         ticular sound level for x% of the workday, the rest of the time being "quiet'*.
    —    time distribution over the workday of mainly one sound level (e.g. 3 minutes
         exposure to  x  dB(A), followed by 5 to 10 minutes  of "quiet")-  Unfortu-
         nately, only a few authors,  from which the data were included  in the
         analysis, give the numerical values of the sound levels  during the quiet inter-
         vals. Most authors describe the quiet intervals in a qualitative way.
Through  this stringent procedure of  data  selection,  out of about hundred papers,
(listed at  the end of this paper under the heading: further literature consulted), eleven
papers, dealing with 20 groups of subjects, could be included in the analysis.

Presentation of data

Unfortunately, insufficient data have been included in the papers about the  spread in
the hearing levels to permit any analysis of this very important subject. Therefore, the
following refers only to median and  average hearing levels and median and average
noise-induced hearing losses.
    All median (or in a few instances, average) hearing levels presented in the papers
have been converted to ISO standards, if necessary. The values given by Spoor  (Spoor
1967) and shown in figure 5 of the  age-dependent median hearing levels of non-noise-
exposed otologically normal people have been subtracted from the median and average
hearing levels of the groups, to calculate the median noise-induced hearing losses.
    Since the mean ages of the several groups are mostly around 40 years, only small
values had to be subtracted from the actual hearing levels (see Table I).
    As most noise data were presented in sound levels in dB(A), those given in octave
band sound pressure levels have been converted into sound levels in dB(A) too.
    An attempt has been made to give a classification of the noise exposures based  on
details given in the papers. Although it is realized  that intermittent noise is included in
varying noise, for the purpose of this paper a more specific definition of varying and
intermittent noise  is given.  Intermittent noise is here defined as noise with  a large
difference (at least 20 dB(A) or so) between the highest and lowest sound levels, and
where  sound levels between these  levels  are present during a negligible tune only.
Ygrying noise is here defined as noise in which several sound levels occur in the course
of time and where sound levels between the highest and lowest sound levels are  present
during a  considerable time. Looking at the times during which the sound levels remain
at a given level, the noise exposures may be grouped in exposures with short and long
times at  a  given sound  level. The limit of short-time noise exposure has been, quite

                                    179

-------
                                        FREQUENCY  IN   Hz
                           500      1000
2000     £000     8000
       3000     6000
                       ""30
                         35
MEDIAN   HEARING  LEVEL

Figure 5. Median hearing levels of otologically normal men. not exposed to noise during working hours, as a
function of frequency. Age in years is parameter.
    arbitrarily, chosen to be  5 minutes. None ot" the noise exposures to be considered here
    have exposures at a given level of less than 2 minutes (taken as the CHABA-limit for
    short noise  bursts).  The 4 resulting  classes are  illustrated in  figure 6. To give  an
    indication of the variations involved, for the varying noise exposurcs the difference is
    given between the sound level exceeded  for  27r of the  time and the sound level
    exceeded  for 5(KT of the time. For the intermittent noise exposures, the total times per
    workday at the highest sound levels are given (see Table I >.
                                     180

-------
      INTERMITTENT  NOISE
      SHORT  INTERVALS
          P.    n.  _
    INTERMITTENT  NOISE
    LONG  INTERVALS
           H	1—I Mill!	1	1—I Mill
00
                                       • 10dB(A)
H	1—I  MINI	1	1—I  I I I I II
      VARYING  NOISE
      SHORT INTERVALS
    5min
    VARYING  NOISE
    LONG  INTERVALS
           Figure 6. Illustration of classification of noise exposures into intermittent and varying noise exposures with
           long and short periods at a given sound level.

-------
         From  the noise data, for each group, the A-weighted energy equivalent sound
     level has been calculated according to

                                      !  t __LA(t')/10
Leq=101ogi/ 10            dt'
              o
     where t is the total daily exposure time (480 minutes) and LA(tr) is the momentary
     sound level at time tf.
         When  the  various levels are given in n classes, then the above formula can be
     rewritten as
                                       i= 1

     where Lj is the average sound level of class i (assuming the class widths are small) and t j
     is the total time per workday, during which  the sound levels are within class i.
          An analogous formula, although modified for a total exposure time of one week
     (40 hours) has been used in the ISO Recommendation  1999  (Assessment of noise
     exposure during work  for hearing conservation  purposes), to determine equivalent
     continuous sound levels. Essentially, the formulae given above calculate the A-weighted
     total sound energy per reference period t, converted into a sound level.
          Relevant data have been included in Table I.

Analysis of data

     Although not necessary, it seems advantageous to compare data for varying and inter-
mittent noise with data for constant noise. The data given here permit comparison with two
sets of results for more or less constant noise, namely those of Burns and Robinson (1968)
and those of Passchier-Vermeer.
     The results of the report by Passchier-Vermeer relate to continuous 8-hour exposures
to constant steady -state broadband noise for exposure times of at least 1 0 years. It is merely
a compilation of data found at that time in the relevant literature  (Bums (1964), Gallo
(1964), Subcommittee Z 24-X-2 (1954), Rosenwinkel (1957), Nixon (1961), Taylor (1961),
Kylin (1960), and Van Laar (1964).  Without going into any  detail here, the results are
shown in  Fig. 7, for  the frequencies 2000, 3000 and 4000 Hz and for an exposure time of
1 5 years.  This exposure time has been chosen  as being the average exposure time of the
groups under consideration here (see table I). To make a comparison possible, the median
noise-induced hearing losses due to the varying and intermittent noise exposures had to be
calculated for an exposure time  of 15 years  as well. Although this involves mostly slight
corrections, these corrections have been applied to the values given in the Table. It should
be noted here too that in order to take into account age-effects on hearing, the values given
by Spoor have been  applied  to both sets of data. The results are shown in Figs. 8 to 1 3 for
the frequencies 500 to 6000 Hz.

                                        182

-------
                                                            Table
Author     number    mean      average
           of                  e::p.
           subjects  age       time
                     in  years   in years
equivalent  indication of   exposure long
sound  •     intermittent or  or short times
level in    varying noise   at  the prevail-
dB( A)                       ing sound levels
average or median noise-induced
hearing loss (age-corrected
according to Spoor) at
500Hz lOOOHz 2000Hz 3000Hz 4000Hz 6000Hz
Cohen
Cohen
Taylor
Feiser
Feiser
Flach
Oo Kuiper
U> 1
Kuiper '
Ephraim
Holmgren
Holmgren
Pressel
Schneider
Schneider '
Satalof f
Jirak
Jirak
Jirak
Jirak
Jirak
18
14
40
16
16
233
144
23
97
59
26
230
69
26
26
23
25
41
40
50
37.5
30
35
45
38
57
44
41
38
33
35
45
42
39
36
38
40
40
12.5
17.5
4
7.5
12.5
15
20
40
10
10
9
4
10
20
15
17.5
12.5
15
15
15
86
86
82
102
102
82
99
99
95
95
98
94
83
83
ll'i
103
88
100
74
108
int. 120 min
int. 120 min
int. 75 min
vary. 20dB(A)
vary. 20dB(A)
vary. 25dB(A)
vary. l5dB(A)
vary. l5dB(A)
vary. l4dB(A)
vary. l4dB(A)
vary.!2.5dB(A)
int. 120 min
vary. 15dB(A)
vary. !5dB(A)
int. 180 min
int. 190 min
int. 115 min
int. 280 min
int. 250 min
int. 212 min
short
short
short
long
long
long
long
long
long
long
long
long
long
long
short
short
short
short
short
short
1
-1
-
11
12
1
-
-
2
2
-
0
12
-
-
-
-
-
2
1
-1
14
14
1
-
-
2
5
-
0
1
17
_
-
-
-
-
3
7
2.5
21
26
0
16
27
15
4
8
10
0
29.5
5
0
3
0
11
3
13
4
26
33
-
30
35
30
16
14
12
0
1
34
-
-
-
-
_
5
10
5.5
30
34
2
38
35
34
18
16
20
2
3
36.5
27
12
20
5
28
12
20
7.5
19
26
-
22
25
25
18
17
21
6
14
38.5
-
-
-
-
-

-------
     The results of Burns and Robinson do not only refer to eight-hour continuous expo-
sure to a constant sound level; noise exposures with fluctuating sound levels on a short time
scale (in the order of seconds) were also involved in their analysis. The authors relate
noise-induced hearing loss to noise immission level E^2> this quantity being equal to LA2 +
10 log T, where T is the total exposure time (e.g. in years) and L^2 is  the sound level
exceeded in 2% of the time during a workday. In their survey, L^2 ~ ^A5Q ranged from 0 to
10 dB(A), but was as much as 15 dB(A) in exceptional cases. As their report states:  "It is
more convenient to express the noise level in dB(A) in the  usual way (by meter reading)
rather than in  the form L^2> and the  average relation between the two (based  on 280
specimen noises in the survey) may be taken as L^-LA = 3.7 dB(A). Since it was found
that Lj\2 corresponds better with hearing loss than LA, using  LA instead of LA2 will be less
exact  for strongly fluctuating  noise environments". All in  all, for a constant or nearly
constant noise level, noise immission can be expressed in the form LA + 10 log T, where LA
is determined by sound level meter reading and is, according to the measuring characteristics
of sound level meters, equal to Leq.
     To permit comparison between the results of Passchier-Vermeer and those of Burns
and Robinson, for both surveys the median NIHLs at 2000, 3000 and 4000 Hz have been
plotted against Leq for an exposure time of 15 years (figure  7). Although it does not seem
the right occasion  here to go  into details about agreements and differences between the
results of both surveys, nevertheless some attention has to be paid to this subject now.
     There  is a good agreement between the median  NIHLs at 2000  Hz and lower fre-
quencies and, to a lesser extent, at 6000 Hz. However, at 3000 Hz and especially at 4000
Hz, large differences can be shownr at the highest sound levels considered. According to
Bums and Robinson, this might be mainly due to different subject-selection criteria used by
Bums and Robinson and  by the authors  of the several papers from which the values given
by Passchier-Vermeer have been derived.
     Only the fourth selection criterion used by Bums and Robinson was not used in any of
the other studies, namely that  audiograms should be compatible with clinical findings. I am
at a loss to indicate the consequences of this criterion. Also for a few of the groups used in
Passchier-Vermeer's analysis, the selection criterion  concerning military service was not as
stringent as in others,  but median results from these groups could not be distinguished from
the other median NIHLs. Moreover, subject selection does not explain that the differences
do increase with sound level. If subject-selection  were the appropriate cause, one would
expect that the  curves should not be divergent but parallel. Looking further for a possible
explanation of the  differences found (noise measurements, variance of hearing levels), none
seem to explain these differences.
     However, left  with two sets of curves for exposure to constant  noise, it seems advan-
tageous to compare the median NIHLs from the exposures to intermittent and varying noise
with both sets of curves.  As already mentioned, in figures 8 to 13 the median NIHLs from
intermittent and varying noise exposures have been plotted against Lgq,  along with  the
curves given by Passchier-Vermeer for constant noise.
     In, Figs. 14 to 19 the median NIHLs are plotted against Lgq + 10 log T, along with the
curves given by Burns and Robinson. Since in Burns and Robinson's survey, median NIHLs
have been calculated from the  actual median HLs, by subtracting median HL of non-noise

                                        184

-------
        MEDIAN
        NOISE-INDUCED
        HEARING  LOSS
            50
            dB
                 PASSCHIER
               _ ROBINSON.
        4000 Hz
                                                        4000 Hz
                                                      / 3000 Hz
                   EXPOSURE TIME
                      15 YEARS
            30
                                   90
100
SOUND
 110 dB(A)
LEVEL
Figure 7. Median noise-induced hearing losses at 2000, 3000 and 4000 Hz, caused by continuous exposure
to steady-state broadband noise for 15 years, as a function of the sound level in dB(A). Curves presented by
Burns and Robinson (1968) and Passchier-Vermeer (1968).
                                 185

-------
    MEDIAN
     NIHL
     I
 20
dB
                EXPOSURE TIME    15 YEARS
       10
            • VAR
            • INT
500 Hz
                                                     PASSCHIER
               80
                                  90          100         110 dB(A) 120
                               EQUIVALENT  SOUND  LEVEL
Figure 8. Median noise-induced hearing losses at 500 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.

Figure 9. Median noise-induced hearing losses at 1000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.
                                       186

-------
  MEDIAN
  NIHL
       30
EXPOSURE TIME:  15  YEARS
                    80
         90          100         110 dB(A) 120
       EQUIVALENT  SOUND  LEVEL
Figure 10. Median noise-induced hearing losses at 2000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.

                                    187

-------
    MEDIAN
    NIHL
      50
EXPOSURE TIME:  15 YEARS
                               90         100         110 dB(A) 120
                          •^EQUIVALENT  SOUND   LEVEL
Figure 11. Median noise-induced hearing losses at 3000 Hz from exposure to varying and intermittent noise
for 15 yean, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968} for exposure for 15 years to steady-state broadband noise.

-------
   MEDIAN
   NIHL
     50
      EXPOSURE  TIME :   15 YEARS
      70
80
  90          100
EQUIVALENT  SOUND
110  dB(A) 120
LEVEL
Figure 12. Median noise-induced hearing losses at 4000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermee
(1968) for exposure tr»r 15 years to steady-state broadband noise.
                                    189

-------
MEDIAN
NIHL
                   EXPOSURE  TIME:  15 YEARS
                                               PASSCH  ;R
                           90          100         110 dB(A) 120
                          EQUIVALENT SOUND  LEVEL
Figure 13. Median noise-induced hearing losses at 6000 Hz from exposure to varying and intermittent noise
for 15 years, as a function of the equivalent sound level in dB(A). Curve presented by Passchier-Vermeer
(1968) for exposure for 15 years to steady-state broadband noise.

                                190

-------
cq.
                                                             L°9
Figure 14. Median noise-induced hearing losses at 500 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (L^  + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

-------
Figure 15. Median noise-induced hearing losses at 1000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (l_eq + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

-------
Figure 16. Median noise-induced hearing losses at 2000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (L^ + 10 log T. where T  is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

-------
                                                 105           115
                                                LCq.  +  10  Log  T
Figure 17. Median noise-induced hearing losses at 3000 Hz from exposure to varying and intermittent noise.
as a function of the noise immission level (!_ + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or tess constant noise.

                                            194

-------
           MEDIAN
           NIHL
                                                            4000 Hz

                                                       ROBINSON
                                            105          115
                                            Leq. + 10 Log  T
Figure 18. Median noise-induced hearing losses at 4000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (L™ + 10 log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

                                        195

-------
             MEDIAN
             NIHL
                                               105          115
                                               Leq   +  10 log  T
125
Figure 19. Median noise-induced hearing losses at 6000 Hz from exposure to varying and intermittent noise,
as a function of the noise immission level (Lgq +  10  log T, where T is the exposure time in years). Curve
presented by Burns and Robinson (1968) for continuous exposure to more or less constant noise.

                                           196

-------
exposed people that differed from those given by Spoor, the values given in the Table have
been adjusted in order to be comparable with the curves given by Burns and Robinson. In
general, corrections of only a few decibels had to be applied.

Discussion

     Let us first consider the relation between median NIHLs at 500, 1000 and 2000 Hz
and the equivalent sound level or noise immission level. For equivalent sound  levels up to
100 dB(A) or noise imission levels up to 110, there is a reasonable agreement between the
curves for exposure to constant noise and the results for varying and intermittent noise. For
higher levels,  however,  the  median NIHLs from exposures to intermittent noise  are lower
than those from constant noise.
     For 3000, 4000 and 6000 Hz, the same applies as far as it concerns the curves given by
Passchier-Vermeer for constant noise exposure. When comparing the  median NIHLs due to
varying and intermittent noise exposures with those  for constant noise exposures given by
Burns and Robinson, it turns out  that all values are above the curves, except those from
intermittent noise exposures with noise immission levels above 110, which lie near or under
the curves at 3000 and 4000 Hz.
     Comparing the median noise-induced hearing losses from intermittent noise exposures
with those from varying noise exposures, without referring to any constant noise  exposure,
it is quite clear from  the figures that intermittent exposures below an  equivalent sound level
of 100 dB(A) agree quite  closely with varying noise exposures, but that exposures to
intermittent noise  with equivalent sound levels of at least 100 dB(A) do cause  median
NIHLs equal  to those caused by varying noise exposures  with equivalent sound levels of
about  10 dB(A) lower.  Therefore, it can be  stated  that at least some intermittent  noise
exposures are less harmful than would be expected from their equivalent sound levels, and
although not necessarily following from this, it seems reasonable to accept that intervening
quiet periods  during  work time do reduce NIHLs. Then, however, the question remains why
the exposures to intermittent noise with equivalent sound levels below 100 dB(A) do not
cause less NIHL than should be expected from their equivalent sound levels. A solution is
not found by considering the lengths of the intervening quiet periods during the  day, since
these are on  the average longer for the equivalent sound levels below 100 dB(A) than for
those above 100 dB(A). Also, the number of exposure cycles  per workday does not con-
tribute to a possible solution, since it is not  systematically different for exposures  with
equivalent sound levels below and above 100  dB(A) respectively. The only, at  the moment
unverifiable, explanation may be that the sound levels during quiet intervals were higher, or
above  a certain sound level, for the exposures with equivalent sound levels below 100 dB(A)
than those  for the equivalent sound levels above 100 dB(A). In agreement with this explana-
tion is the fact that the people exposed to  the high equivalent sound levels have  their
profession  in the mining industries, which mostly coincides with low background sound
levels.
     Rating the intermittent noise exposures according to the CHABA damage risk contours
and considering their criteria for NIPTS (10 dB at 1000 Hz and below, 15 dB at 2000 Hz
and  20 dB  at 3000 Hz and above) with exposure time equal to 15 years instead of their 10

                                        197

-------
years, it turns out that all  exposures, except one, are rated correctly safe or unsafe. The
exception  concerns the only intermittent exposure considered here with noise  bursts of
more  than  5  minutes, which is incorrectly rated safe. However, the correct rating of the
intermittent exposures considered here does not necessarily validate the CHABA contours,
since almost all noise exposures are fairly well below or above these damage risk contours.
    Returning now to the varying  noise exposures, it  seems quite  difficult to decide
whether these exposures can be rated according to their equivalent sound levels or not.
Although median NIHLs at 2000 Hz and below can be estimated with reasonable accuracy
from the equivalent sound  levels, no firm decision can  be made  for the whole frequency-
range. However, in many instances, the limit of safe noise has been fixed at a sound level of
at most 90 dB(A) for a continuous exposure for 8  hours to steady-state noise. Figures 8 to
19 show, for equivalent sound levels of  at most 90 dB(A) or noise immission levels up to
about 100, that varying noise exposures can be rated according to their equivalent sound
level in order to estimate the median noise-induced hearing losses with reasonable accuracy.
    Finally,  it should  be pointed out that all data show that the equal-energy concept is
not overprotective for varying  noise exposures. Therefore,  revision is indicated of those
varying noise exposure limits that are based on the rule that for each halving of the exposure
time an increase of 5 dB(A)  in sound level is allowed.

SUMMARY

    A review is given  of proposed limits for exposure to intermittent and varying noise,
based on temporary threshold shift (TTS) measurements. Revision is suggested of the allow-
able sound levels during quiet periods in intermittent noise.  Recent TTS-experiments show
that TTS2 (TTS measured two minutes after the end of exposure, being the basis of all noise
limits considered) is not a consistent measure of a daily noise exposure, since retarded
recovery occurred after some intermittent noise exposures. Although TT$2 itself is only to
some extent dependent upon the total sound energy of the noise, the reviewer could show
that recovery from TTS after exposure is dependent upon the total sound-energy referred to
an 8 hour noise exposure.
    By using data from several papers, median noise-induced hearing losses at 500 to 6000
Hz, from exposures to varying and intermittent noise, have been related to their A-weighted
energy-equivalent sound level (Leq). Comparisons have also been made with published rela-
tions for exposure to constant noise.
    Relations between Lgq for varying noise,  with Leq  between 80 and 102 dB(A), and
median noise-induced hearing losses at frequencies up  to 2000 Hz, agree reasonably with
those  for constant noise. The same applies for median noise-induced hearing losses at 3000,
4000 and 6000 Hz, due to exposures to equivalent sound levels up to 90 dB(A); no decision
could be made for median noise-induced hearing losses at these frequencies, due to expo-
sures to higher equivalent sound levels.
    Results  for  intermittent noise  exposures, with equivalent  sound levels below  100
dB(A), agree  closely with the results for varying noise exposures. However, exposures to
intermittent noise, with equivalent sound levels  above 100 dB(A), are considerable less
harmful to hearing that should be expected from  their  equivalent sound levels. A possible
explanation is given in terms of low sound levels during quiet  intervals.

                                        198

-------
REFERENCES

W.D. Ward, A. Glorig, D.L. Sklar. Dependence of temporary threshold shift at 4 kc on
    intensity and time. J. Ac. Soc. Am. 30 (1958) 944-953
W.D. Ward, A. Glorig, D.L. Sklar. TTS  from octave-band noise: applications to damage-risk
    criteria. J. Ac. Soc. Am. 31 (1959)  522-528
W.D. Ward, A. Glorig, D.L. Sklar. TTS  produced by intermittent exposure to noise. J. Ac.
    Soc. Am. 31 (1959) 791-794
W.D. Ward, A. Glorig, D.L. Sklar. Relation between recovery from TTS and duration of
    exposure. J. Ac. Soc. Am. 31 (1959) 600-602
W.D. Ward, A. Glorig, W. Sellers. TTS in a changing noise level. J. Ac. Soc. Am. 32 (1960)
    235-237
W.D. Ward. The  use of TTS in the derivation of damage risk criteria for noise exposure. Int.
    audiol 5 (1966) 309-313
A. Cohen,  J.R. Anticilglia; P.L. Carpenter. Temporary threshold shift in hearing from expo-
    sure to different noise spectra at equal dB(A) level. J. Ac. Soc. Am. 57 (1972, II) 503 -
    507
W.D. Ward. Studies on the aural reflex. II Reduction of TTS from intermittent noise by
    reflex activity; Implications for damage risk criteria. J. Ac. Sdc. Am. 34 (1962) 234 -
    241
W.D. Ward. Recovery from high  values of  temporary threshold shift. J. Ac. Soc. Am. 32
    (1960)497-500
K. Schroder, E. Rempt.  Untersuchungen zum Larmpausen problem. Larmbekampfung 6
    (1962) 142-143
H.H. Jones, chairman. American Conference of governmental industrial hygienists proposed
    threshold limit value for noise. Am. Ind. Hyg. Ass. J. 29 (1968) 537-540
F. Schwetz, R.  Donner, G.  Langer, M.  Haider. Experimentelle Horermudung und ihre
    Riickbildung  unter  Ruhe- und  Larmbedingungen.  M.  Schrift Ohrenheilk.  Laryngo-
    Rhinol 104 (1970) 162-167
E. Lehnhardt, J. Bucking. Larmpausen  - eine Moglichkeit zur Prophylaxe der Larmschwer-
    gehorigkeit. Int. Arch. Gewerbepath, Gewerbehyg. 25 (1968) 65-74
W.D. Ward. TTS and damage-risk criteria for intermittent noise exposures. J. Ac. Soc. Am.
    48 (1970) 2 (part 2) 561-574
W. Klosterkotter.  Vorsorge- und Ueberwachungsuntersuchungen  bei Larmarbeitern,  nach
    der VDI-Richtlinie 2058 Blatt 2. Kampf dem Larm 18 (1971) 29-33
"Guidelines for noise  exposure control" prepared by Intersociety Committee on Guidelines
    for noise exposure control, Intersociety  Committee Report. Sound  and Vibr.  1970,
    21-24
G. Holmgren, L. Johnsson, B. Kylin, O. Linde. Noise and hearing of a Population of Forest
    Workers. Br. Ac. Soc. Spec. Vol. no. 1 (1971)
W. Passchier-Vermeer. Steady-state and  fluctuating noise, its effect on the hearing of people.
    Br. Ac. Soc. Spec. Vol. no.  1 (1971)
J. Sataloff, L. Vassalo, H. Menduke. Hearing loss from exposure to interrupted noise. Arch.
    Environ. Health 18 (1969) 972-981
M. Flach,  E.  Aschoff. Zur  Frage berufsbedingter  Schwerhorigkeit beim Musiker. Laryn-
    gologie 4.? (1966) 595-605
                                       199

-------
W. Feiser, R. Hauf, U. Heuft. Larmmessungen und audiometrische Untersuchungen in der
     holzverarbeitenden Industrie. A.S.A. 2 (1968) 38-43
W. Taylor, J. Pearson, A. Mair. The hearing threshold levels of dental practitioners exposed
     to air turbine drill noise. Br. Dent. Journ. 1965, 206-210
D.W. Robinson. The  relationships  between hearing  loss and noise exposure. NPL Aero
     Report Ac 32, 1968
G. Pressel. Horschaden durch Larm bei Ladearbeitern eines grossen zivilen Flughafens. Int.
     Arch. Arbeitsmed. 26 (1970) 231-249
K.D. Kryter, W.D. Ward, J.D. Miller, D.H. Eldredge. Hazardous exposure to intermittent and
     steady-state noise. Journ. Ac. Soc. Am. 39 (1966) (3) 451-464
J.H.  Botsford.  A new method for rating noise exposures.  Am. Ind. Hyg. Ass. J. (1967)
     431-446
E.J.  Schneider, I.E.  Mutchler, H.R.  Hoyle,  E.H.  Ode, B.B. Holder.  The progression of
     hearing loss from industrial noise exposure. Am. Ind. Hyg. Ass. J. 31 (1970) 368-376
A. Cohen, J.R.  Anticaglia, H.H. Jones. Noise-induced  hearing loss. Arch. Environ. Health 20
     (1970)614-623
W. Burns, R. Hinchcliffe, T.S.  Littler.  An exploratory  study of hearing loss and noise
     exposure in textile workers. The Ann. of Occ. Hyg. 7(1964) 323-333
R. Gallo, A. Glorig. P.T.S. changes produced by noise exposure and ageing. Am. Ind. Hyg.
     Ass. Journal 25 (1964) 237-245
The  relations of hearing loss to noise  exposure. A report by subcommittee Z 24-X-2 (1954)
     34
N.E.  Rosenwinkel, K.C.  Stewart. The  relationship  of hearing loss to steady-state noise
     exposure. Am. Ind. Hyg. Ass. Quart. 18 (1957) 227-230
J. Nixon, A. Glorig. Noise induced P.T.S. at 2000 and 4000 Hz. J.A.S.A. 33 (1961) 904-913
W. Taylor, J. Pearson, A.Mair, W. Burns. Study  on noise and hearing in Jute Weaving.
     J.A.S.A. 37 (1964) 113-120
B. Kylin.  TTS  and auditory trauma  following exposure to steady-state noise.  Acta Oto-
     Laryng. Suppl. 152(1960)
F. v.  Laar. Results of audiometric research at some hundreds of persons, working in differ-
     ent Dutch factories. Publication:  A.G./S.A. C 23  of N.I.P.G. TNO (1964)
A. Spoor. Presbyacusis values in relation to noise-induced hearing loss. Int. Aud. 6 (1967)
     48-57
W. Burns, D.W. Robinson. Hearing and noise in industry. London: Her  Majesty's Stationery
     Office 1970
J.P.  Kuiper.  Gehoorbeschadiging bij  machinale houtbewerkers.  Report from "Arbeidsin-
     spectie" 1968
A.C. Ephraim. Discontinu lawaai.  Report from "Arbeidsinspectie" 1970
ISO  Recommendation  1999. Assessment of noise  exposure during  work for hearing con-
     servation purposes 1971
W. Passchier-Vermeer. Hearing loss due to continuous exposure to steady-state broadband
     noise from "Institute of public health engineering" 1968
Z. Jirak, B. Mautner, J.  Kostal,  A. Andel,  C.  Losert.  Larnihorschaden bei Bergleute des
     Ostrau-Karwiner Kohlenreviers. Int. Arch. Arb. med. 28 (1971) 49-61

                                       200

-------
              EVALUATION OF THE HEARING DAMAGE RISK FROM
                 INTERMITTENT NOISE ACCORDING TO THE ISO
                               RECOMMENDATIONS

                       B. Johansson, B. Kylin and S. Reopstorff
                         Stockholm and Sandviken, Sweden

     Comprehensive international investigations, carried out during the last 15-20 years,
have finally  resulted in the publication of an international recommendation (ISO Recom-
mendation R 1999), which lays down the guiding principles for the risk of hearing damage
from noise. On the basis of this recommendation a Swedish standard, SEN 590111, has been
developed, and has been in force since February  1972.
     The assessment of  irregular noise is  in accordance  with  that specified in the ISO
Recommendation, and the damage risk limit  of 85 dB(A) has been taken as a satisfactory
.degree of safety.
     With regard  to regular  continuous noise, this recommendation is  based on reliable
empirical data. However, regarding the effect of irregular noise on hearing, from the point of
view of damage, sufficient data are still not available.
     In an earlier  Swedish investigation (Holmgren et al,  1971), which, comprised studies on
the hearing status of forestry workers who were exposed to very irregular noise from power
saws, it was  found that the conversion method  recommended apparently overestimates the
risk of hearing damage.
     The purpose  of the present study was to investigate in industry the irregular noise that
Usually occurs, to estimate this in accordance with  the recommendations, and to relate the
results to the hearing status found in the persons exposed. The study also includes experi-
ments on a  laboratory scale  in  which a comparison was made  between the temporary
threshold shifts which occur after  exposure to  both regular and irregular noise.

A.   FIELD  STUDIES

     Materials and Methods.  About 170  employees in  the  engineering industry were
selected. The subjects,  between the ages of 20 and 35 years, had to be exposed, without ear
defenders, to intermittent noise,  from which, however,  extreme impulsive noises were ex-
cluded. Exposure must have been  for not less than 2 years. Moreover, their working environ-
ment must, to a great extent, have remained unchanged since the beginning of the exposure
period.
     Routine otoscopy preceded the hearing examinations. Persons with a case history of or
with objective signs of hearing damage due to disease or accidents were excluded.
     The employees selected had  to  wear ear  muffs with an attenuation of about 25-30 dB
from the beginning of the working day up to the time the audiogram was made.
     The recording of the relevant noise for  the employee was made at  his place of work
with regard to his various jobs. The recording of the noise was continued for not less than
one  typical  working cycle. The  most usual  types of work were  sheet-metal production,
turning, welding and chiselling.

                                       201

-------
     The recorded noise was analyzed, in the laboratory, for spectral frequency and the
time distribution of the sound level  during a typical working cycle. The results of the
distribution analysis were subsequently used to calculate, in dB(A), the equivalent con-
tinuous sound level, according to the given ISO Recommendations.
     Results: The investigated employees  were distributed according to exposure level, in
the form of an equivalent continuous sound level in dB(A), ECNL, and the exposure time,
which is shown in table 1.
     On the basis of the composition of the material, the employees were divided into two
groups with regard to exposure time, namely less than and more than 5 years exposure time.
The mean exposure time was 3 and 9 years respectively.
     The mean age of the persons investigated was rather similar both for the short and the
long exposure time, with the exception of the group in the lowest noise level, less than 85
dB(A), (table 2.)
                                     Table 1.
      EMPLOYEES, NUMBER OF YEARS EXPOSED TO NOISE. WITHOUT EAR DEFENDERS,
                            AT DIFFERENT EXPOSURE LEVELS.
dB(A)
^85
85-90

90-95
>95
Years
2-3
7
11

12
5
4-5
7
19

6
4
6-10
17
30

19
8
11-15
4
12
Total
35
72
i
10 j 47
7
24
                                      Table 2.
AGE DISTRIBUTION FOR GROUPS WITH DIFFERENT EXPOSURE LEVELS AND EXPOSURE TIMES.
' dB(A) Exposure time
i
i >&
! 85-90 <'5
j
i '
\ 90-95 <5
i ->5
!>95 <5
: :5
n
14
21
29
42
18
29
9
15
Age, mean, years
24.7
29.2
29,0
29,9
29.4
31.2
30,4
29.7
                                      202

-------
     The employees' hearing loss was positively correlated both with increasing noise level
for the same exposure  time,  and with increasing exposure time at the  same noise level
(figure 1.). For the exposure level below 85 dB(A), there was slight deterioration in the
hearing status in relation to increasing exposure time.
     The mean  age of the persons in the group with less than 5 years exposure time was,
however, about 6 years lower than that of the corresponding group where the exposure time
exceeded 5 years.
     As regards the exposure level between 85 and 90 dB(A) the average hearing was normal
from a clinical point of view, i.e., hearing loss did not exceed 20 dB for the group that was
exposed for less than 5 years. For the group with more  than  5 years exposure time the
hearing status was slightly worse. This average hearing loss, however, does not interfere with
the so-called speech zone of the Swedish language. The mean audiogram for the noise level
90-95 dB(A) showed a further slight deterioration of the hearing, and the hearing loss now
affects the speech zone of the  Swedish language. Thus, the results show that an equivalent,
continuous noise  level between  90-95  dB(A) is a hearing-damage risk level, if the same
definition of hearing damage is applied as in the Swedish standard, SEN 590111.
     Table 3 lists  detailed data  on the composition of the material with regard to age,
exposure level, exposure time, and hearing status.
     A comparison between the results of the present investigation and those of a previous
investigation (Kylin,  1960), which dealt with the relation between the exposure level and
the hearing status of employees who were exposed to continuous noise, shows that exposure
to a continuous noise at a level  between  85-90  dB(A)  causes a hearing loss which corre-
sponds to that on  exposure to irregular noise with an equivalent, continuous noise level
between  90-95  dB(A). Thus the results support a previous observation (Holmgren  et al.
1971), that the  recommended conversion method for irregular noise causes a shift in the
damage risk limit corresponding to about 5 dB.

B.  LABORATORY STUDIES

    Material and Methods. Twenty male normal-hearing subjects between the ages of 18
and 27 years (mean age 23 years) were subjected to a series of noise-exposure experiments
through earphones. The temporary threshold shifts that occurred after exposure was meas-
ured by means of a Bekesy audiometer.
    The noise exposure consisted of both regular and irregular noise. The following condi-
tions were observed.
    a) The irregular noise corresponded  to the  type  of noise  that usually occurs in the
engineering industry.
    b) In the irregular noise the relation between the lowest sound levels and the highest
sound levels was such, that a certain restitution effect could be expected after the individual
peaks (figure 2.).
    c) The frequency spectrum for the regular noise was similar to that which occurred in
the high sound levels in the irregular noise (figs. 3 and 4). As suitable noise of the irregular
type, a two-minute  interval of a recording from a shipyard (fig. 2) was played. The peak
noise was produced by welders in the immediate vicinity who, during short periods, ground

                                        203

-------
                -M


                  10


                  :
             o

             f    1
             c
             z    w
                                         tr«gutncy H:


                           in     i»o     too     1000     noo    ax     MOO



, — >
X
^
\
V
\
s
\
s
\



<• »«.
85-90
00-95
> 95




,
^


X
X
N
^

HA IA\ i
dB (A) •
dB (A)
dB (At


>•>

i

**» ^


"-^


	
! Hi


ytors 01

	 ,
££^=







-



noiM

ir*-'











t

.
"-<










kpot


fc
\



_ • —





—

s

V
,3
\
\










^
— Jf3
*^

^^
rf*
^









/
s
r~

)
r*










v
\
1
/
/
S
s
S






                 -JO


                 -10


                   :
             r.
                  K


                  H
                 100
Tt»t (rtautncv Hi
in rso MO A


*«•
**






•*
/•
^^
?


,'








N
\
\
1
;
/
s
s
*





Figure 1. Average hearing loss (best ear) for group exposed to different levels of irregular noise. The area

marked with broken  lines covers  the  so-called  speech-zone of the Swedish language (G. Fant,  1948 and

1949).
                                                204

-------
                                  Table 3.
AVERAGE HEARING  LOSS (BEST EAR) FOR GROUPS WITH DIFFERENT EXPOSURE LEVELS
AND DIFFERENT EXPOSURE TIMES (Qlf Q2, Q3 INDICATE QUART!LES).

dB(A) ECKL
(mean)
^85
(83,5)



*85
(84.0)



85-90
(88,6)



85-90
(88.5)


90-95
(93.5)



90-95
(92.4)


>95
(98.9)


>95
(96.9)


Kumbers
Years of of
noise exp, subj.
<5 14 raean

Q!
Q2
Q3
>5 21 mean

Q!
Q2
Q3
<5 29 mean

Q!
Q2
Q3
>-5 42 mean
Q1
Q2
Q3
<5 18 mean

Qj
Q2
Q3
>5 29 mean
Ql
Q2
Q3
<5 9 mean
Ql
Q2
Q3
>5 15 mean
Q1
Q2
*3
Frequency

0.5
7.9

5.0
10,0
10,0
6.3

0.0
5.0
10.0
10.0

5.0
10,0
10,0
9.3
5,0
10,0
10.0
11,4

5,0
10.0
20,0
8.4
5,0
10,0
10,0
10,0
5,0
10.0
15,0
3.7
10.0
10,0
10.0

1
5.4

0,0
5.0
10.0
14,0

0,0
5.0
5.0
5.9

2,5
5,0
10.0
6,6
5.0
5.0
10. 0
7.5

5,0
5.0
10,0
6.6
0.0
5.0
5.0
5.6
3.8
5.0
10.0
8.3
5.0
10.0
10.0

2
5.

5.
5.
10.
4.

2.
5.
10,
7.

5.
s.
10,
9.
5.
10.
10.
8,

5,
5.
10.
a.
3.
5.
10.
10.
5.
10.
11.
13.
5.
10.
20.


0

0
0
0
5

5
0
0
1

0
0
0
2
0
0
0
1

0
0
0
3
8
0
0
0
0
0
3
3
0
0
0


7

5
5
10
5

0
5
10
10

5
10
15
15
5
10
20
11

5
15
18
20
5
15
32
15
5
10
25
30
10
30
45
in

3
.5

.0
,0
.0
,3

,0
.0
.0
.4

.0
,o
.0
,3
.0
.0
.0
.7

,o
.0
.8
.0
.0
,0
.5
,6
.0
.0
.0
.7
.0
,0
,0
kHz

4
12.1

10.0
10,0
15.0
10.8

5.0
10.0
15.0
14,8

10,0
15,0
17,5
23.2
10.0
20,0
30.0
19.7

10.0
17,5
25.0
28.3
10,0
30,0
40,0
27,8
10.0
30.0
41,3
46,0
27,5
55,0
60,0


6
11.

5.
10.
20.
15.

7.
10.
25.
16.

10.
15.
25.
22.
15.
20.
30.
25.

10.
25.
35,
26.
15.
25.
35.
21.
13.
25.
31,
51.
37.
50,
63,



n

0
0
0
5

5
0
0
3

0
0
0
4
0
0
0
8

0
0
0
1
0
0
0
7
8
0
3
0
5
0
8


8
10.8

5.0
7.5
15.0
9.7

5.0
10.0
15.0
10.8

6,3
10.0
15.0
16.1
10.0
10.0
25.0
21.4

10.0
15.0
27.5
17.9
5.0
15.0
35.0
17.5
12.5
15.0
22.5
39.4
22,5
35.0
56.3
                                   205

-------
  Noise level dB(A)
  no -
  100 i
  80 -

                           .

                                                             .
                        Figure 2. Example of the irregular noise used.

down the irregular joint between  two steel plates that were welded together. The back-
ground noise was produced by the activities occurring in the other parts of the large welding
hall. For the regular noise a suitable section of the recording of the continuous grinding
noise was played. The above-mentioned recording was subsequently played several times on
to a tape recorder, so that a exposure time of one hour was obtained for each type of noise.
     For the regular noise an exposure level was decided on  which produced a maximal
threshold shift  of 20 dB measured 15 minutes after exposure.  For the  irregular noise the
highest sound level chosen  was that compatible with the subjective tolerance threshold for
some of the more sensitive subjects. In addition to this, two lower levels, in 5 dB steps, were
used for each type of noise in the experimental  noise exposures.
     Each subject had to undergo  noise exposure tests in the  following  order: first, the
lowest level  of the regular noise, followed by the intermediate and highest level respectively:
then the  irregular noise  in the same order. If the threshold shift was greater than 10 dB, the
next experimental-noise exposure test was carried out only after an interval  of at least 48
hours. Otherwise, the  intervals were usually  24 hours (minimum). For audiometry, the
frequencies were  tested  in the following order: First,  2.5 kHz to 8kHz, then from 2.5 kll/.
to 0.5 kHz, the  right  ear  first. Each subject  was trained in  Bekesy audiometry until he
reacted regularly  and reproducibly. For each  ear the hearing  test took 8 minutes. Audi-
ometry was started 15 minutes after the end of the exposure, and was thus  completed 31
minutes after the end of the exposure. A control  audiogram was made before each experi-
mental noise exposure.
                                         206

-------
               08 >•! J.IO  * N-m?
              31 S    U     ITS    350     MO   1000    ?OOO    40OO   ROOD   '6000
               31 i    63     IJ4    740     MO    1000   MOO    «OOO    8OOO    HOOO  M,
Figure 3-4. Frequency spectrum for regular noise (above) and irregular noise (below).
                                          207

-------
     Results: The mean hearing-threshold shifts are shown in table 4.
     A significant (p<0.05) effect of the noise exposure (indicated by underlining) occurred
mainly at the test frequencies 3, 4, and 6 kHz. Although the size of the threshold shifts was
small throughout,  nevertheless there was an evident  trend for  these to increase with in-
creasing noise level.  On exposure to regular noise, a  sound level of 95 dB(A) produced a
definite effect  at the  three  frequencies,  whereas for the irregular  noise,  this effect was
obtained already at  92 dB(A) equivalent continuous  noise level, calculated in accordance
with the ISO Recommendation. However, this threshold  shift was, throughout, less than
that for the 95 dB(A) continuous noise level.
     Moreover, the results indicate that at first the highest noise level for the irregular noise
caused a temporary  threshold shift corresponding to that obtained at 95  dB(A) for con-
tinuous noise.  According to  the conversion method used,  the equivalent level  for the
irregular noise  was  calculated to be 97 dB(A). Consequently,  the  investigation has con-
firmed,  to a  certain  extent,  the results of the field study,  namely, that the risk of hearing
damage is somewhat  overestimated by the conversion method.

DISCUSSION

     In  all retrospective investigations  whose purpose is to look for a correlation between
the degree of influence and exposure to the relevant substance, there are always difficulties
in obtaining  representative and well-defined material. During recent years  propaganda for
the use of ear defenders has been increasingly intensified.  In  the present investigation, this
situation may have led to some uncertainty in assessing  the noise exposure to  which the
-persons investigated  were exposed to. On the whole, however, this uncertainty is the same
for all the groups investigated, and consequently the results for these are comparable.
     The  relatively high correlation between increasing deterioration in hearing status and
increasing noise  level, which agrees also with relevant facts  previously known (Kylin, 1960),
supports the  view that  the estimation of the exposure conditions for the present selection of
the population seems to be pertinent.
     The results of the field study show, however, that the method employed in  evaluating
the exposure level for  irregular noise apparently causes a shift in the risk limit for hearing
damage by about 5 dB compared with the corresponding value for continuous noise. Similar
results were  obtained  in an earlier  investigation on the hearing status of forestry workers
who were exposed to very  irregular noise  which occurs  when  working with  power saws
(Holmgren etal., 1971).
     In the present investigation, the experimental study on the temporary  threshold shifts
after exposure to continuous and irregular noise has also shown indications that the con-
version method  used overestimate the damage risk limit when it concerns the irregular noise.

SUMMARY

     About 170 employees, between the ages of 20 and 35 years, who were exposed in their
work to the  irregular noise that usually ocurs in the engineering industry, were subjected to
a  hearing  examination. The aim of the examination was to  test a proposed method for
assessing the  risk of hearing damage from irregular noise.
                                        208

-------
                                      Table 4.

AVERAGE TEMPORARY THRESHOLD SHIFT IN 20 PERSONS 15-31 MINUTES AFTER EXPOSURE
TO DIFFERENT LEVELS OF STEADY STATE, AND VARYING NOISE. (SIGNIFICANT, PH 0.05
EFFECT OF THE NOISE EXPOSURE IS INDICATED BY UNDERLINING.).




Regular
Noise





Irregular
Koise



dB(A)
90 right
left
95 right
left
100 right
left
d3(A)
87 right
left
92 right
left
97 right
left

0.5
0.5
1.2
0
1.4
1.1
1.0

kHz
1 2
0
1
0
1
MM
1
MB!
0
Equivalent
0
0.3
0
0,2
0.2
1.0
0
0
0
0
0
0
0
.3 1.3
0,8
O. i.o
iJS 1.1
.7 2,2

3
1.2
3^5
2,3
2.1
lal
5.2
continuous
0
0
.3 0
.5 0,5
.7 0,5
.1 0.8
0.9
0,7
1,9
MMMMM
1.5
1,6
•••M^HMI
1.8

*
1.3
0.6
3,1
3,0
4.3
5.8
noise
0,5
1.1
1,7
2.5
2,0
3,1

6
1.5
1.1
2,6
5,0
7.4
8,7
level
2^2
1.1
2.8
2.8
5.0
_!—
    The proposed method  enabled the conversion of an irregular noise to an equivalent
continuous  noise level. The persons investigated were divided into groups with different
exposure levels, in 5 dB steps, between 80 and 100 dB(A).
    The mean hearing loss obtained showed a high correlation with increasing noise level.
The results  also indicated that the conversion method now used for irregular noise seems to
cause a shift in the damage risk limit of about 5 dB.
    In  the experimental study 20  persons with  normal hearing were exposed to both
regular and irregular occupational noise.  The temporary  threshold shift shift which  still
affected the subjects 15 minutes after the completion of the exposure was recorded.  The
irregular noise was converted to an equivalent continuous noise level. The results showed
that the temporary threshold shift after exposure to irregular noise was less than that after

                                       209

-------
exposure to the corresponding noise level for continuous noise. This supports the conclusion
that the conversion method employed  for irregular noise overestimates the damage risk
limit.

REFERENCES

ISO Recommendation R 1999, Acoustics. Assessment of Occupational Noise Exposure for
     Hearing Conservation Purposes. May, 1971.
SEN 590111,  Acoustics. Estimation  of risk of hearing damage from noise.  Measuring
     methods and acceptable values. Swedish Electrical Commission, 1972.
Fant, G.: Analys av de svenska vokalljuden.  LME Report H/P-1035, 1948.
Fant, G.: Analys av de svenska konsonantljuden. LME Report H/P-1064, 1949.
Holmgren, G., Johnsson, L, Kylin, B.  and  Linde, O.: Noise and Hearing of a Population of
     Forest Workers, in Robinson, D.W.  (ed) Occupational Hearing Loss. British Acoustical
     Soc. Spec. Vol. No 1, p 35^2, 1971.
Kylin, B.: Temporary Threshold Shift and Auditory Trauma Following Exposure to Steady-
     State Noise. An  experimental  and field study.  Acta  Oto-Laryngologica,  vol. 51:6,
     suppl. 152, 1960.
                                       210

-------
                      NOISE-INDUCED HEARING LOSS FROM
                        IMPULSE NOISE: PRESENT STATUS

                       R.R.A. Coles, C.G. Rice and A.M. Martin
                     {Institute of Sound and Vibration Research,
                        University of Southampton, England)

    Although  a large number of damage risk criteria have been published and there are
considerable variations in the legislation and codes of practice put forward by national and
state governments  and by numerous occupational authorities, it is evident that the relation-
ships between steady-state noise exposure and hearing loss are becoming a matter of general
agreement in all but  detail. Perhaps the main uncertainty revolves round the question of
intermittent exposures and fluctuating levels of noise,  and whether the energy principle
provides an accurate basis or acceptable approximation on which assessments of hazard can
be made.
    The divergence of opinion on  such exposures is illustrated well by comparing British
and American  attitudes. In the U.K.,  the Department of Employment's Code of Practice
(1972), based on the  research of Burns and Robinson (1970), utilizes the energy principle.
In the U.S.A., the criteria applicable under the Walsh-Healey Act (1969) and the Occupa-
tional Safety and Health Act (1970), together with those of several occupational groups and
those recommended in 1972 by the National Institute for Occupational Safety and Health,
favor a 5-dB correction per halving of daily exposure duration. It is interesting, though, that
the  American National Standards Institute, which of course benefits from the most com-
petent academic advice available and presumably represents the consensus of opinion of
American  industry, is a signatory to  the International Standard (1971) on occupational
noise exposure which utilizes the  energy principle. It is also  noteworthy that the U.K.
evidence is based mainly on Burns and Robinson's PTS (permanent threshold shift) studies,
whereas the U.S.  evidence appears  to be based on  TTS2 (temporary  threshold shift 2
minutes after cessation of noise exposure): this is probably a significant procedural  differ-
ence, and further calls into question the  relevance of TTS studies for derivation of correc-
tion factors for interimttency of exposure or fluctuation of noise level.
    In comparison to the situation concerning steady-state noise, there is remarkably little
legislation or governmental guidance on impulse noise hazards. Indeed, in most instances the
laws, codes and criteria specifically  exclude impulse noise, except perhaps where it consists
of rapidly repeated impacts.
    In looking at the historical aspects  of noise-induced hearing loss, impulse noises have
been prominent causes and have led to such descriptive terms as "gunfire deafness" and the
"boilermaker's notch" (referring to  the  4-kHz dip  in the audiogram). At  the same time,
these noises have  been the most difficult to quantify.  An almost traditional awe has sur-
rounded them, such that the writers of earlier criteria have spoken guardedly of additional
hazards  where a noise contains impulsive components, or they have even suggested addi-
tional arbitrary hazard factors equivalent to an increase in sound level of up to 10 or 15 dB
where such components are present.
    In keeping with the term  "gunfire deafness", the  first attempts to quantify the rela-
tionships between  impulse noise and hearing  loss have come from military fields. The
                                        211

-------
pioneer work was that of Murray and Reid (1946 a, 1946b) and Reid (1946). Almost two
decades passed  before Pfander (1965)  and Rice and  Coles (1965) made further moves
towards establishment of measurement and assessment criteria,  although a number of TTS
studies on various aspects of impulse noise had been published and in one case (Ward, 1962)
led to inclusion in the CHABA recommendations (Kryter et al., 1966) of a 140-dB limit for
impulse noise.
     The first  relatively comprehensive set of recommendations on impulse noise were
published by Coles, Garinther, Hodge and Rice in 1968, after a pooling of research data and
thought derived from studies mainly of gunfire-induced TTS with the British Royal Marines
and the U.S. Army. With minor modifications, these were adopted by CHABA (1968) for its
recommendations on gunfire noise exposure.
     Figure 1  illustrates the Coles et al criterion. A-duration refers to simple (Friedlander)
waveforms and refers to the duration of the positive pressure wave. It is comparatively rare,
and much more commonly there is a succession of pressure fluctuations when B_-duration is
the relevant parameter.  This  refers to  the  total time that the envelope of the pressure
fluctuations, positive  and negative, is within 20 dB of the peak pressure level. Exposure to
100 impulses, whose  physical characteristics fall on the relevant criterion line, will lead to
CHABA (Kryter et al. 1966) levels of TTS2 in 25% of  those exposed. The criterion should
be lowered by 5 or 10 dB if only 10% or 5% respectively are permitted to be affected to this
degree.  By the criterion, as published, the level could be raised by 10 dB for exposures to
one impulse per occasion.
            175
            170
        °   160
         K
            155
         E
         I
ISO
            145
            140
                                                      ....1
                         100
                       1ms       10ms      100 m s

                         Duration
             Figure 1. Damage risk criterion for impulse noise. (After Coles et al, 1968).
                                       212

-------
     In  an attempt to extend the criterion to embrace industrial  impact types of noise,
where the peak levels tend to be lower, the B-durations longer, and the numbers of impulses
per day very much greater, Coles and Rice (1970) proposed a revised and extended series of
corrections for numbers of impulses per exposure occasion. This is  shown in Figure 2. The
arguments for it are given in the reference, but it is worth noting that a major factor in its
construction was the need to  take into account the TTS studies of  Cohen et al. (1966) and
of Walker (1969).
     In  1970 Burns  and  Robinson published the  results of their 10-year government-
sponsored study of the relationship  between  noise  exposure and hearing loss  in British
industry. An important conclusion of this work was that their data  were consistent with an
energy concept;  that is,  the  total A-weighted sound energy received is a representative
measure of noise exposure with respect to injury to hearing. At the conference at which this
work was first presented,  Coles and Rice also gave a general resume of the current situation
on impulse noise hazards. A significant point to future trends came in the discussion, when
Martin (1970) linked the two together and suggested that  the energy  concept might be
extended to include impact noises up to peak pressures of 145 dB. This assertion  was based
on PTS studies in industry published later by Atherley and Martin (1971) and their method
of assessment of impact noise (Martin and Atherley,  1973).
     Recently, Rice and Martin (1973) have made a series of calculations which  show that
an extension of the energy concept gives results which are close to but rather more con-
servative than the Coles et  al (1968) criterion or its modification by CHABA (1968), and
 10

 5

 0

 -5

-10

-15

-20

-25



-35

-40
              - Nixon (1969)
               Tentative estimate by
               Coles et al (196S)
                                  Coles at al (1966) reference point
                                / (DRC expressed for about 100 impulses >

                             ^    
-------
very similar to the modifications by Coles and Rice (1970) and an earlier revision suggested
at its draft stage by Forrest (1967). The results of this analysis are shown in Figure 3. Two
practical points concerning the good agreement between the earlier impulse criterion and
the equivalent continuous noise level (ECNL) extension should be mentioned.
     First,  the  method of  assessment  of  industrial impact noises used  by Martin and
Atherley (1973) needs validation with respect to its application to gunfire noises. Indeed,
research is being carried out to find an easily  practical way to do this, and to adapt a noise
dosimeter for this purpose.  The instrumentation needs have been discussed generally by
Martin (1973).
     Second,  the 90 dBA  ECNL criterion based on an energy concept relates to daily
workday exposures, whilst  Coles  et al. one referred  to  10 or 20 exposure occasions per
annum. In fact, this discrepancy  is more apparent than real. On the  one hand, the ISO
recommendation  and the British government's Code of  Practice reflect  most scientific
opinion in making no allowances for non-habitual exposures. On  the other, the Coles et al.
criterion was derived mainly from TTS  studies and rests on the  generalization that TTS2
indicates the  likelihood of  PTS from habitual exposure. The criterion  was expected to be
applied to military situations, with perhaps  only 10 tor 20 exposure occasions per annum.
This was accepted as the 'norm', and the stated auditory effects were expected in practice to
arise from as  few as 10-20 exposures in many instances. This expectation was based mainly
on  the variable nature of actual gunfire-noise exposure in military situations, which some-
times leads to a more rapid accumulation of hearing loss than expected. On the other hand,
the total amount of hearing loss  accumulating  over a large number of exposure occasions
would not necessarily be markedly greater than that predicted by the TTS measurements on
which the criterion was  largely based, if the physical characteristics of the noise exposures
were strictly controlled. As this cannot be done in practice, and also because unpredictable
variations in susceptibility seem to occur (Reid, 1946), the criterion was considered to apply
to ordinary military conditions of 10 to 20 exposure occasions a year; warning was, how-
ever, given against taking a chance with even  a single unprotected occasion.
     For those individuals and countries who  accept the energy principle, it would seem on
the basis of the work described that we  are now on the threshold of being able to measure
and evaluate the auditory hazard of all types  of noises (from steady-state and intermittent
and fluctuating noises, and also industrial impact noises and high-intensity explosive im-
pulses) with one general principle and possibly with one single instrument The advantages
of this are so obvious, however, that one has to be careful not to lose one's critical appraisal
in face of the claims of expediency: certain further studies are desirable, in addition to the
two items already mentioned.
     Further  PTS  studies in a range  of populations, covering high-intensity impulses of
explosive type to moderate-intensity ones of impact type, are needed in order to validate
further  me agreement  illustrated in Figure 3  and  strengthen the evidence provided by
Atherley and  Martin (1971). These studies will not be easy though, because of the difficulty
in finding suitable populations. In industry, noise levels change with development of manu-
facturing processes, workers are less static in respect of their place of employment than they
used to be, and  hearing conservation  measures are becoming more effective or at least
complicate the  problem of selecting  an unbiased  and fully exposed sample. In military

                                       214

-------
       183
       175
    3
    I
    O
•;  I5S
CSI
S
"  145
    J
    e
       135
       125
                 -(R&M, 1973)
                                               All curvm corrected for 25% affected up lo CHAIA or
                                               equivalent amounts of threshold shift.
                                     Coles et al (1968),modified by Cotei and Rice ( 1970)for 1-100 impuliet
                                                           Rice and Martin (!97i),(ECNL = 90 dBA)
                                                                       CHAM (1968)
                              Coles et al (1968), modified by
                              Coles and Rice (1970) for
                              1000 impulses
          10"1         1          10        10'2        103

           'S'-DURATION » NO. OF IMPULSES, MILLISECONDS.
                                                          10
                                                                    10
 Figure 3. Comparison of damage risk criteria for impulse noise with an equivalent continuous noise level of
 90 dB(A), calculated according to the method of Martin and  Atherley (1973). (After Rice and Martin,
 1973). (Note:  the sound  levels should be lowered by 5 dB where the ears are at normal incidence to the
 impulse sound waves).
situations,  the same  complicating factors apply and there are still greater difficulties in
defining the amount of noise exposure (in terms of both level and numbers of impulses).
     It may be necessary to have recourse to TTS studies. The trouble here is that there are
great doubts as to the validity of the supposed arithmetic equivalence of TTS2 and  ex-
pectancy of PTS. Martin's studies (1970)  suggest that with impact-type noise TTS2 is a poor
quantitative predictor of PTS and, unlike his and other PTS ones, do not agree with  the
energy principle. Ward (1970) has shown that recovery from a given amount of TTS is  not
independent of how the TTS arose, as was originally postulated by Ward, Glorig and Sklar in
1959. Moreover, the rate of recovery from TTS, which must surely be related to risk  of PTS,
is not related in a simple manner to  the  amount of TTS  even from a given pattern of noise
exposure; where the  TTS2 is  over 40 dB, a much slower rate of recovery applies (Ward,
Glorig and Sklar, 1958,  1960). Ward (1970) himself suggested that TTS^Q may be a more
relevant index of risk of PTS,  but the relationship  between this and PTS is not yet  known.
As Passchier-Vermeer (1973) has stated  earlier in  this Congress,  "it may be possible that
recovery  and, hence,  total sound energy  over a workday  plays a more  important role than
can be expected from TTS2 alone".
     There is also a theoretical argument. If the energy principle is postulated for a  damage
risk  criterion, it is evident that TTS2 measurement applied to  a TTS2/PTS relationship
                                         215

-------
cannot be expected to validate the criterion. Consider the time patterns of the two types of
exposure illustrated in Figure 4. Both have four hours of exposure per  day, and  on the
energy principle the risk of PTS must be the same. The TTS2 measurement resulting from
exposure pattern B will be considerably greater than from exposure pattern A however. The
same sort of arguments could be applied to hypothetical patterns of impulse noise exposure.
    On the other  hand, hi spite of all the doubts and uncertainties about TTS2 studies, the
Coles  et al (1968)  criterion for impulse noises, together with Forrest's extension of it, and
the evidence of Cohen et al. (1966) and Walker (1968), were all based on TTS2 measure-
ments. It is these which appear to agree so well with the energy principle and the extrapola-
tions of ECNL 90  dBA into impulse noise fields, both of which were based on PTS studies.
    Finally, there is  the question of interactions  between impulse and steady-state noise
when  combined. On the energy principle, there is of course no problem: the energies from
the two sources are simply summed. However, this would not appear to be the complete
story. Cohen et al. and Walker have both demonstrated protective effects (for TTS) when
the two types of noise  were  added, and attributed these to enhancement of the acoustic
reflex. It should be noted, though, that  these were TTS studies from relatively short-dura-
tion exposures and with subjects who were not habitually exposed to high-level noise. It
seems quite likely that  there would be greater adaptation of the reflex  with longer and
habitual exposures.
    The contrary  and more traditional view is that the presence of impulsive components
hi a steady-state noise involves a disproportionately  greater auditory hazard. Such a view has
found support recently from the work of Hamernik and his colleagues (1972) who observed
the effects of impulse and steady-state noise on the hearing (PTS) and hair-cells (histology)
of chinchillas.
    In turn, these and other animal experiments, together with some electrophysiological
studies, open up much wider questions as to whether  TTS (or even PTS) is in fact a
satisfactory  measure  of  auditory damage.  However, until we have something that  is both
substantially  better and at least as easy to measure (and verifiable objectively when it comes

 NOISE EXPOSURE PATTERN A
                     noise
                   // S
             /
       /no
       ss
                                4
                               hours
             8
 NOISE EXPOSURE PATTERN B
                                             /—q°'
liet-
                                                                TTS measured 2 minutes
                                                                after last noise exposure
                                                                of an eight-hour workday.
Y///S
/noise '
///A
quiet
y///
'- noise S
'//y/

'////.
/noise

-------
to compensation assessment), pure-tone threshold shifts must remain our principal tool for
assessment of auditory damage resulting from noise exposures of all types.
    Acknowledgement: For much of their own material included in this general review, the
authors are indebted to the Medical Research Council for financial support.

                                     Summary

    Until fairly recently, governmental and occupational noise  hazard criteria have either
omitted or specifically excluded reference to impulse noise situations, or merely mentioned
them in guarded and largely non-quantitative  terms. The criteria put forward by Coles et al
in 1968, modified and extended by CHABA and Forrest, have provided some help but refer
mainly to gunfire-type noises. Coles and  Rice  (1970)  offered extensions for  industrial
impact noises,  and Atherley and Martin (1971) studied the problem further showing that
their PTS data gave  support to the equivalent A-weighted sound energy concept as a meas-
ure  of hazard  to  hearing. Recent analyses by Rice and Martin (1973) suggest  that such
equal-energy extrapolations may possibly be extended to cover the high-intensity explosive
type of noise. The highly  desirable prospect of one comprehensive method of measurement
and auditory hazard evaluation for all types of noise is thus revealed. Problems relating to
validation, instrumentation, and other uncertainties are discussed  briefly.
REFERENCES

Atherley, G.R.C. and Martin, A.M. Equivalent-continuous noise level as a measure of injury
     from impact and impulse noise. Annals of Occupational Hygiene, 14, 11-28 (1971).
Burns,  W. and  Robinson, D.W. Hearing and noise in Industry.  Her Majesty's Stationery
     Office, London (1970).
CHABA.  Proposed damage-risk criterion for impulse noise (gunfire). National Academy of
     Sciences - National Research Council's Committee on Hearing, Bio-acoustics and Bio-
     mechanics, Report of Working Group 57 (ed. W.D. Ward), Washington, D.C. (1968).
Cohen, A., Kylin, B. and LaBenz, P.J. Temporary threshold shifts in hearing from exposure
     to combined impact/steady-state noise conditions. Journal of the Acoustical Society of
     Americano, 1371-1380(1966).
Coles, R.R.A., Garinther, G.R., Hodges, D.C. and Rice, C.G. Hazardous exposure to impulse
     noise. Journal of the Acoustical Society of America, 43, 336-343 (1968).
Coles, R.R.A. and Rice, C.G. Towards a criterion for impulse noise in industry. Annals of
     Occupational Hygiene, 13, 43-50 (1970).
Department  of  Employment.  Code of Practice for reducing the exposure of employed
     persons to Noise. Her Majesty's Stationery Office, London (1972).
Forrest, M.R. The effects of high intensity impulsive noise on hearing. Master of Science
     dissertation, University of Southampton (1967).
Hamernik, R.P., Crossley, J.J., Henderson,  D. and Salvi, R.J. The interaction of continuous
     and impulse noise. Personal communication, State University of New York, Syracuse,
     N.Y. (1972).

                                        217

-------
International Organisation for Standardization. Assessment of occupational noise exposure
    for hearing conservation purposes. ISO Recommendation R1999 (1971).
Kryter, K.D., Ward, W.D., Miller, J.D. and Eldredge, D. Hazardous exposure to intermittent
    and steady-state noise. Journal of the Acoustical Society of America, 39, 451-464
    (1966), and Report of CHABA Working Group 46.
Martin, A.M. Industrial  impact noise and hearing. Doctor of Philosophy thesis, University of
    Salford(1970).
Martin, A.M. Discussion on papers in Section 1. Pages 89 and 90, in Occupational Hearing
    Loss, British  Acoustical Society Special  Volume  No. 1, Academic Press,  London
    (1971).
Martin,  A.M. and  Atherley, G.R.C. A  method for the assessment of impact noise with
    respect to injury to hearing. Annals of Occupational Hygiene, 16 19-26 (1973),
Martin, A.M. The measurement of impulse noise. Journal of Sound and Vibration, 342-344
    (1973).
Murray, N.E. and  Reid, G. Experimental observations on the aural effects of gun blast.
    Medical Journal of Australia, 1, 611-617 (1946,a).
Murray, N.E. and Reid, G. Temporary deafness due to gunfire. Journal of Laryngology and
    Otology, 60, 92-130 (1946, b).
National Institute for Occupational Safety and Health. Criteria for a recommended standard
    ... Occupational  Exposure to  Noise. U.S. Department of Health,  Education, and
    Welfare; NIOSH Report HSM 73-11001 (1972).
Passchier-Vermeer,  W. Noise-induced hearing loss from exposure to intermittent and varying
    noise. Proceedings of the International Congress on Noise as Public Health Problem,
    Dubrovnik(1973).
Pfander, von F. Uber die Toleranzgrenze bie Akustischen Einwirkungen. Hals-, Nasen- und
    Ohrenarzte, 13, 27-28 (1965 ).
Reid,  G. Further observations on temporary deafness following exposure to gunfire. Journal
    of Laryngology and Otology, 60, 608-633, (1946).
Rice, C. G. and Coles, R. R. A. Impulsive noise studies and temporary threshold shift. Paper
    B67, in Proceedings of Fifth International Congress of Acoustics, Liege (1965).
Rice,  C. G. and Martin, A.M. Impulse noise damage risk criteria.  Journal of Sound and
    Vibration, 359-367 (1973).
Walker,  J.G. Temporary  threshold  shift from impulse noise. Annals  of Occupational
    Hygiene, 13, 51-58(1970).
Ward, W.D., Glorig, A. and Sklar, D.L. Dependence of temporary threshold shift at 4 kc on
    intensity and time. Journal of the Acoustical Society of America, 30,944-954 (1958).
Ward, W.D., Glorig, A. and Sklar, D.L. Relation between recovery from temporary threshold
    shift and duration of exposure. Journal of the Acoustical Society of America, 31,
    600-602(1959).
Ward, W.D., Glorig, A.  and Sklar, D.L. Recovery from high values of temporary threshold
    shift Journal of the Acoustical Society of America, 32,497-500 (1960).
Ward, W.D. Effect of temporal spacing on  temporary threshold shift from impulses. Journal
    of the Acoustical Society of America, 34, 1230-1232 (1962).
Ward, W.D. Temporary threshold shift and damage risk criteria for intermittent noise ex-
    posures. Journal of the Acoustical Society of America, 48(2), 561-574 (1970).

                                        218

-------
                    HEARING LOSS DUE TO IMPULSE NOISE
                                A FIELD STUDY.

                         Tadeusz Ceypek, Jerzy J. Kuzniarz

                            Otolaryngology Department
                             Silesian Medical  Academy
                                  Francuska 20,
                             40-027 Katowice, Poland

                                 Adam Lipowczan
                              Central Mining Institute
                         P1. Gwarkow 1, Katowice, Poland

    Present  knowledge on consequences of exposure to industrial impulse noise is still
rather scanty (Acton, 1967; Coles, 1970). That was the reason we have started a study on
hearing loss in drop-forge operators, exposed to several thousand impulses each day. For
that purpose we have selected a factory where both the local tradition and a great distance
from another possible place of work made the staff pretty stable over the years. In that way
213, drop-forge operators (426 ears), working up to 30 years at the same place and exposed
to the same kind of noise, were tested.

Characteristics of noise exposure.

    The impulses  are generated by iron drop-forge hammers, weighing 1  to 5 tons, falling
from a height of approximately 1.5 m onto an iron base. Owing to the distance between the
forges (about 5 m) each operator was exposed mainly to impulses from his own forge, from
a distance of less than 1 m. As a rule they did not use ear protectors.
    The impulses  were measured with  a Bruel and Kjaer Impulse Sound  Level Meter type
2204 and 1/2-in. microphone, and recorded on a Kudelski tape-recorder Nagra IV L (speed
19.05 cm/sec., frequency range of 20  - 18 000 Hz, dynamic range 50 dB over the back-
ground level). Using the frequency modulation technique the lower limiting frequency equal
DC was obtained.  The recorded impulses were  measured with the  use of storage oscillo-
scope. (Fig. 1). The following values have been determined:
    (a) Peak pressure level: 127 to 134 dB, independent of the weight of the hammer,
    (b) Rise time: almost instantaneous (a few microseconds),
    (c) Impulse duration (from peak to the ambient level): 100 to 200 msec,
    (d) Level of background noise: 110 dB,
    (e) Repetition rate (during on-time): 0.5 to 2 per sec.
    (f) As drop-forges work in an on-off manner the total number  of impulses have been
calculated from the known  number of strokes necessary to produce each item, and from the
number of items produced  by each drop-forge during the work-day.  Depending on the item
produced there were 3000 to 10,000 impulses a day to which every drop-forge operator was
directly exposed.

                                       219

-------
: -
                      Figure 1. Oscilloscopic picture of impulses produced by drop-forge. Lp = peak level. Lb = background level,
                      Tr • time of repetition, Tj = duration time.

-------
    (g) Frequency spectrum has not been analyzed because of known difficulties with that
procedure (Coles, 1970).

Hearing Tests.

    Hearing examinations were conducted before  work in order to measure the PTS (at
least 16 hours after the last exposure to noise), and after work in order to measure TTS.
Otological  examinations were done beforehand to exclude other possible causes of hearing
loss. Audiometers were calibrated according to ISO standards. Ambient noise did not exceed
the allowable levels.

Results.

    All the workers were divided into 8 groups, according to exposure time in years (Tab.
I). After correction of hearing losses for presbyacusis (Glorig,  1962) a statistical analysis of
the results  was performed.  The scatter of the individual results  was quite large, but dis-
tributed in a near-normal manner (medians and means were equal), so means and standard
deviations  were determined. The results discussed here cover PTS only, as the study on TTS
has not been finished yet (Fig. 2-5)
     Some  characteristic  features of the hearing  losses in that  population may be  sum-
marized as follows:
1.   The most prominent hearing loss during early  exposure (e.g. in groups "under 1 year"
     and "1-2 years of exposure") occurs at 6000 Hz (Fig. 2). It was a "leading frequency"
     in our series.
2.   Hearing loss at 4000 Hz was next in frequency of occurrence and magnitude and after
     5 years of exposure becomes as large as that at 6000 Hz.
3.   The greatest drop in hearing threshold at 6000  Hz and 4000 Hz appears during the first
     two years of exposure (the average rate was 20  dB/year).
4.   Fully-developed hearing loss at 4000 and  6000 Hz appears during the first 5 years of
     exposure  (average rate 10 dB/year). A further drop in hearing threshold occurs slowly
     (on the average, 5 dB over a period of 10 years).
5.   The growth of hearing loss at the lower frequencies was much smaller in the first two
     years (at  2000 Hz about 10 dB/year on the average); a steady slow progression during
     the following years was observed, at an average rate of 5 dB in 10 years. (Fig. 5).
6.   After  2 years  of exposure not a single worker with normal threshold of hearing could
     be found, but individual variations in hearing loss were quite large.

Comment.

     The pattern of noise exposure which has  been found in the present study is different
from a laboratory  type of impulse noise, but typical for industry: the impulses are super-
imposed on a rather high noise background which, by itself, might cause damage to hearing.
But it was probably  the impulses that have  induced a quicker development of PTS in
comparison to  steady-state noise exposure:  fully-developed PTS occurred after 5 years of

                                        221

-------
                                                       Table 1.
                     NOISE-INDUCED HEARING-LOSS IN dB VS. YEARS OF EXPOSURE (MEAN AND STANDARD
                                                DEVIATION - S.D.)
Exposure
time
years
<1
1-2
3-5
6-10
11-15
16-20
21-25
26-30
<1
1
2
3
4
5
srrsirsrrssriEEjscsssErsiz:!:
Age
/meantS.D./
24 ±5
26 ±6
27 ±5
30 ±5
39 ±6
46 ±6
47 ±6
49 ±5






N
/ears/
34
84
98
68
60
22
42
18
34
40
44
34
34
30
Hearing level / dB ISO / - mean and S.D.
500
Mean S.D.
10.3 8.2
12.1 10.3
12.4 9.0
13.2 11.6
14.3 10.9
17.8 15.5
17.8 11.2
18.2 12.3






1000
Mean S.D.
12.42 8.7
17.3 10.1
18.0 8.-
24.9 14.8
26.8 14.8
30.9 20.0
28.8 13.9
31.9 18.6
0 - 5 years






2000
Mean S.D.
17.4 8.2
25.0 15.7
27.4 12.8
34.9 17.6
33.0 19.0
47.2 15.8
36.7 14.9
42.7 13.9
17.4 8.2
18.6 10.5
25.0 16.8
25.6 13.0
24.5 8.-
31.8 15.4
4000
Mean S.D.
24.8 13.3
37.6 23.6
44.6 22.3
49.7 23.2
49.1 17.9
56.3 12.4
48.9 16.5
55.2 17.1
24.8 13.3
27.1 15.3
46.2 26.1
45.5 19.6
41.0 23.0
49.1 23.0
6000
Mean S.D.
28.1 15.7
43.4 23.6
44.6 22.1
50.5 20.2
48.3 16.3
57.0 16.1
50.2 15.2
61.6 24.6
28.1 15.7
31.8 19.0
51.2 25.5
45.7 22.5
40.7 23.0
45.6 20.7
8000
Mean S.D.
24.8 14.2
40.8 22.5
42.4 22.7
48.9 21.1
48.3 17.6
57.9 18.0
50.1 17.8
60.2 26.5
24.8 14.2
31.6 16.1




K)

-------
                tup. hmt < 1y*or.    N* 34 ears.
                     Exp. fimf 1- 2 years.  N * 84 ears.
• i
• -
,.
                                        WOO   6000 Hz
                 Exp.  timt  3-5ytor3.  N* 98 ears.
                     599
9000 Hi
                         500    WO   HOO    WO   MOO  Hz
                     Exp. time  6-10 years,  tf- 68 tars.
WOO  6000  HZ
                  Figure 2. Noise-induced hearing loss (Hearing Levels corrected for presbyacusis) during the first 10 years of
                  exposure. Mean: thick line; S.D.: thin line.

-------
! <

' '

i-
             Cxp. time  11-15 years.   N'Meors
                                              Exp. time n-20 years.     H*2? ears
                              fOOO  tOOO  9000 Hz
txp. time #-?5 years.    N* tf ears
                              two   wo  oooo H:
    500    1000  ZOOO  ¥300  6000 Hi

ftp. time 26-30 years.   N* 18 ears.
                              —i
                                                                           9000  Hz
                 Figure 3. Noise-induced hearing loss (HL corrected for presbyacusis) after 10 years of exposure. Mean:
                 thick line; S.D.. thin line.

-------
to
K)
            10.

            20.

           so
Exposure  time in years
                    A—A   1-2
                    m—•   3-5
                    x—x   6-10
                   500
          WOO
                    Exposure  time in

                       •—•   11 -  15
                       A	A  16 -  2(1
                       m—•   ^J-^5
                       X—x   ffi-30
WOO 6000 6000  Hz    500
1000
tooo
4000 6000 9000
                    Figure 4.  Mean noise-induced hearing loss  (KL adjusted for  presbyacusis)  years of exposure as the
                    parameter.

-------
           10
to
to
co

-------
exposure while it takes approximately 10 years in steady-state noise exposure (Glorig et al.,
1961; Nixon et al., 1961). Similar observations were made by Sulkowski et al. (1972).
    The occurrence of the greatest impulse-noise-produced hearing loss at 6000 Hz has also
been reported by other authors (Loeb and Fletcher, 1965; Salmivalli, 1967; Gravendeel et
al., 1959; Zalin, 1971).
    As far as DRC are concerned, the criteria proposed for gunfire impulses (Coles et al.,
1968) cannot be applied to industrial noise, as Coles and Rice (1970), Walker (1970), and
Martin  et al. (1970) supposed. The impulses of peak level of 125 to 135 dB and duration
100 to  200 msec were clearly harmful when repeated several thousands times  a day for
several  years. The protective action of intraaural muscles seems to be  questionable during
such long exposure, although Cohen et al (1966) reported less harmful effect of impulses
when superimposed on steady state noise. The new concept of the evaluation of risk caused
by impulses reported here by Coles seems to be very promising.
    The large variations in the extent of hearing loss caused  by impulse noise may be
attributed to individual susceptibility: some workers in our series show 50 dB of hearing loss
after 1  or 2 years of exposure, while others have only 20 dB or less after 25 years of work.
The cause  is still obscure,  so routine audiometric tests should be advocated in order to
exclude persons with "tender ears" from  further exposure  or to introduce appropriate
preventive measures (ear protection, enclosures, etc.).
Conclusions.

1.   Fully-developed PTS occurs after 5 years of exposure to impulse noise, which implies
     that rapid transition from TTS to PTS in that type of exposure is quite probable.
2.   The fact that the earliest and greatest change in threshold occurs at 6000 Hz indicates a
     slightly different pattern of development of hearing loss in  impulse noise than in
     steady-state noise exposure.
3.   A very slow increase in hearing loss after 5 years of exposure indicates that PTS caused
     by impulse noise stabilizes  with time, as has been found in  exposure to steady-state
     noise.

                                      Summary

     Results are presented  of hearing examinations in 213 drop-forge operators who were
exposed, for  1  to 30 years, to 3000 to 10000 impulses a day, the impulses having a peak
level of 127-134 dB and duration of 100 to 200 msce. The impulses are superimposed on a
background of 110 dB of steady-state noise.
     The maximum PTS during early exposure (1-2 years)  occurred at 6000 Hz, at a rate
of 20 dB a year. Hearing loss at 4000  Hz was smaller, but after 5 years of exposure  they
become equal. PTS at 6000 Hz and 4000 Hz reached maximum in first 5 years of exposure
(about 50 dB on the average); further increase was rather slow. At 2000 Hz a smaller PTS
during early exposure is observed: about 10 dB per year during first 1 - 2 years, later about
SdBin 10 years.

                                        227

-------
     Fairly large individual variations in threshold shift were observed, which resulted in a
20-dB standard deviation from mean values of PTS (the scatter of values was close to
"normal curve").

                                  REFERENCES

Acton, W.I., A review of hearing damage risk criteria. Ann.  Occup. Hyg.,  10,  143 - 153
     (1967).
Cohen, A., Kylin, B., La Benz, P. J., Temporary threshold shifts in hearing from exposure to
     combined impact/steady-state noise  conditions. /. Acoust. Soc. Am., 40, 1371-1380,
     (1966).
Coles, R.R.A., Garinther, G.R., Hodge, D.C., Rice, C.G., Hazardous exposure to impulse
     noise./. Acoust. Soc. Am.,43, 336-343 (1968).
Coles, R.R.A., Rice, C.G., Towards a criterion for impulse noise in industry. Ann. Occup.
  .   Hyg., 13,43-50(1970).
Glorig, A., Nixon, J., Hearing loss  as a  function of age. Laryngoscope, 72, 1596-1611
     (1962).
Glorig, A., Ward, W.D., Nixon, J., Damage risk criteria and noise-induced hearing loss. Arch.
     Otolaryng., 74, 413-423  (1961).
Gravendeel, D.W., Plomp, R., The relation between permanent and temporary noise dips.
     Arch. Otolaryng., 69, 714-719 (1959).
Loeb, M., Fletcher J.L., Benson R.W., Some preliminary studies of temporary threshold
     shift with an arc discharge impulse noise./. Acoust. Soc. Am., 37, 313 (1965).
Martin, A.M.,  Atherley, G.R.C., Hempstock, T.I., Recurrent impact noise from pneumatic
     hammers. Ann. Occup. Hyg.,  13, 59, 67 (1970).
Nixon, J.C., Glorig, A., Noise-induced permanent threshold shift at 2000 cps and 4000 cps.
     /. Acoust. Soc. Am., 33, 964-968 (1961).
Salmivalli, A., Acoustic trauma in regular army personnel. A eta Otolar.,  (Stockh.), Suppl.
     222(1967).
Sulkowski, W., Dzwonnik, Z.,  Andryszek, Cz.,  Kipowczan,  A., Ryzyko Sawodowych
     uszkodzen sluchu w halasie ciaglym, przerywanym i impulsowym. Med. Pracy, (Pol.),
     23,465-481 (1972).
Walker,  J.G.,  Temporary threshold shift from impulse noise. Ann.  Occup. Hyg.,  13, 51-58
     (1970).
Zalin, H., Noise induced hearing loss. Unmasking other pathology. Proc. Roy.  Soc. Med., 64,
     187-190(1971).
                                       228

-------
                  HEARING DAMAGE CAUSED BY VERY SHORT,
                        HIGH-INTENSITY IMPULSE NOISE

                                   H.G. Dieroff
                          HNO-Klinik, University of Jena
                         Jena, German Democratic Republic
Introduction
    Much has  been reported in recent years about the harmful effect of impulse noise-
especially of very short impulses-on hearing (DIEROFF, GARINTHER and MORELAND,
KRYTER, POCHE, RICE and  COLES, STOCKWELL and ADES). Findings  of damage
caused by ultra-short impulses such as produced by certain toys have also become more
frequent (GJAEVENES, HODGE and McCOMMONS, etc.).
    Rather striking cases of hearing loss are often found in  persons occupied in metal-
working trades  and working in places where continuous sound pressure levels hardly exceed
the crucial intensity of about 90 dB (A). Again and again, individuals are found to suffer
from hearing loss that seems to bear no relation to the sound pressure level (SPL) prevailing
at their work place. The frequently-encountered hearing impairment  in  welders is repre-
sentative of the situation to be described here, as Fig. 1 demonstrates. There is no doubt
that impulse noise, even only a few high-intensity impulses, produced when burrs and slag
are removed from welds  or when the acetylene flame is ignited, largely account for the
extent of the damage. Measurements of such noise reveal very irregularly scattered impulses
of varying quality. Level frequency counts permit a considerably better  assessment of
auditory stress, but even they are not sufficient to establish a relationship between the
actual noise stress  and  the  hearing  loss  detected.  Therefore, today's still-inadequate
techniques of measuring the actual stress involved in densely pulsed industrial noise has to
be regarded as  a major reason for the apparent discrepancy between noise levels and the
degree of hearing impairment. Moreover, the question arises whether impulse noise involves
a damage mechanism different from that caused by continuous noise, which in the end leads
to the familiar severe hearing damages.

Theoretical considerations

    The author's investigations, especially in the metal-working industry, have  shown that
the clashing of two metal parts frequently produces SPL peaks between  150 and 160 dB.
Peaks of that intensity may cause hair-cell damage by way of a heavy electrical discharge of
the hair-cells when the hairs  touch the  tectorial  membrane,  which  leaves a  scar in the
hair-cell's microstructure. A hair-cell thus damaged may sooner or later suffer degeneration.
    Damage after steady-state acoustic stimulation has been described as a deficiency in the
supply  of nutrients to hair-cells; but  assumptions  primarily relying on  the same type of
metabolic disturbance to account for impulse-noise-induced  hearing damage  carry little
conviction.
    A  research  team  consisting  of   BIEDERMANN,  GEYER,   GUTTMACHER,
KASCHOWITZ, MEIER and  QUADE, in addition  to the author, investigated the issue of
                                       229

-------
                                      £r$chweisscr
                      '.".  'Vt IikX  >.',." *«'* r'
                                                                    Ql,r
Figure 1. Single audiograms (air conduction) of welders, each curve with age and number of working years
in noise {after DIEROFF, 1962).

-------
whether the exposure of guinea pigs' ears to very short sound impulses causes functional
hearing changes as well as structural deformations of and metabolic damage to the organ of
Corti. We  hoped  to determine  to  what  extent  there is  any correlation between  the
functional behavior and the microscopical and histochemical findings.

Methods

     1. Functional behavior test series, carried out by BIEDERMAN and KASCHOWITZ:
Guinea pigs weighing 250 to 450 g and showing normal PREYER reflexes were exposed to
acoustic stimulation in  a low-reflection sound chamber of about  1 x 1 x 2.2 m in size. A
spark-discharge sound generator was used as the source; sound pressure peaked around  162
dB with a standard  deviation of ± 1 dB. Impulse widths varied from 200 to 400 jus. The
functional behavior  of  the guinea pigs' ears was  measured in terms of the microphonic
potential  (MP) measured at the round  window. MP was measured exclusively during the
experiment.
     2. Metabolic behavior test series, carried out by Geyer, Quade und Guttmacher, Meier:
Experimental  subjects were guinea pigs weighing  250 to 450 g and exhibiting positive
PREYER reflexes. The  animals were placed  two at a time in a sound chamber containing
very narrow  cages. Acoustic measurements  demonstrated a  uniform sound  pressure
distribution in the area of the animals' heads. Sound pressure levels of  135 or 158 dB ± 1 dB
were  used. Investigation  included  the microstructure and  the behavior  of succinic-
dehydrogenase (SDH) activity in the cochlea.  The same spark-discharge sound generator was
used for  both test series. In either series, some control animals  served  as a  basis of
comparison for the pathological changes to be detected.

Results

     In the functional  behavior experiments, the  guinea  pigs were exposed to 1, 3 or 5
impulses at intervals of about 3.5 seconds, which was sufficient to allow the middle-ear
muscle reflex aroused by the preceding impulse to relax. The first  MP measurement was
made 5 minutes after the start of each  experiment. The MP amplitude, which was recorded
for 120 minutes after stimulation (Fig. 2), showed a noticeable  attenuation after a single
impulse, decreased more rapidly after  3 impulses, and continued to fade after 5  impulses
with a tendency toward an asymptote. Throughout the observation  period, no increase in
MP was detected. Extending the period any  longer was not practicable,  since the MP
measurements were carried out with anaesthesized animals.
     In contrast to the functional behavior described,  the second test-series required a much
longer period of acoustic stimulation before a distinct decrease in SDH could be found. The
first SDH changes were observed only after 8 days, with a daily sound exposure time  of 9
hours alternating  with  15 hours of rest, and a pulse rate of 16 per minute. A marked
increase was detected after the period of sound exposure had been extended to 14 days. On
the other hand, acoustic stimulation of these animals did not produce any MP or any Preyer
reflexes prior to sacrificing that would be indicative of a residual function.
     The decrease in SDH activity extended over  the entire cochlea  and also included the
nerve endings. SDH activity decreased  more in the outer than in the inner hair-cells. No

                                        231

-------
micro-structural changes could be detected. Other histochemical details are not considered
here.

Discussion

    The two series of experiments showed a large discrepancy between the behavior of MP
and that of SDH.  A change in MP can be noticed after the very first impulse, whereas a
decrease in SDH activity will occur only after extensive exposure to impulse sound. How-
ever, histochemical changes after  stimulation by impulse sound are identical to those
observed after continuous stimulation, since the results are well in agreement with observa-
tions reported by VOSTEEN (1958, 1960, 1961), VINNIKOV and TITOVA (1958, 1963)
and QUADE and GEYER (1972). As the only divergence, VOSTEEN in his  experiments
found no decrease of enzyme in the inner hair-cells and nerve endings after minor sonic
stress. According to VOSTEEN (1958), the decrease of SDH activity is to be taken as a stage
preceding the disintegration of hair-cells, with the decrease in SDH activity corresponding to
a state of exhaustion after functional over-exertion.
     Summarizing  the results of the  two test series, we find  that sound impulses of high-
peaked SPL cause  a reduction of MP  that was irreversible for the  duration of the measure-
ment, whereas the impulse  sound stress required for histochemical changes is much longer
than the continuous sound stress that would produce the same changes. The reduction of
MP suggests the very rapid occurrence of a functional over-exertion or  damage due to a few
very short pulses  of some hundreds of microseconds  duration. These can  probably be
explained  by  the  electron-microscopic  observations  made  by SPOENDLIN. After
                                                           70S (min) IX
 Figure 2. The behavior of microphonpotentials of guinea pigs after 1,3 and 5 sparks depending from time
 (after BIEDERMANN and KASCHEWITZ, 1973).

                                       232

-------
continuous  acoustic stimulation  with wide-band  noises  ranging from  125 dB upward,
SPOENDLIN  found changes in the ultrastructure of the outer, and later of the inner,
hair-cells and of the nerve endings, which changes progressed with growing sound intensity.
These changes were of a partially mechanical and partially metabolic nature. The influence
exerted on MP by a few single impulses might be due primarily to purely mechanical damage
to the ultrastructure of the organ of Corti, caused by direct contact between the tectorial
membrane and the outer hair-cells, with a possibly stronger affection of the sensory hairs.

Summary

    A  few  single impulses (1, 3 or 5) having  sound pressure peaks of  162 dB  and pulse
widths  of up  to 400 /its induce a permanent reduction of  microphonic potential in guinea
pigs.  A considerably longer  period  of acoustic  impulse stimulation (Le. 8 to 14 days) is
required to  detect histochemically  the same SDH decrease as that found after continuous
sound stimulation, a decrease assumed to be a stage preceding hair-cell degeneration. What
changes in the organ of Corti are responsible for the observed microphonic potential reduc-
tion remain to be identified.

                                    References

Biedermann and Kaschowitz: Personal communication 1973
Dieroff, H.G.: Zur Problematik der Schlagimpulse  in Industrielarm.  Arch. Ohr.,-Nas.-
    u.Kehlk.-Heilk. 179, 409 (1962)
Dieroff, H.G.: Schlagimpulse und Larmschwerhorigkeit. Intern. Audiology 2 (1963)
Dieroff, H.G.: Horschaden durch impulsreichen  Larm und dessen Erfassung. Kongre/3ber-4.
    Akustiche Konferenz Budapest 1967, Teil I
Dieroff, H.G.: Der gehorschadigende Impulslarm in  der  Industrie  und seine Erfassungs-
    probleme. Kampf dem Larm, H. 2 (1969)
Dieroff, H.G.: Zur besonderen Larmsituation bei Schlossern,  Schwei/Jern. Pressern sowie
    Stanzern und den  zu erwartenden Larmhorschaden. Praxis der Larmbekampfung, AICB
    KongrejS Baden-Baden 1966 Verlag fur Medizin - Technik, K.H. Walter, Baden-Baden.
Dieroff, H.G.: Einige  spezielle Fragen der Gehorschadigung durch Impulslarm in  der
    Industrie. Proc.  Vol.  Ill  S.  37. 3.  Konferenz Kampf gegen Larm und  Vibration
    Bukarest 1969
Dieroff, H.G.: Der Einflu(3 von Schlagimpulsen  auf das AusmajS der Horschadigung eines
    Larmarbeiters. Z. ges. Hyg. H. 2 (1973)
Garinther, G.R., and J.B. Moreland: Transducer  Techniques of Measurin the Effect of Small
    Arms Noise on Hearing. U.S. Army Human Eng. Labs., Aberdeen Proving Ground, Md.,
    Tech. Mem No. TM 11-65
Geyer,  G.,  R. Quade und  H.  Guttmacher: Histochemischer  Nachweis von Succinode-
    hydrogenase-Aktivitat   im   Cortischen  Organ   vom  normalen   und  beschallten
    Meerschweinchen. Z. ges. Hyg.  19, 25 (1973)
Gjaevenes, K.: Measurements on the  Impulsive Noise from Crackers and Toy Firearms. J.
    Acoust. Soc.  Am. 39, 403 (1966). Damage-Risk Criterion for the Impulsive Noise of
    Toys. J. Acoust. Soc. Am. 42, 268 (1967)

                                       233

-------
Hodge, D.C. and R.B. McCommons: Acoustical Hazards of Children's Toys. J. Acoust. Soc.
    Am. 40, 911 (1966)
Kryter, K.D.: The Effects of Noise on Man. Academic Press New York and London 1970
Meier, Ch.:  Personal communication 1973
Poche, L.B., Ch.W. Stockwell and H.W. Ades: Cochlear Hair-Cell Damage in Guinea Pigs
    after Exposure to Impulse Noise. J. Acoust. Soc. Am. 46, 947 (1969)
Quade, R.  und G. Geyer: Der Succinatdehydrogenase-Nachweis mit Hilfe der Perfusion-
    stechnik an der Cochlea des Meerschweinchens  unter Normalbedingungen  und nach
    Dauerlarmeinwirkung. Acta Otolaryng. (Stockh.) 75, 45 (1973)
Rice, C.G. and R.R.A. Coles: Impulsive Noise Studies and Temporary Threshold Shift. Proc.
    Intern. Congr.  Acoust. 5th, paper B 67 (1965)
Spoendlin, H.: Primary Structural Changes in the Organ of Corti after Acoustic Overstimula-
    tion. Acta Otolaryng. (Stockh.)  71, 166  (1971)
Vinnikov, J.A. and L.K. Titova: Intern.  Rev. Cyt 14, 157 (1963)
Vosteen,  K.H.:  Die Erschopfung der Phonorezeptoren nach funktioneller Belastung. Arch.
    Ohr.,-Nas,-und Kehlk.-Heilk. 172,  489 (1958). The Histochemistry of the Enzymes of
    the Oxygen Metabolism in the Inner Ear. Laryngoscope 70, 351 (1960). Neue Aspekte
    zur  Biologic und Pathologic des Innenohres. Arch. Ohr.,-Nas.-u.Kehlk.-Heilk. 178, 1
    (1961)
                                       234

-------
                SESSION 3





 NOISE-INDUCED HEARING LOSS-MECHANISM



Chairmen: H. G. Dieroff (DDR), R. Hinchcliffe (UK)
                   235

-------
          BEHAVIORAL, PHYSIOLOGICAL AND ANATOMICAL STUDIES
                      OF THRESHOLD SHIFTS IN ANIMALS1

                  Donald H. Eldredge, James D. Miller, John H. Mills
                              and Barbara A. Bonne2
                  Research Department, Central Institute for the Deaf
                             St. Louis, Missouri 63110

    In a Congress on  Noise  as a Public  Health Problem we  are really very much more
concerned with man  than  we are with animals.  So what is it that we would like to know
about noise and man  that we may learn from animals? As representatives from an Institute
for the Deaf we would like to know how  to evaluate hazards for hearing and the ear from
information about exposures to noise. Exposures to noise are commonly specified in terms
of the  level  and spectrum of the noise, and the temporal pattern  and total duration  of
exposures. For man, the effect on the ear is usually specified in terms of temporary and/or
permanent shifts of the  auditory thresholds for pure tones. In animal  experiments we would
like to learn as much as possible about the above relations and in addition to learn some-
thing of the pathological physiology and pathological anatomy.
    From animal experiments we  have long known  that exposures to sound can  lead  to
degeneration of the hair cells of the organ of Corti with associated loss of neurons. Most
such experiments  have involved  short  exposures  at high  levels  and  have  served  to
demonstrate  the susceptibility of the ear to sudden acoustic trauma. We are all familiar with
these experiments, and they will not be of much concern to us today because these studies
have not been so  clearly  oriented toward the  human situation. The exposures that are
important for man  do not seem to injure with a single event. Instead, loss of hearing follows
only after repeated exposures over relatively long periods. Accordingly we wish to review
some of our  recent work that emphasizes prolonged exposures at lower levels and exposures
that are not  immediately associated with permanent loss  of threshold  sensitivity for tones.
On the basis of this work we can already state several important relations between exposure
to noise  and loss of hearing in the chinchilla. The measured relations for  the chinchilla
correspond so well to  similar relations for man that we believe common principles apply.
    Advantages of Animal Experiments'. Animal experiments can have certain rather clear
advantages. First, and possibly foremost, it is possible to restrict activities and exposures in
such a way  as to reduce individual variability in responses to exposures to a minimum.
Secondly, it is  possible to look systematically at some  limiting conditions that are not
always  reasonable for man. Thirdly, it is reasonable and possible to  examine the  ears of
animals  under  the  microscope  at appropriate times  and with  good  preservation  of
anatomical detail.
    Advantages  of Chinchillas: The choice of an animal is another important step. We
chose the chinchilla because:
     1)  it can  be  easily  trained,  using  standard  shock avoidance  techniques, to give
behavioral responses to tones;
     2) its threshold of audibility has a sensitivity and a frequency range similar to those of
man;

                                        237

-------
     3) it is relatively healthy and free of diseases of the middle ear;
     4) it has a long life, at least 10 years and up to 20 years; and
     5) three turns of its cochlea are surgically accessible to electrodes for the recording of
cochlear potentials (Miller, 1970).
                                 Behavioral Studies

     Acquisition of Threshold Shifts: Let us begin by asking what happens to the behavioral
auditory thresholds when exposures to noise are continuous. Carder and Miller (1971, 1972)
showed that the threshold for a 715-Hz  tone increased with duration  of exposure to an
octave band of noise centered at 500 Hz for only about 24 hours (1440 min.) and then
remained at a plateau or asymptotic value as exposures were continued for 7 days and even
for 21 days. Four  examples of this kind of growth of threshold shift to asymptotic values
are shown in the left panel of Fig. 1. Here we see that the effect of changing the level of the
band of noise from 75 to 85, 95, and  105 dB SPL is to shift the asymptotic threshold shift
from about  17 dB to 31 dB, 49 dB, and 63 dB, respectively. Note that for each exposure the
asymptote is reached in about the same time and that rate of growth of shift has increased
correspondingly with level.
     At this point  a prominent footnote is required. One of our most important findings is
that the terms "temporary threshold shift" (TTS) and, "permanent threshold shift" (PTS)
are not enough to distinguish among the  operationally important features of all of the
various threshold shifts (TS) that we have encountered. At a minimum we  will need a new
term similar to "asymptotic threshold shift" (ATS). All of the evidence is not yet available
and  we are not yet prepared to recommend a definitive set of terms. The notations we have
used in the  past reflect the fact that an animal must be removed from the noise to measure
his TS. Some recovery from the level of TS present in the noise begins as soon as the animal
has been placed in the quiet, but  this is usually small and the animal is returned to the noise
as soon as  thresholds have been measured. Measurements of  threshold twice at a single
frequency can be made at an average time out of noise  of 4  minutes and this time may be
given as a subscript; e.g., TS4, TTS4. Measurement of thresholds for an audiogram of five or
six frequencies can be made at an average time out of noise of 11 minutes; e.g., TTSj j. In
addition we have earlier  expressed ATS4 as TTS4°° or the TS4 measured with the TS has
stopped growing.
     Part of the data in the left panel of Fig. 1 are replotted in Fig. 2A to show the relation
of ATS4 at  715 Hz to the level of the band of noise. Note that the relation is well described
by the equation

                               ATS4=1.6(OBL-65)

That is, for  every decibel that the band level exceeds the subtractive constant 65 there will
be a 1.6-dB increase in ATS.

                                        238

-------
to
OJ
so
          dB

          70
       (/>
       
-------
     TTS4   =1.6(081-65)
                                                          70
                                                       N  60
                                                       I
                                                          40
                                                     V)
                                                     g
                                                     X
                                                       2 30
                                                     tr
                                                     <
                                                     CL
                                                     S
                                                     UJ
                                                          20
                                                          10
                                                                      B.
                 TTS4fl0= 1.6 (OBL-47)
50      60     70      80      90     100     110
  SPL OF OCTAVE BAND  CENTERED  AT 0.5kHz
          30     40      50     60      70      80     90
           SPL OF OCTAVE BAND CENTERED  AT 4.0 kHz
 Figure 2A. The relation of threshold shift at asymptote (ATS^ for a test-tone of 715 Hz as a function of
 level of noise in an octave band centered at 500 Hz. Filled circles are means of the data in Fig. 1. The open
 square is from Miller, Rothenberg and Eldredge (1971) and the open circle from unpublished data of Miller,
 Eldredge and Bredberg (after Carder and Miller, 1972).
 Figure 2B. The relation of threshold shift at asymptote (ATS^ for a test-tone of 5.7 kHz as a function of
 level of noise in an octave band centered at 4 kHz. The open circles are from the data in Fig. 4A. The filled
 .squares are from other experiments (after Mills and Talo, 1972).
      The curvilinear extrapolations in both  Fig. 2A and Fig. 2B are based on the equation
                   ATS4 = 1.6
10log10l-£j—£)
where le is the square of the
                                           •c
 sound pressure of the noise and lc is a constant such that 10 log^n. lc = 65 for Fig. 4 A and 10 log-jg
 for Fig. 4B.

-------
    The threshold shifts (ATS j ]) as a function of frequency for the above exposures are
shown as audiograms with exposure level as the parameter in Fig. 3A. Here we see that the
largest shifts were at 715  Hz and that there was significant spread of ATS^ to higher
frequencies. Parenthetically we find  the shifts at 715  Hz to be easily replicated. There is
always some spread of ATS to higher frequencies, but it is often not so much as shown here.
    Von Bismarck (1967)  has measured the ratio of sound pressure at the ear drum (P^) of
the chinchilla to that in the free field (Pf) as functions of frequency, angle of incidence, and
static pressure in the middle ear. For the frequencies included in an  octave band centered
around 500 Hz he found the ratio P^/Pf to be about 3 dB. For the frequencies included in
an octave band centered around 4 kHz he found the ratio Pd/Pf to be 15-20 dB. These
differences are similar to those reported for man by Wiener and Ross (1946) and for the cat
by Wiener, Pfeiffer and Backus  (1966),  and represent sound pressure transformations
produced by combinations of acoustic diffractions about the head and pinna and resonances
in the outer ear  canal. Carder and Miller reasoned that the physiologically important
measure of exposure level will be P
-------
to
          dfi
            o
           10
           20
           30
40
           50
           60
                  OCTAVE  BAND NOISE
                  cf= 500 Hz
                        85 dB
                                  (A)
                        AFTER CARDER 8
                        MILLER
                                                                         I   I
OCTAVE BAND  NOISE
cf = 4.0 kHz

AFTER MILLS & TALO
                                                                                   5.7   11.4
                                                                                    I   I    I   I
               0.25     0.715      2.0   4.0   8.0   0.25  0.5    1.0   2.0   4.0   8.0   16.0

                                       FREQUENCY IN  KILOHERTZ

            Figure 3A. Mean audiograms at asymptote (ATS-ji) across ears and days at asymptote for the ears and
            conditions described in Fig. 1. (after Carder and Miller, 1972).
            Figure 3B. Mean audiograms at asymptote (ATS^) across ears and days at asymptote for the ears and
            conditions described in Fig. 4A. (after Mills and Talo, 1972).

-------
to
       60
    1	T-	J	T
       50 -
     Kl
     X
     JC
       30
    m
20
     (ft
        10
                   GROWTH
        *..^*—+J

     f    57 dB	*
                             65 dB-
                                                        1	P
                                               80dB
                                        m ^•^•^'•~*

                                        **'
                                          72 HB	••
                                              EXPOSURE:
                                              OCT-BAND NOISE
                                              cf - 4000 Hz
                                                                           T	p
      DECAY
                                                                                                                           T-f>-
        10
                                                                            60
                                                                                           t  t  t 1  1    t  f  1  t  t  t  1   t
                                                                                      12 1824 36
                         I
                                                                     MINUTES
                                                                                    HOURS
347  8 12 19  29  89
        'DAVS
                 6           12
                    DAYS OF  EXPOSURE
                                                 IS
24   4       50   100         500  IK         5K  10K  20K
                   TIME  AFTER NOISE  !N  MINUTES
                50K"lOOK
                    Figure 4A. Growth of TTS4 at 5.7 kHz. The levels of noise were 57,65, 72, and 80 dB SPL. For each level
                    of noise, the duration  is six days; the total duration is 24 days. Each point is the mean lor four animals
                    {after Mills and Talo, 1972).

                    Figure 4B. Decay of TTS at 5.7 kHz. The filled circles show the delayed recovery measured following the
                    exposure in Fig. 4A. The other data points are from other experiments  in which the more characteristic
                    slow recovery was observed (after Mills and Talo, 1972).

-------
     Up to this point in our review we can conclude:
     1) Continuous exposures to noise lead to losses of threshold sensitivity for pure tones
that increase for about 24 hours to an asymptotic threshold shift that may be maintained
for many days.
     2) The apparent stress on the  ear as indicated by the size of the ATS grows by 1.6 to
1.7 dB for each decibel increase in level above some critical level.3
     3) The critical level, if measured at the eardrum, is probably  only weakly related to
center frequency of the exposure-band over the range from about 0.4 kHz to about 6.0 kHz.
When measured in the field, the critical level is strongly related to the center frequency of
the exposure-band because of the enhancement of sound pressure level produced by
acoustic diffractions around the head and resonances of the ear canal.
     Recovery from ATS: Once the state we have called ATS4 has been reached, exposures
may be terminated and we  can study the recovery of sensitivity for pure tones. When this is
done we learn that ATS is different from some other forms  of temporary threshold shift in
that when ATS4 exceeds about 10-15 dB and the exposure has been nearly Continuous, then
recovery to normal will take several days. Under these circumstances recovery is never rapid
or even prompt. In the right half of Fig.  1 we see that recovery to normal threshold at 715
Hz required 2 to 6 days following exposures to the octave band centered at 500 Hz. There is
a slight trend such that with greater  ATS more time is required for recovery.
     Two  trends for recovery of sensitivity at 5.7 kHz following exposure to octave-band
noise centered at  4 kHz and at 80 dB SPL  are shown in Fig. 4B. The open symbols,
collected from several experiments  in which miscellaneous parameters were being explored,
show the same slow recovery observed at 715 Hz in Fig. 1. The filled symbols show an even
slower or delayed recovery  that was observed following the exposures and ATSs shown in
Fig.  4A.  Recovery to thresholds that were not significantly different from  pre-exposure
thresholds required at least  12 to  28 days and, as we shall see, these exposures produced
small changes in cochlear potentials  and small cochlear injuries.
    Permanent Threshold  Shifts:  The preceding paragraphs and  figures summarize  the
important trends in those relations  of behavioral threshold shifts to noise exposures that we
have already published. From the  point  of view of environmental safety these exposures
have already gone far enough. The  results of subsequent  studies now in press or in progress
show that increases in either dimension of exposure spell environmental hazard.  For example,
continuous exposures for 9  days  to an octave band of noise centered at 4 kHz  and at 80 dB
SPL have not produced permanent  shifts  in behavioral thresholds and are accompanied by
only minor changes in  cochlear potentials  and minor injuries to the inner ear. However,
when the level of exposure is increased to  86 dB or when the exposure to 80 dB is extended
for 90 days, we  see permanent threshold shifts. The recovery of  behavioral  threshold
sensitivity is delayed or even prolonged to more than 60 days and is not complete. When
there are permanent threshold shifts of this kind we have always found reduced cochlear
potentials and injuries to the inner ear.

                                Physiological Studies

     In our descriptions  of the behavioral  studies we have  alluded to changes in cochlear
potentials and to injuries to the inner ear. In the context  of the  present conference the
                                        244

-------
physiological studies are important because we find physiological changes that are quantita-
tively large enough to account for all of the behavioral loss of sensitivity. Thus we need not
invoke any central nervous system components to account for the threshold shifts.
     In a small rodent such as the chinchilla the three scalae of the individual turns of the
cochlea are readily  accessible for the insertion of electrodes. Cochlear microphonic (CM)
potentials can be measured differentially in each of three turns by recording the electrical
differences between scala  vestibuli and  scala tympani in the manner of Tasaki, Davis and
Legouix (1952). The  classic N^-N2 waveform of the whole-nerve action potential (AP)
response  can be recorded  by taking the average potential difference between the pair of
electrodes in the basal turn and a ground reference in the tissues of the neck wound. This
response appears most clearly for clicks  or  the onsets of tones and can be grossly analyzed
into smaller components  by observing  the changes  produced by masking noises with a
computerized version of the method reported by Teas, Eldredge and Davis (1962). It is also
possible to measure the endocochlear DC potential with pipette electrodes inserted in scala
media of each turn.
     Cochlear Potentials Early in Recovery: Benitez et al. (1972) exposed chinchillas to an
octave band of noise centered at 500 Hz and at 95 dB SPL for two to three days. This is
long enough to  assure that the state of ATS had been reached. Then DC, CM and AP were
measured at the  times indicated on the right half of Fig.  1. Two hours after the end of the
exposures the DC potentials were within our range of normal values.
     The  changes in CM responses to tones at 200 Hz are shown in Fig. 5. Panel A shows the
normal growth of CM voltage with increasing sound pressure level at the eardrum for each of
the three cochlear turns. At low levels the CM measured in the third, or apical, turn (CM3) is
larger than CM2 and CM j  in a manner consistent with the envelope of the Bekesy traveling
wave at this frequency. However, the three functions become nonlinear at different levels
and the  CM arising more basally  continues to  increase at higher levels. At maximum
response  the rank  order is reversed so that CMj gives the  largest voltage. In panel B the
average of CMj functions after  5  hours of recovery is compared to the mean control
function  for CMj.  There is 12 dB loss of sensitivity, 6 dB loss of maximum voltage, and the
SPL required to  produce maximum  CMj is shifted about 6 dB higher. Panels C and D show
similar comparisons for CM2 and CM3 after recovery  for  5 hours.  This low-frequency
exposure has produced a  clear gradient of increasing loss from less near the base to more
near the apex.
     The  recovery  of loss of sensitivity for CM as measured 5, 24,  and 48 hours after
termination  of exposure is shown in  Fig.  6 along with  the recovery of  mean behavioral
threshold  sensitivity at 715 Hz.  The physiological data are the means of three different
groups of five ears each at the three different times.  The trends for recovery are the  same for
all measures.
     For  more than 6  hours after the termination of exposure at 95 dB SPL to the band of
noise centered at 500  Hz it was not possible to elicit  an AP response to a wide-band click
Jeyen at levels 90 dB above normal visual detection levels  for this AP. A brain-stem evoked
response  could be  elicited at levels consistent with the behavioral and CM losses of sensi-
tivity. These two observations imply changes in degree of synchrony as well as changes in
sensitivity of neural responses. We  are not now prepared to interpret these changes any
further.
                                         245

-------
         3K
         IK
              CONTROLS
            -  200 Hz
    a.
     i
    a.
       300
   UJ
   CD
   
-------
       N
      X
      o
      o
      CM
      00
to
      00
      2
      LU
      00
           dB
           50
40
           30
           20
10
                           CM
           BEHAVIORAL
           ITS       x,
           (715  Hz)
                CM
                                  1
                                                          1
1
                                  5              10                  24

                           RECOVERY  TIME AFTER  NOISE   (HOURS)

             Figure 6. Recovery of CM sensitivity as a function of time after termination of exposure. The shifts in
             sensitivity at the lOO-^V level on the input-output functions for 200-Hz tones are shown by the ordinates.
             The parameter is the cochlear turn. Recovery from behavioral TTS for tones at 715 Hz is shown by the
             dashed line for comparison (from Benitez et al., 1972).
                                                                         48   HRS

-------
     Cochlear Potentials Late in Recovery. Even  when behavioral thresholds recover to a
sensitivity that cannot be statistically distinguished from normal pre-exposure values (s.e.m.
about 4 dB) we often find residual changes in cochlear potentials (Eldredge et al., in press).
For example, there were four ears in the group used for the four levels in Fig. 4. The CMj
responses were measured in each of these ears after the course of exposures and the recovery
shown. In each ear, CMj was found to be about two standard deviations below mean normal
values throughout most of the  dynamic range. The whole-nerve AP responses had normal
thresholds and in one ear grew quite normally; however, the other three ears gave responses
at suprathreshold levels that were up to more than two standard deviations below normal
means.

                                Anatomical Studies

     Early in these studies involving ATS we were  surprised to  find evidence for some
injuries to the organ of Corti and loss of hair cells with little or  no change in behavioral
thresholds. The first examples were briefly noted by Miller et al. (1971) and by Carder and
Miller (1972). These early indications of injury have now been confirmed repeatedly.
     Changes Seen Early in Recovery: Benitez et al. (1972)had restricted their exposures to
only two or three days in an attempt to achieve ATS without the minor loss of hair cells
reported by Carder and Miller (1972). Nevertheless, 23 of 30 ears so exposed and examined
by conventional histological methods showed unmistakably-missing hair cells and in  13 of
these ears the number of missing cells clearly exceeded  those we had seen in unexposed
control ears. The samples provided by only every fifth section did not allow more precise
evaluation of the injuries.  Also, the fixation of cell structure in  decalcified, celloidin-
embedded sections was not adequate for evaluation of more subtle cytological changes. For
this reason we adopted the  osmium-fixed, araldite-embedded, flat preparation described by
Bohne (1972) for our subsequent studies.
     We have  already been handsomely rewarded by this extra  care in our anatomical
preparations. Bohne (in preparation) has reviewed ears exposed continuously for 48  hours
to an octave band centered at 500 Hz with levels  at 75, 85, or 95 dB SPL and to an octave
band centered at 4  kHz at 80  dB  SPL.  Ears that are fixed one to two  hours after the
termination  of exposure show, by phase  contrast  microscopy, outer  hair cells  with
thickened walls and with intracellular accumulations of homogeneous-appearing material.
These cells are seen primarily in the lower part of the third  turn after exposures to the
octave band centered at 500 Hz and in the lower part of the first turn after exposures to the
octave band centered at 4  kHz. The incidence of outer hair cells with the above changes
decreases with level of exposure.
     When the outer hair cells are examined  by electron microscopy, it becomes clear that
there has been a marked proliferation of the cisternae of the smooth endoplasmic reticulum
which comprises the peripheral  membrane  system of the cells. The top of Fig. 7 shows part
of a normal outer hair cell sectioned at a horizontal angle. Eight rows of flattened cisternae
are present just inside the plasma membrane (at the arrow) of the cell. There are 30 rows of
cisternae hi the peripheral membrane system in the outer hair cell from the third turn which
is shown in the bottom part of Fig. 7. This cell is from an animal exposed to octave-band

                                        248

-------
                                         ."           .

         "

Figure 7. Electron micrographs of portions of two  outer  hair cells showing  the  smooth endoplasmic
reticulum which forms the peripheral membrane system of the cells. Top shows eight  rows of cistemac in a
normal cell. The plasma membrane is at arrow. Bottom shows proliferation of the cisternae to form 30 rows
in a cell from a noise-exposed chinchilla with an asymptotic threshold shift.

-------
noise centered at 500 Hz at 95 dB SPL; the specimen was collected 1-2 hours after the end
of the exposure. The homogeneous-appearing material seen in the outer hair cells by phase
contrast microscopy was found to consist of extensions of the cisternae of smooth endo-
plasmic reticulum into the central portions of the cells. These changes are slowly reversible.
After 7 days of recovery the membrane systems in the outer hair cells show a return toward
average proportions. However, some scattered outer hair cells still contained small areas of
the proliferated cisternae even in ears allowed to recover for 70 days.
     Changes Seen Late in Recovery: Both  early and late,  scattered loss of up to a few
hundred hair cells has been a  regular finding. On closer  examination some of these ears are
showing changes that  persist more than three months after  exposure, at a time when
recovery to stable behavioral  thresholds is complete.  For example,  some of the outer hair
cells in places where outer hair cells have been shown to develop extra rows of cisternae in
the peripheral membrane systems may have membrane systems of average dimensions, but
the shape of the cells may be irregular and the peripherally-arranged mitochondria reduced
in number (Bonne et al., in press). Also, the afferent  nerve fibers under the inner hair cells
and  in  the  outer spiral bundles may show increased numbers of vesicles, vacuoles, and
strands of endoplasmic reticulum.

                         Parallels Between Man and  Chinchilla

     Behavioral measures of  thresholds are the only ways we can compare man to  the
chinchilla. For the chinchilla we have demonstrated the phenomenon of asymptotic thres-
hold shift. Mills et al.  (1970) and Mosko  et al.  (1970)  have demonstrated asymptotic
threshold shifts in man. For both man and chinchilla, slow recovery of-threshold sensitivity
after the end of exposure has been associated with the state of ATS. In chinchilla we have
also seen recoveries delayed even longer than those  we have  characterized as slow. This
finding invites closer comparisons along the dimension of rate of recovery. We are not  yet
ready to specify quantitatively the different rates of recovery that we believe are opera-
tionally important. Nevertheless, we believe that it may be necessary to distinguish among
different rates which may be qualitatively  described as: a) rapid,  b) prompt, 
-------
                                  Table 1.
CITATIONS SUPPORTING DIFFERENT RATES OF RECOVERY FROM AUDITORY THRESHOLD SHIFTS
 Rate of

 Recovery
Chinchilla
 A.  rapid         Peters (1965)

    ^  3 hours*


 B.  prompt        Saunders, et al.

    ^  16 hours*   (in preparation)
 C.  slow

    > 1 day*
Carder & Miller  (1971)

Miller et al.  (1971)
 D.  delayed       Mills & Talo  (1972)

     > 1 week*


 E.  prolonged     Mills  (in press)

     > 2-3 weeks*
Man
                          Ward, Glorig & Sklar (1958)
                          Ward, Glorig & Sklar (1958)
Davis et al.  (1950)

Ward (1960)

Mills et al.  (1970)

Ward (1970)
 * times are not yet well  established but are similar to examples shown.
                                   251

-------
perspective to the data we have acquired at lower levels and reviewed here. Table II sum-
marizes for the octave bands centered at 500 Hz and 4 kHz a) the exposure levels implied by
the subtractive constants in Fig.  2A and 2B, b) the exposure levels required to produce an
ATS4 of about 50 dB, and c) the exposure levels that have been shown to produce in 3.5
hours injuries that progress to total loss of organ of Corti over some distance along the
basilar membrane.
     The subtractive  constants,  65 and 47, of Fig.  2A and 2B imply that exposures to
corresponding levels could be prolonged indefinitely and that the ATS4 would be only 0-5
dB. The functions in this figure state  that above these  levels the stress on  the ear, as
measured by ATS, grows rapidly at about 1.6 dB per decibel increase in level. For exposures
that are about 30 dB above the levels implied by the subtractive constant, ATS4 is about 50
dB and behavioral thresholds can recover completely after exposures for a few days. But
there will usually be  some loss of hair cells after these exposures. We know that when these
exposures are continued beyond 9  days to 90 days (Mills, in preparation) permanent loss
appears. We know that an increase of only 6 dB can lead to permanent threshold shift (Mills,
in press) after exposure for only 9 days. Experiments in progress suggest that exposures for
only 6 hours daily with 18 hours in quiet will only delay the acquisition of injury and loss.
The dynamic range between levels that appear to be entirely safe and those that are clearly
injurious is only 30 dB.
     When  exposure levels are increased by an additional 30 dB, severe cochlear injuries, loss
of cochlear potentials, and permanent threshold shifts follow exposures of only 3.5 hours.
dearly such exposures are excessive and their equivalents are to be avoided by man at all
costs.
     It is tempting to make a final extrapolation to  man.  The data  collected on one man
(Mills et a/., 1970) suggest that the subtractive constant for  the octave band centered at 500
Hz is about 75 and that continuous exposures at 75 dB SPL should produce asymptotic
threshold shifts  less  than 5 dB. The same data  and  some unpublished data provided by
Melnick indicate that for man ATS4 also grows by about  1.6 dB for each decibel increase in
exposure level.

                                Concluding Remarks

     We believe  the results of our studies will support the following assertions concerning
hearing and noise exposure for both man and chinchilla:
     1) Sound Levels: Above some critical level, manifestations of stress on the ear increase
by about 1.6 dB3 for each decibel increase in level of exposure.
     2) Frequency Spectrum:  We can account for most of the differential hazard related to
frequency  spectrum by the differences in the ratios of sound pressures at the tympanic
membrane to the sound pressures in the sound field.
     3) Stress on Hearing:  There are combinations of level and durations of exposure to
noise that  produce temporary threshold shifts characterized by slow recovery to normal
thresholds.
     There is a fourth assertion that we know is true for chinchilla and that we suspect may
also be true for man.

                                        252

-------
                              Table 2.
              CRITICAL EXPOSURE LEVELS FOR CHINCHILLA
Completely Safe Levels,

     ATS4 = 0-5 dB no matter

     the duration
                                           Center Frequency of

                                               Octave Band

                                        0.5 kHz          4.0 kHz
  65 dB SPL
 47  dB SPL
Borderline Levels,

     ATS4 ^ 50 dB and is

     temporary
  95 dB
 77-80 dB
Level for Severe Injury after

     3.5 hour-exposure
120-128 dB
108 dB
 * All values are approximate,  i.e. "t 3 dB
                               253

-------
    4) Cochlear Injury: The pathological anatomy and physiology observed in cochleas of
chinchillas following exposures that were characterized by slow or delayed recovery from
ITS have in all or nearly all instances shown destruction of hair cells and permanent loss of
cochlear potentials.

1 Supported in part by Grant No. NS-03856 from the National Institute of Neurological Diseases and Stroke
 to the Central Institute for the Deaf and in part by Grant No. NS-01791 from the National Institute of
 Neurological Diseases and Stroke to the Department of Otolaryngology, Washington University School of
 Medicine.
2 also Department of Otolaryngology, Washington University School of Medicine.
3 As more data have been accumulated in unpublished repetitions and extensions of measurements of ATS,
 a slope of 1.7 dB per decibel tends to describe the relations better than 1.6 dB per decibel.

                                    References

BENITEZ, L.D., ELDREDGE, D.H., and TEMPLER, J. W., Temporary threshold shifts in
     chinchilla: Electrophysiological correlates. /. acoust. Soc. Amer., 52 1115-1123 (1972)
BOHNE, B.A., Location of small cochlear lesions by phase contrast microscopy prior to thin
     sectioning. Laryngoscope LXXXII, 1-16 (1972).
BOHNE, B.A., ELDREDGE, D.H.,  and MILLS, J.H.,  Electrophysiology and electron-
     microscopy for the study of small cochlear lesions. Ann. Otol  Rhinol Laryngol. (in
     press).
BOHNE, B.A. (in preparation) Anatomical correlates of asymptotic temporary shift of the
     threshold of hearing.
CARDER, H.M. and MILLER, J.D., Temporary threshold shifts produced by noise exposure
     of long duration. Trans. Am. Acad. Ophthal Otolaryng., 75, 1346-1354 (1971).
CARDER, H.M. and MILLER, J.D., Temporary threshold shifts from prolonged exposure to
     noise./. Sp. Hear. Res., 15, 603-623 (1972).
DAVIS, H., MORGAN, C.T.,  HAWKINS,  I.E., Jr., GALAMBOS, R., and  SMITH, F.W.,
     Temporary deafness following exposure to loud tones and noise. Acta otolaryngol.,
     Supp.88, (1950).
ELDREDGE, D.H., MILLS, J.H.,  and BOHNE, B.A., Anatomical, behavioral and electro-
     physiological observations on chinchillas after long exposures to noise.  Proceedings of
     the International Symposium on Otophysiology, Ann Arbor, Mich., Univ. of Michigan
     Medical Center, May 20-22, 1971, (in press).
MELNICK, W., Personal communication.
MILLER, J.D.,  Audibility curve of the chinchilla. /. acoust.  Soc. Amer., 48, 513-523
     (1970).
MILLER, J.D., ROTHENBERG, S J., and ELDREDGE, D.H., Preliminary observations on
     the effects of exposure to noise for seven days on the hearing and inner  ear of the
     chinchilla./, acoust. Soc. Amer., 50, 1199-1203 (1971).
MILLS, J.H., Temporary and permanent threshold shifts produced by nine-day exposures to
     noise. /. Sp. Hear. Res. (in press).
MILLS, J.H. (in preparation) Threshold shifts produced by a 90-day exposure to noise.


                                        254

-------
MILLS, J.H., GENGEL, R.W., WATSON, C.S., and MILLER, J.D., Temporary changes of
    the auditory system due to exposure  to noise for one or two days. /. acoust. Soc.
    Amer., 48, 524-530(1970).
MILLS, J.H. and TALO, S.A., Temporary threshold shifts produced by exposure to high-
    frequency noise./. Sp. Hear. Res.,  15, 624-631  (1972).
MOSKO, J.D., FLETCHER, LTC J.L., and LUZ, CPT G.A., (1970). "Growth and Recovery
    of  Temporary  Threshold Shifts Following  Extended Exposure  to ,High-Level,
    Continuous Noise," Rept. No. 911, (Exp. Psych. Div., U.S. Army Med. Res. Lab., Ft.
    Knox, Kentucky).
PETERS, E.N.,  Temporary shifts in auditory thresholds of chinchilla after exposure to
    noise./, acoust. Soc. Amer., 37, 831-833 (1965).
SAUNDERS, J.C., MILLS, J.H. and MILLER, J.D. (in preparation) Threshold shift in the
    chinchilla from daily exposures to noise for six hours.
TASAKI, I., DAVIS,  H., and LEGOUIX,  J.P.,  The space-time patterns of the  cochlear
    micro phonics (guinea pig) as recorded by differential electrodes. /. acoust. Soc. Amer.,
    24,502-519(1952).
TEAS, D.C., ELDREDGE, D.H., and DAVIS, H., Cochlear responses to acoustic transients;
    an  interpretation of whole-nerve  action potentials.  /. acoust. Soc. Amer.,  34,
    1438-1459(1962).
VON BISMARCK, G. (1967), The sound pressure transformation function from free-field to
    the  eardrum of  chinchilla.  M.S.  thesis, Massachusetts Institute of Technology,
    Cambridge, Mass.
WARD, W.D., GLORIG, A. and SKLAR, D.L., Temporary threshold shift from octave-band
    noise: Applications to damage risk criteria./, acoust.  Soc. Amer., 31, 522-528 (1959).
WARD,  W.D., Recovery from high values of temporary threshold shift. /. acoust.  Soc.
    Amer., 32,497-500(1960).
WIENER,  P.M.  and  ROSS,  D.A.,  The pressure distribution in the auditory canal  in  a
    progressive sound field. /. acoust. Soc. Amer., 18, 401-408 (1946).
WIENER,  F.M.,  PFEIFFER,  R.R., and  BACKUS, Ann S.N., On the sound  pressure
    transformation by the head  and auditory  meatus of  the act. Acta otolarying., 61,
    255-269 (1965).
                                      255

-------
       PRESBYACUSIS IN RELATION TO NOISE-INDUCED HEARING-LOSS

                                     A. Spoor
                        ENT Department, University Hospital
                                Leiden, Netherlands

Introduction

     The well-known phenomenon that hearing deteriorates with age is commonly called
presbyacusis and has in this sense a very broad meaning. On this point we must be more
precise. Presbyacusis is  the hearing loss caused by  the pure process of aging itself, but aging
can happen under more or less favorable conditions. Unfavorable conditions can arise from
genetics,  drugs, noise,  nutrition,  stress, illness, climate and  maybe even  from  unknown
circumstances. These conditions differ widely  and therefore it  can be expected that the
hearing acuity is different in different populations and even may change in time. So we have
to choose as a definition of presbyacusis the process of deterioration of hearing under
circumstances that are  normal for the group under consideration. Perhaps for this kind of
hearing loss a better name would be presbya-socia-cusis. However, it is impossible to have
many different values for the hearing  loss caused by presbyacusis and therefore we  will
present here data based on several field surveys.  The reason why we  want  to know the
hearing level at different ages is that we want to have a basis on which we can  judge the
influence of some special factor, which is for this congress the factor "noise". The influence
of noise on hearing-level is measured  as the amount of noise-induced hearing  loss (NIHL).
With regard to the interaction between presbyacusis and NIHL there are a few questions to
be considered:
1.   Is the interaction between presbyacusis and NIHL additive or  nonadditive and  is
     sensitivity to NIHL dependent on age?
2.   Is the influence of noise on hearing level comparable with that of other factors?
3.   Is it possible to arrive at standard values for presbyacusis, i.e. is it possible to  give
     audiometric zeros for different ages?
4.   If the latter is possible, what is the spread of these values?

1.   Interaction between presbyacusis and noise-induced hearing-loss.

     The pathology of the influence of noise on the organ of hearing is pretty well known.
The destruction confines itself to the hair-cells  of  the basilar membrane and in severe cases
the ganglion  cells are  affected also.  According  to Gacek and Schuknecht (1969) the
pathology of presbyacusis is very complex. The hearing-loss is of the sensorineural type and
may involve one or more of four types: a) sensory presbyacusis, i.e.  degeneration of the
organ of Corti; b) neural presbyacusis with auditory neuron degeneration;  c) metabolic
presbyacusis with atrophy of the stria vascularis and d) mechanical presbyacusis by restric-
tions in the mobility of the basilar membrane.
     One can expect that the interaction between  presbyacusis and NIHL is dependent on
the type  of presbyacusis and simple addition can only be expected for the fourth type of
presbyacusis,  but in general it is not. Some authors, however, report summing without

                                        257

-------
influencing each  other, e.g. Mollica (1969). The  reason  should be that in presbyacusis
mostly ganglion cells are affected but it is not well understandable that this is sufficient to
explain addition.  Mollica observed in groups of people with the same noise exposure that
after deduction for age the hearing loss  is the same at different ages, which can be explained
by addition. He concludes that older people have more hearing loss not because of greater
sensitivity to noise but because of old age phenomena. On  the other hand, Gallo and Glorig
(1964)  observed that for a noise-exposed group, the hearing level at 4000 Hz is constant
after 15 years of  exposure. In a non-noise-exposed group, the hearing level continued to fall
and this suggests  that there is  no simple addition. Although the answer to question one is
not clear in general, the conclusion  must be that the interaction between presbyacusis and
noise-induced hearing loss is not purely  additive.

2.  Influence of noise on hearing compared with other factors.

    There are strong suggestions that the influence on hearing from noise is accompanied
by precapillary vasoconstriction. Precapillary vasoconstruction in man in response to various
types of noise was demonstrated by Jansen et al.  (1964), while Lawrence et al. (1967)
demonstrated  vasoconstriction histologically in  animal experiments  for the spiral vessels
underlying the basilar membrane  of  the cochlea. Friedman  et al.  (1967) showed that
atherosclerosis was greater in noise-exposed animals. Rosen (1969) concluded from hearing
surveys in  the Mabaans, in Finland, in Crete and the Bahamas that accelerated loss of
hearing with age  is strongly correlated with atherosclerosis and coronary diseases and not
with vascular  hypertension and  cerebro-vascular incidents. From these findings  one may
conclude that the influence of noise  on  hearing runs parallel with a factor like diet by means
of the blood supply even to the cochlea.
    A  subquestion might be: Is there an essential difference in hearing-level  between
different populations? The findings of Rosen (1962) in the Mabaan-tribe in the  Sudan at
first suggested that  there were  people with essentially better hearing. Bergman (1966) how-
ever demonstrated in a critical  analysis  of the Rosen data that the hearing of the very young
Mabaans was the  same as that  of the very young people from  cities in other countries and
also that the 10% best hearers  in the Mabaans were equal  in HL with the 10%  best hearers
from  other populations. These findings indicate that the Mabaans preserve their hearing
better, especially for the high frequencies.

3.   Presbyacusis values.

     In order to  evaluate the  influence of noise on hearing of people we must know the
normal hearing levels as a function of age. The latter will  be called presbyacusis values. We
already saw that  there will be  many differences but we must see how far we can get. In a
working group on noise influences of the Organization for  Health Research of the Organiza-
tion for Applied  Scientific Research in the Netherlands (T.N.O.), studies have  been under-
taken to analyse the results of several hearing surveys in terms of both hearing levels and the
spread in hearing levels (Spoor, 1967; Spoor and Passchier-Vermeer, 1969). For determina-
tion of the presbyacusis values, eight hearing surveys in the literature have been compared

                                        258

-------
and analyzed: Hinchcliffe (1959), Corso (1963), Jatho and Heck (1959), Johanson (1943),
Beasley (1938), A.S.A. Report (1954), Glorig (1957) and Glorig (1962).
    Table I gives some details of these surveys: The number of people involved, selection or
not, the kind of population  and the kind  of values given for the hearing-loss (median or
mean).

                                       Table 1.
                         INVESTIGATIONS USED FOR ANALYSIS
author
Hinchcliffe, 1959
Corso, 1963
Jatho and Heck, 1959
Johansen, 19^3
Beasley, 1938
A.S.A. Report, 195^
Glorig, 1962
Qlorig (WSF), 195^
number
men
326
493
399
155
2002
—
2518CR)
1724
women
319
75^
361
155
2660
—
—
17^1
selected +
non-select. -
+
+
+
+
•f
-
-
-
kind of
population
rural
-
-
-
random
-
professional
industrial
value
given
median
median
mean
mean
mean
mean
median
median
     At first glance, there are great differences in the values for the hearing losses, but it was
proven that they can easily be compared by taking the age group around 25 years in each
survey as a reference. A further analysis starts with the assumption that hearing levels for
the age group of 25 years may be equated in order to evaluate the influence of age.
     The  data  from the survey mentioned were brought together in this way  for each
frequency as a function of age. Figure 1 gives an example: the frequency is 4000 Hz and the
values concern men. It is clear that the differences are small except for the Wisconsin State
Fair data of Glorig (1957) and it is easily possible to draw a best-fitting curve. This has been
done for several frequencies for both men and women. In the article mentioned (Spoor,
1967) it is proposed that these best fitting curves can be given by an equation: log (HL+ c)
= b log  (age) - a in which a, b and c are  frequency- and sex-dependent constants. With the
aid of these equations, the hearing levels can be calculated for any age at different frequencies
(Fig. 2). From these data, the normal audiograms for different ages can be drawn for men
and for women. Fig. 3 shows these audiograms for men. It was also discussed that these data
can be considered  as median values; they have been brought together in table 2 and table 3.
It can be repeated that  these data have been calculated with respect to levels for the 25 age
group, but the standard zero level for audiometers is also based on this group and therefore
the data given may be considered as median hearing levels for different age groups of
non-noise-exposed  people.  It may be mentioned that  the hearing  level for men at the
frequency of 4000 Hz can be given with the following rule of thumb:
HL =  inn  - 6; for women this value has to be multiplied by 2/3.
                                        259

-------
Figure 1. Data points derived from 8 investigations giving hearing level as a function of age with respect to
the hearing level in the 25- or 21 Vz-year-old age group in the same investigation. Frequency 4000 Hz. Solid
curve according to the equation (see text).
     After completion of our work two other surveys came to our attention: Riley (1961)
and Glorig and Roberts (1965). The data of Riley for selected people hardly differed from
our data. The Glorig and  Roberts data are  from the  1960-1962 United States National
Health Survey and differ more, but the data for the better ear come pretty close to our data.
In  any  case it can be concluded  that our calculated data would not have been essentially
different when these other data had been included. Naturally, our data show the well-known
age, frequency and sex dependency.

4.   Spread of presbyacusis values of non-noise exposed people.

     From the surveys already mentioned, three could not be used for calculation of the
variability: A.S.A. Report (1954),  Glorig (1957) and Glorig (1962). From the remaining five
surveys the values M-QU and Qj-M have been calculated for each frequency and age group,
where M is the median hearing level and Qu and Qj are the upper and lower quartile values,

                                         260

-------
          -t—
Figure 2.  Curves giving  the relation between hearing level  and age for different frequencies for men
according to the calculated data.
i.e. the hearing-levels respectively not exceeded in 25% and 75% of the people in the age
group. Fig.  4 gives values of M-QU and Qj-M in the age groups of 25, 35, 45, 55 and 65
years in men for different frequencies. Here also the well-known fact  is found  that the
spread increases with frequency and age.  There is little difference for the sexes  and also
there is little difference  between M-QU and Qj-M.  The values of Qu  and Qj  are given  in
table 2 and  3 together with the M values. At last we can compare the quartile values and
median values for the different age groups and then it can approximately  be concluded that
the lower quartile value of one age group coincides with the median value  of the next higher
age group of ten years while the upper quartile value coincides with the median value of the
next lower age group. This can also be formulated as: from a certain age  group of ten years
50% of the people have hearing-levels in between the median hearing-levels of the next lower
and the next higher age group. This is illustrated in fig. 5.

                                          261

-------
                • 0,75-  1    • n.-s
Fig. 3.  Curves giving the relation between hearing level and frequency at different ages for men according
to the calculated data.
Summary:

     In this review article the following points are stressed. Presbyacusis is the deterioration
of hearing caused by the process of ageing. The circumstances for this ageing process can be
more or less favourable. Unfavourable circumstances can arise from genetics, noise, drugs.
nutrition, stress, illness and climate.  In  general, the interaction between presbyacusis and
NIHL  is not  additive. There is no essential  difference in hearing-level between different
populations. The sensitivity for NIHL  is not  clearly dependent  on age. The  influence of
noise to arrive at standard values of presbyacusis: data are proposed. The variability of these
median hearing-levels is also given.

                                          262

-------
                               Table 2.
  MEDIAN HEARING LEVELS AND UPPER AND LOWER QUARTILE VALUES FOR MALES
                       IN AGE GROUP OF 10 YEARS.
age
frequency
HZ
250
500
1000
2000
3000
4000
6000
8000
20-29
«,,

-4
-4
-4
-4

-6
-6
-6
M

0
0
0
0

-------
 dB

 10-
M-0
                                                                        M-
(),'_»5
  ]() -I
 dB
                                                                                 35

                                                                                 45
                                                                                65
                                                                                age
    Fig. 4. Curves giving M-Q,, and Q-j-M as a function of frequency for male groups with age the parameter.
 References:

  1. A.S.A. Report Z-24-X-2, The relation of hearing loss to noise exposure, 1954.
  2. Beasley, W.C.:  National Health Survey, 1938; data quoted from: Sunderman, F.W.
     and Boerner, F.: Normal Values in Clinical Medicine, Philadelphia, 1949.
  3. Bergman, M.: Hearing in the Mabaans. Arch. Otolaryng. 84: 411-415(1966).
  4. Corso, J.F.:  Age and sex  differences in pure-tone  thresholds. Arch, Otolaryng., 77:
     385-405(1963).
  5. Friedman, M.; Byers, S.O. and Brown, A.E.: Plasma lipid resppnses of rats and rabbits
     to an auditory stimulus. Amer. J. Physiol. 272: 1174-1178 (1967).
  6. Gacek,  R.R.  and Schuknecht, H.F.: Pathology of presbyacusis. Int. Audiol. 8: 199-209
     (1969).
                                         264

-------
                       0,5
4
I
6
I
                                                                            M ( 35
   •4O —i
                                                                            M (75
 Fig. 5. Comparison of Qu (45) with M (35), Q1 (45) and Qu (65) with M (55) and Q1 (65) with M (75) at
 various frequencies for male groups.
 7. Gallo, R.  and Glorig, A.: Permanent  threshold  shift changes produced by noise
    exposure and aging. Amer. Industr. Hyg. Assoc. J. 25: 237-245 (1964).
 8. Glorig, A.  e.a.: Wisconsin State Fair Hearing Survey, American Academy of Ophthal-
    mology and Otolaryngology, 1957.
 9. Glorig, A.  and Nixon, J.:  Hearing-loss as a function of age. Laryngoscope 72:
    1596-1610(1962).
10. Glorig, A.  and Roberts,  J.: Hearing levels of adults by age and sex, United States
    1960-1962. National Center for Health Statistics: Series 11, number 11  (1963).
11. Hinchcliffe, R.: Threshold of hearing as a function of age. Acustica 9, 303-308 (1959).

12.  Jansen, G. e.a.: Vegetative reactions to auditory stimuli. Trans. Amer. Acad. Ophthal.
    Otolaryng.: 68: 445-455 (1964).
                                         265

-------
13.  Jatho,  K. und  Heck, K.H.:  Schwellenaudiometrische  Untersuchungen uber  die
    Progredienz  und  Characteristik , der  Alterschwerhorigkeit  in  den  verschiedenen
    Lebensabschnitten. Zeitschr. Laryng., 38: 72-88 (1959).
14.  Johansen, h.: Den Aldersbetingede Tunghorhed, Munksgaard, Kobenhavn (1943).
15.  Lawrence, M.; Gonzales, G. and Hawkins, I.E.: Some physiological factors in noise-
    induced hearing-loss. Amer. Industr. Hyg. Assoc. J., 28: 425-430 (1967).
16.  Mollica, V.: Acoustic trauma and presbyacusis. Int. Audiol., 8: 305-311 (1969).
17.  Riley, S.C. e.a.:  Ten years' experience with Industrial Audiometry. Amer.  Industr.
    Hyg. Assoc. J., 22: 151 (1961).
18.  Rosen,  S.; Bergman, M.; Plester, D.; El-Mofty, A. and Satti, M.H.: Presbyacusis study
    of  a relatively  noise-free  population in the Sundan.  Ann. Otolaryng., 77: 727-743
    (1962).
19.  Rosen, S.: Epidemiology of hearing loss. Int. Audiol. 8: 260-277 (1969).
20.  Spoor, A.: Presbyacusis values in relation to noise-induced hearing-loss. Int. Audiol. ^:
    48-57 (1967).
21.  Spoor,  A. and Passchier-Vermeer, W.:  Spread  in hearing-levels of non-noise  exposed
    people at various ages. Int. Audiol. 5:328-336 (1969).
                                        266

-------
                   NOISE EXPOSURE, ATHEROSCLEROSIS AND
                         ACCELERATED PRESBYACUSIS

                              Z. Bochenek, W. Bochenek
                      Otolaryngological Dept., Medical Academy
                                   Warsaw, Poland


    In  1968, we presented the results of studies, undertaken in the Audiological Labora-
tory of  the Central Research Institute of the Polish State Railroad Health Service, for a
group  of  engine-drivers  concerning  noise  exposure,  atherosclerosis  and  accelerated
presbyacusis.
    The investigations concerning this group were multidimensional and included, besides
the examinations of the ear and  hearing, ophthalmological, neurological, psychological,
electrocardiographic and biochemical examinations. The aim of these investigations was to
evaluate the ability of an engineer to do further work in his profession.
    It  is generally  understood that  the  work  of a railroad engineer involves chronic
exposure to noise, nervous tension,  and irregular eating and resting schedules.
    The results of the ophthalmological, neurological and psychological examinations are
not included  here, since they are not directly connected with the subject of hearing loss. We
shall give only the results of the audiological examinations and tests bearing on the existence
or absence of atherosclerosis.
    The state of hearing was evaluated by means of pure-tone thresholds.
    The evidence for atherosclerosis was classified as either "definite" or "probable." As
definite  symptoms, the  following were  included:  I. heart infarction, 2.  coronary disease,
and 3.  intermittent  claudication  with decreased  oscillometric  deviations in the lower
extremities.
    As  "probable"  symptoms, the following  items of evidence were accepted:  arterial
hypertension, hypertrophy of the left ventricle, accentuation of the second aortic sound in
patients with normal blood pressure, systolic  murmur  heard at  the base of the heart,
diminished elasticity  of the radial arteries, asymmetric or absent pulse in the dorsal pedis or
posterior tibial  arteries, asymmetric oscillations  in the lower  extremities,  non-specific
changes  in the distal part of the ventricular ECG complex and cardiac rhythm disorders in
the form of atrial fibrillations, or multiple premature extrasystoles in an individual without
clinical evidence for heart disease. When the examined person showed two or more probable
signs of atherosclerosis, these  individuals were then assumed to have "definite"  athero-
sclerosis. The results of tonal audiometry were then compared with curves proposed by
Aubry (BBL) as typical for the age group. Aubry also  differentiates  two forms of pres-
byacusis: i.e.  "pure" or "physiological" and "accelerated".
    The group of 110 engineers, aged 51-60 years, with noise exposure in their professional
Hfe of about  30 years—where the intensity of noise in the cabin reached, in certain periods,
values up to  112 dB  SPL—was divided into two subgroups, one with signs of atherosclerosis
and the  other without such signs. All individuals with otoscopic abnormalities or history of
chronic  inflammation of the ears, and persons with a history of skull trauma or who had
been treated with ototoxic drugs were excluded.

                                        267

-------
     The  results are demonstrated in Table I. These results indicate that the accelerated
form of presbyacusis is more common in the group of persons with symptoms of athero-
sclerosis than in the group without any detectable signs of this disease.
Sroup

 I
II
 Age

51-60
                                        Table I
Atherosclerosis present /+/
                absent /-/
Atherosclerosis /+/
Atherosclerosis /-/
                                  The number of
                                  observations
                                      59
                                      51
The form of presbyacusis
pure
22 /37.V
25 /47.V
accerelated
37 /€3.v/
2? /53.V
                               according to Aubry s curves
     For comparison, the examination of hearing was performed on a group of 45 engine-
 drivers aged 41-50 years. The results are shown in Table II. The comparison indicates that
 in this  age group the accelerated  form of presbyacusis was encountered in a relatively
-smaller number of individuals than in the age group 51-60 years. However, it should be
 mentioned that in  the group of engine-drivers aged 41-50 years, presented in Table II, no
 systematic examination for atherosclerosis was performed. It may be only speculated that
 definite signs of atherosclerosis appeared less frequently in this group of individuals than in
 the former group, that is, in the persons aged 51-60 years.
    Group

     I
    II
 Age

41-50
51-60
                                 Table II

           The  number of observations

                      45
                     110
                                    The form  of presbyacusis
                                        pure
                                    27 /60,-V
                                    46 /42jS/
accerelated
13
64
                                   according to Aubry'o curves
     In order to establish  the function of hearing in  persons who are not professionally
 exposed to noise, but demonstrate definite signs of atherosclerosis, studies were recently
 undertaken in the Department of Otolaryngology of Warsaw Medical Academy.
     Preliminary results include the following: The studies concerned 32 men, aged 41-60
 years, who were under  the care  of  the Institute of Cardiology of the Warsaw Medical
 Academy because of at least one previous heart infarction. None of the 32 men examined
 was professionally exposed to noise. The remaining criteria of selection were the same as in
 the previous group. Patients with diabetes, diseases of kidneys, etc. were eliminated.
     Table  III presents the results of tonal audiometry in this group. A slightly different
 system of classification was used  than in the  two former tables, since the norms given by
 Aubry for so-called "pure" presbyacusis seem to be rather elevated, at least according to the
 norms given by other authors, such as Glorig, Hinchcliffe, Jatho and Heck, Leisti, Spoor and
 van Laar.  It  appears  from this Table III that,  particularly in the age group 51-60, an
 accelerated form of presbyacusis was encountered. A comparison of the hearing levels of
                                         268

-------
                                       Table III
Group

 I
II
  Age

41-50
51-60
The number of
observations
     16
     16
Audiometric
better than "pure"
2
3
"pure"
0
2
A. **. A -~ *~ 	
threshold
worse than "pure"
       7
       3
accerelated
     4
     8
                               in comparison to Aubry  a  curves of pure and
                               accerelated form of  presbyacuais
this group (Table III 51-60 years) with the hearing levels for the same age group, but
exposed to noise (Table I) shows that the  accelerated form of presbyacusis appears more
frequently in the group exposed to noise.
                                         269

-------
               HIGH-FREQUENCY HEARING AND NOISE EXPOSURE*

                                  John L. Fletcher
                              Department of Psychology
                    Memphis State University, Memphis, Tennessee

    Little or no research or clinical concern had been shown toward the measurement of
hearing for frequencies above 800 Hz until fairly recently. There were many reasons for this
disregard for quantification and study  of high-frequency hearing. A very real and practical
reason was that early efforts to test high-frequency hearing were highly unreliable, at least
partly because of the problems arising in the coupling of the ear with the transducer used to
deliver the acoustic stimulus. With  higher-frequency signals and the shorter wave lengths
that  are  associated with  higher  frequencies,  the placement  of  typical  over-the-ear
transducers, earphones, is highly  critical and  the thresholds might reflect more upon the
exact placement of the earphone than upon the actual ability of the Subjects (5) to hear the
signal. For that and other technical reasons, then, little was known .until recently about high
frequency hearing in humans, nor were norms or standards available to allow us to evaluate
or compare high-frequency hearing among various S's. The breakthrough in this area came
when Rudmose, after pondering upon the problem, came  up with a simple but effective
solution. He utilized a small microphone with a conical probe tip as an earphone, inserting
the tip of the probe into the ear.  This provided a reliable and effective coupling of the ear
and the transducer. He then proceeded to examine  the hearing of several young healthy
non-noise exposed  male and female high school students, pooling their data to provide an
interim biological baseline that could  be used with caution to evaluate hearing data from
future persons tested. Fletcher (1965), using an early model of the Rudmose high frequency
audiometer, determined that reliability  of the technique compared favorably with that of
conventional audiometry (Fig. 1). Zislis and Fletcher (1966) then tested male and female
non-noise-exposed  6th, 9th, and  12th grade students to establish whether high-frequency
hearing varied within such age limits.  Essentially, they found that it did not, that females
were better than males, and that probably Rudmose's original data were from too select a
population and people  do not hear quite as well as he had supposed. Northern et al. (1972)
in a later standardization study coordinated with earlier studies, emerged with proposed
high-frequency hearing standards. It is now apparent that a technique is available for reliably
testing hearing for frequencies above 8000 Hz. With a technique available for valid and
reliable testing of high-frequency  hearing and with tentative norms or standards proposed,
efforts have begun to determine practical implications of high-frequency hearing.
    In an early study, Fletcher et al.  (1967) found that meningitis  patients who had been
categorized as seriously ill during  the course of the disease had significant losses of hearing
above 8000 Hz compared to those who  had not been seriously ill, while neither group had
noticeable losses of hearing for the conventional frequencies. These results suggested that
high-frequency hearing might possibly be  a sensitive index of possible trauma within the
cochlea. This hypothesis was supported by research conducted by Jacobson et al. (1969)
into the effects of ototoxic drugs on  high frequency hearing. In  a study begun before
chemotherapy on tuberculosis patients,  they  found that ototoxic drug effects upon high-

                                        271

-------
UJ
2  40


«  30
ca

55  20
UJ

~  10


s   o
   -20
	 Ist SESSION
	7n« SESSION

	3" SESSION
                                                            I
                                 9      10     11      12     13
                                 FREQUENCY  IN  KILOCYCLES
                                      14
15
16     II
                 Fig. 1.  Repeatability of high frequency thresholds over three sessions.


 frequency hearing were detected from  41-76 days earlier than effects could be detected
 within the conventional frequency range (Fig. 2). Hearing losses were most apparent first in
 the frequency range from 9-13  kHz. Kanamycin was the drug found most likely to have
 caused the hearing loss, although the patients were also receiving other drugs.
      The indications of the usefulness of high-frequency hearing for the early detection of
 ototoxic drug reactions, and of the sensitivity of high-frequency hearing as an indicator of
 damage from disease,  suggested that high-frequency hearing might also provide an early
 warning system of noise-induced hearing loss, or might differentiate  hearing levels at an
 early stage between populations exposed to various  non-occupational noises such as rock
 music, sport shooting, drag racing, etc.
      Conventional and high-frequency hearing of rock band members, rock spectators, sport
 shooters, drag racers, and motorcyclists was tested for a population of  18-21-year-old males
 and females. A normal or control population of 18-21-year-old males and females was also
 tested to  provide  a  basis for comparison (Fletcher,  1972). In this study (Figs.  3  and 4)
 motorcyclists were found to have incurred the greatest loss of hearing, followed by drag
 racers. Sport shooters had much less loss than expected, probably because most shooters are
                                          272

-------
 0
_C
"c
 *
 O
 v
 O
M
•o
 c
 c
 9
          O •
          10-
         70.
         30-
         4O-
         50-
         60-
         70-
 0 -
 10-
20-
30-
40-
50-
60-
7O-
                                                           C.E.
                                                   8-16-67
                    I      I      i      I      I       I
 0-
 10-
 20-
 30-
 40-
 50-
 6O-
 70-
                                                                     J.G.
                  1-2-68
                                 12-19  67
                                                               — 10-3-67
                           I      I
                                     I      I
                                                                     F.J.
                                                                     8.8.67
       - 1 - 1 - \
        46      8      9
                                               i - 1 - 1 - 1 - 1 - 1 - T
                                           10     11      12     13     M     15     16
                                        Frequency  (kHx)
 Figure 2. Changes over time in high frequency hearing of ototoxic drug patients (subjects C.E.. J.G., and
 F.J.).
                                            273

-------
                                      DRAG RACERS
   ic:

   ec

   t r


   7C

  . 6'
c
z
AGE
 ie •
 19
 20»
 21
                                ROCK BAND MEMBERS
   5C
          AGE
            13 ••
            19 -
            2C-
            21 •
                   2    3   4    6    8   9    10   II   12   13   14   15   16   18
                                     FREQUENCY IN KHZ

                       Figure 3. Percent responses by exposure category.
Acutely  aware of the hazard to hearing and utilize some form of ear protection while
engaged in shooting. Hearing losses among musicians in rock bands were surprisingly small
but not totally unexpected according to the author. He attributed the small losses exhibited
by the musicians to a sampling flaw, saying, "Many times, in trying to schedule known rock
band members or drag racers, we were told they were out of town playing an engagement,
or driving at some track, and repeated calls received the same answer. It would appear that
those at the upper level of experience and skill, if they desire, can spend a great deal of time
at this activity and  make reasonably good money. Therefore,  it could well be that our
sample is missing many,  if not most of those at this level.. .those  who would be expected to
have suffered the greatest exposure and therefore the largest losses".
     Another study presently underway to determine whether high-frequency hearing is a
useful early detector of noise-induced hearing loss involves aviators  (Fletcher, 1973).  The
                                        274

-------
                                      MOTORCYCLISTS
   ICC


   9C


   sc


   7C

o
Z.  6C
O
Z
o  50
V
a:
>-
Z
Ul
           AGE
            !8 	
            !<>	
            20 —-
            21 	
c
u
ICC




e:

7C

6C

50
                                     SPORTS SHOOTERS
           AGE
            18  	
            19  '—
            20
            21  —
                         3     4     6    8    9    10   II    I?    13   U   IS   16   18
                                         FREQUENCY  IN  KHZ


                       Figure 4.  Percent responses by exposure category.
 conventional and high-frequency hearing of a large number of professional pilots was tested.
 Pilots were found who, predominantly, flew jet aircraft, propeller-driven airplanes, or rotary
 wing (helicopter) craft. Also, a sufficiently large sample of each type of pilot was obtained
 to study hearing from entry into aviation training through up to about 9,000 hours flight
 time. Results of the  study to  date (Figs. 5-9,  inclusive) show earliest losses at  the higher
 frequencies, as expected, with  a gradual erosion of hearing with continued exposure, with
 lower and lower frequencies progressively becoming involved with continued exposure and a
 larger percent of the  S's  not hearing the higher  frequencies. In fact, percent of S's hearing a
 frequency appeared a more sensitive index of exposure than average hearing level.
     Propeller-driven  planes  appeared  to  cause the greatest loss of  hearing, followed by
 helicopters,  with jets the least  hazardous (Fig.  9). The results for  percent response, by
 comparison, showed a drop beginning at 4 kHz for the prop  pilots (Fig. 5), whereas for jet
                                          275

-------
   too

   »5

   90

   Bi

   80


   7S
o
5;  /o
w

-------
u
w
_l
u
VI
a
M
O
2  40
W
K
Z
                                               10   11    12    13   U   IS    16   18
                                   FREQUENCY IN KHZ
                      Figure 6.  Mean hearing levels by flight time—prop.
both sides  of the shooter randomly.  The  non-noise-exposed male population  used for
comparison had hearing levels an average of about 26 dB better at the frequencies 10, 12,
14, 16, and 18 kHz than  those for the boys' rifle team members. For the females, the girls'
rifle team member average hearing levels were some 17 dB higher (less sensitive) than those
of the levels found for the non-exposed females.
    These  data  do demonstrate rather  clearly the usefulness of tests of high-frequency
hearing for the early detection of at least some types of noise-induced hearing loss.
    In summary,  data have been  presented suggesting the usefulness of high-frequency
hearing testing in early detection of not only noise-induced hearing losses, but also ototoxic
responses, and losses attendant upon certain types of illness. Tests of high-frequency hearing
have been shown to be reliable, equipment for  such testing is now commercially available,
techniques for testing are no more complex than those of conventional tests, nor is the
required testing environment as quiet as that necessary for conventional testing. Therefore,
it is recommended that serious consideration be given to use of high-frequency hearing tests
in the early detection of noise-induced hearing loss.
                                         277

-------
    100
    90
    85
    80
o
z

S
z
o

Si   70
UJ
tE
    65
vj
«E
60


55


50


45
                AVG. MRS.
    35
                                                       10     II    12    13    U    15    16    IB
                      Figure 7.  Percent responses by frequency and hours—Jet
                                              278

-------
  ul
  z
  VI
  <
  a
  a)
  •a
  1/1
  a
  VI
  hi
                                 6    6     9    10    II   12   U
                                   FREQUENCY IN KHZ
                                                                  U
16
                      Figure 8. Mean hearing levels by flight time—jet.
                                    References

1.   Fletcher, John  L. Reliability  of high-frequency  thresholds. /. Aud. Res., 1965, 5,
    133-137.
2.   Zislis, T. and Fletcher, John L. Relation of high frequency thresholds to age and sex. /
    Aud. Res., 1966,6, 189-198.
3.   Northern, Jerry L., Downs, Marion P., Rudmose, Wayne,  Glorig, Aram, and Fletcher,
    John  L. Recommended high-frequency audiometric threshold levels (8,000-18,000
    Hz)./. Acoust. Soc. Amer., 1972,52, 585-595.
4.   Fletcher, John  L., Cairns,  Adrian B.,  Collins,  Fred G., and Endicott, James. High
    frequency hearing following meningitis. /. Aud, Res., 1967, 7, 223-227.
5.   Jacobson, Edward J., Downs, Marion P., and Fletcher, John L. Clinical findings in high
    frequency  thresholds during known ototoxic drug usage. US Army Med.  Res. Lab.,
    Rep. 820, Mar. 25, 1969.
                                        279

-------
Ul
<
at
    -5


     0




    10




    JO
 m  30
 o"
 X
 in
 ui
 IE
    40
    50
    60
                     HOUttS  N
               JET VI   U?2  \7
               HELO VI  Jlftt  22
               PROP VI  4!7I  29
          .5    1    2    3    t    6    8    9    10   II    12   13   14   15   16   18
                                   FREQUENCY  IN KHZ


          Figure 9. Selected comparisons of mean hearing levels by aircraft type & flight time.
6.   Fletcher, John L. Effects of non-occupational noise exposure on a young adult popula-
     tion. Final Rep., Contract HSM-099-71-52, NIOSH, Dept. HEW, October 13, 1972.
7.   Fletcher, John L. Conventional and high frequency hearing of naval aviators. Office of
     Naval Res., Contract N00014-71-C-0354, Progress Rep., February, 1973.
8.   Gasaway, Don. Personal communication, 1972.
9.   Simonton, Jane K. Results of high frequency tests of members of Denver public
     schools high school rifle teams during Dec., 1972. Personal communication,  1972.
                                        280

-------
                        SUSCEPTIBILITY TO TTS AND PTS*

                                   W. Dixon Ward

        Department of Otolaryngology; Department of Communication Disorders
                               University of Minnesota
                               Minneapolis, Minnesota

    More  than 140  years  ago,  John  Fosbroke  noted  that  people differed in their
susceptibility to hearing loss,  as they did in so many other ways, and speculated about the
underlying constitutional factors responsible for these differences. About a century later, it
occurred to Jakob Temkin that it might not be necessary to determine individual differences
in how well sound was conducted to the inner ear, in the elasticity of cochlear structures, in
blood circulation, etc., or to institute monitoring audiometry in noisy industries (which he
did favor, however), in order  to forestall hearing loss. Instead, one could merely expose the
ears of all the workers to a moderate  sound, and measure the auditory fatigue produced.
The ear showing the greatest effect  surely should be the most susceptible  to permanent
hearing loss. This  idea was developed by Alfred Peyser, who proposed the first formal
susceptibility test: the change in the  amount of time a 250-Hz tuning fork could be heard
immediately after half an  hour of exposure to the noise from a  "Klappermaschine aus
Metall"  (1930). In  the intervening period, literally millions of man-hours have been
expended in a test  of Temkin's hypothesis, as exemplified in at least 20 different proposed
susceptibility tests  (reviewed  by Ward,  1965). I am indeed sorry that Dr. Temkin, who is
still active in the field of noise in Moscow, is unable to attend the Congress and participate
in a discussion of the present status of his idea.
    As we  all know, the relation between auditory fatigue, or temporary threshold shift
(TTS), and permanent threshold shift  (PTS) did  not turn out to be  as simple  as Temkin
hoped. Already 24 years ago, when Walter Rosenblith got Ira Hirsh and myself interested in
TTS, Theilgaard (1949) and Greisen (1951) had shown that since there was little correlation
between TTSs produced by pure tones of different  frequencies, susceptibility could hardly
be a unitary function. The same independence of susceptibilities was implied by Theilgaard's
(1951) finding that there was no  consistent correlation between the TTS produced by a
1500-Hz pure tone and the magnitude of the hearing loss at 4 kHz in a group  of 59 weavers.
And of course Flugel had shown in 1920 that the two ears of a given individual differed in
fatigability.
    However,  there  was  and is  no  question  that there  were large  differences in
susceptibility to PTS,  as Borge Larsen (1952) pointed out while discussing the implications
of Theilgaard and Greisen's work. He  cited two persons tested by him and  found to have
normal hearing, despite employment histories of 15 years as a boilermaker or  14 as a riveter,
respectively.  The same point was made a few years later by Shapiro (1956), who found a
normal-hearing drop-forge operator with 28 years of experience. Clearly, such workers are
unusually resistant.
 •Preparation of this manuscript was supported by Grant NS-04403 from the Public Health Service, U.S.
 Dept. of Health, Education and Welfare.

                                         281

-------
     The reverse, however, does not follow—i.e., that an individual who has a hearing loss is
 necessarily more susceptible than the average man, although this assumption is sometimes
 made (e.g. Harris, 1965). He might be more susceptible, but he might also have been merely
 more unlucky in  being over-exposed  on a  particular unusual  day-that  is, in having
 experienced  a  noise dose that would give anyone a large PTS. In any actual situation, the
 group of workers with elevated Hearing Levels will include both the susceptible and the
 unlucky.
     This fact causes all sorts of trouble when one tries to validate a susceptibility test based
 on ITS  with a cross-sectional study of a group of  men who have been working in noise for
 many years and who have a wide range of hearing losses. One can, of course, give these men
 the same susceptibility test, measure the TTSs produced, and then determine the correlation
 between these  TTSs and the existing Hearing Level (HL). However, a source of bias exists if
 this procedure is followed. Men with hearing loss at the frequency whose shift is taken as a
 susceptibility index will, on the average, show less  ITS than normals, and so there will be a
 significant negative correlation even  if  the susceptibility test per se is worthless. In extreme
 cases, this is intuitively obvious. When  the loss is due to conductive factors, then of course
 the effective level of the sound reaching the cochlea will be reduced and so less effect would
 be expected. On the other hand, when the loss represents a sensorineural deficit, then there is
 less shift possible.
     Figure  1 shows the curve relating average TTS at 4  kHz to the resting HL of some
 seasoned workers in a planet with uniform levels throughout the working area of about 100
 dB(A).  It can be seen that the average TTS decreases with HL in a linear fashion, with an
 intercept at about 80 dB HL.
     The individual results of three workers are also shown. Workers A and B have resting
 HLs of  10 and 50 dB, respectively, but  both show TTSs of 20 dB." Worker C, with a
 threshold of 40 dB HL shows a TTS of 10 dB.
     It is clear that despite the equivalence of TTSs, worker B should be considered to  show
 a greater "effect" than A; he displays considerably more TTS than the average man with a
 50-db loss, while A shows less than the average of his group. What is not so clear is whether
 C  is more susceptible than A, or vice versa. Both show a TTS that is  5 dB less than the
 average for their respective  HL groups. However,  one might say,  "Yes, but C shows only
 67% as much TTS as the average (10  vs.  15), while A shows 80% as  much (20 vs. 25).
 Therefore A is the  more susceptible." This is supported by considering also the variability
 involved. The variance of TTSs for HLs of 40 dB will be smaller than for HLs of 10 dB, so
 the 5-dB departure  from the mean does indeed imply that worker C is "farther"-i.e.,  more
 standard deviations-below the average line than is A.
     This is the procedure that was used by Bums et al. (1970) in their recent study of 218
workers, one of the few recent results that offers  much encouragement for susceptibility
tests. Not only were all individual TTSs-in their case, the  TTSs produced by the workers'
own normal working day—converted to standard scores based on the use of average TTSs and
their respective variances, but HLs themselves were also normalized to scores reflecting how
the individual's HL compared to  the HLs of men  of the  same  age and cumulative noise
exposure; that  is, corrected  for both average noise-induced PTS and for presbyacusis plus
sociacusis. With this procedure, they were at  least  able  to'show a correlation of 0.34

                                        282

-------
     LJ —
     1-0
     Lu Z
        LJ
        o:
2
    HO:
           '°
                                                                 I
                   0         20         40
                    HEARING  LEVEL  IN
                                             60          80
                                             DB  (ISO)
Figure 1. Dependence of ITS on the resting threshold. The noise had a sound level of about 100 dBA
Crosses indicate hypothetical results from three workers (see text).
between the normalized TTS averaged for 1 and 2 kHz and the normalized HL averaged over
3,4 and 6 kHz.
    Correlation coefficients of about one-third or smaller seem to be the rule, being about
the same as that found by Jerger and Carhart  (1955), in  their study  of the changes in
hearing suffered by  178  airmen at a  school for jet mechanics during a 10-week training
course. The time to recover to within 20  or 10 dB of original threshold at 4500 Hz,
respectively, after a 1-min exposure  to a 3000-Hz tone at 100 dB SPL, was correlated
against the  average shift in HI at the end of the course (presumably PTS). The correlation
coefficient was 0.36 for the recovery-to-20-dB-TTS criterion, though only 0.23 for recovery
to 10 dB.
    These results are about what one would expect on the  basis of the extensive study of
the intercorrelation among different types of susceptibility indices done in our laboratory
several years ago (Ward, 1965, 1968). In these experiments, 49 college students with normal
hearing, 24 men  and 25 women, were exposed  to a host of different short susceptibility
tests: TTSs from exposure to pure tones and octave-band noises for various times and at
various levels, measurement of the intensity of noises and of simulated gunfire  required to
produce a given TTS, perstimulatory adaptation, and contralateral remote masking were
measured, using earphones or free-field exposure, in one ear or two.
                                       283

-------
     The test-retest correlation was about 0.65, even when there was a 6-month intervening
period; this increased to 0.77 if the average of TTSs at the three frequencies most affected
were used instead of only the frequency at which the maximum TTS occurred. Correlations
among  TTSs  from different tests involving exposure in the same frequency range had a
median value of about 0.55. The  correlation between TTSs produced by different ranges of
frequency was smaller,  being statistically insignificant (less than  0.3) even for two 3-min
exposures to 1000-Hz octave-band noise at 120 dB SPLand to a 2000-Hz octave band noise at
116 dB, for example. Between the former and the TTS from a 15-min exposure to 500-Hz noise
at  120 dB SPL, however, the correlation was 0.5. Factor analysis  of the correlation matrix
implied that there is a common thread of "general susceptibility" to auditory stimulation,
but that this  would account  for only about a third of the communality  in  the matrix.
Varimax analysis indicated that one should really speak of susceptibilities to TTS-to low-,
medium- or high-frequency noises or tones,  or to impulse noise.  However, exposure to
broad-band  noise did tend to produce TTSs that agreed well with  those produced by the
appropriate single octave band presented alone, so it appeared that the best course of action
would  be  to  use broad-band noise as the fatiguer, but  measuring  TTS  at  the  various
relatively-independent frequency ranges.
     It appears, then, that a quick test that will reliably measure all aspects of susceptibility
even to TTS is not  at hand, much less one that produces an effect-either the TTS or the
recovery time—that is a valid predictor of eventual PTS. However, it is clear that some of the
validation studies so far attempted, such as those of Jerger and Carhart (1955) and Burns et
al. (1970),  suffer from the fact that  the true noise exposure  was  only estimated or was
assumed to  be the same for all workers. Other attempts at validation have another problem
as well—namely, that PTSs may be produced in so few ears that the correlation between the
results of a  susceptibility test given at the beginning of employment and the change in HL is
meaningless. For example,  Sataloff et  al. (1965) in 1951 gave a 2-kHz 95-dB-HL (ASA) test
to 105 Ss,  and then in J 962  examined the hearing of the 33 still at this jet-engine test
facility. Unfortunately,  no important PTSs had been produced in this time, so the absence
of a significant correlation proves nothing about the test per se.
     There  is  clearly only  one socially-acceptable solution to the lack of control over the
noise exposure of the test subjects in  all tests of PTS involving humans, and that is to use
experimental animals whose auditory history is completely known. Toward this end, we
have for the past few years been conducting studies using chinchillas. They are trained, using
a conditioned  shock-avoidance technique, to jump across a barrier when they hear a sound.
Temporary  and permanent threshold shifts thereby can be measured. However, We find that
if our animals all come from  breeders who maintain a quiet environment,  the differences
among them in resting thresholds are so small that they barely reach statistical significance
even after weeks of  testing. Furthermore, when the exposures are done, as ours have been,
by restraining them  in a head-holding device in a fixed position in front of a loudspeaker,
which gives  good control of the exact noise dose they receive, the amount of TTSs and PTSs
produced are  also nearly  the same.  While occasionally there is an animal who shows a
significantly greater TTS  than the rest; a repetition of the exposure usually will  fail to
substantiate his indicated higher susceptibility. Similarly, the occasional animal who shows
less PTS than the average has always showed a normal (in this case, average) TTS.

                                        284

-------
    The results just cited are of course subject to the limitation that we are at the moment
still not sure that we did not have an artifact of some sort in our testing situation. Animals
who had been given TTSs of 70 to 80 dB apparently recovered completely in about two
weeks, yet histological examination showed extensive, even complete destruction of the hair
cells in the basal turn of the cochlea (Ward and Duvall,  1971). Furthermore, the degree of
destruction could  be  quite different for two animals that showed the same TTS. These
results are difficult to explain. We are only now beginning a set of experiments to confirm
or deny the accuracy of these observations, our laboratory having been inoperative for the
last 6 months because it was necessary to move it.
    In view of the fact that the exposure that produced equal TTS and PTS in all animals
did not produce the same histological damage, we rather expect to find that individual
differences in TTS and PTS are not as small as thus far indicated, so that we can proceed to
study individual susceptibility. Even if we do not, however,-if, for example, the histological
differences are artifactual instead-this  will not  be particularly disheartening. If these
animals, kept free of sociacusic influences such as avocational noise, blows to the head, and
middle-ear infection, really all do have the same susceptibility, at least it  will not require a
very large number of animals in order to establish the group relations between TTS and PTS,
especially in, regard to intermittent noise, which is our other chief area  of activity at the
moment.
    Other laboratories have not had much success with different animals either. Herman
and Clack (1963) got a zero correlation between TTS and PTS from white noise in the rat,
and Luz et al.  (1971)  had a similar outcome exposing monkeys to high-intensity impulse
stimuli. However, it must be noted that in the Luz et al. study, test-retest correlation for
TTS was significantly negative, so one  can hardly expect any correlation with anything else
to be significant.
    In man, only a few recent papers  claim to have established a positive relation between
differences in TTS and PTS. Sulkowski  (1969) claims  that in a*sj;udy of 127 beginning
workers in a textile mill, a combination of a Peyser-type test (4 min  of 4000 Hz at 90 dB
HL) and a tone-decay test predicted the degree of hearing losses developing in the next two
years,  Strubinski (1970)  also  reports success with  such a  combination in forecasting
hearing-loss development  in 33 diabetics. Pfander (1968) exposed  100  recruits to three
susceptibility tests, two involving white noise,  the third being exposure to five shots from an
ordinary military weapon (161 dB peak). He  indicates that five soldiers who showed TTSs
from the rifle shots that required 3 to  6 days for full recovery had permanent losses at the
end of their training. However, it is not made clear in the article how many of the men
originally had high values of HL to begin with, how many dB constituted a "significant"
loss, or even whether or not these 5 were the only cases showing permanent shift, so some
uncertainty still exists.
    No one  seems to have followed up a study of telephone operators by Kuroyanagj
(1960). Despite the fact that there are consistent differences in susceptibility between the
two ears, the median correlation between ears for TTS is 0.63 (.Ward, i 968), which indicates
that the two ears of a given observer are generally quite similar. If, then,  one ear is always
used for telephone listening, and that ear is exposed to intense noises (.or clicks and buzzes,
in the case at hand), then the more susceptible will end up with a hearing loss, but in only

                                        285

-------
one ear. Kuroyanogi found that in 914 telephone operators, 18% had "some" loss in the ear
used. It is indicated that those with such loss showed more TTS than the average on the
normal side.  Unfortunately  for science,  though not for operators, hearing losses are no
longer being produced by telephone noises (Glorig et al., 1969), so this source of possible
validation of susceptibility tests may no longer exist.
    If the relation between susceptibility to TTS and the susceptibility to PTS seems
uncertain, it is no more so that the question of what it is that determines either one. Many
articles have been  written  about the relation of mastoid pneumatization  and PTS in
highly-exposed  workers, for example. Link  and Handl (1955) and Ceypek et al. (1956)
found a statistically significant though not large indication that well-pneumatized ears were
less susceptible  than poorly-pneumatized ones. Kosa and Lampe (1967), however, found no
correlation whatever, and Kubo, who also concluded that pneumatization was a factor of no
importance,  actually got data that show conclusively that men with the most  pneumatiza-
tion had  more loss than those with the least. I tend to believe that Kosa and Lampe have hit
upon the truth.
    Another equivocal area, although this always seems odd, is in the area of middle-ear
problems. The  natural inclination is to believe  that anything that interferes with the
conduction of sound must act as a protection and thus reduce susceptibility to both  TTS
and PTS. Although this does seem to be true of ears with ordinary otosclerosis (Gerth,
1966), even after stapedectomy (Fletcher and King, 1963), other middle-ear problems such
as otitis  media do not seem to reduce TTS or PTS to a degree commensurate with the
conductive deficit that can be measured. Results reported by Paparella et al. (1,970) on 279
ears with otitis media actually led them to the conclusion that otitis media can cause
sensorineural loss by somehow invading the cochlea; however, it may merely be that this
condition has its effect by changing the susceptibility to noise damage rather than by direct
action. The  question of the effect of middle-ear pathology on susceptibility  is still open.
However, I   am  glad tto  report that no  one has yet  challenged Kristensen's (1946)
demonstration that there is no relation between hearing loss and body type.
    The role of the middle-ear muscles  has also received considerable study; inoperative
middle-ear muscles would be expected to produce an ear that was unusually susceptible to
low-frequency stimulation, particularly to intermittent sounds. Such ears are occasionally
found; in; one  that displayed the symptoms just cited, impedance measurements indeed
implied that the reflex was inoperative (Ward, 1962). In a study of 40 Marines who had just
completed training, Coles and Knight (1965) found that men with "poor high-tone hearing"
had a higher reflex threshold than those with "good" or "fair". So in this field the evidence
seems to  point at least in the  same—and, happily, the expected—direction.
    Theoretically, individual differences  in susceptibility to acoustic stimulation having a
particular frequency can be ascribed to a near-infinite number of parameters, not only those
dealing with transmission of sound to the cochlea but also with inferred characteristics of
the cochlea itself such as blood circulation, thickness of the various membranes, etc. A most
unusual correlation has recently been reported by Tota and Bocci (1967)j for example, who
report that bhie-eyed persons showed twice as much TTS as brown-eyed ones (27 dB vs. 13
dB) following a 3-min exposure to a  1000-Hz tone at  100 dB  (HL, presumably), and so
conclude that melanin plays an important role in protecting the ear from hypoxia.

                                       286

-------
     There seems to be a consistent relation between individual differences in the degree of
vasoeonstriction  caused by  a noise and  the resultant TTS, persons with the  smaller
vasoeonstriction  showing  the greater  TTS,  provided  that  one  uses as  the  index of
vasoeonstriction the value observed 20 sec after onset of the noise, not at the end of a long
exposure (Jansen, 1970). This may account for the fact that Oppliger et al. (196) found no
consistent relation. It must  be pointed  out,  however, that this correlation  does not
necessarily mean that vasoeonstriction is causing a greater TTS; if the vasoconstrictive effect
depends on the loudness of the sound producing it,  then in those persons for whom more
energy is reaching  the inner ear,  the sound  will appear louder, hence produce  more
vasoeonstriction, and will also produce more TTS, whether or not the vasoeonstriction has
anything to do with, say, the accumulation of fatigue products.
     Both eye color and vasoeonstriction should be investigated further. There continues to
be no convincing evidence that ears ever get "tough"-more resistant to damage-because of
habitual exposure, or that younger or older persons are more susceptible than young adults,
or even that any sort  of medication can decrease susceptibility to TTS and  PTS, although
there is no doubt that it can be increased by the administration of certain ototoxic drugs,
even though the doses of drug are by themselves subtoxic. An unreported study by John
Park in our laboratory found no differences in TTSs from noise in the chinchilla to be
caused by injection of hydergin or adenosine triphosphate (cf. Plester,  1953; Faltynekand
Vesely, 1964). Dextran, which is reported by Kellerhals (1972) to reduce TTS, also was
without  effect;  however, in  this  case  the  dosage  turned  out  to be, less than  that
recommended by Kellernals, so additional experiments are planned.
     The concept of a "critical intensity" for a given ear seems to have  died a natural death
as it became clear than an intensity  level that was "critical" for one duration of noise was
not "critical" for a different duration. The notion of a "critical energy"—a measure of noise
exposure, not just  of noise—may have some merit, but if it merely denotes the  "breaking
point" above which permanent damage will result, it is not usually worth looking for.
     On  the other  hand, there are characteristics of noise as such whose perception varies
from individual to individual.  Perhaps determination  of the LDL (loudness discomfort level)
is worth  intensive study hi regard to susceptibility (Hood, 1968)—that  is, provided we can
agree on  the exact instructions to the listener,  in this rather instruction-sensitive task.
Perhaps someone will be stimulated to  extend some results reported by van Dishoeck and
Spoor  (1958)  15 years ago, who claim to have gotten a "good criterion of the  individual
sensitivity of noise" by asking subjects  to "indicate  at which intensity a pure tone acquires
an impure and sharp character".
     Nowadays,  the possibility must be kept in mind  that perhaps PTS may not be the
correct validating  index for susceptibility in  the  first place.  It  may be that auditory
sensitivity is not much affected  until all of the hair cells in  a certain area of the basilar
membrane are destroyed (Eldredge  and Miller,  1969; Ward and  Duvall, 1971). If, then, a
given exposure destroys only  a few hair cells, then the full recovery of the TTS that ensues
will incorrectly imply that the exposure was innocuous. It may be that in man a much lower
TTS should be permitted than we now  deem to be safe; however, in the chinchilla a 40-dB
shift—at least if caused by a short exposure—is known to be safe. Figure 2 shows the course
of recovery from TTS in two groups of chinchillas. One group of 5 animals was exposed to

                                        287

-------
 114 dB SPL of 700-2800-Hz noise for 10 min, the other for 20 min. As can be seen, full
 audiometric recovery occurred after one day in the 10-min group, but required 4 days in the
 other. Inspection of the cochleas showed no missing hair cells in the 10-min group, but a few
 (3 to 7) hi each ear of the 20-rnin ones. We have not yet run the experiments to answer the
 question of whether or not  a limit of 40  dB of TTS is also safe if it was produced by a
 longer or an intermittent exposure. In view of the longer persistence of such TTSs (Ward,
 1970), it is certainly not to be taken for granted that this will be the case.
     On the other hand, perhaps hair-cell destruction is itself no better a validating criterion.
 Both Elliott (1961) and Hunter-Duvar (1971), among others, have produced  tonal gaps in
 their experimental animals (cats and monkeys, respectively), yet upon examination the hair
 cells all appeared normal.
     We are truly in a difficult position at the moment. We cannot rule out the possibility
 that what are  clearly innocuous exposures to noise, in the sense that recovery from any
 measurable effects is complete within a  few hours and that no cumulative effects can be
.seen over  a period of many weeks, may nevertheless be producing latent damage—changes
 that are unobservable in the intact organism.  It is not enough, in this case, to argue that a
 difference that makes no difference is not a difference. If noise exposures that cause no
 change in  auditory function are nevertheless gnawing away at nature's safety factor, then
 the  fact that no functional deficit is observed is not completely relevant. There are many
 who support such a cautious viewpoint.                                       >
     However, if this were the case, then one would expect that years of work  in a noise
 that is slowly destroying hair cells one at a time—the essence of the "microtrauma" theory
 of Gravendeel  and Plomp(I960)—would suddenly produce severe impairment of sensitivity,
 as the last few hair cells in a given area finally succumbed. I know of no evidence for such
 sudden growth of impairment in men working for a long time in uniform noise; on the other
 hand, I  doubt that anyone has looked very  carefully for it. Nevertheless, the burden of
 proof,  in  my opinion, still rests  on those  who  assume such  latent  (or  residual but
 immeasurable) effects  to  be occurring. Otherwise we may be forced to adopt absurd
 damage-risk  criteria, as for example that proposed, apparently in  all seriousness,  by
 GePtischcheva and Ponomarenko (1968) in Russia: a measurable TTS at any frequency!
 Because a  1-hr exposure to a 500-Hz octave band of noise at 75  dB (SPL, presumably), to a
 1-kHz  octave  band  at 65  dB, or  to  a  4-kHz  octave band  at 60  dB  produced in
 15-to-16-year-olds  a  3-dB TTS measured  within the first minute after exposure, they
 propose that adolescents be protected from anything higher than these levels.
     At the  moment, therefore, I shall continue to conduct my research as if it were true
 that if recovery  from TTS induced by a daily exposure is complete before the next day's
 exposure begins, no hazard to hearing exists. I hope this is correct.
                                        288

-------
    70
    60
    50
 g  40
 ^  30

 £20
 H
    10


    0
       EXPOSURE:
700-2800-HZ NOISE
                                                    0 M1N
                             1
           I
I
I    I   I   I   I
               IH           4H     8H         ID      2D
                     TIME   SINCE   EXPOSURE
                                     4D    7D
Figure 2. Recovery of chinchillas exposed only once to noise: either a 10- or a 20-minute exposure to
700-2800-Hz noise. The 20-minute group showed a few missing hair cells; the 10-minute group did not
                                References

Burns,  W.,  Stead, J.  C.  and  Penney,  H.  W., The  relations between temporary
    threshold shift and occupational hearing loss. In: Hearing and Noise in Industry,
    W. Burns  and D. W. Robinson,  Eds., Her Majesty's  Stationery Office,  London,
    183-210 (1970).
Ceypek, T., Lepkowski, A. and Szymczyk, K., Sensitivity  to  acoustic traumas and
    pneumatisation of  the mastoid, Otolaryng. Pol 10, 329-334 (1956) ref. ZHNO
    57, 219 (1957).
                                    289

-------
Coles,  R.  R.  A.  and  Knight,  J.  J., The  problem of  noise  in  the Royal  Navy  and
     Royal Marines.,/. Laryng. Otol,  79,  131-147 (1965).
Dishoeck, H. A. E.  van  and  Spoor, A., Auditory fatigue and occupational deafness.,
     Laryngoscope, 68, 645-659 (1958).
Eldredge, D. H. and Miller, J. D.,  Acceptable noise  exposures—damage risk  criteria.
     In:  Noise  as  a Public Health Hazard; ASHA Reports  No. 4,  W. D.  Ward  and
     Fricke,  J. E.,  Eds. Am.  Speech Hearing Assoc., Washington, D.C.
Elliot,  D. N.,  The  effect of  sensorineural  lesions on pitch discrimination  in  cats,
     Ann. Otol. Rhinol. Laryngol., 70, 582-598 (1961).
Faltynek, L.,  and  Vesely,  C.,  Zur Restitution  der  Mikrophonpotentiale  des  Meer-
     schweinchens nach Kurzfristiger Larmbelastung, Arch.  Ohren— Nasen— u. Kehlk-
     opfheilk, ver. Z. Hals-,  Nasen- u.  Ohrenheilk, 184,  109-114 (1964).
Fletcher,  J.  K. and  King,  W. P.,  Susceptibility of stapedectomized patients  to noise
     induced  temporary threshold  shifts, Ann.  Otol. Rhinol. Laryngol, 72,  900-907
     (1963).
Flugel, J. C., On  local fatigue in  the auditory system., Brit.  J.  Psychol., 11,  103-134
     (1920).
Fosbroke, J., Practical observations the pathology and  treatment of deafness. Lancet,
     1,  645-648 (1830, 1831).
Gel'tishcheva, E. A. and Ponomarenko, I. I., Noise standards for adolescents, Hygiene and
     Sanitation, 33, 199-203(1968).
Gerth, B. Otosklerose bei  Larmarbeitern. HNO (Berlin), 14,  205-208 (1966).
Glorig, A.,  Hearing studies of telephone operating personnel, /. Speech Hearing Res.,
     12,  169-178 (1969).
Gravendeel,   D.  W. and Plomp, R., Micro-noise trama? AMA Arch. Otolaryngol,  71,
     656^663 (1960).
Greisen,   L.   Comparative   investigations  of  different  auditory  fatigue  tests.  Acta
     OtO'Laryngol, 39, 132-135 (1951).
Harris,  J.  D.,  Hearing-loss  trend  curves  and  the  damage-risk  criterion  in diesel-
     engineeroom personnel, /. Acoust. Soc. Am., 37, 444-452  (1965).
Herman,  P.  N. and Clack, T. D.,  Post-stimulatory shift of the Preyer  reflex threshold
     in the  white  rat, /. Aud. Res.,  3,  189-204 (1963).
Hood,  J. D.,  Observations  upon  the  relationship of  loudness  discomfort level  and
     auditory fatigue to sound-pressure level and  sensation level, /. Acoust. Soc. Am.
     44,  959-964 (1968).                                               r
Hunter-Duvar, I. (Henry Ford), unpublished thesis, Wayne State U., (1971).
Jansen,   G.,  Relation  between temporary  threshold shift and  peripheral  circulatory
     effects   of  sound, In: Physiological Effects  of Noise,   B.  L. Welch  and  A.  S.
     Welch,  Eds., Plenum,  New York,  67-74 (1970).
Jerger,  J.  F.  and  Carhart,  R;i  Temporary  threshold shift as  an  index of noise
     susceptibility, /. Acoust. Soc.  Am.,  28, 611-613 (1956).
Kellerhals,  B.,   Acoustic   trauma   and  cochlear microcirculation, Adv.  Oto-Rhino
     Laryng.,  18, 91-168  (1972).
                                        290

-------
Kristensen, H.  K.,  On the  relation  between  the hearing of  weavers  and the body
    type, Acta Oto-Laryngol,  34,  82-94 (1946).
Kosa,  D.  and   Lampe,   I.,  Uber   die   Pneumatisation  des  Warzenfortsatzes  bei
    Larmarbeitern,  HNO  (Berlin)  15,  324-325 (1967).
Kubo,  M.,  Acoustic  trauma-its  development  and  prevention  in  transport workers,
    Proc.  16th   General  Assembly of the Japan  Medical  Congress,  Osaka,  (April
     1963).
Kuroyanagi, S., A contribution of  the  so-called telephone-deafness, J. Oto-rhinolaryng.
    Soc. Jap.. 63,  1438-1448  (1960),  ref ZNHO 69, 242  (1960/61).
Larsen,  B., Occupational deafness, Acta  Oto-Laryngol, 41,  139-157  (1952).
Link,  R.  and  Handl,  K.,  Die  Pneumatisation  des  Schafenbeins,  ein Schutz  gegen
    Larmschwerhorigkeit,  Arch. Ohr. Nas. Kehlkopfheilk,  167,  610-613 (1955).
Luz, G. A.,  Fletcher, J.  L.,  Fravel,  W.  J,  and  Mosko,  J. D.,  The  relationship
    between  temporary  threshold shift  and  permanent  threshold shift  in  rhesus
    monkeys  exposed  to  impulse noise,  US  AMRL  Report  No.  928,  U.S.  Army
    Med. Res. Lab.,  Fort Knox, Ky.  April  (1971).
Oppliger, G. C., Schulthess, G. V. and Grandjean, E., Uber die Beziehung der Gehorermu-
    dung zur bleibenden traumatischen Schwerhorigkeit, Acta Oto-Laryngol, 52,  415-428
    (1960).
Paparella,, M. M.,  Brady, D.  R.  and  Hoel, R., Sensori-neural  hearing loss in chronic otitis
    media and mastoiditis, Trans. Am. Ophthalm. OtolaryngoL, 74, 108-115 (1970).
Peyser, A., Gesundheitswesen u. Krankenfursorge. Theoretische und experimentelle  grund-
    lagen des personlichen schallschutzes. Deutsche  med.  Wochenschrift 56,  150-151
    (1930).
Pfander, F., Hone der  temporaren Schwellenabwanderung (TTS) in Audiogram und "Ruck-
    wanderungszeit"  gerausch- und  knallbelasteter  Ohren als Test  knallgefahrdeter
    Hororgane.^rc/i.  Klin. Exp. Ohren-, Nasen- Kehlkopfhlk. 191, 586-590 (1968).
Hester,  D., Der  Einfluss vegetativ wirksamer  Pharmaka  auf  die  Adaptation  bezw.
    Horermudung. Arch. Ohren-, Nasen- Kehlkopfheilk, 162, 473-487 (1953).
Sataloff, J., Vassallo,  L.,  and Menduke, H.,  Temporary  and permanent  hearing loss. A
    ten-year follow-up, Arch. Env. Health 10, 67-70 (1965).
Shapiro, S. L., Deafness following short-term exposure to industrial noise, Ann. Otol.  etc.
    68, 1170-1181 (1959).
Strubinski,  A., The  significance of the  tests of hearing adaptation in diabetes, Otolaryng.
    Pol 24, 311-318(1970).
Sulkowski, W., Badania nad przydatnoscia wybranych prob zmeczenia i adaptacji sluchu w
    profilaktyce  przemyslowych urazow akustycznych. Research on the value of certain
    hearing fatigue and hearing adaptation tests in the prevention of occupational deafness,
    Medycyna pracy (Warsaw),  20, 354-368 (1969).
Temkin, J., Die Schadigung des Onres durch Larm und Erschutterung, Mschr. Ohrenheilk. u.
    Laryngo-Rhinologie 67,  257-299, 450-457, 527-553,705-736, 823-834 (1933).
Theilgaard, E., Testing of the organ  of hearing with special reference to noise prophylaxis,
    Acta OtolaryngoL 37,  347-354 (1949).
                                        291

-------
Theilgaard,  E., Investigations in auditory fatigue in individuals with normal hearing and in
     noise workers (weavers)., Acta Oto-Laryngol. 39, 527-537 (1951).
Tota, G.  and Bocci, G.,  The importance  of the  color of the iris in the evaluation  of
     resistance to auditory fatigue (in Italian), Rivista Oto-Neuro-Oftalmologica (Bologna,
     Italy) 42, 183-193(1967).
Ward,  W. D.,  Studies on  the aural reflex, HI. Reflex latency as inferred from reductions  of
     temporary threshold shift from impulses, /. Acoust. Soc. Am. 34,  1132-1137 (1962).
Ward,  W. D., The  concept of susceptibility to hearing loss, /. Occup. Med 7, 595-607
     (1965).
Ward, W.  D., Susceptibility to auditory  fatigue. In: Contributions to Sensory Physiology, 3,
     W. D. Neff, Ed., Academic Press, New York (1968).
Ward,  W. D., Temporary threshold  shift and damage-risk criteria for intermittent noise
     exposures, /. Acoust, Soc. Am. 48, 561-574 (1970).
Ward,  W. D. and Duvall, A. J. HI, Behavioral and ultrastructural correlates of acoustic
     trauma, Ann. Otol. Rhinol Laryngol. 80, 881-896 (1971).
                                        292

-------
     GROWTH OF ITS AND COURSE OF RECOVERY FOR DIFFERENT NOISES;
                       IMPLICATIONS FOR GROWTH OF PTS

                                 Wolfgang KRAAK
              Technical University of Dresden, Sektion Informationstechn.
                          8027 Dresden, Germ. Dem. Rep.
    Temporary threshold shift (TTS) is often taken as a measure of noise effects that are
detrimental to hearing. Here it is mainly the TTS2 - i.e., the temporary threshold shift 2 min
after noise  exposure - which serves as criterion for designating the accumulating stress on
hearing. But there are various severe  objections to such a procedure. Noise exposure has
shown that after prolonged periods the TTS2 approaches a limiting value. The reduction of
the TT$2, the course  of recovery, depends on how long exposure is continued after this
limiting value has been reached.
    Fig. 1 shows the growth and course of TTS recovery for wideband noise having a sound
pressure level of L^ = 100 dB at test frequencies of 0.5,  1, 2, 4 and 8 kHz as measured in 20
young people (40 ears) of normal hearing [ 1 ]. This clearly  demonstrates the approaching
of a limiting value and the strongly delayed course of recovery after prolonged retention at
this asymptotic value. A similar asymptotic behavior  and delay in recovery time  at the
retention on the limit  value has also been established for intermittent noise [2] and pulse
sequences [ 3 ].
    Because there is no unambiguous connection between  the  time needed for TTS re-
covery  and TTS2 the  latter alone proves to be inadequate for characterizing  the stress on
hearing.
    Knowledge of the biochemical processes in the inner ear, especially in the hair cells [4]
and the course of TTS at noise exposure suggests the introduction of
                                      S= /(TTS) dt                          (1)
                                          •X

as a measure  of physiological stress where the integral  over TTS has to cover the period
during and after noise exposure as well.
    For further investigation, the TTS at 4 kHz has been selected to characterize the stress
on hearing.
    G. FUDER and L. KRACHT [2]  have studied the physiological stress at different noise
exposure according to Fig. 2. Each series of measurements has been made on 15 to  25
young people of normal  hearing (students and apprentices). It can be shown that for steady
noise  (types a and  b in Fig. 2), there is a simple relation between cause (noise) and effect
(physiological stress S) if noise exposure is expressed by

                                             (t)| dt.                          (2)

Here the integral of the magnitude of the A-weighted sound pressure should cover the entire
period of noise exposure.

                                        293

-------
dB

25

20-

 15-

 V

 5
        773P
              10
 dB
 25\  irr*2

 20

  15
 iO

 5*
 dB
 25
11
 20

  V

  10

  5"
       rrs
                      100
                                  30min
                            •fOOOmin
                              tE*12Omin
                                 -I-
                                10OO rrtin
                       W
                                 •+•
                               •COOmin
Figure 1. Growth of ITS and 'course of recovery as a function of time of noise exposure. Sound level L
100 dBA. Mean ITS of 20 persons.
                                        294

-------
     Steady noise
     L - 90... 105
     t=Bs... 8h
 Steady noise
 intermittent exposure
L = $3.  105 dBA
ft = 8s...40 min
if 8s... 40min
                           Impulse noise
                           t=KO... «5  dB
                           tz = 0,1ms ...3ms
                                    .  00
Figure 2. Types of noises used for noise exposure during the laboratory tests. The values below the figures
indicate the range in which the tests take place. Not all combinations could be realized.
     Figure 3 shows the averages S§ ^/(TTS)  dt = f (B§) obtained from 54 different noises
of types a to b. This correlation can be expressed by the linear equation
where
                                                (3)

                                                (4)
                                    PS"1"J(/ubar)^

For single pulses and pulse trains (type c) the studies of H. ERTEL [3]  - performed with
peak levels of L ^ 140 dB and pulse spacings of tp ^ 3 sec. - resulted in
                                     Si=0iB!                                     (5)
where

and
=Jp2 (t) dt

   1     dB
                                         15  (/Ubar)2-
                                                (6)

                                                (7)
The time  integral  of the squared sound pressure  in Eq. (6) again covers the whole noise
exposure time. With a low pulse spacing of tp ^ 0.5 sec. as obtained, for example, with
automatic small arms,  the physiological stress will reduce. If the pulse rate is sufficiently
high, the effective value ft reduces to approx. 10%, an effect that, in all probability, can be
attributed to the acoustic reflex.
     But a measure of the noise-dependent physiological stress on hearing  proves to be
useful only in the event that a connection can be established to the noise-induced hearing
                                          295

-------
»•
• ~-
I
«
1
I
V
1
•
4
2
»'
I
f
4
I
«'
1
t
t
t
•0'









	

















































» V











































y
y[
o
B
oy
/
*
















»
>
'


















/



















/
1















.
*
7.
^

















y

i


















^







































/
/
















0
jf
V










































t
»


































y
^
















.^
f



















/















































Sj -/(Trs; vo*M a/am IfH-urm tntirmtntnt t*f>oiurr to «*od> noat
             TTS • tut AvfMOT} 4 000 Mi e-20penorLi pur tut itntt
                  •/ type o andtt n F,g 2
                     Figure 3. TTS - integral as a function of noise exposure.
loss. Because of the linear relationship of Eq. (3) it was merely the expression PTS = f (85)
for steady  noise that has been searched for (PTS = permanent threshold shift). Apart from
some other data, the material collected by PASCHIER-VERMEER (5]  was used for evalua-
tion based  on the following hypothesis:
     The loss of hearing with increasing age (presbyacusis)  shall be expressed by a load
variable that is equivalent to Eq. (2), viz: -
                                          Padt=Pa(tL-tE)
(8)
                                           296

-------
where
                             tg = Overall duration of noise exposure as covered by Eq.(2)

                             pa = 0.4/ubar (= La = 66 dB).

As a rule, work published in the relevant literature about noise-induced hearing loss gives the
PTS after presbyausis correction. Here the threshold shift measured in subjects suffering
from loss of hearing has been reduced by the probable age-dependent threshold shift. In our
evaluating for the literature about noise-induced hearing loss, this  correction had  to be
revoked.
    The result of evaluating the data  collected  by PASCHIER-VERMEER [5]  and other
researchers is shown in Fig. 4 where the ordinate contains the prospective actual threshold
shift at 4 kHz (caused by noise  and age) plotted against the function
                                = BS+Ba
In close approximation this yields: -

                              PTS=551g- dB
with
                                         B
                                           o
                              B0 = 2.5 • 108/ubar • s.
                                               (9)



                                               (10)

                                               (11)
If it holds true that S of Eq. (1) is equal to the physiciological stress, a hypothesis can be
worked out on the loss of hearing, both for steady and impulsive noises
with
where
   PTS = 55 lg4-dB
               PC
                                            f(p)
                                                                           (12)
                     dt                        (13)

         = instantaneous value of A-weighted sound pressure
In a range of 100/ubar <
been determined yet.
   pc   = 22.5 ubar

   |3C   =  1.1-107

   f(p) -  1 for (PA(0  < 100/ubar

   f(p) = 2 for (PA(0  > 2 • 103/ubar

(t)| < 2 • 103/ubar, f(p) follows a transfer function that has not
                                         297

-------
PTS
       •c
      H
      M
      >C
      :c
  Sound level L in

 •>• -m   105
 *--^   101
 •   •   100
                                                           .-»   9V
                                                                         *- -4   9Z
                        H
                        IS
                        71
r-3 . f(P).     ! • ', » ra
(nee tq.C>). (0) and (?)).
Fvery point represents the "
at 40CC lit of i« to rtfjrly ?
ployeee «ho rere tipo-e^1 to
noiien (after
                                                                          i.-ls due
                                                        hypothetic .1 oound IcTel of
                                                        (PTS . t(*ft) after
 tr .
-------
                                    References

1.   Funder, G., and Kracht, L., Der Einfluss von Grenzwerten der TTS auf die Beurteilung
    der gehorschadigenden Wirkung stationaren Larms.
    VII. AICB-Kongrej3 Dresden Proc., 87-91, 1972.
2.   Fuder,  G., and  Kracht,  L.,  Ph.D.-THESES  at  Techn.  Univ. Dresden. Paper in
    preparation.
3.   Ertel, H., Ph.D.—THESES at Techn. Univ. Dresden. Paper in preparation.
4.   Vosteen, K. H., Neue Aspekte zur Biologie und Pathologic des Innenohres.
    Arch. Ohr.-Nas.-und Kehlk.-Heilk. 178, 1-104, 1961.
5.   Passchier-Vermeer, W., Hearing loss due to exposure to steady-state broadband noise.
    Research Institute for Public Health Engineering. Delft, The Netherlands,  Report 35,
    April 1968.
6.   Coles, R. R. A., and Rice, C. G., Assessment of Risk of Hearing Loss due to Impulse
    Noise.
    In Kryter, K. D., Effects of Noise on men. Academic Press, 71-94, 1970.
                                        299

-------
      EXPERIMENTS ON ANIMALS SUBJECT TO ACUTE ACOUSTIC TRAUMA

                                  Wiktor Jankowski
                       Clinic of Otolaryngology of the Medical
                            Academy in Wroclaw, Poland.

    It is well known that the  development of temporary hearing loss in operators working
under bad acoustic conditions during the  8-hour shift is not linear with respect to the time
of work. The greatest amount of that loss occurs during the first hour or hour and a half of
the initial time period.
    Thus in our experiments  made on animals our particular attention has been paid to
that period of acoustic exposure.
    The development of hearing loss in our animals has been observed by us on the basis of
the measurement of the loss of cochlear microphonics (CM) and action potential (AP). The
concept of using the decrease of  both these potentials, occurring during the acoustic expo-
sure, as  the measure of the drop  in hearing  sensitivity is not a new one. Other authors have
long been making such experiments; still, they were applying mostly only short (lasting only
a few minutes) acoustic exposures.
    In  our experiments we have been using white noise at levels of 80,85, 90,95 and 100
dB, with the exposure times of 5,15, 30, 60 and 90 minutes.
    In  one of our experiments we have conducted 25 test series, each of them embodying
four to  five individual tests - thus making jointly in that part of work more  than 100
individual tests.
    The values of CM have been measured for six pure tones and the mean value of the CM
voltage for all six tones has been calculated. In our experiments, the values of both potentials
have been expressed as a percentage of the potential's initial value, i.e. the value obtained
prior to exposure. The obtained results are illustrated on the diagram. The exposure time
periods are marked off on abscissa, while  the loss values  of both potentials are marked off
on the ordinate. The solid curves  represent the behavior of the CM and the dashed lines that
of the AP. On the right hand side of the diagram the exposure level is marked off.
    I should like to direct your attention only to the relation between the CM loss and the
AP loss. As may be seen on the diagram, the respective curves are not parallel  to each other.
After  short exposures, the loss of CM  exceeds that of AP. That can  be best seen when
employing high exposure levels. I should like to bring your attention to the variation of the
AP with these high intensities.  With  95 dB, the AP value falls below the lowest value that
could be measured after ninety minutes of exposure. At the 100-dB level, the  AP disappears
already after 60 minutes of exposure. That never has been observed in the case of the CM. It
is known that after an animal dies, the AP disappears at once, while the CM endures-to be
sure, at a lower voltage—still for several minutes.
    We have been also observing the recovery of the losses of both potentials, after termi-
nation of the exposure, during the  next  90 minutes.  In general, the smaller the loss, the
quicker  the recovery of the potential loss. After terminating the 90 minute exposure to the
100 dB level, during the next 90 minutes, the reappearance of AP has not been observed.
     The above losses of both potentials constitute a result of metabolic unbalance. We
have tried to prove that assumption by administering cytochrone C (which promotes the

                                        301

-------
                                  Figure 1. Jankowskj
oxidation processes) before exposure, as well as ATP (which promotes the energetic processes)
or both of these substances simultaneously, and then observing their effect on the formation
of losses appearing during exposure.
     As the most prominent differences between  the shape or both potentials have been
occurring in response to 100 dB, we have repeated these experiments, except that the above
mentioned substances have been given before applying the load. The prophylactic giving of
cytochrome C, ATP or both had a most beneficial effect on the reduction of the loss of CM
following exposures lasting 5,  15 and 30 minutes. However, for the longer exposures, the
beneficial effect of pre-administration of the above substances has not been so distinct: the
prophylactic administration of the above preparations had not so advantageous effect on
reduction of the AP loss; for  durations exceeding 60 to 90  minutes the  effect of these
preparations on the AP-curve shape was negligible.
     How are the results of these experiments to be explained? It is known that, as the
sensory cells of the organ of Corti operate under rather unfavorable oxygen supply condi-
                                        302

-------
tions, even a weak disturbance of the exygen or energetic economy provokes a distinct drop
of CM, and the administration of cytochrome C or ATP distinctly improves the efficiency of
the  sensory cells of the organ of cort. The drop  of AP occurs after the longer and stronger
exposures because its supply with metabolites necessary for keeping up its efficiency is, so
to say, better, because they are drawn directly from the blood circuit system. In that way
could be explained in some measure the differences between the behavior of both poten-
tials.
     It might  be asked if the behavior of both  potentials  is  characteristic only for the
acoustic trauma?
     Doc. dr  Zb. Ziemski, my  associate, has noticed a  similar behavior of both potentials
when he  was poisoning  guinea  pigs with  some organic solvents (polyethylene glycol,
propylene glycol,  dimethyl formamide, dioxane,  four-hydrofuran) by way of inhalation.
During the initial period of poisoning the drop of CM exceeded that of the AP. As the time
of poisoning was extended, the loss of AP exceeded the loss of CM.
     The same results have been obtained by Ziemski and myself when poisoning animals
with sodium salicylate. During the initial periods  of poisoning the losses of CM were greater,
but with extension of the time of poisoning, the losses of AP exceeded those of CM.
     Similar results have been obtained by Preibisch-Effenberger and Ziemski in the com-
bined work in which they have been exposing guinea pigs to ultra sonic noise (800 kHz at an
intensity of 7 watts/cm2). After 5 min of exposure, the  loss of CM greatly exceeded that of
AP, while after 30 min the values of the loss of AP approach those of CM.
     Similar differences in the  behavior of  these two potentials have been observed by
Deutsch, who, using a series of hypoxia periods,  found a greater loss of AP than of CM.
Simmons et al.,  after a strong and long-lasting  exposure, have found that the loss of AP
exceeds greatly that of CM.  Silverstein  has obtained, in  case of the salicylate poisoning,
results much like ours.  Spoendlin has noticed, in his ultramicroscopic studies after severe
exposures,  greater changes  in the mitochondria  of the  nerve endings  than in the
mitochondria of the hair cells.
     All above cited studies stress the difference in behavior of these potentials ascribed to
disturbance in the metabolism of the hair cells of  the organ of Corti and in the metabolism
of the aural nerve.
     If we  were perhaps to propose a general law, it would be that "any disturbance of the
metabolism of the organ of Corti or of nerve metabolism  will at first produce the strongest
effect on the CM, while more prolonged exposure to  the disturbing factor will lead to a
greater diminution of AP."
     At any rate it seems that, in view of the experimental results, it would be worthwhile
to continue investigations regarding the different behavior of both morphologic elements.

                                     References

 1.   Aubry, M.,  Pialoux, P.,  Burgeat, M.: Influence d'une stimulation acquistique intense
     sur la response de la cochlee. Acta Otolarying.  (Stock 60, 191-196, (1965).
 2.   Burgeat, M.,  Burgeat, C.: Study of changes  in the  cochlear microphonic potentials
     arising after intense auditory stimulation. J. Physiol  (Paris) 56, 225-232, (1964).

                                        303

-------
 3.  Deutsch, E.: Cumulative effects of oxygen lack on the electrical phenomene of the
     cochlea. Ann Otol Rhinol Laryngol. 73,348-357, (1964).
 4.  Faltynek,  L.,  Vesely,  C.:  Zur  Restitution  der Mikrophonpotentiale  des Meer-
     schweinches nach Kurzfristiger Larmbelastung.  Arch. Ohr.-Nas.-KehlkJHeilk.  184,
     109-115. (1964).
 5.  Gerhardt, H. J.,  Wagner, H.: Die Wirkung dosierter Ger&usch-belastung auf Mikrophon-
     potentiale der Meerschweinchenschnecke.  Arch. Ohr.-Nas.-Kehlk.Heilk. 179, 458-472,
     (1962).
 6.  Gisselsson, L., Soerensen, H.: Auditory adaptation and fatigue in cochlear potentials.
     Acta Otolaryng.  Stockh. 50,39M05, (1959).
 7.  Jankowski, W., Ziemski, Z., Cyrulewska-Orowska, J.: Pharmacological tests of Preven-
    tion  the acoustic trauma in  electrophysiologic  experiments. Archium Akustyki 5,
     129-143(1970).
 8.  Jankowski, W., Ziemski, Z., Birecki, W., Cyrulewska-Orowska, J., Praga, J., Kowalew-
  .   ska,  M.: The effects of acute acoustic trauma and urethan anesthesia on the patterns of
     biopotentials of internal ear in guinea pigs. Otolaryng. Pol., 25,253-265, (1971 A).
 9.  Jankowski, W., Ziemski, Z.: Differences in Performance of Cortis Organ sensual cells
     and  of the Aural Nerve at Experimental  Disturbances of their Efficiency.  Archiwum
     Akustyki. 3, 233-231, (1971B).
10.  Kumagai, K.: Studies on  the  microphonic action of the cohchlea in auditory
     disturbance. Excerpta Med. sect. XL, vol. 13, 904, (1960).
11.  Meyer zum  Gottesberge, H.: Rauch, S., Koburg, E.: Unterschiede in Metabolism us der
     eincelnen Schneckenwindungen. Acta Otolaryng. (Stockholm) 59, 116-123,  (1965).
12.  Preibisch-Effenberger R., Freigang  B., Seidel, P., Ziemski Z., Reczek-Krauss H.: The
     effect of ultrasounds and cytochrome C  on functional  disturbances of Corti's organ
     and  cochlear nerve in animals. Otolaryng. Pol. 24, 21-26, (1970).
13.  Sekula, J.,  Trabka, J., Miodonski, A.: Noise Influence on Biopotentials  of Hearing
     Organ. Pamietnik XXVII Zjazdu Otolaryngologow Polskich. Katowice  1968, Warszawa,
     PZWL,(1970)24.
14.  Silverstein, H., Bomstein, J. M., Davies,  J.: Salicylate ototoxicity a  biochemical and
     electrophysiological study. Ann of Otol Rhinol Laryngol, 76, 118-129, (1967).
15.  Simmons, F. B.,: Middle Ear Muscle Protection from the Acoustic  Trauma of loud
     continuous Sound. Ann Otol Rhinol Laryngol. 69,  1063-1071, (1960).
16.  Spoendlin,  H.:  Submikroskopische Veranderungen an Cortischen Organ des Meer-
     schweichens nach  acustischen  Belastung.  Practica Otolaryngologica.  20, 197-201,
     (1958).
17.  Wagner, H.,  Gerhardt, H. J.: Die Wirkung  dosierter Reintonbelastung auf die Mikro-
     phonpotentiale  der Meerschweinchen-schnecke.  Arch. Ohr.-Nas.—Kehlk. Heilk.,  181,
     82-106, (1963).
18.  Ziemski, Z., Jankowski,  W.: Ototoxidity of sodium salicylate. Otolaryng. Pol. 26,
     391-395. (1972).
                                        304

-------
                         SESSION 4 A
INTERACTION OF NOISE WITH OTHER NOXIOUS AGENTS IN PRODUCTION OF
                        HEARING LOSS

                   Chairman: E. Lehnhardt, BRD
                             305

-------
             INFLUENCES OF CHEMICAL AGENTS ON HEARING LOSS

                                 M. Haider, Vienna

    The ototoxic effects of chemical substances have been known for a long time. There
are otologic reports on carbon  monoxide  poisoning that were written in the seventeenth
century (van Helmont,  1667).  Other known ototoxic industrial chemicals  include lead,
phosphorus, halogenated hydrocarbons, mercury, carbon disulfide, etc. An extensive review
on the industrial health aspects of these substances is given by Lehnhardt (1965).
     Besides industrial products there are  many drugs for which ototoxic side-effects have
been described. At  the  moment the best  known are  some antibiotics like streptomycin,
dihydrostreptomycin, kanamycin, and neomycin. Some other ototoxic drugs are salicylates,
quinine and substances with similar effects (e.g. chloroquine), arsenic, alcaloids (strychnine,
morphine, scopolamine), oleum chenopoii, etc.  Finally, the ototoxic effects  of stimulants
(nicotine, alcohol), narcotics and endogenous intoxications (e.g. during some infections and
other diseases) are worth mentioning. Huizing (1966) gives a summary of these problems.
    This paper will mainly review the combined effects of ototoxic substances and noise.
On principle, the combination of chemical agents and noise can be:
    (a) indifferent (combination does not  differ from its most effective component),
    (b) additive (combination corresponds roughly to sum of both factors),
    (c) synergistic (effect of combination  is higher than sum of individual components).
    (d) antagonistic (effect of combination is less than most effective component), or
    (e) protective (substances make the ear less susceptible to hearing loss).
    Data  on such combined effects may  be derived  from  different sources, e.g.: clinical
case studies, systematic field studies, experimental research with animals  and  experimental
research in connection with temporary threshold shift.
    There are many clinical observations on combined effects of intoxication (e.g.: under
CO, C$2,  Nitrobenzol)  and noise.  For twelve  years, the working conditions of a man
described by Wagemann (1960) contained  noise  of 80-90 Phon,  but the clinical symptoms
(audiogram, vestibular signs, etc.) seemed to indicate a CO-intoxication.  It  is possible that
both noxes had a combined  effect. One  case, described by Lehnhardt  (1965) showed a
similar combined effect of trichlorethylene and noise. Some of the experimental studies on
drugs are based on clinical  observations. Darrouzet (1967) mentions a case of hearing loss
after streptomycin treatment and surgical intervention on one ear with noise-stress through
a milling-machine. Dayal et al. (1971) designed their experiments according  to the common
clinical situation of premature babies in incubators (generating noise of about 68-72 dB)
receiving kanamycin treatment.
    Field studies show the frequency of hearing loss that has to be expected under the
influence of certain  chemical substances. Examples are given in the Scandinavian reports on
hearing loss caused  by "chronic" carbon monoxide poisoning.  Lumio (1948 a, b) found
hearing deficiencies  in 78% of his patients. 44% of them he assumed as typical for carbon
monoxide poisoning.
  With financial support of the Austrian "Fonds zur Forderung der wissenschaftlichen Forschung" and
   the "Allgem Unfallversicherungsanstalt".

                                        307

-------
     Some other examples are given in reports on hearing loss caused by carbon disulfide.
Zenk (1970,1971) found hearing losses under this condition to have a higher incidence than
control groups of the same age. In such cases it seems difficult to rule out some industrial
noise effect. Unfortunately none of the reports I  found in the literature tried to single out
the influence of hearing loss in connection with chemical agents versus hearing loss due to
industrial noise. It could be interesting to compare equivalent groups of individuals with and
without  noise exposure to groups with or without the influence of chemical  agents.
     Animal experiments have shown that the combined effect of noise and chemical agents
may be  synergistic. Darrouzet (1962)  demonstrated that antibiotics (kanamycin) might
sensitize the cochlea to the damaging influence of noise. If noise was given before the drug,
no synergistic effect occurred. Quante et al. (1970) reported a potentiating effect of noise
(90, 100, 110 dB) with  kanamycin treatment. Jauhiainen et al. (1972) examined harmful
effects of noise  (115 dB) and neomycin both electrophysiologically and microsopically. In
the guinea pig they found a synergistic effect in hair cell damage as well as in amplitude
reduction of cochlea microphonic potentials as demonstrated  in Fig. 1. The authors con-
cluded that there is greater susceptibility to noise-induced hearing loss in persons treated
with these antibiotics.
     Many authors have shown  that ototoxic effects damage  mainly the  outer hair  cells,
maximally at  the basal end. Combination with noise seems to extend the damage further
towards the apical end. Dayal et al. (1971) found that even a low-level noise (68 to 72 dB at
125 Hz) combined with low dosage of kanamycin may have produced a synergistic effect.
                             toe,
                             10
u
x
o
a
u
»  M
                          •4
                          a
                                    • 10
                                                 -10
                                                        -4.0
                                                           dB
                                 DECREASE  OF  VOLT ASK

                                    OF  CM -PO IE N T I A I
Fig. 1.  Relation between percentage outer hair  cell damage and average loss in cochfear microphonic
potentials. Point A refers to animals exposed to neomycin, point B to those exposed to noise alone, and
point C to those exposed to both noise and kanamycin. (From Jauhianinen et al., 1972).
                                         308

-------
Two factors which are ineffective (subthreshold) by themselves may in such way give rise to
manifest damage. In this case the changes of the outer hair cells were seen primarily in the
apical turn of the cochlea. This is demonstrated in Fig. 2.
     The authors assume that the hair cells of the  apical turn were sensitized to damage by
the low frequency noise.
     A possible cumulative effect for carbon monoxide and noise has been assumed by Zorn
(1968), who found a delaying effect on carbon monoxide elimination with 65 dB noise.
Klosterkotter (1972), however, could not verify these results.
     Some of the possibilities of protective influences by various chemical agents will only
be mentioned here briefly. A combination effect of neomycin and noise of 120 dB together
with a protective effect of adenosintriphosphate (ATP) has been described by Faltynek and
Vesely (1969). The course of normalization of cochlear microphonic potentials after com-
bined neomycin and noise influence was favorably affected by ATP. Protective effects of
vitamin B and aminoacids have been described amongst others  by Darrouzet (1962, 1963,
1967). The possible protective effects of  vitamin  A have been under discussion since the
                 ooooooooooooooooooooooooooooooooooooo
           3*  o<
JOOOOOOOOOOOOOOOOOC
'^OOOOOOOOOOOC
 3OOOOOOO»OO«
3OOOOOOOOOOOOOOC
3OOOOOOOOOOOOC
X3«OOOOOOOOOOC
                 ooooooooooooooooooooooooooooooooooooo
               oooooooooooooooooooooooooooooooooooooooooc
                OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOC
               oooooooooooooooooooooooooo«o*oooooooo*«
                                      JOOOOO
                 ooooooooooooooooooooooooooooooooooooo
                     JOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
                     3OOOOOQOOOOOOOOOOOOOOQOOOOOOOOOOOOOOOOOOOOOOO
                     JOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
                 ooooooooooooooooooooooooooooooooooooo
            2 OOOOOOOOOQOOOOOOOOOOOOOOOOOOOC
               ooooooooooooooooooooooooooooc
              oooooooooooooooooooooooooooooc

Fig. 2. Cochleogram of guinea pig exposed to incubator noise and receiving kanamycin 15 mgm/kgm body
wt, per day for 5 weeks. Dark circles indicate damaged hair cells. The damage is predominant in the OHC of
the apical turn. (From Dayal et al., 1971)
                                      309

-------
early reports of Willemse  (1952) and Ruedi (1954). A positive effect of nicotinic acid is
described in Sheehy (1960) and Nowak et al. (1971). As one example of the combination
effects of chemical agents and  noise on temporary threshold shift I  will mention some
preliminary  results of our own experiments. Eighteen normal-hearing students were tested
under two conditions. In the experimental condition they were exposed to 200 ppm CO for
4 hours. Before and after  the exposure, auditory thresholds were measured with a Bekesy
audiometer. After half an hour and after two and a half hours, 100 clicks were given and the
evoked potentials were computer-analyzed out of the EEG (Vertex-Mastoid). In the control
condition, the subjects performed the same tests without the influence of CO. The situa-
tions were rotated systematically in  a double-blind  design. The threshold measurements  at
2000, 3000 and 4000 c/s showed no systematic change and no statistical significant dif-
ferences.
     For the auditory evoked potentials no significant latency changes but amplitude reduc-
tion under  CO-exposure of the  main negative-positive  peak to peak amplitude could be
demonstrated. This is shown in Fig. 3.
    - There was also a diminution of amplitudes from the 05-hour exposure to the 2.5-hour
exposure. This diminution occurs in both situations but it is significant only under the
CO-condition. Some authors have shown similar results for the late part of the visual evoked
potentials for animals under the CO-condition (Xintaras et  al, 1966). In an earlier experi-
ment, we demonstrated that the slow electric brain potentials (expectancy waves) show a
marked reduction under the influence of even  lower CO-concentrations in the air (Groll-
Knapp et al. 1972). So it seems that even very low concentrations of CO in the inhaled air
may change the brain reactions evoked by acoustic signals.
    €0

    50-

    W-

    30-

    ao-
          of  E?
             ew  crz
            Control
EM Etl
ioorr,co
Latenet)
msec.
410-
410-
400-
90-
*>•
10-


b








































EM EfZ EM ffl
CoKtrol iOOfrCO
Figure 3.  Amplitudes (relative values) and latencies (msec) of auditory evoked potentials before and after 4
hours CO-exposure (200 ppm) compared to the control condition.
                                        310

-------
     To get some information on possible combination effects of CO and noise, the subjects
were  exposed to a 105-dB octave-band noise with a middle frequency of 2000 Hz for 15
min in both situations. One result of TTS-measurements at 4, 8, 16, 32, and 64 minutes
after exposure is shown in Fig. 4.
     The TTS values, measured with  a test-tone of 3000 c/s are slightly higher immediately
after  the CO-exposition than under the  control-situation. But  the late TTS values are
                                Control
CO	
                                                         33t
.Figure 4. TTS at 3000 HZ after  15-min exposure to octaveband noise (center frequency  2000 Hz)
following a 4-hour CO exposure (200 ppm) compared to control condition.

                                         311

-------
practically identical. There is no statistical significant difference between both situations. It
must be concluded that ««hrer the circumstances described no synergistic effect of CO- and
noise exposure could be demonstrated.
                                  REFERENCES

(1) Darrouzet, J. and E. de Lima-Sobrinho: Oreille interne kanamycine et Traumatisme
     acoustique. Etude experimentelle: Rev. Laryng.Otol. 83, 781-786, (1962)
( 2) Darrouzet,  J.: Essais  de protection  de 1'organ  decorti centre 1'otoxicite. Rev.—
     Laryng. (Bordeaux) 84,459-478 (1962).
( 3) Darrouzet, J.: Essais  de  protection de 1'organ decorti centre 1'ototoxicite des anti-
     biotiques. Table Ronde: Rev. Laryng. (bordeaux), 84,478-488 (1963).
( 4) Darrouzet, J.: Essais de protection de 1'organe de  corti centre 1'ototoxicite  anti-
     biotiques. Etude experimentelle: Acta otolarying. (Stockholm) 63, 49-64 (1967).
( 5) Dayal, S., A. Kokkainen and P. Mitchell: Combined effects of Noise and Kanamycin.
     Ann.Otol. 80,897-902 (1971).
( 6) Fairy nek,  K. and C. Vesely: Einflu/3 des Neomycin auf  die  Funktion  der Meer-
     schweinchenschnecke. Mschr.f. OHrhk. 103, 545-547 (1969).
( 7) Groll-Knapp, E., H. Wagner, H. Hauck and M. Haider: Effect of low Carbon Monoxide
     Concentrations on  Vigilance and Computer-Analysed Brain Potentials. In: Carbon
     Monoxide-Origin,  Measurement and Air Quality Criteria, 116-119, Dusseld. (1972).
( 8) Huizing, E.  H.:  Toxische Schaden des Hororgans.  In:  Berendes:  HNO-Heilkunde-
     Lehrbuch, HI/3 (1966).
( 9) Jauhiainen, T., A. Kohonen and M. Jauhiainen: Combined effect of noise and neomy-
     cin on the cochlea. Acta otolaryng. 73, 387-390 (1972)
(10) KlosterkStter, W.:  Der Umweltfaktor Larm als Komponente kumulativer Umwelt-
     wirkungen. Arb.-Med., Soz.-Med., Arb.-Hyg. 10,281-286 (1972).
(11) Lumino, J. S.: Otoneurological Studies of Chronic  Carbon Monoxide Poisoning in
     Finland. Acta oto-Iaryng. (Stockholm), Suppl.67,65-112 (1948 a)
(12) Lumino, J. S.: Hearing deficiencies caused by carbon monoxide (generator-gas).  Acta
     oto-laryng. (Stockholm), Suppl. 71, 1-112 (1948 b).
(13) Lehnhardt, E.: Die Berufsscnaden des Ohres. Arch.f.Ohr.-Nas.us. Kehlk.-Heilk. 185,
     11-242(1965).
(14) Nowak,  R.,  B.  Kleinfeld und D.  Dahl:  Tierexperimentelle Untersuchungen  zum
     Verhalten  der  Mikrophonpotentiale  nach  Nikotinsaure-gabe am  durch  Schall
     vorgeschadigten Ohr. Mschr. Ohr.Hk. Wien, 11, 495-498, (1971).
(15)Quante,  M.,  H.   Stupp  und J.  P. Brun:  Ototoxikosen  unter Larmbelastung
     Arch.klin.exp. Ohr.-Nas.-Kehlk.Heilk. 1960, 233-237. (1970).
(16) Ruedi,  L.: Wirkungen des Vitamin A im menschlichen  und tierischen Gehorgang.
     Schweiz.med.Wschr. 84,1411-1414 (1954)
(17) Sheehy, J. L.: Vasodilator therapy in sensorineural  hearing loss. Laryngoscope 70,
     885-914. (1960).

                                       312

-------
(18) Wagemann, W.: Das otologische Bild der Kohlenoxidvergiftung. Zschr. f.Lar.Rhin.Oto.
    38,691-702(1960).
(19) Willemse, C.:  Protection against occupational deafness: Role played by Vitamin A.
    Acta oto-laryng. belg. 6, 319-324, (1952).
(20) Xintaras, C., B. L. Johnson, C. E. Terrill and M. F. Sobecki: Application of the Evoked
    Response Technique in Air Pollution Toxicology. Toxikol. and AppLPharm. 8, 77-87
    (1966).
(21) Zenk, H.: Schwefelkohlenstoffeinwirkungen auf die rhino-otologischen Funktionen der
    Beschaftigten in der Kunstfaserindustrie. Int.Arch.ArbMed. 27, 210-220 (1960).
(22) Zenk,   H.:  Zum   rhino-otologischen   Bilde   chronischer  Schwefelkohlenstoffein-
    wirkungen. Z.Laryng.-Rhinol. 50, 170-175 (1971)
(23) Zorn, H.: Die kombinierte Wirkung physikahscher und chemischer Noxen, aufgezeigt
    am  Beispiel  der Schadigung durch Larm und CO. Schriftenr.  Arb.Med.Soz.Med.
    Arb.Hyg. Stuttgart, Bd.27, (1968)
                                        313

-------
       HEARING LOSS OF FOREST WORKERS AND OF TRACTOR OPERATORS
                    (INTERACTION OF NOISE WITH VIBRATION)

                                    Istvan Pinter
               State Institute of Occupational Health, Budapest, Hungary

    In our investigations to be presented an answer was sought to the question, whether
simultaneous exposure to noise and vibration has any influence on the dynamics of the
development of hearing loss.
    The starting point of the investigations is, on one hand, the equal energy principle, i.-e.
that exposure of the same magnitude results in the same degree of hearing loss and, on the
other hand, the interrelationship between the PTS and TT$2, namely that the value of ITS
of healthy young people measured 2 minutes after the end of a daily noise exposure equals
the average PTS caused by a ten-year exposure.
    In order to make a judgement of the question possible,  the HLs found in tractor
operators and forest workers were compared to those of control persons, who were exposed
to noise only, and for whom the basic principle mentioned above could be proven to hold,
on the basis of data  in the  literature and of the results of investigations of our own.
Accordingly, the persons investigated were grouped as follows:
    1.) Tractor drivers with noise exposure and exposed simultaneously to vibration of a
frequency around 10 Hz; workers of the furniture industry served as their controls.
    2.) Contra - Stihl power  saw operators of the forest industry with noise exposure and
with a  simultaneous exposure to vibration of frequency in the range of 125-350 Hz; their
controls were workers in the textile industry, exposed to the same level of noise.
    3.) Forest workers with clinically verified vibration damage (alterations in the locomo-
tion and vascular systems Raynaud-syndrome); "healthy" forest workers served as their
controls.
    The tractor drivers and the workers of the forest and furniture industry are the em-
ployees of the  same forest company at the same geographical location near the capital; the
textile industry workers are employees of a metropolitan plant.
Noise analysis

     Noise exposure was determined according to the R 1999 ISO noise-measuring recom-
mendations on the basis of the equal-energy principle. For this, the noise in the different
working cycles was picked up for a period of time according to the regulations at the height
of the ears of the persons working on the working places, and recorded on magnetic tape.
For the recording on magnetic tape, a precision noise level meter (Bruel & Kjaer type 2204)
and  a portable measuring tape recorder (Nagra III) were used. The noise recorded was
analyzed  in the laboratory (Bruel & Kjaer Real-time 1/3 Octave  Analyzer type 33-47, Level
Recorder type 2305) and was evaluated with mathematical methods and the Lgq value was
determined with a dosimeter (Bruel & Kjaer type  4423).
     The results  of the noise analysis  in case of several typical occupational activities are
shown in Fig. 1 and 2.

                                        315

-------
                100
                •'.
                s:


                II


                70
              g  M
                «o
              K
              tf 30
M


II

           WOOD WOfiK
     ;  !  TRACTOR
     i  •
                                                 r
                                                          110 dBA
                H
               IM
                                                                        ••— » »»-
                                                         ,..,,,...,.
Figure 1. Percent distribution of sound level, Lgq and 1/3 octave spectra measured for tractor driving and
at working places in the furniture industry during appropriate working cycles.
     The upper parts of Figures 1  and 2 demonstrate the  percent distribution of the noise
levels and the corresponding Leq -values in the period of time investigated, in case of forest
workers (Contra-Stihl power saw operators), in relation to work done on the tractor, in the
                                           316

-------
100
 '.
 K

 U

 70
              —
              U
                40
                30
                10
n
                          TEXTILE WORK
                        i  STIHL-SAW
                                  >-r
                                     80
                           *•
                             100     ifa  dBA
                            ...... 1f •.,,.,,
                                       I
                                       ST^-S*.                  :::|  B. I
Figure 2. Percent distribution of sound levels, L^ and 1/3 octave spectra measured at the power saw and
at working places in the textile industry during appropriate working cycles.
furniture factory and in the  textile plant; the lower parts of the figures show the l/3-octave
spectra of these same noises.
                                          31 7

-------
     The equivalent sound levels of the different workers vary, depending upon the working
phase and upon the location of the measurement, as follows:

                   Forest workers:     96-100   dB(A)
                   Tractor drivers:     90-98    dB(A)
                   Workers in the furniture industry:    90-98  dB(A)
                   Workers in the textile industry: .      97-101 dB(A)


Audiometric investigations

     The  hearing investigations were performed with a Peters Type AP-6 clinical audio-
meter, standardized according to  the  1964 ISO recommendation /R 389/. The measure-
ments were done 16 hours after the last noise exposure in an anechoic chamber correspond-
ing to the ANSI SI. 3-1960 standard.
     By the evaluation of the audiograms the ISO R 1999 and the AAOO 1970 recommen-
dations were taken into account.                             .         -•••••
     The  de facto noise-induced hearing loss was determined from the audiograms taking
into consideration the sociacusis according to sex (Spoor 1967, Passchier-Vermeer 1968),
thus the audiograms can be compared on ground of the exposure  times and levels only,
without the necessity of taking the variations due to age and sex into account.
     From the audiograms corrected according to the procedure above, the mean values of
the hearing curves (059) were calculated in case of all the 4 groups, at  1-4, 5-14, 15-24 and
finally,  at precisely 10 years  of  exposure. Exposures of less than one year were not con-
sidered. At the mean  value curves the average width of the field of scatter was calculated
according  to  frequencies; from this the relative deviation  was computed,  which varies
between 0.53 and 0.20 in the frequency range of 3000-6000 Hz. Accordingly, they are
within the values evaluable mathematically.
     The audiograms (right ear) of only such workers are entered into the evaluation, whose
hearing loss can be ascribed with high probability  on the  basis of the anamnesis and of
otological investigation, to noise exposure of occupational origin only.
Results of the investigations
     On  the  basis of the criteria outlined above, in the first  approximation the average
hearing thresholds (059) were determined as the function of the type of noise exposure only.
     This is shown by Figure 3.
     No. 1 shows the D$Q audiograms of forest workers (ContraStihl power saw operators),
No. 2 those of the corresponding control textile industry 'workers, No. 3 demonstrates the
curves of tractor1 operators working in the forest and No. 4 those of their controls, workers
in a furniture factory.
     It is to  be seen on the figure that, on  the one  hand,  the hearing loss in the high-
frequency range of both the tractor drivers and forest workers is greater than that of the
controls and, oil the other hand, that the hearing threshold in the low frequency range (250
Hz -1 kHz) of tractor drivers and of forest workers is higher than usual.

                                        318

-------
                                                                       1 FELLERS
                                                                       2 TEXTILE
                                                                        WORKERS
                                                                       3 TRACTOR
                                                                        DRIVERS
                                                                       4 WOOD
                                                                        WORKERS
               2SO
1000
2000    3000  tOOC    6000  8000   Hz
             Figure 3.  Average hearing loss due to noise as a function of noise exposure.
    In order to follow the development of  this type of audiogram, the D$Q values were
compared also in terms of exposure time. The audiograrns corresponding to  1-4, 5-14 and
15-24 years of exposure, of tractor drivers and of furniture industry workers are seen in Fig.
4.  In case of tractor drivers the group with 1-4 years of exposure is not depicted, because
there were only two workers in this category.
    Beyond the course of the increase in the hearing loss, the  figure clearly  demonstrates
the different character of  the curves that is correlated with doubling the exposure time, and
within this, especially the difference at the low frequencies.
    The same comparison in case of forest- and of textile-industry workers is demonstrated
in  Fig. 5.
    Here an even  more marked difference between the two groups can be seen. In case of
the textile-industry workers the  equal-energy principle is valid, and according to this the
hearing threshold increases with exposure time.
    This is, however, not  observed in forest workers; on the contrary, there  is hardly any
change in  the already-developed hearing loss between 2 kHz and 8 kHz  with  an increase in
the exposure time. It can, however,  be seen also in case of forest workers that the hearing
thresholds are higher between 250 Hz and  1 kHz.

                                         319

-------
••
-









:?













































•*













ta

9!.





^
- - " '












T.,f
-
S-H






-*;



'»«;
-







! ' N 1
_J
•
•.





^



s 1
a • K
. c
- - m
1








	 	





,,
>
N










Vvv,
S.
S
^







k
— ""
l»^
N
s
\








k
A
\
k
N
1





\
\
\
\


ft
4








X
•N
> •*
\
\
\
,



"~ — ' — _

1 	



* «M

IHACTOfi




r
i
1 c
t *
i

i 1
y

WOOD
WORKERS


....
,— - t.

, '

<
,'

                         a
                                 •ODD
                                          i
                                                  *OOD
Figure 4. Average hearing loss due to none of trader drivers and of workers in the furniture industry for
exposures of 1-4, 5-14 and 15-24 years, respectively.
     The object of our further investigations will he the nowadays accepted interrelation-
ship between the TTS2 and PTS on 4000 Hz in case of a 10 year exposure. The upper part
of  Figure 6  shows the D5Q-curves which developed in the various groups in case of an
exposure of  IO years.  The lower part demonstrates  the mean values measured  at 4000  Hz
and  the TTS2  values expected 2 minutes after a one-day  exposure and calculated on the
basis of the Ward et al. (1958) and Nakamura (1967) equation.
     It can be seen from the curves that the equal-energy principle is acceptable in relation
to the controls (furniture- and textile-industry workers), but it does not seem valid in case
of tractor drivers  and  forest workers. The  situation is similar at  an equal noise exposure
when comparing the controls and the groups corresponding to them. In the equation  for
calculating the  TTS2 for the investigation of the interrelationship between the PTS and the
                                         320

-------



                                                                FEU c»s
                                            .^".
                                             - V
                                                *



                                                      :
                                                           (COD WOO  M;
Figure 5. Average hearing loss due to noise of forest workers and those of textile industry for exposures of
1-4, 5-14 and 15-24 years, respectively.
T]$2 tMC values of K  the on-traction  was calculated from the daily exposure times found
on the basis of field investigations of several year duration. Comparing the PTS and TTS2
values we - as many others - found good agreement in case of the controls. In case of forest
workers and  tractor drivers, however,  the  measured  values exceeded the calculated TTS->
values. Correcting all the values on the basis of an equal value for R, it turns out even more
pregnant that, in  case of forest workers and tractor drivers the found PTS values exceed the
expected values.  Hence, if we take the daily exposure  times into consideration, then the
discrepancy having displayed itself by the curves is solved on 4000 Hz.
     Finally,  the  audiograrns of power  saw  operators suffering from clinically verified vibra-
tion disease (alterations in the locomotor or in the vascular system) were compared to those
of healthy power  saw operators exposed to noise  tor the same length of time.
                                          321

-------
             -••
              :
1 FEUERS
1 TEXTILE
 WORKERS
                                                               3 TRACTOR
                                                                DRIVERS
                                                               X WOOD
                                                                WORKERS
                                             N.
                                             ••x 's
                                             V-^
             U
                                 nco
                                           MOO
                                                          tooo eon HI
                            TTS,-  1.06R(S-85) log
&TOUP
I
2
1
4
U
4104
58
55
T.,,
•*
s
D
•
V
o
R
0.7
1.0
03
0.9
4000 Hz
PIS
30.0
35,6
28.8
24.0
TTS1
22.0
34.0
24.G
24.0
n
ncofr
0.7
0,7
0.7
0.7
400QHz«*r
PTS
J0.0
24.9
20.2
16.8
TTS,
22.0
23,8
16.8
16,8
                              I  2  )  i
      Figure 6. Average hearing Ion in case of 10 yean exposure and the TTS calculated at 4000 Hz.
     The average hearing threshold and the range of deviation for all the forest workers are
Jepicted by  the curves in  the figures. It can be seen  that the  hearing losses of patients
oiffering from the damage of the locomotor organs only (the points in Fig,  7) correspond to
hose of the "healthy persons". In contrast to this, the hearing loss of those  with vascular
iamage (Raynaud  disease) at 3000-6000 Hz is always greater than that of the healthy
                                          322

-------
                                              GROUP I
                         	
                   '-•
                   •
                   -

                                            ODD
                                                      WOO    3000 4030    6000 MOD Hi
Figure 7. A Hearing  loss in the range of 2000-6000 Hz due to noise in patients with  vibration damage of
the locomotor system.
                           B. X-ray photograph of typical vibration damage.

                                                323

-------
persons.  Figure  9 shows the audiogram and X-ray angiogram  of a patient  with  Raynaud-
syndrome.
     The  D5Q-curves found in  the various groups correspond  quantitatively as  well  as
qualitatively  to  the  results of investigations upon workers of similar occupation  published
by Passchier-Vermeer(1968), Kylin(1971), Burns(!964). Robinson (1970), Dieroff (1963)
and others.
     On  the  basis of the data presented we  feel justified in concluding that the simultaneous
effects of vibration and noise result  in an influencing of the dynamics of hearing  loss,
differing from that elicited by noise alone.
                                         iROUl •
                                       BOO
                                                2000   xoo ton    no torn HI
   Figure 8. Hearing loss in the range of 2000-8000 Hi due to noise in patients with Raynaud syndrome.


                                         324

-------

                                                       XOO     JOCC OOD    tOOO tOOOnr
Figure 9.  A. Hearing loss of patient D.M. suffering from Reynaud-syndrome following a 10-year exposure.
                                    B. The angioram of O.M.




                                               325

-------
Discussion

     Acoustic stimuli reach the inner ear normally by air conduction and thus, according to
the general opinion accepted at present, vibration of the whole body plays only a secondary
role in the development of hearing losses.
     The measurements of Bekesy  (1960) proved that hearing loss develops by way of bone
conduction, too.
     Wittmaack (1928, 1934), Popov (1928), Jokoyama (1963) and Morita (1958) proved
on the basis of animal experiments that the simultaneous effect of body vibrations and noise
is more marked on the upper turns of the cochlea corresponding in the low tones, contrary
to the anatomical effect of noise alone, which manifests itself at the base.
     Our  investigations verify that the results of experiments with animals are  valid for
humans too.
     From the investigations  demonstrated it can undoubtedly be concluded that vibration,
acting together with noise, has a potentiating effect which, depending on the frequency
range of vibration, exerts different influences upon the dynamics of hearing  losses. Vibra-
tion in the subacoustic frequency range (tractor drivers) damages  the hearing at the low
frequencies (250 Hz -  1 kHz) and,  at the high frequencies (3-6  kHz) it  shows an increase of
the threshold, increasing with the increase of the exposure time and anyhow exceeds the
values of the controls. Vibration in the audible range (125-350 Hz)  similarly causes damage
in the range of the low tones, but also, the loss brought about in the range of the high tones
is greater than that in case of the controls, with the limitation that here the damage develops
during an exposure of 1-4 years  duration, and hardly changes with an increase  of the
exposure  time. The potentiating effect is even more  marked in patients with Raynaud-
syndrome.
     The data demonstrated prove  that the change in the dynamics of hearing loss in case of
tractor drivers and forest workers is affected  by the simultaneous effect of  vibration and
noise, but they do not give any hints as for the course of pathomechanism. The clarification
of this needs further investigation.

                                    References

American Standard Criteria for Background Noise in Audiometer Rooms S -1.  3. (1960).
Bekesy, G., Experiments in hearing. McGraw Hill, New-York (1960).
Burns, W., Hinchcliffe, R., Littler, T.  S.,  An exploratory study of hearing loss and noise
     exposure in textile workers. Ann. Occ.Hyg. 7, 323-333 (1964).
Burns, W. and Robinson, D. W., Hearing and Noise in Industry. H.M.S.O. London (1970).
Dieroff, H. G., Die Larmschwerhdrigkeit in der Industrie, Johann Ambrosius Barth, Leipzig,
     (1963).
Guide for Conservation of hearing  in Noise. Supplement to the Transactions  of the Ameri-
     can Academy of Ophtalmology and Otolaryngology, (1970).
Holmgren, G., Johansson, L., Kylin, B. and Linden, O., Noise and Hearing of a Population
     of Forest Workers. Cit. Robinson, D. W.: Occupational Hearing Loss, Academic Press
     London and New-York (1971).

                                       326

-------
International Organization for Standardization. "Standard Reference 200 for the Calibra-
    tion of Pure-tone Audiometers"; ISO Recommendation R 389, Geneve, (1964).
International  Organization  for  Standardization.  "Assessment of  Noise  Exposure during
    Work  for  Hearing  Conversation  Porposes"; ISO Draft Recommendation R.  1999,
    Geneve, (1970).
Lehnhardt, E.,  Die  Berufschaden  des Ohres.  Arch.Ohn-Nas.-Kehlk.-Heilk., 185, 11-242
    (1965).
Kryter, K. D., The effect of noise on man. Academic Press, New York and London, (1970).
Morita,  M., Experimental  studies  on  the  acoustic  influence of  vibration  and  noise.
    OtoLFukuoka, 4, Suppl. 5, 327-332 (1958).
Nakamura, S. and Katano, Y., Relationship between temporary threshold shift and duration
    of noise exposure J.Aud.Res. 7,401-411 (1967).
Nitschkoff, St., Kriwizkaja,  G., Larmbelastung, akustischer  Reiz  und neurovegetative
    Stdrungen. VEB Georg Thieme Leipzig, (1968).
Nixon, J., Glorig, A., Noise Induced P.T.S. at 2000 and 4000 Hz., J. Acoust.Soc.Amer., 33,
    904-913(1961).
Passchier-Vermeer, W., Hearing loss due to exposure to steady-state broadband noise. Insti-
    tute for Public Health Engineering TNO, Netherlands, (1968).
Pinter, I., Gehorschaden  der Arbeiter verschiedenartiger Betriebe. International Congress of
    the Hungarian Society of Occupational Health, Budapest, (1971).
Pinter, I., A textilipar zajproblemainak munkaegeszsegugyi vetulete. Magyar Textiltechnika,
    24, 329-332, (1972).
Popov, N. F., cit by Nitschkoff and Kriwizkaja (1968).
Spoor, A.,  Presbyacusis  values in relation to noise-induced hearing loss. Int.Aud. 6,  48-57
    (1967).
Ward,  W. D., Glorig, A. and Sklar, D. L., Dependence of Temporary Threshold Shift at 4000
    cps on Intensity and Time. J.Acoust. Soc.Amer., 30, 944-954 (1958).
Wittmaack, K.,  Vergleichende Untersuchungen uber die Luftschall  Luftleitung und Boden-
    schwingung und Korperleitungschadigungen des akustischen Apparates. Arch-Ohr.-Nas.
    u. Kehlk.-Heilk. 102, 96-100 (1928).
Wittmaack, K.,  Uber die Wege der Knochenleitung mit besonderer  Beriicksichtigung der
    Schallschadigung durch  Kochenleitung Acta oto-larying., Stockholm, 19, 105-11
    (1934).
Yokoyama, T.,  Studies  on the occupational deafness.IV. J.Oto-rhino-laryng.Soc.Jap., 66,
    777-783(1963).
                                        327

-------
                           INFRASOUND AND HEARING

                        Charles W. Nixon and Daniel L. Johnson
                        Aerospace Medical Research Laboratory
                      Wright-Patterson Air Force Base. Ohio USA

                                  INTRODUCTION

     Airborne acoustic energy in the frequency region below 20 Hz is arbitrarily described
as infrasound. Human hearing is insensitive  to infrasound  except at exceedingly intense
levels. Technical knowledge on the effects of infrasound on  man is rather sparse; however,
limited information obtained from  a few real life experiences(2 > *7) and experimental in-
quiries clearly suggests that  infrasound  exposures  at high  levels might adversely  affect
man(3'9). The extent to which  infrasound experienced during routine living and occupa-
tional activities might influence human performance, health  and wellbeing is an open ques-
tion. Infrasound is generated by various events in nature as well as numerous  man-made
systems and activities and is experienced by all of us to varying degrees (Table 1).
     Infrasound occurs in nature at  relatively low levels as a result of actions such as winds,
air turbulence, thunder, volcanic activity, storms, large waterfalls and even the impact of
waves on beaches (4>14). Natural activities such as  walking, jogging, and swimming must
theoretically produce  the same low  frequency  pressure fluctuations  as infrasound on the
auditory system. For example, walking or jogging in a way which causes the head  to vary 15
cm in altitude at each step is equivalent to  approximately 90 dB.  Swimming in such a way
that  the ear becomes submerged in  7.5 cm of water during part of the stroke (and  not
submerged otherwise) is equivalent to 141 dB.
     A variety of adverse effects of  naturally  occurring infrasound on human behavior has
been speculated, however essentially no objective  data relating  human response to the
infrasound exposure have been generated. A single study does report^5) a correlation (0.5)
of infrasound exposure with activities such  as automobile accidents, absenteeism in school
children and in unskilled workers during a period of high infrasonic exposure  in a metropoli-
tan area. Although these data are not conclusive, they do sustain the possibility that such
relationship and effects might exist for low level exposures.
     The incidence of  infrasound from man-made sources appears to  be growing both in
terms of intensity and in number of exposures.  Infrasonic energy is found in a wide variety
of sources including air heating and cooling systems, occupational environs  in which com-
pressors, pneumatic devices, air turbulence,  and the like, are  found, in essentially all forms
of transportation systems including the high powered propulsion systems for space vehicles,
and many more(ls'16l  Man-made  infrasound  generally occurs at much higher intensity
levels than  that found  from natural causes, consequently the threat of potential adverse
effects on people is also much greater. Subjective reports of  effects of infrasonic exposure
from other than natural sources have included disorientation, nausea and general unpleasant-
ness as well as a variety of other symptoms  (2>! 5). A  comprehensive study  by Mohr et
al,(9) which examined intense infrasound and  low  frequency effects'on humans  demon-
strated clearcut adverse symptoms which are summarized in Figure 1 (! °). The nature of the
observed behavior indicates that human subjective tolerance  limits for these short duration

                                        329

-------
                                       Table 1
           REPRESENTING SOURCES OF INFRASOUND FOUND IN NATURE AND IN
                               MAN-MADE ACTIVITIES

                               SOURCES OF INFRASOUND
Nature
Source Est Freq Est Max SPL
Thunder
Earthquake
Ocean Waves < 1
Wind: lOOKm/hr 135 dB
25Km/hr 110 dB
Atmospheric Pressure

Fluctuations < 1 100 dB

Volcano







Man -Made
Source
Free Field
Jet Engines
Helicopters
Large Rockets
Diesel Engines

Activities

Running

Swimming
Riding in:
Aircraft
Submarines
Rockets
Automobiles
Helicopters
Est Freq

1-20
1-20
1-20
10-20



<2

<2

<10
5-20
1-20
'' 1-20 '
5-20
Est Max SPL

135
115
150
110



90

140

120
140
145
120
130
(~2 min) exposures may have been very close. The extent to which these symptoms may
disappear as the levels of the exposures decrease has not been defined.
    Auditory system response has long been a criterion for measuring the acceptability of
noise  exposures in terms of both subjective and objective effects. Although infrasound is
inaudible, except at high intensity levels, its potential effect on the human auditory system
and function must  be fully  examined.  Due to such factors as the harmonic contents of
overtones of infrasound, the fact that infrasound rarely if ever occurs free from higher
frequency sounds and the fact that the ear itself causes distortion at high pressure levels, it is
                                        330

-------
                                          EXPOSURE
                                    OBSERVED BEHAVIOR
                                   HOUR
       FREQUENCY
    20
HERTZ
                                            TEST
                                             (10)

                                             (n)
                                Middle ear discomfort, pressure
                                build up; Tickle in ear;
                                One subject had mild
                                Abdominal wall vibration

                                No TTS  1 hour later
                                                          All tolerable; But subjective
                                                          sensations rose rapidly
                                                          for exposures above 145 dB
                                                                No Protection
                                                          One hour exposures;
                                                          No protection; No subjective
                                                          effects observed
                                                            Adapted From
                                                            Mohr et al
                                                            Reference 9
Figure 1. Representative Low Frequency and Infrasound Noise Environs and Human Subjective Responses.
desirable that maximum permissible exposure conditions for infrasound be defined relative
to human auditory function. Knowledge of infrasound and hearing, based on the few studies
reported in the technical literature and on results of investigative efforts recently completed
or underway in our own laboratory, is discussed herein.
               GENERATION AND MEASUREMENT OF INFRASOUND

     Infrasound lies below the range of frequency operation of many high quality items of
instrumentation typically  used in the study of acoustics and psychoacoustics. In order to
accurately measure and analyze infrasound some  additional  instrumentation performance
characteristics must be employed and certain precautions must be exercised (8>16). The
transducer must be capable of responding to DC to  insure acquisition of the total signal. The
remainder of the system must reflect an equivalent low frequency response, especially tape
recorders which must  operate in the FM range. Frequency analysis should be accomplished
on a "per cycle" basis or a small percentage bandwidth filtering of about 5% or less. Sound
level meters are not appropriate for assessing infrasound. It is extremely important that
technically acceptable instrumentation be  used and the frequency response be accurately
                                         331

-------
described. The higher harmonics generated with high level infrasonic signals will likely spill
over into the lower audio frequency regions where they are well above threshold of hearing
(Figure 2). It is clear that this energy is much louder than the fundamental infrasound and
may, in fact, determine the response of the experimental subject. Unless the experimenter is
fully aware of all of the energy present in his test conditions, as described by technically
appropriate  measurement instrumentation, human  responses to these  higher frequency
energies may erroneously be attributed solely to infrasound.
     This matter becomes more important as one considers the design and fabrication of
generators of infrasound for experimental investigations in  which humans will serve as
subjects. On the basis of threshold of hearing values for infrasound shown in Figure 3, a 10
Hz signal at 120 dB would be approximately 18-20 dB above threshold, on average. In order
for this signal to be inaudible at 20 Hz it must be more than 40 dBdown and at 100 Hz it
must be greater than 80 dB down. The very great difficulty in generating an infrasonic signal
CM
   140
    130
 0120
 0)
 miio
 UJ
LJ
en
en
o
I
en
    90
    80
    60
REPRESENTATIVE
   INFRASOUND
THRESHOLD  OF HEARING
"ACTUAL"SOUND
 SOURCE SHOWING
   " SPILLOVER"
                            "IDEALIZED"
                            SOUND SOURCE
                                4   5  6  7  8 9 10
                                 FREQUENCY  IN HERTZ
                                                    20
             30   40 50
 Figure 2. A Graphic Display of the "Spillover" or Upward Spread of Energy Into the Audiofrequency
 Region Frequently Encountered When An  Infrasound Source is Used for Experimental Purposes With
 Humans.
                                       332

-------
    140
"E  130
 3-120
 CM
 S NO
uJ 100
tu
uj 90
a:
V)
to
LJ
or
o.

170
    60
                                            I   I  I  I
                  tx.
            —  BEKESY, MAP
             O   YEOWART, MAP
             O   YEOWART, MAP
             *   YEOWART, TONE THRESHOLD
             A   YEOWART, NOISE THRESHOLD
                                                 I	1
                           3    4   56789K)
                                FREQUENCY   IN  HERTZ
                                                               20
30   40  50
Figure 3. Hearing Threshold Levels for Maximum Audible Pressure (MAP), Minimum Audible Field (MAF)
and for Bands of Noise.

with harmonics  below such values is rather obvious. Consequently, it is not only essential
that infrasound be accurately measured and analyzed, but in addition that interpretation of
human responses be made with knowledge of the total frequency response of the specific
test signal to which the observers were exposed. This problem of "spillover" or upward
spread  of the signal is not as great for measurement of hearing threshold levels as for the
high level energy required for studies of temporary threshold shift (TTS) due to infrasound.
    One of the reasons why so little research has been accomplished at these frequencies
may well be the difficulties encountered in the generation of infrasound signals.  Although
the  number of infrasound investigators is small, a variety of systems have been used for
infrasound generation, ranging  from small pistonphones coupled to an ear to large special
purpose pressure chambers which enclose the whole body. The characteristics of a number
of systems are summarized in Table 2. Data collected from the use of these generators are
sufficient to allow research on  infrasound  to be summarized  in a technical discussion as is
done in this paper.
    The purpose of this paper is to present a review of technical literature and experience
which represents the state of the  scientific understanding of the influence of infrasound on
                                       333

-------
                                                 Table 2

                 TYPICAL EXPERIMENTAL SOURCES USED IN INVESTIGATIONS OF
                                     INFRASOUND AND HEARING
 INVESTIGATOR

Bekesy  (1)
     FACILITY

Thermophone
BENOX, von Gierke
(18)
Mohr. etal  19)
Nixon (lit
Leventhall and
Hood (8)
Yeowart (22)
Pistonphone and
Mercury Manometer
Whole Body
Enclosures; and
Free Field:  Jet
Engine; High
Pressure Air Source;
Low Frequency Test
Chamber

Pistonphones
Whole Body Pressure
Chamber 3x4x6
feet
Monaural'Binaural
Headphones From
0.3m Diameter
Loudspeakers

Whole Body 1200
Litre Cabinet
    GENERATOR

Loudspeaker
Coupled Via
Manometer
Motor Driven
Pistonphone; Manual
Control Manometer

Hydraulic Loud-
speakers; High
Velocity Air Flow;
Low Frequency
Siren
Coupled to Ear via
Closed Tube
Four 15 meter Diam.
Loudspeakers: 300 W
Amplifier
Earmuffs; Drivers
Worked Into Volume
of 1 Litre
                                          Six 0.46m Diameter
                                          Loudspeakers on
                                          Sides of Chamber
  OPERATION

Beating of Two AC
Inside Thermophone
Capsule; Frequency
Response Down to
IHz

Alternating
Pressures Up to
50 Hz

Normal Operating
Modes for Various
Devices and
Facilities
Motor Driven
Alternating
Pressures

Operates as a
Helmholtz
Resonator Tunable
Over Range of 3 Hz
to 18 Hz

Response is Flat up
to 200 Hz
                       Electrodynamic
   PERFORMANCE

104 - 105 dynes/cm2
Sound Pressure
165 dB Alternating
Pressure;  180 dB
Static Pressure

Discrete Tones and
Bands of Noise at
Levels of 150
154 dB
165 dB + Alternating
Pressure
145 dB for Single
Frequency; 126 dB
For Noise Band
Maximum SPL is
ISOdBat IHz
                        Maximum SPL is
                        140 dB
Johnson  (7)
Whole Body Chamber
55 Cu. Ft.
Hydraulic Driven
6 Ft. Piston and
1.5 Ft. Piston
Alternating
Pressures, 0.5-
30 Hz
6 Ft. Piston: 172 dB
(0.5-10Hz). Falling
to!58dB($30Hz

1.5 Ft. Piston:
144 dB d- 10 Hz)
Falling to 135 dB
(?)20 Hz
                                                   334

-------
the human auditory system. The body of the report, which comprises the basic data review,
is  organized into four functional areas, (1) hearing threshold  levels  for infrasound, (2)
temporary hearing loss (TTS) due to infrasound, (3) additional effects of infrasound on the
auditory mechanism, and (4) infrasound effects on the  speech  reception aspect of voice
communication. These data are discussed  in terms of exposure guidelines, and tentative
limiting noise levels for infrasound exposures are recommended.

                          AUDITORY SYSTEM RESPONSE

Hearing Threshold Levels

     In order to evaluate effects of infrasound on the human auditory system and function,
the nominal  response  of the system must first be determined.  Perhaps one of the most
long-standing descriptions of hearing threshold levels (MAP) for acoustic energy below 20
Hz is that  of Bekesy('X A number of other investigators, at different times and using a
variety  of  instrumentation have  independently measured infrasound  hearing thresholds.
Most recently this has been accomplished by WhittleO9) and YeowartC20). The hearing
threshold levels from a series of studies by Yeowart are  compared in Figure 3. The agree-
ment among  the  various values is very good and  it provides confidence that the general
sensitivity curve for the human ear for this frequency region has been described.
     Measured MAP  values for infrasound are  also  contained in Figure 3. The classical
MAP-MAP  difference of 3 dB for audio frequencies is also present for infrasound (22).
Hearing threshold level values for  noise bands in this frequency region are also shown (2').
It  is evident  that no significant  difference between tone and  noise  threshold data are
observed between about 30 Hz and 100 Hz. The noise thresholds are significantly lower, by
about 4 dB, for frequencies below about  16 Hz. The greater sensitivity of the ear for the
bands of noise is attributed to detection of the peak factors present in the noise signal.

Temporary Threshold Shift (TTS)

     A quantitative relationship between human exposure to infrasound and hearing loss is
not well established.  Very few investigations of this phenomenon are found in the technical
literature, partly  because of a general low level of interest in this frequency region  and
partly because of the problems associated with the measurement of hearing thresholds for
infrasound  as well as the general inability to produce infrasound exposures free of audible
overtones. A  few investigators who have ventured into infrasound research with cognizance
of the latter problems are identified in Table 3.
     For threshold  determinations the general  approach has  been to measure effects of
infrasound  on  the standard audiometric test frequencies instead of for lower frequency
signals.  The question of audible overtones is not at all clear because most reports do not
contain spectral representations of the test signals. It cannot be determined from these
reports if the stimulus was truly infrasound  or if it was a multiple-component signal with the
fundamental  an infrasonic frequency. Nevertheless, it is a reasonable assumption that the
adverse effects of the infrasound alone would be no worse than those of the infrasound plus

                                         335

-------
overtones. This review considers the data as it is reported and does not attempt to critically
analyze the acoustic exposure while recognizing that actual observed effects may have been
highly influenced by energy above 20 Hz.
     Some early observations of possible infrasound effects on hearing were not described in
terms of audiometric test frequencies. Tonndorf C1 7) reports on the effects of infrasound in
the diesel rooms of submarines on the hearing of crew members. Depression of the upper
limits of hearing were demonstrated by decreased time  periods during which tuning fork
tests were audible. Recovery occurred after various intervals of time outside the diesel room.
Mohr et al(9) exposed subjects to infrasonic signals, both pure tones and noise bands, for 2
minutes or less at levels of 150-154 dB. Audiometry was not performed immediately follow-
ing the  exposures; however, measures taken about one hour later showed no ITS. In the
latter work an exposure signal was experienced only  once, while on-board submarine expo-
sures were experienced daily.
     Jerger (6) exposed 19 males to repeated three-minute signals of from 2-12 Hz at levels
of 119 dB to 144 dB SPL. ITS in the range of 3000-8000 Hz was observed in 11 of the 19
subjects for exposures of  137 to 141 dB. All ITS values were small, ranging from 10-22 dB.
The author indicates  that  the 7-12 Hz signals at 120-144 dB did produce considerable
masking over  the 100-4000 Hz range. It seems likely that some of the measured TTS was
caused by the masking signal.
     In  our laboratory a  number of studies of TTS and infrasound have been conducted
using  pistonphones and a large pressure chamber as signal generators. Using a pistonphone
coupled tightly  to the ear via an earmuff, Nixon 0 1) investigated effects of 14 Hz at 140 dB
and 18 Hz at 135 dB for 30 minute exposure durations on hearing threshold levels. Some
subjects experienced no changes in hearing due to the exposures  while others showed
various amounts of TTS, with one subject showing 20-25 dB at one test frequency.
    In another series of investigations, Johnson (7) measured the effects of auditory expo-
sures of 135 dB to 171 dB  at 0.5 Hz to 12 Hz and whole body exposures of 135 dB to 144
dB  at 1 Hz to 20 Hz. A pressure chamber which provides whole body exposures to infra-
sound at levels as high as 172 dB was used to generate  the stimuli. Exposure durations varied
from 26 sec of 7 Hz at 171 dB to 30 min of 4, 7 and 12 Hz at 140 dB. The various exposure
parameters, effects on hearing, if any, and recovery are itemized in Table 3.
    It is clear from the data contained in  Table  3 that TTS has been measured following
infrasound exposures at moderately intense levels. The observed changes in hearing thresh-
old  levels have  been small and recovery of pre-exposure hearing levels has been rapid for
the  few situations in which TTS did occur.
    Susceptibility,  Susceptibility of ears to infrasound-induced TTS appears to be gen-
erally the same  as TTS induced by higher-frequency energy. Amount of TTS induced by a
specific  exposure or whether or not TTS occurs,  both within and between subjects, show
about the same variability  as for audio frequency  exposures. Data are not available to
determine if susceptibility to infrasound due to age or to sex is different from that due to
exposures to audio frequency energy.
    Middle Ear Ventilation. As will be discussed later, infrasound exposure at levels suffi-
cient  to induce TTS also produces retraction of eardrum  membrane. The  efficiency of
middle ear transmission of energy to the inner ear is reduced when this system is retracted.
                           j
                                        336

-------
As  an investigator, one  must consider the advisability of having experimental subjects
periodically  ventilate  the middle ear system during studies of infrasound since different
effects would be expected from exposure of a retracted vs non-retracted drum-membrane
system.  Regardless of the exposure, it is critical that the middle ear system be adequately
ventilated prior  to measurement of post-exposure hearing threshold levels. A retracted
middle ear system will show reduced sensitivity which may be attributed to sensorineural
effects.


Other Effects on the Auditory System

     Infrasound  may  stimulate the auditory system at rather low levels  so as to be un-
detected or  at levels of sufficient  magnitude to cause aural pain. During whole body and
aural exposures to infrasound, particularly below about 5 Hz, at levels of 120 dB and above,
subjects may report a sensation that the eardrum membrane is being mechanically massaged.
At the lower intensity levels, perception of the sensation is not unpleasant and it becomes
less noticeable after a little time. At  higher intensity levels, subjects may report the sensa-
tion as being quite unpleasant and disliked; however,  this too appears to dissipate during
continuation of the  exposure period. At one time, pneumatic massage of the eardrum-
middle ear system was a common otological practice. It has been reported that mild massage
under properly controlled conditions is beneficial to the ear. However,  excessive mechanical
massage could be  detrimental,  presumably  because the mechanical  displacement due to
intense infrasound is so much larger than during typical listening situations. Massage of the
drum membrane system by infrasound, at rather high levels and/or for long duration expo-
sures, has produced effects clearly recognized at the drum membrane by investigators^' ' l).
    Pressure Build-up. Experimental subjects  almost  universally  describe a sensation of
pressure buildup in the ear shortly after initiation of infrasound exposure. This sensation is
reported by  many subjects at 126 dB and by virtually all persons at 132 dB. This fullness is
experienced  for  both  aural and whole body exposures. The sensation remains throughout
the exposure and persists for some time afterwards in many subjects. Ventilation of the ear
during exposure  may relieve the sensation of fullness; however, it is only temporary, for the
feeling of pressure quickly returns. This phenomenon appears  to occur a little earlier than
injection of the drum membrane is observed.
     Vascular Infection. A vascular injection  of the eardrum  membrane may be observed
during and following exposure. This injection is  similar to  that produced by therapeutic
massage  of the  drum membrane. The degree of injection may be slight to severe in which
case congestion appears all along the  handle of the malleus and in the folds. Although slight
injection may be caused by many different factors, it is not considered "abnormal". Severe
congestion must be recognized as a positive indication of overexposure.
     Drum Membrane Retraction.  The  cyclical displacement  of the  drum membrane, in-
ward  phase, during infrasound exposure appears to force gases from the middle ear cavity
out through the collapsed Eustachian tube. The negative pressure created by this action is
not automatically equalized  on the alternate phase of the cycle and drum membrane retrac-
tion will likely  occur. The effect of retraction on hearing is to reduce transmission of
acoustic energy to the inner ear and  in this mode is likely beneficial during exposure if not

                                        337

-------
allowed to become severe. There is some evidence^1'»'7) that drum membrane retraction
acquired during infrasound exposures may not be due entirely to a negative pressure in the
middle ear. During exposures to 14 Hz at  140 dB, all subjects  exhibited mild retraction.
After termination of the exposure each individual ventilated the middle ear system via
Valsalva, under the observation and guidance of an  otologist. Following confirmed ventila-
tion or pressure equalization, the retracted drum condition remained and persisted for some
time. An unexplainable function of the middle ear muscle system is believed to account for
the persistent retraction; however, this assumption  remains  to be investigated. One conse-
quence of this retraction phenomenon is that even with positive ventilation following expo-
sure, prior to audiometric testing, some retraction  may remain and indicate a slightly
elevated threshold.
     Drum Membrane Scar Tissue. The potential adverse effects of long duration exposures
of this nature have not been examined in the laboratory. Tonndorf (! 7),  in describing the
condition of drum membranes  of submariners exposed daily to diesel  room infrasound
found the formation  of cicatrized  tissue and loss of elastic fibers. Although a direct causal
relationship was not established, the incidence of these conditions among  the exposed crew
members clearly exceeded what  would normally be  expected in that  population from non-
infrasound and noise factors.
     Pain. Pain in the ear is easily recognized by individuals and is related to mechanical
displacement of the middle ear system beyond its limits of normal operation. Aural pain is
not related to sensitivity, as evidenced by the fact that normal and hard-of-hearing persons
have the same average aural pain thresholds.  Likewise,  it is not associated with sensorineural
hearing loss which can become  very  severe without any experience of pain. At infrasonic
signals, pain may be experienced at levels which pose no risk to the hearing function.
     Thresholds for aural pain are summarized in Figure 4. The data of Bekesy (* )and from
the BENOX report (13) are highly consistent. One datum on the figure represents a condi-
tion where one subject reported pain from a pistonphone exposure of around  140 dB at 14
Hz, which disappeared after a few minutes exposure at that level. It  appears that the pain
threshold might be elevated a few dB, 2-3 dB, with high-level exposure experience of the
subject.  The pain threshold appears to be about 140 dB around 20 Hz rapidly increasing to
about 162 dB at 2  Hz. Pain produced by static  pressure  on the ear, either positive or
negative, appears between 175 dB  and 180 dB.  Any form of aural pain during infrasound
exposure must be considered  an indicator that the  tolerance  limits of the mechanical
systems  have been reached and the exposure  should  be terminated and avoided in the
future.

                  VOICE COMMUNICATION-SPEECH RECEPTION

     Effects of infrasound exposure on voice communication are  generally considered to be
those affecting the talker. High-intensity  infrasound may influence various organisms and
functions involved in speech production. Amplitude modulation during speech production is
obvious  at very low frequencies as a result of the respiratory cage or chest being driven by
the infrasound, and is reported by essentially all exposed persons. Choking, coughing, gag
sensations,  chest wall vibration and modulation of respiratory rhytm have been reported for

                                        338

-------

-------
Hearing Protection

     Human exposure to intense infrasound may occur at hearing tolerance limits where
potential hazards exist or at lower levels which pose no  hearing risk but are subjectively
disagreeable. Effective hearing protection  is highly desirable in each of the situations
described. Classically, insert hearing protection  has provided good performance across the
audio  frequency range whereas earmuff protector performance decreases with  decreasing
frequency.
     Subjective reports  of ear protector effectiveness in intense  infrasound indicate that
good insert-type earplugs provide appreciable attenuation of the acoustic energy. Earmuff
type protectors appeared  to provide  negligible  protection  and on occasion appeared to
amplify  the noise under the muff. Earmuffs, which are  suspended from lightweight spring
tension headbands,  were noticed to visibly vibrate against the sides of the subject's head
during infrasound  exposure.  When worn over insert earplugs, earmuffs appeared to  add
attenuation obtained by the wearer.
     An experimental investigation of earmuff effectiveness in infrasound, using both a
subjective and a physical method, confirms the subjective observations reported above(* 2).
Good earmuff protectors provide about 10 dB of sound  protection between 20 Hz and  100
Hz and very little protection in the infrasound region.  For optimum protection in sound
fields below  20 Hz, good  insert earplugs are  recommended for intense exposures of long
duration.


Limiting Levels of infrasound

     Limiting levels of infrasound exposure effects on the auditory system must consider in
addition to potential hearing loss,  mechanical effects on the middle ear system-including
pain, speech reception and discomfort. The available knowledge from which limiting levels
may be  formulated  comes from  experience in intense infrasound and  from  laboratory
investigations.
     The  purpose  of defining relationships of infrasound  exposure  to  auditory system
effects, is  to allow potential risk  to  be determined on  the basis of descriptions of the
physical stimulus, There is somewhat of a problem in depicting the three stimulus variables
of importarice-SPL, duration and  frequehcy-in a simple fashion. Consequently, we have
adopted  a method  of representing exposures in terms  of level and  of number of cycles
(frequency times time) as parameters. Although the utilization  of this procedure does not
extend to the extremely low frequencies or to high frequencies it does appear to be a very
good approximation for representing exposure for the range 1 Hz to 20 Hz.
     The proposed equal risk formulation based upon adoption of this method is:

                         SPL = 10 log t+10 log f+base SPL

     A number  of experimental subjects have experienced exposures to 10 Hz at 144 dB for
durations of 8  minutes, via auditory  only or whole body presentation of the stimulus.
Although not necessarily enjoyable, no adverse effects  have been observed which would

                                       340

-------
indicate that these exposure conditions are threatening or harmful.  Accepting this set of
conditions as a base acceptable exposure, the formulation becomes:
                                           t           _f_
                    Limiting SPL = 10 log 8 min + 10 Log 10+144

     Various infrasound exposures conducted in  our laboratory are displayed in Figure 5
along with the curve which  represents the Limiting-SPL formulation. The Limiting-SPL
curve shows reasonable agreement with the experimental data collected to date, as well as
the proposed criterion. The same data are presented in a more conventional form in Figure
6. It is clear that exposure durations of 8 minutes for levels up to 150 dB caused essentially
no TTS. Actually, over 100 ear-exposures are shown for this duration range and  only two
experienced a mild TTS of 8 dB, with immediate recovery.
     Limiting levels for frequencies of 0.5 Hz to 20 Hz and  exposure durations of 0.5 min
to 1440 minutes in terms of "Limiting  SPL" are displayed in Table 3. On the basis of
experience  to date and lack of more complete data, it is essential that the table  values be
qualified. It is clear that non-auditory, whole body effects of infrasound occur at levels of
UJ

UJ
3j 140
v>
V)
UJ
SOUND PR
g
   120
                                        t   ^
                                   CHINCHILLA
                                        ( DRUM MEMBRANE
                                         AND MIDDLE EAR
                                           DAMAGE)
                                                          O WHOLE BODY EXPOSURES
                                                          A AURAL EXPOSURES
                                                          O SOME  TTS OBSERVED
                                                                     SUBMARINERS
                                                                    (TONNDORF)
        EXPOSURE LEVEL     ._
        LIMITED BY           <->
        NON-AUDITORY EFFECTS
                                 A

                                 AO

                                 A
                                      i
                      10
                                    100
1000
                                           111
                                           60
10,000
 Figure 5 Various Laboratory Infrasound Exposures in Terms of Level and Number of Cycles (frequency X
 time) and a Limiting Sound Pressure Level Curve Based on the Formulation:
                     t         _f_
 Limiting SPL = 10 log 8 min + 10 log 10 + 144
                                         341

-------
    30
 CO
 LJ
rr
D
Q
iu
o:
z>
CO
p
Q_
X
LJ
    25
    20
     15
     10
     0
             EXPOSURE
                      LEVELS
          A GREATER THAN ISOdB
          O 140-149 dB
          Q !30-l39dB
          O 120- 129 dB
         FILLED SYMBOLS INDICATE
         THAT SOME TTS OCCURRED
       D
CSD
  mn
                                                                        QD
                               3     4    5678910
                           FREQUENCY  IN  HERTZ
                                                                  15   20
                 30
Figure 6 A  Conventional Display  of Individual Exposures Recorded in Our Laboratory in Terms of
Frequency and Duration With Level as the Parameter. Solid symbols indicate that some TTS was observed;
the vast majority of exposures show no TTS.
150 dB and above. Consequently, whole body effects impose limitations at levels considered
to be safe for the auditory system.
    The proposed limiting noise levels, in more general terms, which may be considered as
acceptable are 150 dB at 1 Hz - 7 Hz,  145 dB at 8 Hz - 11 Hz and  HOdBat 12 Hz- 20 Hz.
These levels apply to discrete  frequencies or octave bands centered about the stated fre-
quencies. Maximum exposure duration is eight minutes with 16 hours rest between expo-
sures.  The use of good insert earplugs may increase the permissible  levels by 5 dB for the
same exposure times by reducing the aural contribution to the overall response. Earplugs are
strongly recommended for all  intense  infrasound exposures to minimize subjective  sensa-
tions.  Levels above 150 dB should be avoided even with maximum hearing protection until
additional technical data are accumulated.
    The normal levels at which aural pain is induced  by infrasound correspond  closely to
the limiting  values shown above for the 10 Hz  and  20 Hz frequency regions. At  2 Hz the
value is much higher at about 162 dB. Consequently, the threshold  regions for aural pain are
compatible with the proposed values and do not  impose any additional limitations.
                                        342

-------
                                             TABLE 3

       A SUMMARY OF STUDIES OF TEMPORARY HEARING LOSS FOLLOWING EXPOSURE
                                        TO INFRASOUND
INVESTIGATOR
 Tonndorf (17)
Mohr. et al  (9)
Jerger, et al  (6)
Nixon (11)
Nixon (11)
Johnson (7)
    EXPOSURE
Submarine Diesel Room
10 Hz -20 Hz.  No Level Given
Discrete Tones; Narrow Band
Noise in 10 Hz - 20 Hz Region.
150 - 154 dB Exposures of
About 2 Minutes

Successive 3 Minute Whole
Body Exposures. 7-12 Hz;
119- 144 dB

Pistonphone Coupled to Ear
viaEarmuff.  18 Hz at 135 dB.
Series of 6,  5 Minute
Exposures Rapid in Succession

Pistonphone Coupled to Ear
viaEarmuff.  14 Hz@140dB.
Six Individual Exposures of
5,  10,  15, 20, 25 and 30
Minutes

Ear Only:   Pressure Chamber
  Coupled to Ear via Tuned
  Hose  and Muff
171 dB  (1 - 10 Hz) 26 sec.  Is
168 dB  (7 Hz) 1 min, Is
155 dB  (7 Hz) 5 min, 2s
140 dB  (4,  7, 12 Hz)
  30 min, Is
140 dB  (4,  7, 12 Hz) 5 min.
  8s
135 dB  (.6.  1.6, 2.9 Hz)
  5 min. 12s
126 dB  (.6.  1.6, 2.9 Hz)
  16 min, 11s
  HEARING RESPONSE
Depression of Upper Limits of
Hearing as Measured by Number
of Seconds a Tuning Fork was
Heard - No Conversion to MAP

No Change in Hearing Sensi-
tivity Reported by Subjects;
No TTS Measured About One Hour
Post Exposure

TTS in 3000-6000 Hz Range
For 11 of 19 Subjects (TTS of
10dB -22dB)

Average TTS of 0 - 15 dB After
30 Minute Exposures
Three Experienced Subjects
NOTTS in One; Slight TTS in
One; 20-25dBTTS in One
    RECOVERY
Recovery in Few Hours
Outside of Diesel Room
Recovery Within Hours
Recovery Within 30
Minutes
Recovery With in 30
Minutes
                                               NOTTS
                                               NOTTS
                                               NOTTS
                                               14- 17dBTTS

                                               8dB TTS for 1 Subject

                                               NoTTS

                                               NOTTS
                               Recovery Within 30 min

                               Recovery Within 30 min
                 Whole Body:  All Exposures. 2s:
                   8 min at 8 Hz at SPL'sof        NoTTS
                    120. 126, 132. 138
                   8 min at 1, 2, 4, 6, 8, 10       NoTTS
                   .Hz at 144 dB
                   8 min at 12.  16. 20 Hz at        No TTS
                    135 dB to 142 dB
                                               343

-------
     Injection and retraction of the eardrum membrane may occur at values well below the
 limiting levels shown above. No potential risk to the auditory system is expected to develop
 because of the brief durations of exposure permitted by  the  limits. No change in the
 proposed values is indicated with respect to injection and retraction.
     General face-to-face speech reception in  experimental noise exposures at  the  same
 levels  as the limiting values was  considered acceptable by participants in those studies.
 Headphone reception of speech during laboratory studies has been observed to involve some
 slight  difficulties at levels of about 145 dB. Data  are insufficient at this time to justify
 lowering the proposed  levels at exposures of  7 Hz and below  on  the basis of potential
 interference to speech reception with headphone listening.
     The limiting values proposed  in Table 4 are estimates based upon the various kinds of
 responses made by the auditory system to infrasound. Individuals observing these limita-
 tions may be expected  to display no symptoms of overexposure or abuse of the human
 auditory system.


                            FUTURE  CONSIDERATIONS

     The few studies-reported in the literature  and  a series of studies accomplished in our
 laboratory have been used to formulate tentative exposure criteria for infrasound. In order
 to  more  firmly establish  limiting levels additional information is required  on specific
 responses of the auditory system in infrasound.
     The  parameter of exposure duration appears to be a significant one  requiring better
 definition. The majority of work completed on TTS has been limited to exposures of 8
 minutes and less duration. Essentially no adverse effects have been observed. Conversely a
 small number of exposures of 30 minutes show a reasonably high incidence of TTS. The role
 of duration between  10 minutes and 30 minutes certainly requires better definition for
exposures up to 150 dB.
     Voice communication in infrasound, both speech reception and production, requires
 investigation. Speech reception via headphone listening has already been identified as a
 potential  problem.  Face-to-face communication  in exposures  longer than  one  to  two
 minutes  have not  been  considered.  An important  aspect is the degree to which  the ear
 distorts during intense  infrasound  exposure and its effect on  normal speech  reception
 (300-2000 Hz range).
     Of particular interest to the authors is the  observation that drum membrane retraction
 which  was incurred  during infrasound  exposure  did not disappear following confirmed
inflation of the retracted middle ear system. This implies retraction due to some mechanism
other than negative middle ear pressure, possibly the middle ear  muscle system. Identifica-
 tion of the mechanism sustaining retraction is in  order.
     Additional information is desirable on relative effects of aural vs whole body exposures
 to specific stimuli. Although it is asserted that  middle ear pressure equalization will occur
 for whole body exposures  but not for  aural exposures, this appears unlikely due to the
 relatively  slow action of the eustachian tube in its automatic mode.  It is conceivable that
 individuals who experience  relatively long duration exposures will learn to equalize pressure
 buildup as it occurs in essentially an unconscious manner.

                                        344

-------
                                         Table 4

     LIMITING VALUES FOR INFRASOUND EXPOSURE AS A FUNCTION OF FREQUENCY
                                    AND DURATION
                                     MAXIMUM PERMISSIBLE EXPOSURES
                                                  FREQUENCY in hertz

.5
1
2
4
8
10
20
30
Ihr. 60
120
480
1 day 1440
0.5
169
166
163
160
157
156
153
151
148
*>S-^ «,
|s$
i
166
163
160
157
154
153
150
148
145
t *. •.
•«£ ff r * *f '••'<
'^131'*"
*>! Si
2
163
160
157
154
151
150
147
145
142
'•/' - ?>
* 139
^ 133
V*128./
4
160
157
154
151
148
147
144
142
139
*- 136''
130,
- v?%
8 10
159 156
154 153
151 150
148 147
145 144
144 143
141 140
139 138
136 135
^ 133 132
:" 127, -126,
-, 122 / I?, &!
12 16 20
155 154 153
152 151 150
149 148 147
146 145 144
143 1.42 141
142 141 140
139 138 137
137 136 135
134 133 132
- 131 v 130 * 129
- 125, * }' 124 ^ 123 ,v ,
;;'/i2o;^ ,"ii9 j)iiy ^
  Table of Recommended Maximum Permissible Exposures Based on SPL    = 10 log   t      +10log_f_  + 144
                                                      max      Trtim"         10

   No whole body exposure is recommended over 150 dB for frequencies greater than 0.5 Hi.  Shaded area is an
   extrapolation and should be used with care.
     Improvements in instrumentation and test facilities for evaluating all infrasound effects
on man are also needed, particularly for infrasound exposure signals which typically have
higher frequency energy at levels well above  threshold. It is equally as important to accu-
rately and conveniently  measure hearing at these very low frequencies in order that possible
changes due  to  exposure, which will be missed by testing only audio frequencies, may be
identified.
                                          345

-------
                                    SUMMARY

     Infrasonic energy, both natural and man-made, is present at varying levels in a wide
variety of environments occupied by man. Efforts to describe natural and potential adverse
effects of these exposures are just beginning. The small amount of data available have been
reviewed.
     (1) Nominal infrasound hearing threshold levels are reasonably well defined.
     (2) Hearing function for infrasound appears equivalent to hearing function for audio-
frequency.
     (3) Tentative limiting levels of infrasound exposure are recommended on the basis of
measured  effects on hearing threshold and on other characteristics of the auditory system.
     (4) Improvement in the  tentative  criteria and expansion of its scope,  will require
additional research on factors such as TTS, voice communication and instrumentation.
     This  report has been limited to acoustic energy below 20 Hz; however,  most of the
questions  may relate equally as well to energy from 20 Hz to 50 Hz and even 20 Hz to 100
Hz. Also,  the review has been restricted  to effects only on the auditory system. The overall
technical area of infrasound effects on man is equally or even more in need of information
than is the specific auditory system effects area.
                                  REFERENCES

 1. von Bekesy, Georg, Experiments in Hearing. McGraw-Hill, 1960.
 2. Bryan, M.  E., Annoyance Effects Due to Low Frequency Sound. Proceedings of Fall
    Meeting of British Acoustical Society, 71.109, November 1971.
 3. Evans, Margaret J., Infrasonic Effects on the Human Organs of Equilibrium, Proceed-
    ings of Fall Meeting of the British Acoustical Society, 71.104, November 1971.
 4. Fehr, Vri, Measurements of Infrasound from Artificial and Natural Sources. Journal of
    Geophysical Research, Vol. 12, No. 9, pp. 2403-2417, May 1967.
 5. Green, J. E. and F. Dunn, Correlation of Naturally Occurring Infrasonics and Selected
    Human Behavior. Journal of ASA, 44(5), 1456, 1968.
 6. Jerger, J., B. Alford, A. Coats and B. French. Effects of Very Low Frequency Tones on
    Auditory Thresholds. Journal of Speech and Hearing Research, 9, 150-160, 1966.
 7. Johnson, D. L (Unpublished Data.)
 8. Leventhall, H. G. and R. A. Hood. Instrumentation For Infrasound, Proceedings of Fall
    Meeting of British Acoustical Society, 71.101, November 1971.
 9. Mohr, G. C, J. N. Cole, E. Guild and H. E. von Gierke, Effects of Low Frequency and
    Infrasonic Noise on Man. Aerospace Medicine, 36, 817-824, 1965.
10. Nixon Charles W., Some  Effects of Noise on Man. Proceedings of 1971 Intersociety
    Energy Conversion Engineering Conference, Boston, Mass, August 1971.
11. Nixon, C. W. (Unpublished Data.)

                                       346

-------
12.  Nixon, C. W., H. K. Hille and L. K. Kettler, Attenuation Characteristics of Earmuffs at
    Low Audio and  Infrasonic Frequencies,  Aerospace Medical Research Laboratory
    Technical Report No. 67-27, May 1967.
1-3.  Pickett, Ji  M.,t Low Frequency Noise and Methods  for Calculating Speech Intelligi-
    bility. JASA, 31.9, 1259, Sept 1959.
14.  Stephens, R. W. B., Natural Sources of Low Frequency Sound. Proceedings of Fall
    Meeting of British Acoustical Society, 71.105, November 1971.
15.  Stephens, R. W. B., Very Low Frequency Vibrations and Their Mechanical and Biologi-
    cal Effects,  Seventh International Congress on Acoustics, 26G1, Budapest, 1971.
16.  Tempest, W., Low Frequency Noise  in Road Vehicles. Proceedings of Fall Meeting of
    British Acoustical Society 71.106, November 1971.
17.  Tonndorf, M. D., The Influence of  Service on Submarines On The Auditory Organ.
    (Personal Notes.)
18.  von Gierke, H. E., H. Davis, D. H. Eldredge and J. D. Hardy., Aural Pain Produced by
    Sound, Benox  Report, Contract N6  ori-020, Task Order 44, ONR Project Nr. 144079,
    University of Chicago, December 1953.
19.  Whittle, L.  S.,  The Audibility of Low Frequency Sounds, Proceedings of Spring Meet-
    ing of the British Acoustical Society,  April 1971.
20.  Yeowart, N. S., M. E. Bryan  and W. Tempest, The Monaural M.A.P.  Threshold of
    Hearing at  Frequencies From 1.5 to 100 c/s. J. Sound and Vibration, 6(3), 335-342,
     1967.
21.   Yeowart, N.  S., M. E.  Bryan and  W. Tempest.  Low Frequency  Noise Thresholds.
     Journal of Sound and Vibration, 9(3), 447-453,  1969.
22.   Yeowart, N.  S.,  Low Frequency Threshold Effects. Proceedings of Fall Meeting of
     British Acoustical Society, 71.103, November 1971.
                                        347

-------
      THE EFFECTS OF AIRBORNE ULTRASOUND AND NEAR-ULTRASOUND

                                  W. I. ACTON,
                    Wolfson Unit for Noise and Vibration Control,
                      Institute of Sound and Vibration Research,
                       The University, Southampton, England.

INTRODUCTION

     Ultrasonic devices are now widely used in production industries for a variety of pro-
cesses, including drilling, dicing, soldering, cleaning, welding plastics, emulsification, mixing
liquids, initiating free-radical chemical reactions and so on.  Relatively low ultrasonic fre-
quencies in the range 20 to 40 kHz are generally employed for mechanical reasons, although
small apparatus has been encountered operating at a frequency as low as 16 kHz. Measured
sound pressure levels  at the operator's working position rarely exceed  110 to 120 dB
(ACTON,  1968, GRIGOR'EVA, 1966a, KNIGHT,  1968).
     These sources invariably emit air-borne noise, not only at the operating frequency and
its harmonics, but also at sub-harmonics which may be audible. Furthermore, processes
involving liquids, e.g.'washing, mixing and using a liquid-suspension of abrasive powder, are
accompanied by the phenomenon of "cavitation". This is thought to involve the formation
of bubbles of gas previously held in solution around nuclei such as the abrasive particles in
suspension or dirt on objects being cleaned. The bubbles grow  until they reach  a resonant
size, when they oscillate with an increasing amplitude until they implode. Non-linear radial
and  surface oscillations  of the  gas-filled bubbles may be responsible for more tonal noise,
and  the violent collapse of cavities is responsible for the generation of high levels of random
noise at frequencies of approximately 3 kHz upwards (WEBSTER, 1963).
     Ultrasonic frequencies used in medicine for cell destruction are generally in the range 1
to 3 MHz and for  diagnosis in the range  1 to 20 MHz. Diagnostic exposures were not
considered likely to be  potentially harmful by HILL (1970). As these frequencies do not
appear to have found widespread industrial application yet, they will not be  considered
further.


HISTORICAL REVIEW

     When jet aircraft  were introduced, the term "ultrasonic  sickness" was coined (DAVIS,
1948, PARRACK,  1952)  to  cover a complex  of  symptoms which included excessive
fatigue,  headache, nausea, vomiting, etc., exhibited by  personnel working in their vicinity.
ALLEN,  FRINGS and RUDNICK (1948) observed a loss of the sense of equilibrium  or
slight dizziness on exposure to  intense (160 to  165 dB) high  frequency, audible sound, and
unsteadiness and dizziness have been reported in personnel exposed without ear defenders
and  at close range to the noise  from the air intake of jet engines (DICKSON and WATSON,
1949, DICKSON and CHADWICK, 1951). The latter authors suggest that this  might be due
to vestibular  disturbances  caused  by intense acoustic stimulation. In any case published
analyses of jet engine noise show that radiated airborne ultrasound is not present at signifi-

                                       349

-------
 cant intensities. As ultrasonic frequencies are rapidly absorbed by air, intense ultrasound
 would only be encountered in  regions where approach was  normally barred by safety
 considerations (DICKSON, 1953, GUIGNARD, 1965). Finally, PARRACK (1966) stated
 that "ultrasonic sickness" was "largely psychosomatic in origin", although, of course, the
 other effects had been real enough.
     Then  followed a period when  the possibility of effects  from exposure to airborne
 ultrasound was dismissed. DAVIS, PARRACK and ELDREDGE (1949) stated that there
 was no evidence that airborne ultrasonics themselves constituted a hazard to the hearing,
 and, in general, high-intensity audible noise was potentially more hazardous. PARRACK
 (1952) concluded that there was no hazard from laboratory sources of airborne frequencies.
     A note of caution was introduced in the mid-1950's. CRAWFORD (1955) reported
 that some laboratory workers had suffered unusual fatigue,  loss of equilibrium, nausea and
 headaches which persisted after the exposure had ceased, and "some loss of hearing in the
 upper audible frequencies", although this was not substantiated by audiometry and was
 probably based on purely subjective observations. Systematic  research into the biological
 effects of ultrasound  was started in Russia in the late 1950's (GORSLIKOV, GORBUNOV
 and ANTROPOV, 1965),  but some of the translations and reviews available in the West
 should be viewed  critically,  as effects observed with liquid- or solid-coupling to the ultra-
 sonic source have apparently been attributed to airborne ultrasound.
     There have been a number of audiometric temporary threshold shift investigations
 involving  laboratory  (DOBROSERDOV,  1967),  PARRACK,  1966, SMITH,  1967) and
 industrial (ACTON and CARSON, 1967) exposure, and at least one retrospective permanent
 threshold shift investigation (KNIGHT,  1968). Subjective effects have been correlated with
 measured exposure levels by SKILLERN (1965) and ACTON and CARSON (1967). Finally,
 a number of exposure criteria for the prevention of both  auditory and subjective effects
 have been proposed, and these do not differ widely (ACTON, 1968, GRIGOR'EVA, 1966a,
 1966b, GORSLIKOV et al, 1965, ROSCIN et al, 1967).

PHYSIOLOGICAL EFFECTS

     One difficulty in reviewing the physiological effects is  to be certain that the exposure
was to airborne ultrasound. Another difficulty is that many,  often bizarre, effects have been
reported without  the exposure  level being quantified. Consequently, this section of the
review has been limited to references which specifically stated exposure conditions, and
these have been summarized in Fig. 1.
     In the case of airborne ultrasound, the acoustic mismatch between the air and tissue
leads to a  very poor transfer of energy. The effects on small fur-covered afnimals are more
dramatic because the fur acts as an impedance-matching device, they have a greater surface
area to mass  ratio, and they have a much lower total body  mass to  dissipate the  heat
generated than man.  Furthermore, the lower ultrasonic frequencies may well be audible to
these animals,  and the exposures have  been to high sound pressure levels. Therefore, the
effects on small laboratory animals cannot be extrapolated directly to the human species.
     Mild biological changes  have been observed in rats and rabbits as a result of prolonged
exposure to sound pressure levels in the range 95 to 130 dB at frequencies of from 10 to 54

                                       350

-------
  HUMAN

 Death (calculated)
Loss of equilibrium

Dizziness  —^—.
Mild warming
     (body surface)
Mild heating     _
     (skin clefts)
No physiological
     changes
Industrial exposure
no hearing loss
            SMALL  ANIMALS
                       180
160
..I-
                        120
                        100
     dB
           Death  (rabbits)
           Body  temperature rise
           (hairless mice)

           Death (mice, rats,
           guinea-pigs)


           Body  temperature rise
           (haired mice)
        — Mild biological changes
           (rats, rabbits)
             Figure 1 Physiological effects of ultrasound

                          351

-------
kHz (ANTHONY and ACKERMAN, 1955, BUGARD,  1960, GORSLIKOV et al, 1965,
GORSKOV et al, 1964, and others). Where the sound was audible to the animals, these
represented relatively high sensation levels and the biological changes were typical of any
stress condition in many cases. Actual body heating in mice was not measured until a level
of 144 dB at 18 to 20 kHz was reached. With hairless mice, the corresponding level was 155
dB,  indicating  the role  of the fur  in absorbing energy (BANNER,  ACKERMAN and
FRINGS, 1954). The deaths of mice and guinea pigs as a result of exposure to a level of 150
to 155 dB at 30 kHz (DICKSON, 1953), of rats and guinea pigs to 144 to 157 dB at 1 to
18.5 kHz (ELDREDGE and PARRACK, 1948), and of rabbits to 160 to 165 dB at 22.5 and
25 kHz (BUGARD, 1960, ROMANI and BUGARD, 1960) have been reported also.
     In man, there are reports of both a drop  (ASBEL, 1965) and an increase in the blood
sugar level (BYALKO et al, 1963) and electrolyte balance  changes in the nervous tissues
(ANGELUSCHEFF, 1957) as a result  of exposure to ultrasound, although neither sound
levels or frequencies were reported. However, BATOLSKA et al (1969) rightly pointed out
that many  of the effects attributed to ultrasound are also typical of exposure to other
physical and toxic conditions at their places of work, and conclusions should not be drawn
without comparison of results with a control group. GRIGOR'EVA (1966a) failed to find
any significant physiological changes as a result of one hour exposure to 110 to 115 dB at
20 kHz in a comparison with control subjects.
     Slight heating of skin clefts was observed by PARRACK  and  PERRET  (1962) as a
result of exposure to ultrasound at levels of 140 to 150 dB. At 159 dB there may be a mild
warming of  the body surface (PARRACK, 1951). Loss of equilibrium and dizziness oc-
curred at levels of 160 to 165 dB at 20 kHz (ALLEN et al, 1948). The calculated lethal dose
for man is at least 180 dB (PARRACK, 1966).
AUDITORY EFFECTS

    The ear constitutes an efficient impedance matching device for high frequency airborne
sound, and it seems likely  that any hazard from airborne ultrasound will manifest itself
initially as a hearing loss or an associated psychological effect.
    An investigation to determine if the noise from  industrial ultrasonic devices caused
auditory effects was described by ACTON and CARSON (1967). The hearing threshold
levels  of 16 subjects (31 ears) were measured in the frequency range 2 to 12 kHz before and
after exposure to the noise over a working day. No significant temporary threshold shifts
were detected (Fig. 2). On the assumption that if a noise exposure is not severe enough to
cause  a temporary threshold shift, then it cannot produce permanent damage, it was con-
cluded that hearing damage due to exposure to the  noise from industrial ultrasonic devices is
unlikely. A parallel retrospective investigation by KNIGHT (1968) on a group of 18 young
normal subjects using ultrasonic devices showed a  median hearing level within 5 dB of that
of a matched control group of hospital staff except at 4 kHz where the departure was 7 dB
(Figure 3). It was concluded that it would have  been difficult to attribute this exposure
solely to ultrasonic radiation. In addition, no abnormal vestibular function  test (caloric test)
results were found.

                                      352

-------
        -2

        -1

          0
        dB
          1

          2

          3

          4

          5
                     I	1	1	
                     248

                                    Frequency  KHz

                 Figure 2. Temporary threshold shift due to industrial exposure
12
    Some temporary threshold shifts have been reported as a result of exposures to ultra-
sound under laboratory conditions, and the exposure conditions have  been summarized in
Figure 4. (PARRACK, 1966, DOBROSERDOV, 1967, SMITH,  1967).
    The exposures used  by Dobroserdov were at high audible frequencies, and those by
Smith contained high-audible-frequency noise. The results due to Parrack are interesting in
that he exposed subjects to discrete frequencies mainly in the  ultrasonic region, and meas-
ured temporary  threshold  shifts  at subharmonics of one half of the fundamental and
occasionally at lower subharmonic frequencies after 5-minute  exposures to discrete fre-
quencies in the range 17 to 37 kHz  at levels  of 148 to 154  dB. Subharmonic distortion
products have  been reported  in the cochlear-microphonic  potentials  of guinea pigs
(DALLOS and LINNEL, 1966a) and have also been monitored in the sound field in front of
the eardrum using a probe-tube microphone (DALLOS and LINNEL, 1966b). They were
believed to result from nonlinear amplitude distortion of the eardrum, and they appeared at
a magnitude of the same order as that of the  fundamental. This observation may help to
explain Parrack's findings.
    Many sources of ultrasound, and particularly processes involving cavitation, produce
substantial levels of noise in the high  audible range. Reported auditory effects can often be
explained  in  terms of the  audible noise only, and  these references have been omitted
deliberately.
                                        353

-------
   2

dB3


   4


   5

   6

   7
              I
            0.25
0.5
                                     Frequency  KHz
        Figure 3 Permanent threshold shift due to industrial exposure relative to control group
SUBJECTIVE EFFECTS

     It has been mentioned already that early laboratory workers reported suffering unusual
fatigue, loss of equilibrium, nausea, and headaches which persisted after the stimulation had
ceased,  as a result of their exposure  to airborne ultrasound. Complaints of fatigue, head-
aches,  nausea and tinnitus are frequently made by the operators of industrial ultrasonic
devices, but their exposure does not seem to be sufficiently intense to cause loss of equilib-
rium. Observers entering the sound field for shorter periods often experience an unpleasant
sensation of "fullness" or pressure in the ears.
     It has been  shown that  these subjective effects are due to the high levels of high-
frequency audible noise usually produced as a by-product of industrial ultrasonic processes,
and especially those involving cavitation (ACTON and CARSON, 1967). SKILLERN (1965)
attempted to correlate  these effects with frequency, and erroneously concluded that the ear
was sensitive to a narrow band of frequencies centered on 25 kHz. Examination of his data
shows that all frequency spectra he quoted as producing effects contained high levels of high
frequency audible noise as well as an ultrasonic component at about 25 kHz.
                                        354

-------
   160 _
   140 -
 PQ
 •o
 > 120
 V
 3
 o>
 •a
 c
 o
 CO
100 -i
    80  -
                                                 Parrack  1966
                                                     Y/// ////,
                                               Uobroserdov 1967
                                  Smith 1967
                                       I
                                      10

                              Stimulus frequency KHz
                                                     20
30
50
              Figure 4 Laboratory exposures producing temporary threshold shifts
    The effects  are often only reported by young females in exposed populations, but
ACTON and CARSON (1967) showed that this was a function of auditory threshold. Males
employed  industrially  often have high-frequency  hearing losses, probably due to noise-
induced hearing loss and presbyacusis, or "sociacusis". The effects were not considered to
be psychosomatic in origin or due to hysteria.

EXPOSURE CRITERIA

    Unlike audible noise, the area of ultrasonic noise is characterized by a lack of published
exposure criteria. Early Russian workers (GORSLIKOV et al, 1965, ROSCIN et al, 1967)
proposed an overall limit to ultrasonic exposure of 100 dB, regardless of frequency. This
was probably a cautious move made in the absence of data to quantify some of the physio-
logical  effects  being reported at that time. Only two frequency-dependent criteria are
known. The first was due to GRIGOR'EVA (1966a) and is shown in Figure 5. The levels at
and below a frequency  of 16  kHz are based  on the  experimental results of temporary
threshold shift measurements, and at 20 kHz and above as a result of experiments to detect
physiological changes. However, later in the same year,  GRIGOR'EVA (1966b) proposed a
                                       355

-------
level of 110 dB in the 20 to 100 kHz frequency range, and mentioned that this had been
incorporated into a (USSR) Ministry of Health memorandum.
     The criterion due to ACTON (1968), also shown in  Figure 5, was based on experi-
mental evidence  to prevent both auditory and subjective effects  in the greater part of
population exposed over a working day. The author did not feel justified in extrapolating
this  criterion  beyond the one-third-octave band  centered on  31.5 kHz on the  basis of
available experimental evidence.

CONCLUSION

     Many of the reports of effects  due to exposure to ultrasound must be  regarded as
anecdotal rather  than factual. Further confusion has undoubtedly arisen because results
obtained with small, fur-covered  animals have been transposed directly to man, and because
airborne  exposure has not been sufficiently  differentiated from liquid- or solid-coupled
exposure. Nevertheless,  there is  ample evidence to show that  exposure  to high levels of
ultrasound can have some effects  on man.
     In industry,  the  exposure to the high levels  of high-frequency audible sound which
accompanies many ultrasonic processes is more likely  to prove troublesome than the ultra-
   120 -
«  110 _
   100 -
M
2   90
1
.a
v
o
o
    70  -1
                                  Grigor'eva 1966a
                                                              Grigor'eva 1966b
                       Acton 1968
        I
        3
         I
       10

Frequency  KHz
20
 f

30
50
                         Figure 5 Exposure criteria for ultrasound

                                        356

-------
sonic frequencies themselves. Subjective effects include headaches, nausea, tinnitus, possibly
fatigue and so on, and some temporary threshold shifts in hearing have been observed as a
result  of experimental laboratory exposures to ultrasound. Two similar exposure criteria
have been published, both in the 1960's.

ACKNOWLEDGEMENT

    The author wishes to thank Mr. A. E. Crawford for bringing some of the references to
his notice, and Dr. R. R. A. Coles for reading and commenting on the draft.
                                 REFERENCES

ACTON, W. I. (1968) A criterion for the prediction of auditory and subjective effects due
     to  air-borne noise from ultrasonic sources, Annals of Occupational Hygiene, 11, 227.
ACTON, W. I. and CARSON, M. B. (1967) Auditory and subjective effects of airborne noise
     from industrial ultrasonic sources, British Journal of Industrial Medicine, 24, 297.
ALLEN, C. H.,  FRINGS, H. and RUDNICK, I. (1948) Some biological effects of intense
     high frequency airborne sound, Journal of the Acoustical Society of America, 20, 62.
ANGELUSCHEFF, Z. D. (1957) Ultrasonics, resonance and  deafness,  Revue de Laryn-
     gologie, d'otologie  et de Rhinologie, July-August 1957, abstract in CORDELL (1969).
ANTHONY, A. and ACKERMAN, E. (1955) Effects of noise on the blood eosinophil levels
     and adrenals of mice, Journal of the  Acoustical Society of America, 27, 1144.
ASBEL, Z. Z. (1965) The effect of ultrasound and high-frequency noise on blood sugar
     level, Gigena truda i professional  'Nye  Zabolevanija (Moscow),  9, 29.,  abstract in
     Occupational Safety and Health Abstracts 4, 104, (1966).
BATOLSKA, A. et al (1969) Occupational disorders due to ultrasound, Work of the Scien-
     tific Research Institute of Labour  Protection and Occupational Diseases (Sofia), No.
     19, pp 63-69, abstract in Noise and Vibration Bulletin March 1971, p. 97.
BUGARD, P. (1960) The  effects of noise on the organism, the importance of non specific
     effects Revue des  Corps de  Sante des Armees (Paris),  1, 58,  abstract in CORDELL
     (1968).
BYALKO, N. et al (1963) Certain biochemical abnormalities in workers exposed to high
     frequency noise, Doklady Vsesoyvznogo Nauchno - Prakticheskogo Soveshchaniya Po
     Izucheniyu Deistriya Shuma  Na Organizm (Moscow), pp. 87-89, abstract No. 2659 in
     Excerpta Medica 17, 570 (1964).
CORDELL, J. (1968) Physiological effects of airborne ultrasound; a bibligraphy with ab-
     stracts, Labrary  Report No.  4, Commonwealth Acoustic  Laboratories, Sydney, Aus-
     tralia.
CRAWFORD, A. E. (1955) Ultrasonic engineering, Butterworth, London.
DALLOS,  P.  J. and  LINNEL,  C. O.  (1966a)  Subharmonic  components in  cochlear-
     microphonic potentials, Journal of the Acoustical Society of America, 40, 4.

                                       357

-------
 DALLOS, P. J. and LINNEL, C. O. (19665) Even-order subharmonics in the peripheral
     auditory system, Journal of the Acoustical Society of America, 40, 561.
 BANNER, P. A., ACKERMAN, E. and FRINGS, H.  W. (1954) Heating of haired and
     hairless mice in high intensity sound fields from 6 to 22kc.
 DAVIS, H. (1948) Biological and psychological effects of ultrasonics, Journal of the Acous-
     tical Society of America, 20, 605.
 DAVIS, H., PARRACK, H. O. and ELDREDGE, D. H. (1949) Hazards of intense sound and
     ultrasound, Annals of Otology (St. Louis), 58, 732.
 DICKSON, E. D. D. (1953) Some  effects of intense  sound and ultrasound on the ear,
     Proceedings of the Royal Society of Medicine, 46, 139.
 DICKSON, E. D. D. and CHADWICK, D. L. (1951) Observations on disturbances of equilib-
     rium and other symptons induced by jet engine noise, Journal of Laryngology and
     Otology, 65, 154.
 DICKSON, E.  D.  D. and  WATSON, N. P. (1949) A  clinical survey into the effects of
     turbo-jet engine noise on service personnel, Journal of Laryngology and Otology, 63,
     276.
 DOBROSERDOV, V. K. (1967) The effect of low  frequency ultrasound and high frequency
     sound on exposed workers, Gigiena i sanitarija (Moscow), 32, 17, abstract in Occupa-
     tional Safety and Health Abstracts, 5,658 (1967)
 ELDREDGE, D. H., and PARRACK, H. O. (1948)  Biological effects of intense sound, Paper
     presented at 36th meeting of the Acoustical Society of America, Cleveland, Ohio,
     November, 1948, abstract in Journal of Acoustical Society of America, 21,55.
 GORSKOV, S. I. e't al (1964) The biological effects of ultrasonics in relation to their use in
     industry, Gigiena i sanitarija (Moscow), 29,  37, abstract in Occupational Safety and
     Health Abstracts, 3,39(1965).
 GORSJJKOV, S. L, GORBUNOV, O. N. and ANTROPOV, S. A. (1965) Biological effects
     of ultrasound, Medicina, Moscow, review in Ultrasonics, 4, 211 (1966).
 GRIGOR'EVA, V. M. (1966a) Effect of ultrasonic vibrations on personnel working with
     ultrasonic equipment, Soviet Physics - Acoustics 11,426.
GRIGOR'EVA, V.  M.  (1966b) Ultrasound  and  the question  of occupational hazards,
     Maschinstreochiya No. 8, p. 32, abstract hi Ultrasonics, 4, 214.
GUIGNARD, J. C. (1965) Noise, Chapter 30 in  GILLIES, J. A. (Editor) A textbook of
     aviation physiology, Pergamon, Oxford.
HILL, C. R. (1970) Paper presented to conference on "Clinical aspects of ultrasonic diag-
     nosis", London, 10 December 1970, review in Ultrasonics, 9, 172 (1971).
KNIGHT, J. Jl (1968) Effects of airborne ultrasound on man, Ultrasonics, 6, 39.
PARRACK, H. O. (1951) Physiological and psychological effects of noise, Proceedings 2nd
     Annual National Noise  Abatement symposium, Chicago, Illinois, 1951, abstract in
     CORDELL(1968).
PARRACK, H. O. (1952) Ultrasound and industrial medicine, Industrial Medicine and Sur-
     gery, 21, 156.
PARRACK, H. O. (1966)  Effect of airborne ultrasound on humans, International Audio-
     logy, 5, 294.
                                      358

-------
PARRACK, H. O. and FERRET (1962) Effect on man of low frequency ultrasonics pro-
    duced by aircraft, Report presented at meeting of group of experts on struggle against
    noise caused by aircraft, Organisation de  Co-operation et de Development Econom-
    iques, Paris, abstract in CORDELL, (1968)
ROMANI, J. D. and BUGARD, P. (1957) Further experiments on the effect of noise on the
    endocrine system, Acustica, 7, 91.
ROSCIN, I. V. et al (1967) Occupational health hazards of technical applications of ultra-
    sound, Gigiena truda i professional 'Nye Zabolevanija (Moscow), 11, 5., abstract in
    occupational Safety  and Health Abstracts, 5, 657 (1967).
SKILLERN, C. P. (1965) Human  response to  measured sound pressure levels from ultra-
    sonic devices, American Industrial Hygiene  Association Journal, 26, 132.
SMITH, P. E. (1967) Temporary threshold shift  produced by exposure to high-frequency
    noise, American Industrial Hygiene Association Journal, 28, 447.
WEBSTER, E. (1963) Cavitation, Ultrasonics, 1, 39.
                                       359

-------
         SESSION 4 B





PERFORMANCE AND BEHAVIOR




  Chairman: D. E. Broadbent, UK
             361

-------
                 PSYCHOLOGICAL CONSEQUENCES OF EXPOSURE
                       TO NOISE, FACTS AND EXPLANATIONS

                                     Edith Gufian
                                 Institute of Psychology
                                  Bucharest, Romania

     I hate beginning this paper with a pessimistic statement, but I must say that systematic
investigations carried out on the psychological effects of noise since  1950*) are in a rather
controversial state and therefore no firm conclusions can be drawn.
     There is, however, a sound reason for the rather ambiguous nature of the psychological
results, and this will be evident from Figure  1. This reason is the extraordinary complexity
of the various factors which intervene in the effect of noise on man.
     In considering the effects of noise on behavior, one usually assumes  that noise is a
more or less undifferentiated  variable, defined mainly by its intensity. However, the  facts
are that the different possible  interactions between the several intrinsic parameters of noise
x)This paper deals primarily with experimental results since 1960, as there are excellent reviews of former data by Krytei
(1950) and Broadbent (1957).
                   VAIUAIU>:M
•
IS 1 KNSI 1 V <
- I'ltK- I K.SCY^r-
1
i
DI u.uiiiN

- I'YPK-^S^^^^
I-IIMPI.KMTY
XAITKK ^— —
^

; CONSTANT
VAH1AIU.K
- CONST AN f
^ VAIUAIILE


- CON riNl'OTS
- INTERMITTENT <

-TIIAKFIC
•MMilSITUAL
" AIIICKAFT ETC.
                                          IVfKHMEUAItY VAIUAHLES
                                          llnlra-h.d tnlvriiMlividual iliirc'risu-L's>
                                                                  DEPENDENT VAUIAHI.KK
                          PACED vs. I SPACED
                               :s HATE
                          KA.VDOM vsHEGl-LAn
                               :VEUIIAL
                               NON-VERBAL
                      visi AI. t:rc



M



A









0
H
A
X
]
S
M
C

P
S
Y
C
II
O
L
O
G
I
C
A'
L
- UENERAL HEALTH STATE 1
- AUDITORY SYSTEM ,
- OTHER SENSORY SYSTEMS (
y SELECTIVE 1
- A HOUS A U LEVEL^pjppj.gj, ^
- ACCUMULATED PHYSIO LOGICAL
EFFECTS (FATIGUE ETC. I ;

- SENSORY INTERACTION
- AUDITORY PEHCEPTION
- AllOl'SAL LEVEL^ TASK '
'DURING
- PltEVIOUS EXPERIENCE !
. NOISE
WITH < TASK '
- MOTIVATION >
- ATTITUDE *
1
S
-
ACCU.ACY c<
C
3 - SPEK1I
I
I - TOY
! -DISCIIIMlSAilliN
: -OV'I'Pl'T


^
* - IMHUW Al.
)
'
I
C - OIUU'PKKA



    Figure 1. A schematic representation of the ways and modalities through which noise acts on man

                                          363

-------
 (intensity, frequency and complexity) and  its temporal structure give rise to such a great
 diversity of stimulation, that it practically could never be tested in the laboratory.
     The  picture is further complicated  by the other independent variable-the routine
 activity in real-life situations or the laboratory task, to which we shall restrict ourselves here.
 The multiplicity of tasks resulting from the possible combinations of the different elements
 listed above  is in fact infinite and new associations are always possible. If we turn now to
 man himself, we encounter a great number of variables which have to be taken into account
 if noise influence is to be understood. First  there is a biological level which is of interest to
 us insofar as it constitutes  the basis of behavior and yields indications in case the overt
 behavior cannot be specifically interpreted. Secondly, there is the psychological level which
 accounts for a great part of the variance in noise investigations, in particular the role played
 above and beyond the actual stimulation, by the familiarity with both  the noise and the
 type of activity, motivation, attitude etc. Each of these variables can be studied separately
 and yields more or less specific results. However, no one has ever ventured  to look into all of
 them simultaneously and  draw a complete picture of their influence, even though each of
 them mediates in  a particular way the  influence of noise on the final output. The  strong
 interrelationships among these variables is a widely known fact although in empirical studies
 it is sometimes overlooked. Finally, we are faced with  the dependent variables, the final
 concrete output of the various above-mentioned factors-the performance, and the subjec-
 tive annoyance.
     Several  points need to be clarified here.  First of all, performance is a rather vague term;
 it refers to many measures, each having a different meaning as a function of the type of
 task, although they all may be divided into two main classes: those measuring accuracy and
 those measuring speed of performance. What I would like to point out is that the diversity
 of tasks imposes a diversity of  measures which are comparable only  in a very general
 manner.
     Maybe  the inclusion  of annoyance among the  dependent variables seems somewhat
 peculiar, but  my contention is, however—and I shall try to substantiate it—that regardless of
 the final effect of noise on performance (positive, neutral or negative) a certain annoyance is
 always present. Of course, annoyance has a feedback on the psychophysiological  state and
 thus on performance, just as performance level cannot fail to exert a certain influence on
 annoyance level.
     How  then does noise affect the dependent variables? I suggest that there are two main
 ways, depending on man's activity. In case the individual is engaged in an activity, the noise
 acts concurrently with the task  and its effect on performance depends on different task
 characteristics as well as on those of the intermediate variables. On the other hand, if man is
 not actively  engaged  in a professional  task, but is either resting or performing a simple
 relaxing home activity the noise  affects him more directly and its main consequence is a
 psychic disturbance, the subjective annoyance.
     In view of these complex  interrelationships,  the  divergent results in noise  effects
investigations could obviously be  anticipated. In fact, as shown in Figure 2, this is precisely
what happened; performance has been either  impaired, unaffected, or in some cases even
 improved  under noise conditions. The fact  that performance in noise showed no effects,
 detrimental effects or improvements  at the same sound pressure levels (between 50 &  110

                                        364

-------
                          NEGATIVE EFFECTS


                          Jurist, 1-J..9     Mmitoriitn
                          Turn-Pri- and
                          U'isncr, IS62     ViKilam-v


                          Ml., J'JC-I      liatii*
                          Mark and &)., IBtl PsycSomuli.r
                          llritiidhutt and   Visual mu-
                          Crufjory, 19(t.~>   nitarmp
                          Hclrmann and Osier-
T:-VS HMJ
.U Vs 75 uh
                          WilttTshctm a
                                , b.
                                    llwntH-v
Mrnnomann, jyos
WiU-cW, IbTO
NcMtal, 19TI
tu-nlr-sch, 1S71
Uyliuidvr and at..
19T2
POSITIVE EFFECTS
McCralh, 1963
Kirk and Hecot,
19631
Psychamotur
Simple KT
Simple AT
Psyehomotnr
Trackinfi
Attention

VfciUac*
Detection
^:,-io(i
30 -W
yo
7(t-9h

0.4788

72
C4
                          Waiktns. 1964    Detection
                          Samuel, 1964    AtLcnUon
                          Wvinstcin and    »IlK,AATE
                          U^cKvnziv 19M  H'nJpilnt.
                          O*Hallc> Mid Poplvw-
                          hky, 1971     Stroop
                          K'ttima, 1'jTl    Attention

                          SEl'THAL EFFECTS


                          .k-risiwi, l(J-'.7    Vifilant-c
                          Licni-it InJ
                          -l.uifttii, t'JM    AtU-nLii*i
    MU-THItl.


   industrial
     varied
     steady
     Inter rntt.
  «»t. Intermit.
 v» 110 «hitv
     «hitc
     airi-rad
                                                              Arouxal
                                                              Dislrat-tkm
CO
En

pcrfo.

TOT
CD
                   Arousal

                   ArouMl
CitllOn, !!*(.!.

Mi*Cujint IW.L

I.uL:i!« an«lal.l:P7ll
Itl:ii-I.uc11 .ind IV-lt,
1!*71
Sli'%i'Hf*. 1-I7J

Vinilani-t- Tlt-'Kl

Vi|>ilm^i- *
-------
 suggests that the major cause of the disruption is the muscular reflex associated with startle
 (May and Rice, 1971; Thackray and Touchstone, 1970). Tasks involving complex perceptual
 and/or  cognitive processes may be impaired for longer periods-up to 30 sec. (Woodhead,
 1959, 1964)—the detrimental effects of noise resulting from an interference with either
 information processing (Broadbent, 1971) or information reception (Woodhead, 1964).
     While all these  studies show some degree of impairment  associated  with  impulsive
 noise, investigations employing real or simulated sonic booms show divergent results ranging
 from performance decrement (Lukasetal., 1970; Rylander et al., 1972; Woodhead, 1969), to
 generally  non-significant effects (Lukas et aL, 1971; Harris, 1970) to performance improve-
 ment (Thackray et al., 1972).
     Concerning intermittent noise of lower intensity, Teichner et al. (1963) predicted that
 the same  noise  stimulus could either facilitate or disrupt performance. The  predominant
 effect at  any point in time would depend upon duty cycle, the ratio of noise on-time to
 periods of silence between noise presentations. Significant effects  were observed on speed of
 visual  target  detections  for  all  on-off ratios  except the 70% when compared  to  a
 control, no-noise  condition. Warner (1969), testing different noise intensities (80, 90 and
 100 dB) in an attention-demanding task at the 70% duty cycle, found no effect of intensity
 level  on detection time,  and fewer errors as a function  of noise intensity.  In  a  further
 experiment (Warner and Heimstra,  1971), it was found that the particular effect attributable
 to varying ambient noise ratios (0, 30, 70, 100%) on target-detection  time  is dependent
 upon the degree of difficulty of the inspection task.
     Here we  run  up  against a vital question, the fact now I think well-established that the
 deleterious effect  of  noise on performance increases as a function of increasing  task com-
 plexity. Task difficulty can be manipulated in different ways: (1)  One way is by multiplying
 the stimulation sources as was shown by Broadbent (1954) in the difficult 20-dials test and
 in the detection of the easily seen 20-light task in 100 dB noise, and by Jerison (1957,
 1963) in the  three-clock task vs. the one-clock task. (2) Another way is by  changing the
 intrinsic difficulty of  the task. Thus Hsia (1968) using six difficulty levels finds that a 65-dB
 noise exerts a detrimental effect on information processing only when the stimulus material
 is difficult,  while  Houston  (1968), manipulating  two difficulty levels, found that noise
 facilitates  performance in the high-difficulty condition for the more difficult  task but not
 for the easy one. Gulian (1972), comparing percentage of  errors  and reaction time (RT) in
 three  vigilance tasks  of different  difficulty level, under three noise conditions, found  a
 clear-cut interaction between noise and difficulty level, particularly with respect to the very
difficult one (Figure 3) (3) Another source  of difficulty arises from the temporal structure
of the task. It was found that noise acts adversely on performance as a function  of high
variability of  intersignal interval  (Dardano,  1962), or high signal  rate (Broadbent and
Gregory, 1965)  but that  lower signal rates  either improve or  do not change  detection
performance despite high noise level (95 dB) (Davies and Hockey,  1966).
     There is finally a fourth method of manipulating difficulty which seems to detect best
 the effect of noise  the method of simultaneous tasks.
    Indeed, results in an experiment reported by Boggs  and Simon  (1968)  with a two-
complepdty-level four-choice RT task and a secondary auditory monitoring task showed  that
noise produced a significantly greater increase in secondary task errors when the secondary

                                         366

-------
                    100
                                   TOdB  BOdB
                                   Cont  Intermit.
Quiet  7OdB  9OdB
       Cont  Intermit
Figure 3.  Effects of Noise on Performance level as a function of task difficulty
    A — an auditory vigilance task • the most difficult one. S discriminated a pure tone (400 Hz) among 4
different pure tones
    B — a word recognition task of medium difficulty. S reacted to one trigram (doc) among 4 different
Digrams
    C -  an easy visual and auditory reaction time task. S reacted to a visual and an auditory stimulus and
discarded another visual and auditory stimulus.
     The noise was  the same in all  3 experiments:  70 dB white continuous noise  and 90 dB varied
intermittent noise.
task was paired with the complex than with the simple primary  task. Finkclman and Glass
(J970) showed  that while performance on  the primary task is unaffected by predictable vs.
unpredictable noise, only the unpredictable  noise resulted in performance degradation in the
subsidiary  task. Hockey (1°-70 a. b) using a primary tracking and a secondary multisource
monitoring task, showed that the tracking  task improved in noise (100 dB vs. 70 dB) as did
the  location  of centrally located  signals in the monitoring task, but that detection of
peripheral  signals is impaired.
     These and other similar experimental results arc explained by Broadbent by his arousal-
filtering  hypothesis. Noise increases arousal and it  affects  perception only when there is a
competing stimulus from which the reaction stimulus has to be discriminated  that is, when
a filtering  process has  to take place. Arousal affects filtering in the sense  that the aroused
system devotes a higher portion of its time to  the intake of information from dominant
sources and less from  relatively  minor ones. Of course  the higher  the arousal level, the more
                                           367

-------
 adverse the effects of noise, whereas at moderate levels of arousal, performance maintains its
 efficiency.
      That effects of a noise cannot be accounted for only by its physical characteristics or
 by those of the task is emphasized in several experiments which evidence modifications in
 performance, arousal and annoyance through manipulation of the relevance of the stressor.
 In two consecutive experiments, Glass et al. (1969, 1971) showed that adverse postadaptive
 effects following loud unpredictable noise (110 dB) were substantially reduced if the subject
 believed he had control (vs. no control) over the termination of the noise.
      Approaching the problem from another  point of view, Munz et al. (1971) tested
 subjects who had either high or  low involvement in  a pursuit rotor tracking task while
 simultaneously exposed to a task-related or task-unrelated 80 dB noise. They found no
 effects of noise on performance, but highly-motivated  Ss reported experiencing greater dis-
 comfort under task-unrelated noise as compared to the other condition of the control one,
 and their statement was supported by their post-experimental ranking of working condition
 performances.
      Finally,  the meaning of noise is an important variable-and  it has been shown that
 speech impairs performance more than a neutral noise. That is why Wisner (1971) concludes
 that  laboratory experiments using meaningless  noises  cannot explain performance of sub-
 jects working in a place where there is considerable conversation.
      This evidence fully justifies Kryter's suggestion that in one way or another the task and
 its completion are dependent upon the presence of noise  and all the observed effects of
 noise are due to psychological factors related to stimulus and response contingencies associ-
 ated  with the noise by individuals. Individual differences in  reaction to noise arise because
 of inappropriately interpreted stimulus  and response contingencies, but these tend to be
 eliminated with learning and experience.
     To summarize: (1) the level of noise needed to show adverse effects is high-95 dB-and
 high  frequencies  seem to be more noxious  than low ones; (2) the harmful effect of noise
 seems to  be on accuracy rather than on speed; (3) monitoring that requires time sharing
 among several potential signal sources is affected by high levels of noise, as is (4) monitoring
 that requires the operator to translate delayed data.
     If noise and task characteristics only partially explain the various shifts in performance
 efficiency, it follows that their causes should be sought elsewhere, as well. It is suggested
 that level of arousal, peculiarities in auditory perception, and annoyance produced by noise
 are the main variables involved. Individual differences in all  these variables, correlated with
 personality measures, introduce an important cause of variation.
     What evidence is  there to support these assumptions? We refer first to the hearing
mechanism in  order to emphasize that if differences in auditory perception and processing
of sounds are established, they probably are at the basis of differences in noise suscepti-
bility, and ultimately would have consequences on performance.
     Several investigators have linked  personality with two  aspects of the auditory thresh-
 old: the  absolute sensitivity and  the variability of the measures  obtained (see Stephens,
 1972, for a review).
     Eysenck  (1970)  proposed that sensory thresholds, tolerance levels,  and preference
levels for  sensory stimulation will  differ in  introverts and extroverts. Indeed some studies

                                        368

-------
showed that introverts might have more sensitive auditory thresholds (Smith  1968), less
variability of the audiometric threshold (Reed, 1961, Reed and Franci 1962; Farley and
Kumar, 1969; Stephens,  1971), and that less extraneous stimulation is required to produce
uncomfortable loudness level (Stephens and Anderson, 1971) for them, but that extroverts,
whether children (Elliot,  1971) or adults (Hockey, 1972), prefer higher levels of sensory
input.
     Independent evidence of individual differences in auditory  perception correlated with
personality measures comes from Soviet researchers. With respect to differences in sensory
threshold, subjects with a strong nervous system are shown to have higher thresholds than
do those with weak nervous systems (Nebylitsyn, 1957, 1966, Borisova, 1967). Specific
differences were found  in individual loudness functions, which suggests the need to intro-
duce a concept of susceptibility to noise (Barbenza et al., 1970a) which might be correlated
to anxiety (Stephens,  1970) and excitability on the MMPI scale (Barbenza etal., 1970b).
     Obviously more research is necessary in order to assess the influence of these factors on
performance level.
     Arousal. It is widely  assumed that noise, by increasing the  amount of stimulation
reaching the CNS, has the effect of raising the level of arousal, so that the Ss feel more alert
and perform better; but in the extreme, when noise exceeds a certain SPL, Ss become more
tense and in this case arousal may result in inefficient behavior. These relationships are best
expressed by the inverted-U hypothesis but its validity is sometimes questioned, particularly
with respect to the concept of over-arousal which Broadbent considers as lacking in preci-
sion.
     Evidence about  the arousing effect of noise comes  from physiological and behavioral
studies, especially from  studies about interaction of noise with other agents such as loss of
sleep, knowledge of results, alcohol, etc. (Wilkinson 1969; Hamilton and Copeman 1970).
The way in  which noise induces physiological and behavioral arousal can be clearly followed
up in Kryter's (1970) diagram (Figure 4). Even though changes in arousal level  provide a
satisfactory  explanation of the changes in performance efficiency, the diversity of results in
studies  using more or less the  same  task  and noise parameters points to the  need for
additional clarification. This is even more apparent when considering those studies  where no
effects of noise  whatever could be  detected. For instance, in a study  by Gulian (1970),
performance in  a vigilance task of  Ss exposed to noise (70 and 90 Db,  continuous and
intermittent) showed  no overall effects of noise. However, when Ss were divided according
to their arousal  level, established  through several EEC parameters, significant differences
were found  between  hypo- and hyper-reactive Ss not only in performance efficiency but
also in evolution of performance (Figure 5).
     Different individuals manifest not  only a distinctive basal level of arousal, but also
differ in arousability toward noise. Extroverts are thought of as chronically less highly
aroused than introverts, and it is argued that when subjects enter the task situation at a low
level of arousal,  their performance in noise should improve to a greater extent than would
that of subjects who begin work at comparatively high arousal levels.
     Data supplied by different investigators support this viewpoint. Thus, it was found
(Davies and  Hockey,  1966) that  the facilitating effect of high-intensity noise was signifi-
cantly  greater for extroverts than for introverts, that extroverts make significantly fewer

                                         369

-------
                                                No 2 MMMTOKV-AUTONOMIC  7
                                                   NfMVOUS SYSTEM   /
                                                                /
Figure 4. Schematic diagram of primary auditory (hearing) and secondary auditory (nonauditory) systems
(After Kryter, 1970)
errors in a variety of auditory conditions (Davies et al., 1969; Blake, 1971; Di Scipio, 1971),
etc. Yet, when the demands of the task are slight, no significant differences in performance
appear between extroverts and introverts, although significant differences in arousal level
(skin conductance) are apparent (Gulian 1971, 1972). Thus it seems that the lack  of diffi-
culty blurs the differences between introverts and extroverts in performance, although they
remain present at the physiological level. On the other hand, recent studies (Hockey 1970 a,
b) have outlined that in complex simultaneous tasks, introverts tend to emphasize the high
priority demands more under normal environmental conditions. Thus differences in arousal
level certainly act differently on performance level. Perhaps a deeper insight could be gained
in the mechanism of noise-induced  arousal  and its effects on performance level if  the
relationships between selective and diffuse arousal were taken into consideration.
    Annoyance.  With anoyance, at last we enter a field of noise investigation where we no
longer meet with conflicting evidence. Everybody complains about noise.
    Annoyance is the final product of noise, whether or not it impairs performance, but of
course many factors, psychological, educational etc. influence its extent and its expression.
Hawel  (1967)  devised a sophisticated scheme for defining  the  complex  relationships
between activity, noise type  and intensity,  the individual's momentary disposition,  and
specific reactions and then" impact on annoyance level. As was stressed by Anderson (1971)

                                         370

-------
                 HYPOREACTIVE
                             HYPERREACTIVE
    */.
    1001-

  u
  cs»
  SJ  90
  o
  u
     85
     80

                      correct detections
V
                                                         arousal
                     arousal
                                                       correct detections
                                                        110
                                                                              100
          Q     WC     LC     WI     LI
              experimental conditions
                              WC     LC     WI     LI
                           experimental conditions
                                                           c
                                                           a
  o
  i/i
  •D
  O
800
                                                                               70
                                                                               60
                                                         50
Figure 5. Differences in arousal and performance level in hyporeactive and hyperreactive individuals. Q -
quiet; WC - weak continuous noise 70 dB; LC - loud continuous noise 90 dB; WI - weak intermittent noise
70 dB; LI - loud intermittent noise 90 dB.
there appear to be at least four aspects of noise annoyance, including social awareness of
noise, personal sensitivity, annoyance  toward specific noise  in a particular situation, and
annoyance toward a set of specified noises in unspecified situations.
     Borsky (1954) found differences in annoyance due to changes in attitude as large as 6
dB as did other authors (e.g, Sorensen, 1970). Atherley et al. (1970) stated that the subjec-
tive importance of certain noises  would influence the  attitude towards them and induce,
accordingly, changes in physiological measures, while  Hermann et al. (1970) demonstrated,
in a  pseudo-tracking task, that annoyance, muscle tension, and TTS are dependent on  the
Ss' emotional attitude toward noise.
     Anderson  and  Robinson  (1971)  advance a two-factor explanation of annoyance:
annoyance is partly produced by changes in arousal caused by purely physiological response
and partly by a so-called cognitive element.
     Perhaps another factor should  be added, namely, aversion  to noise (Sullivan et  al.,
1970) which depends on anxiety  (Broadbent,  1957; Sullivan, 1969), previous experience
with different  noise environments (Spieth, 1956), the individual subjective definition of
aversiveness (Wolff, 1964), task involvement (Kryter, 1966), etc.
     Several studies  have stated a  considerable intersubject variability in individual annoy-
ance, which led Moreira and Bryan (1972) to claim that there  exists  a noise-annoyance
                                         371

-------
susceptibility and individual noise functions, the greatest difference between noise-sensitive
and insensitive Ss occurring at quite moderate levels of noise (Figure 6). Becker et al. (1971)
emphasize that noise-sensitive persons rate all noises, irrespective of their intensity, as being
more intrusive in their daily activity, and rated everything in their environment much more
unacceptable than did  the  noise-insensitive.  The noise-sensitive  subjects were also more
likely to perceive themselves as being more sensitive than the average person, and believed
that it affected their health.
     The differences in sensitivity to noise annoyance are stable, and do not depend upon
age, sex, education, job responsibility or such personality traits as determined by the EPI
and the MMPI, but are correlated with anxiety, and with various measures of personality as
given by the Rorschach Projection Test (Moreira and Bryan, 1972).
     Clearly more studies  are needed to delimit the  annoyance  produced by noise, its
psychophysiological and personality correlates, its effects on performance—the more so as
people tend usually to consider "noise"  the sounds encountered at work, while the noises
experienced at home are considered as merely sounds.
     Noise is annoying, it is a nuisance. That is why man started to study noise and the ways
of reducing it. And this is  the  only hard  fact, the only undisputable one;  for all  the
                          NOISr ANNOYANCE SUSCEPTIBILITY
                          6O
70          80
 Noise level (dBA)
                                                                9O
 Figure 6. Individual noise functions for 6 subjects (3 of the most noise sensitive and 3 of the most
 insensitive to annoyance by noise). Each curve is the mean  rating of all 3 noises by the subject. Noise
 sensitive: m	m,S,   RBH;»	•.S,  AC>	*£,  EN.  Noise  insensitive:D	a£>
 SC; o	o, S,   MLDu	AS,  DW. (After Moreira and Bryan, 1972)

                                          372

-------
behavioral studies of noise effects disclosed simply that certain noises are harmful, certain
tasks and activities are sensitive to noise, certain persons are more affected  than others by
noise, etc. Therefore,  annoyance appears as the crux of the psychological consequences of
noise  and I  feel that  it deserves  much more attention from the psychologist than it has
received up to now.
    Although  the results  in noise  studies are  quite conflicting and although it is rather
improbable that new studies would produce facts contradictory to those already established
it is certainly worth while to continue the investigations, because many aspects have been
neglected:
    - no data exist on very long-term habituation to noises;
    - only few data exist on sensory interaction and its effect on performance;
    - almost no data, except a study by  Lukas et al. (1971) exist on effects of noise on
persons who  are accustomed to noise and are given a task which usually is sensitive to noise;
    - almost no data exist on effects of noise in performing a habitual activity, which is
usually accomplished in a quiet environment;
    - only  few data are available  about individual differences in response to noise and
about evolution of performance and annoyance level over long periods of time in individuals
with differing reactivities:
    - only few data are available about the individual's compensatory effort, both psycho-
logical and physiological,  in the  process of adaptation to  noise and  its consequences on
subsequent performance.
    Many  more problems  exist  and  qualified answers could have a great practical and
theoretical impact. It is our privilege to seek these answers.
                                     Bibliography

   1. Anderson, C. M. B., The measurement of attitude to noise and noises. NPL Acoustic
     report Ac 52, Oct. 1971.
   2. Anderson, C. M. B., Robinson, D. W., The effect of interruption rate on the annoy-
     ance of an intermittent noise. NPL Acoustic report Ac 53, Oct. 1971.
   3. Atherley, G. R.  C., Gibbons, S. L., Powell, J. A., Moderate acoustic stimuli; the
     interrelation of subjective importance and certain physiological changes, Ergonomics,
     1970,75,536-545.
   4. Barbenza, C. H., de, Bryan, M. E.,  McRobert,  H., Tempest, W., Individual loudness
     susceptibility, Sound, 1970(b),4, 75-79.
   5. Barbenza, CM.,  de, Bryan, M.  E.,  Tempest, W., Individual loudness functions, J.
     sound&vibr., 1970 (a), //, 399-410.
   6. Becker, R. W.,  Poza, F., Kryter, K., A study of the sensitivity to noise, AD - 728 332,
     1971.
   7. Bieri, J., Meyers, B., Differential effects of two types ofaversive arousal on discrimina-
     bility,  Psychon.Sci, 1968,13, 203-204.
   8. Blake, M. J. F., Temperament and time of day, in W. P. Colquhoun (ed.), Biological
     rhythms and human performance, London, Academic Press, 1971.  109-148.

                                         373

-------
  9. Blackwell, P. J., Belt, J. A., Effect of differential levels of ambient noise on vigilance
     performance. Percept motor skills, 1971,32, 734.
 10. Boggs, D.H., Simon, J. R., Differential effect of noise on tasks of varying complexity,
     J. appLpsychol., 1968,52, 148-153.
 11. Borisova, M. N., Individualinye razlichiia v. pbrogah razlicheniia gromkocti zvukov, in
     Tipologicheskie osobennosti  vysshei   nervnoi  deiatelinosti  cheloveka,  Moskow,
     Prosveshchenie, 1967.
 12. Borsky, P, Noise and the community. Chaba Report nr. 4 (p. 11).
 13. Broadbent,  D.  E.,  Some effects  of noise on  visual performance.  Quart,  j.exp.
     psychology, 1954, 6, 1-5.
 14. Broadbent, D.E., Effects of noise on behavior, in C. Harris (ed.), Handbook of noise
     control, New York, McGraw-Hill, 1957.
 15. Broadbent, D. E., Decision and stress, London, Academic Press, 1971.
 16. Broadbent,  D. E., Gregory, M., Effects of noise and of signal rate upon vigilance
     analysed by means of decision theory. Human Factors, 1965, 7, 155-162.
 17. Dardano, J. F., Relationship of intermittent noise,  intersignal interval and skin
     conductance to vigilance behavior, J. appl. psychoL,  1962,46, 106-110.
 18. Davies, D. R., Hockey, G. R. J., The effects of noise and doubling the signal frequency
     on  individual differences  in visual vigilance performance. Britj.psychol., 1966, 57,
     381-389.
 19. Davies, D. R., Hockey, G. R. J., Taylor, A., Varied auditory stimulation, temperament
     differences and vigilance performance. Br.j.psychol.,  1969, 60,453-457.
 20. Diespecker, D. D., Davenport, W. G., The initial effect of noise on a simple vibrotactile
     learning task. Percept. & Psychophis, 1967,2, 569-571.
 21.  Discipio, W. J., Psychomotor performance as a function of white noise and personality
     variables. Perceptmbtskills, 1971, 33, 82.
 22.  Elliott, Colin D., Noise tolerance and extroversion in children. Br.j.psychol., 1971, 62,
     375-380.
 23.  Eschenbrenner, A. J., Effects of intermittent noise on the performance of a complex
    psychomotor task.  Human Factors, 1971,13, 59-63.
 24.  Eysenck, H. J., Personality and tolerance for noise, in Proceed. Symp.Psychological
     effects of Noise.  W. Taylor (ed.), Dept Social &  OccupatMed., Univ. of Dundee,
     1970.
25. Farley, T. H.,  Kumar,  K. V., Personality and audiometric response  consistency.
    Lauditres., 1969, 9, 108-111.
26. Finkelman, J. M., Glass, D.  C, A  reappraisal of the relationship between noise and
    human performance by means of a subsidiary task measure.  J:'Appl.psychol., 1970, 54,
    214-220.
27. Glass, D. C., Singer,  J. E., Psychic cost of adaptation to  an environmental stressor.
    J.personal. & Soc. psycho!., 1969,12, 200-210.
28. Glass, D. C, Reim, B., Singer, J. E., Behavioral consequences of adaptation to control-
    lable and uncontrollable noise. JJexp. social psychol., 1971, 7, 244-257.
29. Gulian, E., Effects of noise on an auditory  vigilance task. Rev.roum.sci.sociales - sene
    de Psychologic, 1966,10, 175-186.

                                       374

-------
30. Gulian,  E.,  Effects  of noise  on reaction  time and induced  muscle tension, Rev.
    roum.sci.$ociales - serie de Psychologic, 1967,13, 371-385.
31. Gulian,  E., Effects of noise on arousal  level in auditory vigilance, in A. F. Sanders
    (ed.), Attention and Performance HI, Amsterdam, North-Holland, 1970, 381-393.
32. Gulian,  E., Psychophysiological correlates of auditory vigilance under noise conditions
    in introverts and extroverts. Rev.roum.  sci.sdciales - serie de Psychologic, 1971,  15,
    125-136.
33. Gulian,  E., Cu privire la ponderea factorului "dificultate  a probei" in determinarea
    niveluluiperformantei In conditii de zgomot.Rev.Psihol., 1972,75,323-333.
34. Gulian,  E., Focusing  of attention and arousal level under interaction of stressors in
    introverts and extroverts.  Rev. roum. Sci.sociales -  serie de Psychologic,  1972,  16,
    153-167.
35. Hack, J. M., Robinson, H. W., Lathrop, R. G., Auditory distraction and compensatory
    tracking. Perceptmbtskills, 1965, 20, 228-230.
36. Hamilton, P.,  Copeman, A., The effect of alcohol and noise on components of a
    tracking and monitoring task. Rr.j.Psychol., 1970, 61, 149-156.
37. Harris, S.  C., Human Performance effects or repeated exposure to impulsive acoustic
    stimulation,  AMRL, 1970, nr.70- 38.
38. Hawel, W., Untersuchungen eines Bezugssystems fur die psychologische Schallbewer-
    tung. Arbeitswissenschaft, 1967, 6.
39. Hockey, G.  R. J., Signal probability and spatial location as possible bases for increased
    selectivity in noise. Q.j.exp.psychol., 1970, 22, 37-42.
40. Hockey, G.  R. J., Effect of loud noise on attehtional selectivity, Q.j.exp.psychol.,
    1970,22,28-36.
41. Hockey, G.  R.  J.,  Effects of noise on  human efficiency and some individual  dif-
    ferences. J.sound & vibr., 1972, 20, 299-304.
42. Hermann, H., Mainka, GM Gummlich, H., Psychische  und physische Reaktionen auf
    Gerausch verschiedener subjektiver Wertigkeit. Psychol. Forsch., 1970,33, 289-309.
43. Houston,  B. K., Inhibition and  the facilitating effect  of noise on interference tasks.
    Percept-motskills, 1968, 27, 947-950.
44. Hsia, H. J., Effects of noise and difficulty level of input information in  auditory, visual
    and audiovisual information processing. Percept, mot. skills, 1968,26,99-105.
45. Jerison, H. J., Performance on a simple vigilance task in noise and quiet. J.acoustsoc.
    Am., 1957, 29, 351-353.
46. Jerison, H. J., Effects of noise on human  performance.  J.appl.  psychol., 1959,43,
    96-101.
47. Jerison, H.  J.,  On  the decrement function in  human vigilance, in Buckner,  D.  N.,
    McGrath, J. J. (eds), Vigilance, a.symposium, New York, McGraw-Hill,  1963.
48. Kirk, R. E., Hecht, E., Maintenance of vigilance by programmed noise. Percept.mot.
    skills, 1963,76, 553-560.
49. Kodama, Habuku, Psychological  effect of aircraft noise upon inhabitants of an airport
    neighbourhood. Proceed. 17-th Int. Cong. Applied Psychol., Liege, 1971.
50. Kryter, K., The effects of noise on man. J.speech dis. Monog., 1950, suppl.l.
51. Kryter, K. D., Psychological reactions to aircraft noise. Science, 1966, 757, 1346-1355.
    1346-1355.

                                        375

-------
 52. Kryter, K., The effects of noise on man. New York, Academic Press, 1970.
 53. Lienert, G.  A., Jansen, G., Larmwirkung und Testleistung. Int.Z.  angew.Physiol.
     einschl. Arbeitsphysiologie, 1964, 20, 207-212.
 54. Lukas, J. S., Peeler, D. J., Kryter,  K. D., Effects of sonic booms and subsonic jet
     flyover noise on skeletal muscle tension and a paced tracing task.  NASA CR - 1522,
     1970.
 55. Lukas, J. S., Dobbs, M.  E., Peeler, D. J., Effects on  muscle tension and tracking task
     performance of simulated sonic booms -with low and high intensity vibrational compo-
     nents. NASA CR-1781, 1971.
 56. Mariniako, A. Z.,  Lipovoi, V.  V.,  Ucet summarnovo vremeni otdelinyh shumovyh
     vosdeistvii prigigienuceskoi otsenke preryvistyh shumov. gig. truda, 1972,5, 15-18.
 57. May, D. N.,  Rice,  C.  G., Effects of startle due to pistol shots  on control precision
    performance. J: sound Avibr., 1971,15, 197-202.
 58,  McCann,  P.  H., The effects of ambient noise on  vigilance performance.  Human
     Factors, 1969,11, 251-256.
 59.  McGrath, J. J., Irrelevant stimulation and vigilance performance, in Buckner, D. N.,
    McGrath, J. J. (eds), Vigilance: a symposium, New York, McGraw-Hill.  1963.
 60. Moreira,  N., Bryan, M.  E., Noise annoyance susceptibility.  J: sound vibr.,  1971, 21,
    449-462.
 61. Munz, D. C., Ruffner,  J. W., Cross, J. F., Reduction of noise annoyance through
    manipulation of stressor relevance. Perceptmbt. skills, 1971,32, 55-58.
 62. Nebylitsyn, V. V., Individualinye razlichiia v zritelinom i sluhovom analizatorah po
    parametru sila - chiuvstvitelinostl Voprosy psihogii, 1957,4, 53-62.
 63. Nebylitsyn,   V.  :D.,  Osnovnye svoistva  nervnoi  sistemy cheloveka.  Moskow,
    Prosveshchenie, 1966.
 64. Nosal, Effects of noise on performance and activation level. Polish Psychol.Bull., 1971,
    2, 23-29.
 65. O'Malley, J.,  Poplawsky, A.,  Noise induced  arousal  and  breadth of  attention.
    Percept.motor skills, 1971,33, 887-890.
 66. Parrot, J., Wittersheim, G., Effets  transitoires  d'echelons  de bruit sur la motillite
    palpebrale et divers criteres de performance,  au cours de  I'execution d'une tdche
    psychomotrice de longue dure'e.  in Problemes  actuels de la recherche en ergonomie,
    Paris, Dunod, 1968.
67. Reed,  G.  F., Audiometric response  consistency, auditory fatigue, and personality.
    Perceptmbtor skills, 1961,12, 126.
68. Reed, G. F., Francis, T. R., Drive, personality and audiometric response consistency.
    Perceptmbtskills, 1962,15,681-682.
69. Rentzsch, M., Einfluss von Larm aufeine Kombination geistig-korperlicher Tatigkeiten
    unter Betriebsbedingungen. Wiss.Z.Techn. Univers-Dresden, 1972, 21, 567-572.
70. Rylander, R. (ed.), Sonic boom exposure effects. J.sound&vib., 1972,20, 477-544.
71. Rylander, R., Sorensen, S., Berglund, K., Brodin, C., Experiments on  the effect of
    sonic-boom exposure on humans, J.acoust Soc.Am., 1972, 51, 790-798.
 72. Samuel, W. M. S., Noise and the shifting of attention.  Quart.j.exp. psychol., 1964,16,
    123-130.

                                        376

-------
73. Sanders, A. F., Bunt, A. A., Some remarks on the effects of drugs, lack of sleep and
    loud noise on human performance. Nederlands Tijdschrift voor de Psychologie, 1971,
    26, 670-684.
74. Smith, S.  L.,  Extroversion  and sensory  threshold. Psychophysiology,  1968,  5,
    293-299.
75. Sorensen,  S., On the possibility  of changing the annoyance reaction to noise  by
    changing the attitudes to the source of annoyance. Nordisk Hyg. Tidsk., 1970, suppl.
    1.
76. Spieth,  W., Annoyance  threshold  judgments  of bands  of  white  noise. J.acoust.
    Soc.AmM 1956, 28, 872-877.
77. Stephens, S. D. G., Personality and the slope of loudness function. Q. J.exp.PsychoL,
    1970,22,9-13.
78. Stephens, S. D. G., The value of personality tests in relation to diagnostic problems of
    sensorineural hearing loss. Sound, 1971, J, 73-77.
79. Stephens, S. D. G., Hearing and personality: a review. J: sound & vibration, 1972,  20,
    287-298.
80. Stephens, S.  D. G., Anderson,  C. H. B., Experimental studies on  the uncomfortable
    loudness level J. speech hearing Res., 1971,14, 262-270.
81. Stevens, S. S., Stability of human performance  under intense noise.  Jlsound & vibr.,
    1972, 21, 35-56.
82. Sullivan, R., Subjective matching of anxiety to intensity of -white noise. J:abnormal
    psychol., 1969, 74, 646-650.
83. Sullivan, R., Warren, R., Dabice, M., Minimal aversion thresholds for white noise:
    adaptation. Am. j. psychol., 1970, 83, 613-620.
84. Tanasescu,  Gh., Stanciulescu, EL, Domilescu, M., lonescu, M., Predescu, M., Catoranu,
    V., Popescu, E., Influenza zgomotului produs de copii in pauza asupra unor procese
    simple de  invafare  ale acestora, m Combaterea  zgomotului si  vibratiilor  - a  2-a
    Consfatuire, Bucuresti, 1964.
85. Tarriere, C., Wisner, A., Effets des bruits significatifs et non significatifsau cours d'une
    epreuve de vigilance. Trav.humain, 1962,25, 1-28.
86. Teichner,  W.  H., Arees,  E.,  Reilly,  R., Noise  and human performance, a psycho-
    physiological approach. Ergonomics, 1963,  6,83-97.
87. Thackray, R. I., Touchstone, R. M., Recovery of motor performance following startle.
    Percept, mot. skills, 1970, 30, 279-292.
88. Thackray, R. I., Touchstone, R. M., Jones, K. N., Effects of simulated sonic booms on
    tracking performance and autonomic  response. Aerospace Med., 1972,45, 13-21.
89. Warner, H. D., Effects  of intermittent noise  on human target  detection,  Human
    Factors, 1969,11, 245-250.
90. Warner, H. D., Heimstra, N. W., Effects of intermittent noise on visual search  tasks of
    varying complexity, Perceptmbtskills, 1971, 32, 219-226.
91. Watkins, W., Effect of certain noises upon detection of visual signals. J.exp.psychol.,
    1964,67,72-75.
92. Weinstein,  A., Mackenzie, R. S., Manual performance and arousal. Percept.mot.skills,
    1966, 22, 498.

                                        377

-------
 93. Wilkinson, R., Some factors influencing the effect of environmental stressors upon
     performance. PsychoLBnlL, 1969, 27, 260-272.
 94. Wisner,  A., Dutheil, R., Foret, J.,  Paraht, G, Signaux  sonores en  milieu bruyant.
     Complexite de la situation reette et experiences de laboratoire. Proceed 17-th  Int.
     Congr.Appl.Psyehol., Liege, 1971.
 95. Witecki, 1C, Determination of the  time of simple reaction to visual and acoustic
     stimuli under  the  conditions of interfering noise. Polsk.tygod. lekarst,  1970,  25,
     1012-1015.
 96. Wittersheim, G., Hennemann, M. C., Effets de deux types de bruits industriels sur la
     precision et la rapidite de choix dans  une tache psychomotrice de longue duree, in
     Problemes actuals de la recherche en ergonomie, Paris, Dunod, 1968.
 97. Wolff, B. B., The relationship of experimental pain  tolerance to pain  threshold: a
     critique of Gelfand's paper. Canadian j. psychol., 1964,18, 249-253.
 98. Woodhead, M. M.,  Effect of brief loud noise on decision making.  J.acoustsoc.Am.,
     1959,31, 1329-1.
 99. Woodhead, M. M.,  Searching a visual display in intermittent noise.  J.sound & vib.,
     1964,7,157-161.
100. Woodhead, M» M., The effect of bursts of noise on an arithmetic task. Am.j.psychol.,
     1964, 77,627-633.
101. Woodhead, Muriel,  M., Performing a visual task in the vicinity of reproduced sonic
     bangs. J.sound. &vib.,  1969, 9, 121-125.
                                        378

-------
          SIMILAR AND OPPOSING EFFECTS OF NOISE ON PERFORMANCE

                                    L. R. Hartley

     Medical Research Council, Applied Psychology Unit, 15 Chaucer Road, Cambridge

     Since the  1950's there have been many reports of effects of continuous broad-band
noise on performance. The general observation has been that performance in levels of less
than about 95  dBC is often improved by noise, but  that performance in levels of greater
than about 95 dBC is often impaired by noise. In this present series of experiments we have
continued to use broad-band noise with a spectrum of equal energy  per octave. In all the
following  experiments it has been presented at 95 dBC (the noise conditions) or at 70 dBC
(the quiet conditions).
     In many studies x>f performance in continuous noise, the adverse  or beneficial effect of
the noise  has been observed most conspicuously towards the end of the test, following
about 15  min of performance in noise. One question arising from these studies concerns
whether this effect of the noise arose because of an interaction with time-on-task or whether
it arose because there was  a cumulative effect of the noise, independent of time-on-task.
One way  we have attempted to answer this question is by comparing the performance  of
four groups of subjects. Two groups were exposed to 10 min of noise  or quiet, respectively,
during which time they did two short  tests. The other two groups were exposed to noise or
quiet, respectively, for a further 20 min and allowed to read magazines, before performing
the same two tests. Hence one group received 10 mm of noise and the other group 30 min
of noise, but both groups were tested for only 10 min.
     The experimental test was a modified version of the Stroop color interference test.
This version of the Stroop test was devised by Ray Adams of the Applied Psychology Unit.
In this test, color names are printed in incongruous hues. The subject is asked to select the
hue of the name on the left of the sheet and then cross out  the name of that hue from the
set of names on the right of the sheet.  He is instructed to work as quickly and accurately as
possible, completing as many lines of the test as he can. To assess the degree of interference
caused by having hue and  name attached to  the same response, performance on the coloured
sheets was compared with performance on a similar  test  where all the colour  names are
printed in black ink. In this latter control test the subject  simply crosses out one of the
names on  the right of the sheet that matches the name  on the left of the sheet. Both tests
are  performed  for 5 min each, and scores are the number  of lines  completed. Order  of
presentation of control and experimental tests is counterbalanced as is the order of pre-
sentation of noise and quiet.
     If, as has been argued by Broadbent (1971), noise  affects the filtering of one stimulus
from another, when both are present at the same time,  selection on the basis  of hue when
the predominant color name is also present should be impaired by noise.
     Figure  1 shows  the number of lines of material correctly completed in the  two tests.
The upper part of the figure shows performance on the control sheets on which the names
were  written  in black   ink. The lower  part of the figure  shows the experimental
(incongruously-colored) sheets. The main point of interest is the comparison between 10

                                        379

-------
 MEAN  NUMBER  OF LINES  COMPLETED  IN NOISE AND QUIET


o
UJ
H. ,
y
j
Q.
5
8
V)
5
HI
u.
O

££
HI
CO
2

-

IOU

I7O

/ ~

I6O

ISO


I4O

I3O


I2O


no
IOO
177
,72 	 -- CONTROL
— — 	 SHEETS
170 (69


.

IO MIM FYPfVllinF

	 	 3OMIN. EXPOSURE

-
132
"""~— - — —
"~* "^ — 	 . |25
-— EXPERIMENTAL
	 T24 SHEETS
^^^- 	 '
115
"

QUIET
                                           NOISE
Figure 1 The number of lines completed  in the experimental interference  and  on the control
noninterference tests in noise and quiet.
minutes of quiet or noise and thirty minutes of quiet or noise. The group of subjects who
had only 10 minutes of noise during the tests showed reduced interference in noise com-
pared to quiet; noise speeded up performance on the colored experimental sheets. In com-
parison, the group who had 30 min of noise in all, 20 min of it prior to the  tests, showed
increased interference in noise as compared to quiet. The longer exposure to noise slowed
down performance on the colored sheets, as compared  to quiet. This interaction between
exposure duration and  interference implies that there  are two different effects of noise
depending on the duration of exposure: an initial beneficial and a later detrimental effect.
These effects appear prominently hi the  experimental test involving perceptual selection,
but not in performance on the control test which is purely a measure of speed. These results
are consistent with  the view  that noise  has  a cumulative adverse effect independent  of
time-on-task. It is this cumulative effect of noise which  appears to impair perceptual selec-
tion and probably makes the organism increasingly distractable.
                                       380

-------
     A further noteworthy feature of this test is that there are no overt auditory cues
associated with performance. Hence the  effect of noise could not have been mediated by
auditory  masking, particularly in view of the different effects of the long and short expo-
sures.
     A large proportion of the results demonstrating  the adverse effect of noise  and its
interaction with incentive and  sleep loss have employed the 5-choice serial reaction test,
shown here  in Figure 2. In this test there are 5 brass discs corresponding to 5 possible light
sources. Using a stylus  the  subject taps  the disc appropriate to the light illuminated.  The
light promptly extinguishes and another source is lit. The  subject works as quickly  and
accurately as possible tapping as many correct discs and making as few errors as he can. A
third score is the number of pauses or gaps greater than \1A sec. between responses. Noise
over 95 dBC has almost  invariably shown  an adverse effect on the number of errors and gaps
made. Figure 3 shows the mean number of errors made at various intervals following the last
response, in noise and in quiet. The figure  on the left shows the mean absolute number of
errors made in noise and in quiet by a group of n subjects. The later part of the distribution of
errors is  quite similar to that of correct responses,  but as the figure shows, a  number of
errors are made with a latency of less than  200 msec. The figure on the right shows  the
errors plotted as a percentage  of the totals in quiet and in noise.  This  shows that noise
increases by an equal amount the number of errors made at each latency following the last
response. Noise  does not selectively increase the anticipatory or the  slower misplacement
type of error.
     Figure  4 shows similar latency distributions of correct responses as a function of noise
on  the left of the figure  and  as a  function of time-on-task on the right of  the figure.
Considering the effect of noise, there is an increase  in the proportion of responses with a
latency of 1000 msec or more  in noise as compared to quiet. Comparing the distribution of
responses in the first and  second halves of a test on  the right of the figure, time-on-task
causes a  similar increase in the proportion of responses with a latency  of 1000 msec or
more. Both  noise and time-on-task are similar in this respect. The gap score in the 5-choice
test records  the number of responses in the extreme tail of these distributions.
     The fact that noise can have at least two different effects on performance depending
on  exposure duration may be related to  qualitatively different aspects of noise; namely the
loudness or annoyance  experienced  in noise on the  one hand and the monotony and per-
ceptual isolation experienced on the other.  The results of the following experiments, involv-
ing  the  5-choice  test,  go  some way to support this dichotomy between loudness and
monotony in noise.
     In the  following experiment, the interaction of noise with headphone and free-field
presentation was  considered.  This  interaction is  of interest  since monophonic noise
binaurally presented over headphones is less variable and more isolating perceptually than
free-field  noise  and the perceived loudness may be less. Noise and  quiet presented  over
headphones was compared with the same sound pressure levels presented in the free field.
Subjects  performed the 5-choice test for 40 min under each of these 4 conditions.  The
subjects wore Knowles  miniature microphones  at the entrance  to each external auditory
meatus in ah1 conditions. Sound pressure  level was adjusted to 95  or 70 dBC in the free-field
and the same sound spectrum was presented over headphones at the same sound pressure

                                         381

-------
 5  Neon  lamps
 5 cm. apart
       Stylus
   18 cm.  long
1-5  cm. diameter
             30'
5 Circular brass discs
3 cm.  diameter  and
9 cm.  between centres
                            Figure 2 The 5-choice test.
levels as recorded by the microphones in the external meatus. Sound pressure level was
monitored continuously and adjusted as necessary during  the test.  Figure 5 shows the
difference scores between quiet and noise presented over headphones or in the free-field.
Gaps are shown on the 'left of the figure  and errors on  the right. The two modes of
presentation have approximately equal adverse effects overall although the pattern of
impairment is rather different. There is  a greater adverse effect of headphone noise upon
gaps and a greater adverse effect of free-field noise upon errors in the test. This interaction
in the-effect upon errors and gaps may be related to the  greater perceived loudness of
                                   382

-------
         24    6    8    IO   12
         RESPONSE TIME x  IOO MSEC
14
                                                                = Q-Q
                                                        Q_
                                                                                  Q N
 24    6     8    IO    12    14
RESPONSE TIME  x IOO MSEC
        Figure 3 Latency distribution of mean errors and mean percent errors in quiet and noise.
free-field as compared  to  headphone noise on the one hand and the greater perceptual
isolation and monotony in headphone as compared to free-field noise on the other.
     In a further experiment, the perceptual isolation and monotony accompanying con-
tinuous noise was reduced and the arousal or annoyance quality of the noise maintained by
presenting  the  noise  intermittently. Subjects  performed the 5-choice test for 40 min  in
continuous quiet, continuous noise and in intermittent noise presented in free-field. Intermit-
tent noise consisted of bursts from 1-5 sec long with an average  duration of 3 sec. Average
length of the quiet intervals was  1.5 sec. One group of subjects performed the test with
immediate knowledge of results and a second group performed without this incentive. The
difference between continuous and intermittent noise lies in gaps as Figure 6 shows. Con-
tinuous noise produced approximately twice the increase in gaps that intermittent noise did,
but overall, intermittent noise produced as large an increase in errors as continuous noise,
although there were minor differences  between  incentive conditions. Reducing the
                                         383

-------
mm
2*T-3
gl-9
1*8
.z
< 1-6
g 1-5
O ''4
1- 1-3
Z
ujl.2
O , ,
it '•'
gio
£ .9
*) -8
0 -7
bl
a -6
n
z
< -3
w
5 2
fi .t
V3
O /-v
•
'
•
.
•
-
.

•

•

• •
1
•

•
•

-

•
_


FL
I










i

2 4
?2-0
_ Z .

Q








1







i



1
-1-8
i _j
Q-Q-Q H,.6
N=N-N H(.S
01-4
















i


-

0
















6 8
H-l-3
Z
ui|.2
0
ail
al-O













Q











«2 Q
1 < ->
,2 -8
0 .7
Ul
CC 6
ffl ^4
Z
p ^ -3
Q!W (^J
1 2 -2
1 —
P O

•
'
•

:
.


•

•
-
•
-
-

.
•

.

•



i

























IM
1





































.


= FIRST HALF IN Q-Q
2=SECOND HALF IN Q-Q

i



















a





H
























mm
a











'





m
| |
«m
i
'• •• %
n
i i
J ;
§• 1
IO 12 14 -» ~ 2 4 6 8 IO 12 14
           RESPONSE TIME x IOO MSEC
RESPONSE TIME x IOO MSEC
      Figure 4 Latency distribution of mean percent correct as a function of noise and time-on-task.
monotony and isolation accompanying noise by varying it appears to greatly reduce the
adverse effect of noise upon gaps but leave the adverse effect upon errors unchanged.
     Finally, subjects performed  the 5-choice test in continuous free-field noise and quiet
with and without ear-defenders on. In this experiment, the perceptual isolation of noise was
maintained but the sound pressure level at the ear reduced by wearing ear protection. The
main finding was that, as Figure 7  shows for the difference scores between quiet and noise,
ear-defenders interacted with noise, greatly reducing the adverse effect of noise in the first
half of the test but causing as large an adverse effect as continuous noise in the second half
of the test. These results appeared  reliably in gaps but in this experiment subjects  failed to
show an effect of the 95  dBC noise upon errors in the test. Hence, reducing  the sound
pressure level by wearing ear protection appears to be beneficial initially, but the accom-
panying monotony and perceptual isolation may nevertheless have a detrimental effect later
in the test.
     In summary,  the results of these experiments are consistent with the view  that noise
has a cumulative, adverse effect, increasing with exposure duration, independent of  time-on-
tiSk. The cumulative effect may, however, have two rather different components corre-

                                         384

-------
   DIFFERENCE IN  GAPS AND ERRORS  BETWEEN  NOISE AND QUIET
    15
    14
    13
    12
    II
    IO
     9
     8
§
cc
£
UJ
DC
O
tn
a.
z
<
UJ
7
6
5
4
3
2
 I
O
                                   13-92
                              HEADPHONES
    IO-5O
        6-92
                                           6-56
                                  6-26
JREE-FIELD
      5-74

HEADPHONES
                                         O-3C«
       FIRST HALF           SECOND HALF   FIRST HALF             SECOND HALF
              TEST  HALVES                       TEST  HALVES
 Figure 5 Difference between noise and quiet with headphone and free-field presented with gaps on the left
 and errors on the right.
 spending to the loudness on the one hand, and to the monotony and perceptual isolation
 accompanying noise, on the other. The effect of loudness on performance may predominate
 in the short exposure, whereas the adverse effect of perceptual isolation and monotony may
 predominate following many minutes of exposure to continuous noise.
                                      385

-------
      MEAN  GAPS  AND ERRORS
         IN QUIET AND NOISE
o.
<
O
UJ
2
24
22
2O
18
16
14
12
10
8
6
4
                       GAPS
GC
O
CL
CL
UJ
 UJ
18
17
16
15
14
13
12
II
IO
9
8
7
6
                       19-28
      IO-I7
          INCENTIVE
                           16-50.
                                            22-61
                                           16-23
                           7-94
                          NO INCENTIVE
                      ERRORS
                  CN
                      12-28
     IO-94
     963
              IN
     8-11
                       7-78
12-83
12-16
                          7-83
    FIRST
    HALF
                 SECOND FIRST
                 HALF   HALF
              HALVES OF TEST
                                          SECOND
                                          HALF
Figure 6 Gaps and errors in continuous (CN), intermittent noise (IN) and quiet (Q).
                       386

-------
    DIFFERENCE  BETWEEN  QUIET
       & NOISE WITH & WITHOUT
            EAR-DEFENDERS
S.
      12
      II
      to
S   8
*   7
3   6
2   5
i   4
<   3
iu   2
2   I
" 9-56
                 NO  DEFENDERS
                                 6-5O
                                 4-28
                 EAR-DEFENDERS
         O-79
      FIRST HALF
                            LAST HALF
               TEST HALVES
     Figure 7 Difference between noise and quiet in gaps, with and without ear defenders.
                  REFERENCES

Broadbent, D. E. Decision and Stress, Academic Press, 1971.

                    387

-------
       THE EFFECTS OF DIFFERENT TYPES OF ACOUSTIC STIMULATION
                               ON PERFORMANCE

                                 C. Stanley Harris
                       Aerospace Medical Research Laboratory
                     Wright-Patterson Air Force Base, Ohio, USA

                                   ABSTRACT

    Most studies conducted in our laboratory on the effects of acoustic stimulation on
human performance have produced results showing: (1) no adverse effects,  (2) transient
effects (the adverse effect did not continue throughout the testing period), or (3) effects so
small that the subjects would be expected to adapt with repeated exposure. However, a few
experiments showed adverse effects. High-intensity broadband noise in which subjects wore
ear protection in  levels to 140 dB SPL adversely affected performance on a rail balancing
task and on a Hand-Tool Dexterity task (HTD). The adverse effect on the HTD task resulted
in part from the noise directly vibrating the test apparatus.
    Although an adverse effect  of high intensity noise  on performance was not easy to
demonstrate when measuring  for short time  periods, a lower  intensity level (105 dB) of
broadband noise presented for a longer  time period was found to adversely affect perform-
ance on a continuous task.

INTRODUCTION

    In the Aerospace  Medical  Research Laboratory  (AMRL), the effects  on  human
performance of several types of acoustic stimuli have been studied. However, the emphasis in
the present  paper was limited primarily to  a discussion of the effects of two types of
broadband noise. The first type was a free-field, low-frequency broadband noise which was
presented  in a  reverberation chamber to subjects  wearing  ear protectors. This noise  was
presented  at intensity levels from ambient to 140  dB SPL,  and is characteristically experi-
enced  by personnel working in the vicinity of operating jet engines. The second type was
relatively high-frequency broadband noise presented through earphones. This noise  was
presented  at intensity levels up to 115  dB. Most studies in  the literature on the effects of
noise on human performance have used noise similar to the second type of broadband noise.
The effects of several other types of acoustic stimuli have been studied as a followup to the
results obtained in  the reverberation  chamber. Research has also been conducted using
impulsive noise and noise combined with vibration. A discussion of the results of the latter
experiments has either been omitted or included under the task used for measuring perform-
ance.  However, in order to give  a complete  picture  of the research effort, all studies are
summarized in Table 1.

HIGH-INTENSITY BROADBAND NOISE (TO 140 dB)

    Nausea, vertigo, incoordination, and fatigue have been reported by individuals exposed
to jet  engine noise  (1, 19).  These  reactions, as well as  mental confusion (4), have been

                                        389

-------
 attributed to vestibular stimulation and to reflexes elicited by vibration of the skin, muscles,
 and joints. Any of these symptoms could lead to a reduction in performance efficiency.
 Furthermore, since many proprioceptive reflexes occur  with little or no conscious aware-
 ness, performance efficiency may be  affected by noise  intensity levels lower  than those
 necessary to elicit subjective symptoms.
     Several years ago the authors of the Benox  Report  (1) summarized the problems and
 explored  the effects of high intensity noise on man. Since that report, few studies have been
 conducted using intense noise. The urgency in understanding the effects of this noise on
 man has been reduced since ear protection is currently worn in intense noise environments.
 Nevertheless, intense  noise may adversely affect the performance of personnel even when
 they wear ear protection. In AMRL performance studies, the noise levels in the ear canals
 (up to  115  dB) of the  subjects were  no higher than have been used by several previous
 investigators. However, extra-auditory  effects of noise at  ambient levels to 140 dB added to
 the auditory stimulation. These studies were conducted in a reverberant noise chamber.
 Subjects either wore the same ear protection in noise intensity levels of 120 dB, 130 dB, and
 140 dB or were exposed to an ambient level of 140 dB with different types of ear protec-
 tion. In every study, each group was presented with four noise conditions in four counter-
 balanced orders. Testing within each condition was for a  brief period of time, usually 5 min
 to  20 min. Eight to 20 subjects were used in each group. A practice session was given to all
 subjects on the day prior to experimental testing. Subjects were also given audiograms and
 those without normal hearing  were rejected. Those accepted as subjects were briefly ex-
 posed to the acoustic stimuli used subsequently in the experiment. The noise presented in
 the reverberation chamber peaked at the low frequency end of the spectrum. Both  the
 intensity levels and the spectra are given in Figure 1 for  the ambient conditions and when
 the levels were attenuated by earplugs or earplugs plus muffs.


Cognitive Performance in Intense Noise

     Dickson and Chadwick asked jet mechanics to  describe their subjective experiences
when standing close  to an accelerating jet engine.  Most  reports were  vague, but were
described  best by "one  of the engineers who said  he experienced a momentary sense  of
imbalance accompanied  by  a lack of power to think (4)." In addition, the authors of the
Benox Report (I) have stated that people in high-intensity noise tend to forget or neglect to
follow instructions and to work hurriedly but require more time to complete a task. Both
studies suggested impairment of cognitive ability during exposure to intense noise.


Short-Term Memory

     The procedure used by Korn and Lindley (17) was  adopted for measuring short-term
memory (STM). This  task required the subject to  remember the order of presentation of
nine consonants. The  consonants were projected  on a screen for six seconds, then removed,
and the subject was allowed 10 seconds to write his answer. Testing was conducted in blocks
of 15 trials, and each  slide presentation of the nine consonants was a trial. This task seemed

                                        390

-------
                     s
150;
130-
110 -
90 •
70 -

7
140-
120-
K)0 ;
80-
60 -
40 r

130 •
110 -
90 -
70 -
50 •
                                   EAHMUFFS 8 PLUGS
                        SO-
                                                    AMBIENT
                                   EARMUFFS 8 PLU6S
                            J	I
EARMUFFS A PLUGS'

 I     I     I
                           OVER 75-
                            ALL ISO
             150-
             300
     300-
     600
600- 1200-
1200  2400
2400- 4800-
4800  9600
                                   OCTAVE PASS BANDS (Hz)
Figure 1 Ambient noise spectra and calculated noise levels in ear canals after reduction of noise by ear
protection.
an appropriate one to use because little or no learning was involved, the instructions were
easy to understand, yet the subjects seldom remembered the position of all nine consonants.
     Four experiments were conducted in noise. The results are summarized in study 9 in
Table  1. The STM  task showed no sensitivity to  broadband noise in the reverberation
chamber either with consonants chosen on the basis of high usage (Group 1) or low usage in
the English Language (Group  2).  Similarly, no sensitivity was  shown,  using high usage
consonants, to broadband  noise presented through  earphones (Group 3) or to broadband
noise that varied in the low cut-off frequency (Group 4).

Discrimination Task

     In this task six symbols were presented in one-inch-square boxes. The subject com-
pared each of four boxes with one centered above them as to whether the same or different
symbols occupied the same relative spatial  position in the respective boxes.  He then noted
the number of differences on a line directly under each block (see  Figure 2). After the
completion of a set of comparisons, the subjects advanced to other sets of five boxes until
the time limit expired. Performance was measured in two 4-min periods, with a 1-min rest
given between periods. The score for each period was the number of boxes completed minus
the number of errors made. A symmetrical exposure group and an asymmetrical exposure
                                         391

-------
                                      0

                                      H
V

C
/ V
S D
I



0 X
S C
N
0
H
V
0
z
                                                                  X

                                                                  C
                         Figure 2 Example of Discrimination Task.
group performed  the task on two practice days and during exposure to noise on four test
days. Noise at 130 dB and 140 dB had a small detrimental effect on the performance of the
asymmetrical group during  the first 4-min period but not  during the second 4-min period
(Figure 3). No adverse effect was obtained for the symmetrical exposure group. Therefore,
only a transient  decrement occurred and only for the asymmetrical group. Prior to the
experiment, the subjects had never been exposed to this type of noise and it seems unlikely
that more experienced  subjects  would  have  shown  even  a transient  decrement  in
performance.

Summary

     At  the  time these studies were conducted, an adverse effect of noise on  cognitive
performance seemed assured. Most of the literature on  the effects of noise on  human
performance did  not seem relevant, and  anecdotal  reports of people  working in noise
strongly suggested that mental processes were adversely affected. However, little evidence
was found to indicate an adverse effect of this type of noise on cognitive performance (see
studies 3, 9, & 10 in Table 1). Testing has, of course, been limited. More sophisticated  tasks
presented for longer periods of time may show more sensitivity  to noise in this range.
Nevertheless,  the  effect of high intensity jet  noise on cognitive performance is  less  than
many have previously believed.

Psychomotor Performance in Intense Noise

     Noise was  expected to have a greater effect on  psychomotor  tasks  than on  cognitive
tasks, because most reports of individuals working around jet engines refer to equilibratory
and  postural  disturbances rather than to mental confusion or disturbances. Parrack  (19)
points out that personnel working in noise levels up to  160 dB report heating of the  skin,
strong sensation of vibration in various parts of the body, sensations of muscular weakness,
                                        392

-------
                       50
                     S
                       45
                     (0
                     s
                     z
                       40
                        CONTFK
                                             O PLUGS ONLY
                                             • PLUGS 8 ONE MUFF
ITROL 120  I3O I4O CONTROL 120 130  140
 NOISE CONDITIONS (dB re. O.OO02 Dyne /cm*)
Figure 3 Mean corrected score of noise conditions for asymmetrical and symmetrical exposures for both
4-minute periods.
and excessive fatigue. Several cases of staggering, falling, and feelings of forced movement
have also been reported (1).  Therefore, high-intensity noise can adversely affect psycho-
motor performance through  effects on human physiology or through mechanical inter-
ference with motor movements.

Rail Task

     The rail task initially used in our studies was based on an ataxia test battery developed
by Graybiel and Fregley (5). Various parts of the task were dropped  because they showed
no sensitivity to noise conditions (6, 12, 18). Since most of the results obtained with the
task have been published, detailed experimental procedures are not presented. To perform
the task, the subject was asked to stand on a narrow rail in a heel-to-toe manner with his
arms folded across his chest and with his eyes open  (see Figure 4). Two rails were used, one
1% inch  wide and the other % inch wide. The balance time on both rails was scored to the
nearest second, beginning when the subject assumed the correct position on the rail, and
ending when he lifted a foot, unfolded his arms, or fell off the rail. The maximum score for
each trial was 60 seconds; if the subject was still balanced on the rail at the end of this time,
the trial  was discontinued. Five trials were presented on each rail, in most experiments.
                                          393

-------

                       Figure 4  Subject performing on the Rail Task.
     The rail task has been shown to be quite sensitive to acoustic stimulation (6). In the
reverberation chamber, decrements were found to be related to intensity of noise stimula-
tion and to stimulus asymmetry (6).  Detrimental effects on performance were found at 140
dB  regardless of the type of ear protection worn. As shown in Figure 5,  the asymmetrical
exposure produced decrements in performance at the 120 dB and  130 dB levels whereas
improvement occurred at those levels with balanced exposures. After-effects of exposure to
noise were found for asymmetrical exposures but not for symmetrical exposures (6). Using
pure tones presented  through earphones, evidence was obtained for a frequency (Hz) sensi-
tivity of the task. Frequencies around 1000 Hz to  1500 Hz seem to have  more detrimental
effects than either the higher or lower frequencies in the range of 100 Hz to 2500 Hz (8).
Furthermore, an asymmetrical stimulus presented intermittently (2.86 pulses per second) at
a frequency  of 1000 Hz has produced  large decrements on the rail task. The decrement
produced by intermittency added to the effect produced by asymmetry' (15) (see Figure 6).
Finally, an experiment has demonstrated that stimulus asymmetry has an adverse effect on
rail  task performance because one ear is  always stimulated more strongly than the other anil
not because the same ear is always stimulated rnore intensely (see study 6 in Table  1).
                                       394

-------
                      3O
                      20
                       IO
                       10
                     -20
                     -30
 AO BALANCED EXPOSURES (plugs only)
 AC BALANCED EXPOSURE (muffs 8 plugs)
- A« UNBALANCED EXPOSURE (plugs 8 one muff *
                               120         130        140
                          NOISE CONDITIONS (dB re. 0.0002 Dyne /cm*)

       Figure 5 Percent change from control means for each type of ear protection (Rail Task).
Hand-Tool Dexterity Task (HTD)

     The HTD was designed to measure motor performance largely independent of mental
factors. The equipment for the task  consisted of three horizontal rows of nuts and bolts
mounted on a wooden stanchion (see Figure 7). The nuts and bolts were of three different
sizes and  four  of the same size were mounted in a row. The subject's task, using two
adjustable wrenches and a screwdriver, was to remove all of the bolts from the left upright
and transfer them to the corresponding rows on the right upright. The score was the time
taken to make the transfer. Two groups were tested on the task during exposure to noise
(see  study 7 in Table  1). Significant increases in the time  taken to  complete the task
occurred at 130 dB  and 140 dB for a symmetrical exposure  group and at  140 dB for an
asymmetrical group  (see  Figure  8). The large difference in the initial  ability  and in the
different rate of learning of the groups made it impossible to determine the relative effects
of symmetrical versus asymmetrical exposures (7). Nevertheless; noise at the 140 dB level
adversely affected the performance of both groups. Furthermore,  the adverse effects may
have resulted from either auditory stimulation or extra-auditory or a combination of both.
The extra-auditory stimulation may have produced adverse effects either  by vibrating the
body  and limbs or  by vibrating the  task.  The latter explanation was supported by the
                                         395

-------
                        10
                     rt
                     <0
                    O
                    C
                    U.

                    g
                        10
                       20-
                        30-
                       4C -
 I            I
 O ASYMMETRICAL
 A SYMMETRICAL
	CONTINUOUS
	INTERMITTENT
                                 1
 1
            105
                                85          95
                                    INTENSITY MB
    Figure 6 Percent change relative to 65 dB for subjects tested using the 1000 Hz tone (Rail Task).
statements of three subjects who mentioned that they were bothered by the shaking of the
smallest bolts in their stations at 140 dB. Direct observations suggested to the experimenter
that the shaking of the nuts and bolts did add a few seconds to the time taken to complete
the task at 140 dB but probably had no effect on the completion time at 130 dB.
     In a subsequent experiment (21) subjects were tested on the HTD task during auditory
stimulation alone. The noise spectra and the levels approximated those delivered to the ear
canals  of subjects  wearing ear protectors  in  a free-field broadband  noise  of 140  dB.
Specifically, the subjects were presented: control, 100 dB, 115 dB,  and 115 dB in the left
and  100 dB  in the right earphone (see study 8 in Table 1).  The noise presented in  this
manner did not adversely affect  performance on the task.

Summary

     The rail task was quite sensitive to acoustic stimulation; however, these results  are
more of theoretical than practical interest since this sensitivity would not have occurred
using wider rails. On Jhe other hand, the Hand-Tool Dexterity  test was more representative
of tasks performed by individuals working in high  intensity noise.  Therefore, decrements
obtained on the HTD test are of greater concern than decrements obtained on the rail task.
                                        396

-------

                                               .,-

                   Figure 7 Subject performing the Hand-Tool Dexterity Test.
We are still uncertain of the manner in which the HTD test is affected by noise. It" perform-
ance was adversely affected by  a direct  mechanical effect of noise  on the task, then the
results are not as important as an adverse effect produced  by interference with the motor
coordination  of man.  The reason is that changes in the  task equipment, dimensions, or
procedures often can  reduce  or eliminate problems caused by  direct mechanic;!, inter-
ference.

BROADBAND NOISE  (EARPHONES)

     In these studies, broadband noise was presented through earphones at intensity levels
to 115  dB. The experiments described earlier in which noise levels and spectra, presented
through earphones,  approximated  the acoustic stimuli in  the ear canals of the subjects
exposed to an ambient noise level of 140 dB wearing different types of ear protection, fit
the category of this  type of noise. Although these were control studies with the purp<
comparing the results  with those obtained in ambient noise in the reverberation chamber,
performance was measured during short  exposures to broadband  noise. No adverse elfects
were obtained on the short-term  memory test, or on the Hand-Tool Dexterity  test; however,
there was a small but statistically significant reduction in the amount of time subjects would

                                         397

-------
                         310
                         300
                       E.290
                         280
                        270
                        260
                                   	ASYMMETRICAL EXPOSURE
                                   -- SYMMETRICAL EXPOSURE
                               CONTROL   120    130     140
                            NCISE CONDITIONS (dB re 0.0002 Dyne/cm2)

Figure 8 Mean time to complete Hand-Tool  Dexterity Test at each  noise level for asymmetrical and
symmetrical exposures.
balance on the rails. This noise was of a lower frequency than the noise used with other
tasks measured under this heading, and there was probably less reason to expect an effect on
performance.

Serial Search Task (SST)

    All two digit numbers from 10 to  99 were used twice in constructing a test sheet for
the serial search task (see Figure 9). The numbers were presented in pairs in six different
columns and 15 pairs were in each column. Each two-digit number occurred once as ,the first
member of a pair and once as the second member of a pair.  The subject started each  sheet
by looking for the  number  10  as the first member of a pair. When he found it, he wrote
down the  number which was the corresponding second number of the pair on a piece of
paper and  proceeded to look for this new number as the first member of a pair, and so on.
    In two experiments noise had an adverse  effect on the serial  search task. In one
experiment the effect was transient (13), and in the other the effect was constant across the
testing period and increased  with days of presentation of the task (14). In the first experi-
ment, the performance of one group of subjects was measured during exposure to a broad-
band noise of 105  dB (see Figure 10), while the other group served as a control. All the
                                        398

-------
12
41
56
51
84
69
66
47
95
20
40
16
71
45
13
61
47
41
83
30
73
35
69
42
74
96
49
76
31
25
79
)0
49
83
50
65
53
21
82
94
80
91
22
44
63
60
88
97
75
92
91
72
50
39
32
22
48
33
64
38
19
23
62
81
26
25
85
37
48
64
96
78
67
98
53
39
93
44
54
40
53
15
94
95
14
16
55
51
56
59
52
14
6)
30
57
77
70
86
74
87
13
29
55
11
72
70
37
36
98
52
57
43
82
23
71
34
90
45
29
46
36
93
27
68
46
60
43
17
97
54
42
73
76
75
59
23
58
11
89
27
62
26
68
87
31
21
20
65
86
84
92
33
99
15
24
33
31
32
88
39
34
28
90
89
35
63
85
G6
19
77
80
24
17
18
67
79
10
78
13
12
                        Figure 9 Sample sheet of the Serial Search Task.
subjects were tested on the task during 5 days, a preliminary training day and 4 test days.
The pattern of testing on the preliminary day was: test 6 min, 1st sheet—rest 3 min—test 12
min, 2nd sheet-rest 3 min—test 12 min,  3rd sheet-rest 3 min-test 12 min, 4th sheet. This
same procedure was followed throughout the remaining four days of testing with the excep-
tion that the noise group was informed  on the  second day  that they would be asked to
perform the task in noise each day after the six minute warm-up. The noise was turned on 1
min before testing began in the  first  12-min testing period. The noise was presented con-
tinuously throughout the remaining testing periods and the rest intervals.
     An analysis of variance calculated on the  data revealed no significant effect for noise;
however, there were significant effects for the  noise x trials interaction, and for the noise x
days x  trials interaction. The noise x trials interaction was due to the large difference
between the noise and control group on the first trial, which was statistically significant (see
Figure 12). The three way interaction was obtained mainly because the difference between
groups on the  first trial was larger on the 4th  day than it was on the preceding three test
days.
     In  the second experiment, three groups of subjects were tested in the same manner as
the previous groups, except that the rest periods were omitted. One group was tested during
exposure to intermittent noise. This group was presented the same spectrum and intensity
level (105 dB) as the continuous noise group; however, the noise was interrupted at a rate of
2.86 times per  second and a duty cycle of 50% was used. The results showed an effect for
noise that approached significance (p<.10) and a statistically significant effect for the noise
x days interaction. This effect is illustrated in Figure 11. The difference between the control

                                         399

-------
                    i  no
                    9
                    §  100
                    d
                    s
                    §  90

                    ij  so
                    Id
                    oc
                    w
                       70
                       60
                       50
                       40
                                I   I    I    I   I
                                       I
                           3L5 63  125  250  500  IK  2K  4K  8K  I6K
                                 OCTAVE BAND FREQUENCY (Htrtz)
                       Figure 10 Octave band levels under the earphones
group and  the  noise groups  increased as a function of days with the largest  difference
occurring on  the last day of testing. No significant differences were obtained on day 1 or
day 2, but both noise groups differed significantly from the control on day 3 (beyond the
.05 level), and  on  day  4 (beyond the .01 level). No significant differences were obtained
between the continuous and the intermittent noise groups. In these experiments, the rest
periods produced more learning on the SST and allowed the subjects in the noise group to
perform more like the control group on the last two trials on each day.


SUMMARY AND DISCUSSION

     High-intensity broadband noise (up to 140 dB with ear protection) had little effect on
cognitive performance.  No adverse effects were obtained with a short-term memory task, a
paper & pencil  maze task (see study 10 in Table 1), and only a transient effect was obtained
using a  discrimination  task. The transient effect was obtained only for an asymmetrical
exposure group and not for a symmetrical exposure group. It may be  that for this type of
noise to adversely affect cognitive performance, the same conditions apply as Broadbent (2,
3) has suggested are necessary for obtaining a detrimental effect on performance at lower
intensity levels  of noise. These conditions are: (a) testing should be continuous  for a mini-
mum of  30 to  60  minutes, (b) noise of 100 dB or above should be used, and (c) the task
should be experimenter paced or one that requires the continual attention of the subject.
     Psychomotor  performance showed  more sensitivity to noise. The rail task was more
sensitive to noise than any other task used. However, the part of the rail task battery used in
                                        400

-------
                       70
                      65
                    O
                    tit
i
8
                      60
                    i
                      55
f
                                   O CONTROL
                                   A CONTINUOUS NOISE
                                   Q INTERMITTENT NOISE
                                         ,
                            ,
                                 I       2       3
                                          DAYS
          Figure 11 Mean number completed per trial on each day for rest and no-rest groups.
most experiments was deliberately selected for its sensitivity and was a difficult task. There-
fore, the practical implications of these findings are unknown. On the other hand, there is
little doubt that Hand-Tool Dexterity performance is adversely affected by noise, and it
seems reasonable to assume that such performance will be adversely affected by intense
levels of jet engine noise. Of course, the conclusion would not be that personnel cannot
perform the HTD task in the noise, but that it would take them a little longer to complete
the task; approximately  10%  longer at 140 dB as indicated  by our studies. Whether the
increase in the time to complete the task was caused by extra-auditory effects of the noise
or by a combination of auditory and extra-auditory effects  cannot be determined at the
present  time.  There was  some indication that the effect  may have been due to direct
interference with the task itself, primarily  because of the shaking of the nuts and bolts in
their stations.  It seems unlikely that this could account for all the decrements observed but
it cannot totally be disputed at the present time.
     In  the studies using the  serial search task we deliberately applied the conditions sug-
gested as necessary for producing a detrimental effect of noise on human performance.
There is no question that noise either adversely affected performance on the SST or slowed
down the rate  of learning/ The effect was one that increased with repeated days of exposure.
If testing had  been stopped  after two days of exposure to the noise then no statistically sig-
nificant effect of noise would have been demonstrated since only on days 3 and 4 did the
                                         401

-------
                       70
                       65
                     o
                     UJ
                     III
                     8 60
                       55
                       50
                     	1	

                      NO REST
                     ./TEST
O CONTROL
A CONTINUOUS NOISE
D INTERMITTENT NOISE
                                 I
                                 I           2           3
                                      TWELVE MINUTE TRIALS

              Figure 12 Mean number completed for trials for rest and no-rest groups.
difference between the noise and control groups reach statistical significance. These results
suggested that if the conditions recommended by Broadbent for producing an effect of noise
on performance are applied, and if a fairly large number of subjects are tested during re-
peated exposure to noise, then a significant effect of noise on performance can be demon-
strated.
                                         402

-------
                       Table 1

NOISE STUDIES CONDUCTED IN AEROSPACE MEDICAL RESEARCH
               LABORATORY, 1965-1973.
/
Study



1




2





3









U





5








6



Author



Ehoenberger
& Harris
(1965)


Nixon, Harris,
& von Gierke
(1966)



Harris & von
Gierke (1967)








Harris St Som-
mer (1966)




Harris (1972)








larris
(unpublished)



Group



1




1





1

2
3

ft



5
1


a


i

2

3


4

1

2


N



16




21





S

12
16

a



8
24


24


10

10

10


10

10

10


Tajis



Tsai- Partington num-
bers, draw a line be-
tween successive num-
bers (1 - 25) randomly
scattered on a paper.
Rail task (CT)
Eyes open (EO)
Eyes closed (EC)
Rail walking (BW)


Rail task (EO & EC)

Rail task (EO & EC)
Rail task (EO 8r EC)

Discrimination task
(DT) (during exposure)
Rail task (after exp-
osure) .
Same as above
RT (EO & EC)


BT (EO & EC)


RT (EO)

RT (EO)

RT (EO)


RT (EO)

Rf CEO.) ' •

RT (EO)

Whole Body
(MB) or
Earphone
Exposure
(&>
EE




WB





WE

WB
WE

WB



WB
EE


EE


EE

EE

EE


EE

EE

EE

A pproxijna t e
Exposure
Time Per
Condition
(Minutes )
30 - 15




25





13

S
8

10



10 '•
10


10


10

10

10


10

10

10

Exp. Conds.
on Same or
on Differ-
ent Day a

Different




Different





Different .

Different
Different

Different



Different
Different


Different


Same

Same

Same


Same

Same

Same


Moise Conditions



(l) Quiet - Quiet
(2) Quiet - 110 dB
(3) 85 dB - 110 dB
(4) 95 dB - 110 dB

(1) Control (70 dB)
(2) Earplugs & muffs in 120 dB
noiae (symmetrical).
(3) Earplugs & 1 muff over right
ear (RE), in 120 dB noise
(asymmetrical) .
Earplugs «t muffs in 70, 120, 130,
and 140 dB.
Earplugs in above noise levels.
Earplugs & 1 muff (RE) in above
noise levels.
Earplugs & 1 muff (RE) in above
noise levels.


Earplugs in above noise levels.
Control & 95 dB to both ears -at
frequencies of 100, 260, 590, 1500,
and 2500 Hz.
Control & asymmetrical exposure of
75 dB (RE) - 95 dB (LE) of above
frequencies.
Symmetrical presentations of 65,
85, 95, 105 dB of 1000 Hz tone.
Asymmetrical presentation (above
tones, presented only to LE).
Symmetrical presentation of above
stimuli with intennittent, 2.86
pulses/second (pps), presentation.
Asymmetrical presentation of
stimuli used for group 3.
1500 Hi tone, 2pps, 65, 85, 95, & E
105 dB. f
Above stimuli with pulses alter- c
nated between ears.

Results



Small but significant effect
when noise level switched from
Quiet to 110 dB (condition).


Small but significant effect on
RT (EO) during asymmetrical
condition.



Effect on RT (EO) at 140 dB.

Effect on RT (EO) at 140 dB.
Effect on RT (EO) at 120 & 140 dB.

Transient effect on DT & margin-
al effect on fiT (EO) at 130 &
140 dB (p< .10).

No effects on either task.
No effect.


Asymmetrical exposure on RT (EO)
1500 Ha significantly different
from control, 100 Hz & 590 Hz.
No effects.

Significant effect, 95 dB & 105
dB different from 65 dB.
Significant effect, 105 dB dif-
ferent from 65 dB.

Significant effect, 85, 95, 105
dB different from 65 dB.
jignificant effects for both
;roups with no significant
Lifference between groups.


-------
TABLE 1 (CONT)
7



8

9











10




11





11




12






13







Harris (1968)



Sooner t Har-
rlB (1970)
Harris
(unpublished)










Harris
(unpublished}



Harris
(unpublished)




Harris
(unpublished)



Harris & Fil-
eon (1971)





Harris (1972)







1

Z

1

1



2

3

-. 4



1




1





2




1



2


1


2

3


16

16-

16

12



12

12

12



12




20





10




15



15


20


20

20


Hand tool dexterity
task (HID).
HID

HID

Short tern memory (STK,
for relative position
of high frequency con-
sonants.
SIM, low frequency
consonants.
SIM, high frequency
consonants.
Sane as above.



Finding path through a
series of mazes, pre-
sented via paper it
pencil. Also finding
most direct route test.
Minnesota Rate of Man-
ipulation (turning).
Medium Taping.
Two Plate Taping.
Purdue Pegboard (both
hands).
Purdue Pegboard (both)
Purdue Pegboard (right)
Purdue Pegboard (left)
O'Connor Finger Dex-
terity test.
Visual search for num-
bers, Serial Search
task (SST).

SST


SST


SST

SST


WB

WB

EE

WB



WB

EE

EE



WB




EE





EX








EE





EE

EE


6

6

6

5



5

5

5



10




6





6




44



44


38


38

38


Different

Different

Same

Same



Same

Same

Same



Different




Same





Same




Different



Different


Different


Different

Different


Earplugs in 70, 120, 130, &140 dB
noise.
Earplugs & 1 muff (RE) in above
noise levels.
Same as group 3, below *

Earplugs in 70 dB It 140 dB noise.
Earplugs & muffs in 1W> dB, ear-
plugs St 1 muff (HE) in 1AO dB.

Same as above.

"Simulations of above noise levels
in ear canals.
105 dB, broadband noise varying
in lower cut-off frequency, 600
Hz, U200 Hz 4 2,00 Hzj control,
300 Hz cut-off presented at 70 dB.
Earplugs in 70, 120, 130, and
140 dB.



1500 Hz, 2 pps, at 65, 85, 95,
and 105 dB.




1500 Hz, 2 pps, at 65, 85, 95,
and 105 dB.



Control group, perform SST for
16 min. with 2, 3 rain, rest per-
.ods between trials (U days of
testing).
ilxp. same as above but 105 dB
noiae presented throughout test-
ing.
Control group, SST presented con-
.inuously on each of i daya for
36 consecutive min. (3 trials).
Same as above, but with 105 dB
iroadband noiae.
Same as above with intermittent
2.R6 pps) 105 dB broadband noise.

Significant effect, 130 
-------
TABLE 1 (CONT)

IJt
15
16
17
18

Sonuner
(unpublished)
Harris & Sch-
oenberger
(1970)
Harris & Som-
mer (1973)
Harris (1970)
Harris (1970)

1
1
1
1
2
1

16
8
12
10
10
6

SST presented for 30
min. at 1.5, 3-5, 5-5,
and 7-5 hours.
Compensatory tracking
(vertical & horizontal
scores) St 2 response
time tasks.
Compensatory tracking
(vertical & horizontal
scores) & 2 response
time tasks.
Pursuit Rotor task
(PR)
PR
PR

WE
EE
EE
WB
WB

480
20
80
30
30
30

Different
Different
Different
Different
Different
Different

Control, 65 dB(A), 90 dB(A), 95
dB(A), aircraft fly over at 1 min.
between peaks.
(1) 85 dB noise (2) 85 dB noise
with 5 Hz vertical vibration at
.25 gz (3) 110 dB noise (4) 110
dB noise with 5 Hz vertical vib-
ration at .25 gz.
60 dB and 110 dB presented alone
and combined with vibration (6 Hz
at .10 gz).
Control, 60, 20 sec. trials on
PR per day.
9 impulsive noise stimuli (112 dB
peak, 400 milliseconds duration)
presented per day over a 4 day
period .
Practice day, 1 test day with
above stimulus.

No effects.
Significant effect on vertical
dimension of tracking task.
Significant effect on horizontal
tracking & approached signif-
icance on vertical tracking.
(p<.10).
Very small but significant effect
which was considerably reduced
by the 4th day of testing.
No significant recovery in sen-
sitivity to impulsive noise
after 5-3 month period.

-------
                                  REFERENCES

  1.  Benox Report: An exploratory study of the biological effects of noise.  ONR Project
     NR 144079, Univer. of Chicago, December, 1953.
  2.  Broadbent, D. E., Perception and Communication. New York: Pergamon Press, 1958a.
  3.  Broadbent, D. E., Effects of noise on an intellectual task. Journal of A coustical Society
     of America, 30, 824-827, 1958b.
 4.  Dickson, E. E. D., and Chadwick,  D. L. Observations on disturbances of equilibrium
     and other symptoms induced by jet-engine noise,/. Laryngol. andOtol, 65: 154-165,
     1951.
 5.  Graybiel,  A., and Fregley, A.  R.}  A New Quantitative Ataxia  Test Battery, BuMed
     Project MR 005.13-6001,  Subtask 1, Report No. 107 and NASA Order No. R-37,
     Naval School of Aviation Medicine,  Pensacola, Florida, 1963.
 6.  Harris, C. S., and von Gierke, H. E., The Effects of High Intensity Noise on Human
     Equilibrium,  AMRL-TR-67-41,  Aerospace Medical  Research  Laboratory,  Wright-
     Patterson AFB, Ohio,  1967.
 7.  Harris, C.  S., The Effects of High Intensity Noise on Human Performance, AMRL-TR-
     67-119,  Aerospace  Medical  Research  Laboratory,  Wright-Patterson  AFB, Ohio,
     1968.
 8.  Harris, C.  S., and Sommer, H. C., Human Equilibrium during Acoustic Stimulation by
    Discrete  Frequencies,  AMRL-TR-68-7,   Aerospace  Medical Research Laboratory,
     Wright-Patterson AFB, Ohio, 1968.
 9.  Harris, C.  S., Human Performance Effects of Repeated Exposure to Impulsive Acoustic
    Stimulation,  AMRL-TR-70-38,  Aerospace Medical  Research  Laboratory,  Wright-
     Patterson AFB, Ohio, 1970.
10.  Harris, C. S., Long Term Adaptation of Pursuit Rotor Performance to Impulsive Acous-
     tic  Stimulation,  AMRL-TR-70-92, Aerospace  Medical Research  Laboratory, Wright-
    Patterson AFB, Ohio, 1970.
11. Harris, C.  S., and Shoenberger, R. W., Combined Effects of Noise and Vibration on
    Psychomotor Performance,  AMRL-TR-70-14, Aerospace Medical  Research Laboratory,
    Wright-Patterson AFB, Ohio, 1970.
12. Harris, C.  S., Effects of Acoustic Stimuli on the Vestibular System, AMRL-TR-71-58,
    Aerospace Medical Research Laboratory, Wright-Patterson AFB, Ohio, 1971.
13. Harris, C.  S., and Filson, G. W., Effects of Noise on Serial Search  Performance, AMRL-
    TR-71-56, Aerospace Medical Research Laboratory, Wright-Patterson AFB, Ohio, 1971.
14. Harris, C.  S., Effects  of intermittent and continuous noise on serial search perform-
    ance, Perceptual & Mo tor SkiUs.  35,627-634, 1972.
15. Harris, C. S., Effects of increasing intensity levels of intermittent and continuous 1000
    Hz tones on.human equilibrium, Perceptual & Motor Skills,  35, 395-405, 1972.
16. Harris, C.  S., and Sommer, H.  C.,  Interactive effects of intense noise and low level
    vibration on tracking performance and response  time, Aerospace Medicine, 1973 (sub-
    mitted for publication).
17.  Korn, J. H., and Lindley, R. H., Immediate memory for consonants as a function of
     frequency of occurence and frequency of appearance. Journal Experimental Psychol-
    ogy. 66, 149-154, 1963.

                                      406

-------
18.  Nixon, C. W., Harris, C. S., and von Gierke, H. E., Rail Test to Evaluate Equilibrium in
    Low-level Wideband Noise, AMRL-TR-66-85, Aerospace Medical Research Laboratory,
    Wright-Patterson AFB, Ohio, 1966.
19.  Parrack, H. O., Eldredge, D. H., and Koster, H. F.,  Physiological Effects of Intense
    Sound, Memo.  Report No. MCREXD-695-71B, U. S. Air Force, Air Material Com-
    mand, Wright-Patterson AFB, Ohio, 1948.
20.  Shoenberger, R. W., and Harris, C. S., Human performance as a function of changes in
    acoustic noise levels, Journal of Engineer ing Psychology, 4 108-119, 1965.
21.  Sommer, H. C., and Harris, C. S., Comparative Effects of Auditory and Extra-Auditory
    Acoustic  Stimulation  on Human  Equilibrium and Motor Performance,  AMRL-
    TR-70-26,  Aerospace Medical  Research Laboratory, Wright-Patterson AFB,  Ohio,
    1970.
                                       407

-------
             BEHAVIORAL EFFECTS AND AFTEREFFECTS OF NOISE1

                                     David C. Glass
                               Department of Psychology
                            The University of Texas at Austin

                                    Jerome E. Singer
                               Department of Psychology
                      State University of New York At Stony Brook
                                     Introduction

     Many aspects of urban life can be viewed as work under stress. People have roles, duties
and tasks to perform  while all around them there is noise, crowding, litter, and traffic. A
number of social critics have commented upon global aspects of these factors, but there is
little  research  of an analytic  nature directed toward ascertaining the specific effects of
urban-like stressors. This paper reports results of approximately two dozen laboratory and
field  experiments,  conducted  over a five-year period, which systematically explored the
effects of stress in man.
     Broad-band noise was the  principal stressor used in our research, and we will, therefore,
limit our discussion to the behavioral consequences of noise exposure. An audio tape con-
sisting of a melange of indistinguishable sounds was  prepared, and, when played back at
intensities up to  108 dbA, served as the stimulus.2 The high-intensity noise thus generated
proved stressful;  the ability of subjects to work  under this stress,  as well as adverse after-
effects of noise exposure were noted.

                                      Adaptation

     The most reliable result  was that people adapted to the noise. When noise was pre-
sented in intermittent bursts over a 24-minute session, few disruptive effects were shown
after  the  first  few noise  trials. This result is, of course, consistent with a good deal of
previous research in the area (cf.  Broadbent, 1957; Kryter, 1970).  Indeed, it is easier to list
the special circumstances under  which noise does produce  an  immediate effect than to
 'Preparation of this paper was made possible by grants from the National Science Foundation (GS-34329
 and GS-33216), Russell Sage Foundation, and the  Hogg Foundation  for Mental Health.
 2The noise consisted of a tape recording of the following sounds superimposed upon one another (1) two people
 speaking Spanish", (2) one person speaking Armenian; (3) a mimeograph machine; (4) a desk calculator; (5)
 a typewriter. We selected this particular concatenation  of sounds as an  analogue of the spectrum of
 complex noise  often present in the urban environment. A sound-spectrographic analysis of the noise
 recording showed that energy did indeed range broadly from 500 Hz to 7,000 Hz, with the mode at about
 700 Hz. Free field stimulation was used throughout  most  of the research.

                                          409

-------
 catalogue the general cases where it does not have an effect. We found that people do not
 adapt to noise under at least two particular arrangements:
      1.   If a person is in a situation of cognitive overload, working on more than one task
          and straining his ability to cope with nonstressful stimuli, the addition of noise
          produces performance decrements.This can be seen, for example, as an increase in
          errors on the subsidiary task in a two-task  situation (e.g., Finkelman and Glass,
          1970).
     2.   If a person is working on a vigilance task requiring constant monitoring or atten-
          tion,  the presence of high-intensity  noise is  apt to be disruptive. Someone who is
          tracking a  pursuit rotor or monitoring a series of dials will do his job less effi-
          ciently under noisy than under quiet conditions, (e.g.,  Glass and Singer, 1972;
          Broadbent, 1957).
     In our research, adaptation was noted by comparing the performance  of people sub-
jected to noise  with  that of people not so subjected on a variety of tasks ranging from the
boringly simple to the  oftentimes challenging and  interesting. Over the course of several
experiments, matching stimulus  configurations with motor movements, clerical aptitudes,
spatial relations, and  driving an automobile simulator all showed adaptation—no decrement in
performance under conditions of loud intermittent noise. A typical set of data are shown in
the first table. These data illustrate adaptation and accompanying lack of adverse behavioral
effects. Under most circumstances, task performance under noise does not differ from task
performance without noise, past the first few bursts.
     Adaptation can be defined in other ways, however (cf.  Lazarus, 1968). In findings
parallel to those obtained with performance measures, people showed adaptation or habitua-
tion (we use the terms interchangeably)3 on several psychophysiological indices, including
phasic skin conductance (GSR), muscle tension in the neck,  and finger vasoconstriction.
These autonomic measures  failed to show continued high reactivity to spasmodic bursts of
noise over a 24-minute noise session.  Figure 1 shows GSR adaptation data as average log
conductance change  scores within each of 4 blocks of noise  trials. There is a significant
decline in GSR on successive blocks in each noise condition. Since initial reactions to loud
noise (108 dbA) were greater than to soft noise (56 dbA), the magnitude of GSR decline is
understandably greater in the former condition. However, the magnitude and  rate of adapta-
tion is virtually identical in the predictable and unpredictable conditions within each noise-
intensity treatment;  that is, subjects  were  equally  reactive at the beginning of the noise
session and equally unreactive at the end.

                        Noise Aftereffects and Unpredictability

     The finding that noise had no routine effects upon task performance in  the laboratory
can be considered an  example of art imitating life, for in the city, despite noise and a host of
other stressors,  work goes on. This finding does not imply that noise has no adverse effects;
to the contrary, our research suggests that it is deleterious to routine functioning subsequent
3See Thompson and Spencer (1966) and Lazarus (1968) for discussions of adaptation, habituation, and re-
lated concepts.

                                        410

-------
                                       Table 1
        AVERAGE NUMBER OF ERRORS ON PART 1 AND AVERAGE DECREMENTS IN
           ERRORS FROM PART 1 TO PART 2 OF THE NUMBER COMPARISON TEST
                                          Experimental  condition

                                  Loud noise      Soft  noise      No  noise
                                   (108  dbA)       (56  dbA)        control
                                   (n =  18)        (n = 20)       (n  = 10)

  Part 1 errors                     3.28            3.30            2.80

  Decrement  in errors             -1.85           -1.48           -0.20
     from Part 1 to
     Part 2
to its  occurrence.  In other words, despite lack  of direct effects, noise had disruptive
aftereffects. These aftereffects occurred whether or not adaptation took place and were
demonstrated on a variety of performance measures. The ability of people to find errors
when proofreading, to continue working on difficult graphic puzzles, and to work effi-
ciently on a competitive-response task were all adversely affected by having been previously
exposed to noisy conditions.
     These aftereffects, surprisingly, were not only a function of the physical intensity of
noise, but  also  depended  upon the social and cognitive context in which noise occurred.
Indeed, the reduction of noise level from 108 to 56 decibels did not have  as large  an
ameliorative effect as any of several cognitive factors. Two of these factors—predictability
and controllability—had a particularly powerful impact on noise aftereffects.
     Exposure  to unpredictable noise, in contrast  to predictable noise, was followed  by
greater impairment of task performance and lowered tolerance for post-noise frustrations.
Despite equivalent adaptation in the two noise conditions, the magnitude of adverse after-
effects was greater following unpredictable noise.  Typical results are shown in Figure 2.
These data are the average number trials taken by subjects on insoluble (though seemingly
soluble) graphic puzzles (Glass and Singer,  1972). Since the task is inherently  frustrating,
fewer trials represent lower tolerance for frustration.
     As can be  seen, subjects showed less persistence on the puzzles following exposure to
loud unpredictable noise than to loud and soft predictable noise. And this effect was true
for both puzzles. It would appear,  then, that lowered tolerance for frustration is a con-
sequence of exposure  to the presumably more aversive unpredictable noise. There is also an

                                         411

-------

     •  -

j-

c
* CJO
u
•
Jj wo
o
a
*l C)0
0
U
5 OJO
• o>0
I
.







































N t
•0^








1



                                                            N 10


             Loud           Soft
         Unpredictable  Molt*
Loud            Soft
 Predictable  Nolee
  Control
No  Nol»e
       Figure 1. Mean log conductance change scores for four successive blocks of noise bursts.
unexpected  tendency for this effect  to  appear even when the unpredictable noise i-
particularly  loud. Soft unpredictable noise was associated with lower frustration-tolerance
than loud predictable noise.
     We have obtained, as I mentioned earlier, similar results with aftereffect measures other
than frustration tolerance. It would appear that unpredictability is indeed a potent  factor in
the production of  noise  aftereffects. The  case for the  existence of this phenomenon is
further strengthened by the range of conditions over which it has been obtained in various
replications  (Glass  and Singer, 1972i These include:  (a) different ways of manipulatin;-
predictability, such  as periodic noise schedules as well as signalling noise onset; (b)different
levels of noise intensity,  such as  108 dbA  and 56  dbA; (c) different subject populations,
such as male and female college students and middle-aged white  collar workers; and  (d)
different laboratory settings.
     Our emphasis on the unpredictability variable is not meant to minimize the importance
of intensity in producing noise aftereffects. We have recently completed a  field  study ot
traffic noise in New York City which clearly demonstrates the importance of the  inteiiMi\
                                         412

-------
     28
     24
     20
      16
      12
E
bl
00
       8
UJ
C5
<
QC
hi
             INSOLUBLE PUZZLES

             • 1st
                  nd
                LOUD      SOFT      LOUD     SOFT    CONTROL

          Unpredictable  Noise    Predictable Noise    No  Noise
               Figure 2. Average number of trials on the insoluble puzzles.


                                  413

-------
 parameter (Cohen, Glass and Singer, 1973). The results produced correlations of the order
 of .45 between impaired auditory discrimination in children (i.e., the ability to differentiate
 speech sounds) and the ambient noise level in their high-rise apartments. These levels ranged
 from a low of about 55 dbA on the 32nd floor to a high of about 75 dbA on the 8th floor.
 Moreover, the magnitude of the  relationship between noise and discrimination was greater
 the longer the children had  resided in  the apartments. There was also evidence linking
 impaired auditory discrimination to deficits in reading achievement.
                   Noise Aftereffects and Perceived Uncontrollability


     Another cognitive  factor mediating noise aftereffect phenomena is the individual's
belief that he can escape or avoid aversive sound, i.e., perceived controllability. In a series of
laboratory experiments (Glass and Singer,  1972), subjects who were given  a switch with
which to terminate noise (Perceived Control Condition) showed minimal aftereffects com-
pared to other subjects exposed to the same noise without the switch (No Perceived Control
Condition). This reduction in aftereffects occurred even though the switch was not in fact
used. Merely perceiving control over noise was sufficient to ameliorate its aversive impact.
     Figure 3 shows the relevant results. It is immediately obvious that tolerance for frustra-
tion was appreciably increased by the perception of control over noise termination. These
effects have been obtained with a number of experimental variations of perceived control,
including the induction  of a perceived contingency between instrumental responding and
avoidance of the stressor.
     But, what specific  stress-reducing mechanisms _are aroused by the manipulation of
perceived control? In answering this question, we reasoned that uncontrollable and unpre-
dictable noise confronts the individual with a situation in which he is powerless to affect the
occurrence  of the stressor and he  cannot even anticipate its occurrence. The individual is
likely to give up his  efforts  at controlling the stimulus under these circumstances, and we
may thus describe his psychological state as one of "helplessness" (cf. Seligman, Maier, and
Solomon, 1971). Perceived Control subjects label their psychological state as one in which
they have control over their  environment, and, thereforei are not helpless. By contrast, No
Perceived Control subjects label themselves  as having minimal environmental control. Task
performance after noise stimulation is affected in a way that is consistent with prior experi-
ence, when control was or was not perceived as available.
     We tentatively conclude,- ,therefore, that unpredictable and uncontrollable noise pro-
duces adverse aftereffects because unpredictability and undontrollability lead to a sense of
helplessness which manifests itself as lowered motivation in subsequent task performance.
David Krantz and I have just completed two experiments designed to test aspects of this
helplessness interpretation. Preliminary analysis of the results indicates that manipulated
helplessness does indeed produce lowered motivation which transfers from one experimental
task to  another.  It should  also be emphasized that the same effect occurred following
exposure to both high and  moderate noise intensities (i.e., 105 dbA and 75 dbA).

                                        414

-------
              45
              40
              35
              30
              25
              20

               15
         M
         IB
               10
                        Insoluble  Puzzles
                        EH  2
                               nd
                               Perceived
                                Control
No  Perceived
   Control
Figure 3. Average number of trials on the insoluble puzzles for perceived control and no perceived control
        conditions.
                                         415

-------
                              Summary and Conclusions

      In summary, noise appears to have few direct effects. People adapt to aversive sound.
 But noise does have disruptive  aftereffects, and these are in large measure a function of the
 cognitive circumstances  in which acoustic stimulation  occurs. These conclusions do not
 mean that  aftereffect phenomena are  the  "psychic price" paid by the individual  for his
 adaptation  to noxious noise (cf. Dubos,  1965; Selye,  1956; Wohlwill, 1970). It is entirely
 possible that noise  aftereffects are  as much post-stressor phenomena  as postadaptation
 phenomena. Further analysis and experimentation enabled us to reach a partial adjudication
 of this theoretical issue. Our  current position  is that after-effects represent  behavioral
 consequences of cumulative exposure to aversive stimulation.  It is not the adaptive process
 itself that causes deleterious aftereffects,  but  the fact  of mere exposure in spite of adapta-
 tion.
                                   REFERENCES

Broadbent, D. E. Effects of noise on behavior. In Ch. 10 of C. M. Harris (Ed.), Handbook of
    Noise Control. New York: McGraw-Hill (1957).
Cohen, S.,  Glass, D. C.,  and Singer, J. E. Apartment noise, auditory discrimination,  and
    reading ability in children. /. exp. soc. Psychol. (1973, in press).
Dubos, R. Man Adapting. New Haven, Conn.: Yale Univ. Press (1965).
Finkelman, J: M., and Glass, D. C. Reappraisal of the relationship between noise and human
    performance by means of a subsidiary task measure. /.  appl. Psychol,  54, 211-213
    (1970).
Glass, D.  C., and Singer, J. E. Urban Stress: Experiments in  Noise and Social Stressors. New
    York: Academic Press (1972).
Kryter, K. D. The Effects of Noise on Man. New York: Academic Press (1970).
Lazarus, R. S. Emotions and adaptation: Conceptual and empirical relations. In W. J. Arnold
    (Ed.),  Nebraska Symposium on  Motivation.  Lincoln, Nebr.:  Univ.  of  Nebr.  Press,
     175-266(1968).
Seligman, M. E. P., Maier, S. F., and Solomon, R. L. Unpredictable and uncontrollable aversive
    events. In F. R. Brush (Ed.), Aversive Conditioning and Learning. New York: Academic
    Press, 347^00(1971).;
Selye,H. The Stress of Life. New York: McGraw-Hill (1956).
Thompson, R. F., and  Spencer, W.  A. Habituation: A model phenomenon f6r the study of
    neural substrates of behavior. Psychol. Rev., 73, 16-43 (1966).
Wohlwill, J. F. The emerging discipline of environmental psychology. Amer. Psychologist, 25,
    303-312(1970).

                                        416

-------
                         EFFECTS OF NOISE ON A SERIAL
                         SHORT-TERM MEMORY PROCESS.

                            G. Wittersheim and P. Salame

                       Centre d'Etudes Bioclimatiques du CNRS
                                 Strasbourg, France
1. Introduction
     The advancement achieved these last 10 years in our knowledge of memory processes
has numerous implications in the design  of communication systems. Working processes
involving short-term memory are  very frequent in industrial work situations as well as in
private life.  Classical examples of this are the telephone-girl's job or the dialing of a tele-
phone number which, after having been picked out from a telephone directory, needs to be
kept at least for a short time in memory until its dialing. Recent ergonomical investigations
have led, on one hand, to  the design of the format, structures and codes for material to be
memorized and, on the other hand, to the design of new keyboards. However, our knowl-
edge concerning the effects  of environmental  factors  such as noise on the information
receiving, storing and transmitting process is still incomplete.
     The cycle of such a process can be split into three main phases:
     1. An acquisition phase for material to be memorized and which, generally, involves a
relatively high perceptual  load, either visual, auditory,  or coming from some other sense
organ;
     2. A retention phase, which can be very short, but during which rehearsal may be
performed;
     3. A reproduction or response phase requiring motor activity, either verbal or manual,
but  also,  in most cases, perceptual  control. In a few situations where the cycles to be
processed are regularly repeated or paced as was the case in our investigation, a fourth phase
has to be added:
     4. An expectation phase before feedback of information as  to the correctness of the
response, during which no active mental operations are required.

2. Experimental conditions

     The aim of the present investigation was to study  the effects of noise on short-term
memory  depending on whether the noise was produced during the first, second, third or
fourth phase
     A 95 dB(lin) pink noise in an open field condition was used in this experiment.  The
spectral composition  of the noise is shown in Figure  1. The noise had been previously
recorded on a magnetic tape and  its emission synchronized with  the onset  and the end of
each phase by means of a device controlled by the signal programming machine.
     The task was a sequential machine-paced memory task. Each session lasted 30 minutes
during which 140 ± 2 cycles were displayed one after the other to the subject (s). The time
structure of one cycle was the following (Figure 2).

                                        417

-------
                          Figure 1  Pink noise Spectral composition.
     - Acquisition phase: Six digits taken from the vocabulary of the 5 first digits were
displayed in a random sequence to  s.  Each digit was displayed for 500 msec and separated
from the next by  140 msec. So the whole acquisition phase lasted for 3700 msec.
     - Retention  phase: This phase  also lasted for 3700 msec. Overt and or covert rehearsal
were allowed. At the end of  that phase, the letter  R (on a cold cathode tube, as the digits)
was automatically turned on,  indicating to s that he should respond.
     - Response phase: This phase lasted for 4480 msec.
     In  case of an error or an omission, a small white light signal flashed at the end of the
cycle, but no error correction had to be performed. After a one-second delay, the next cycle
began.
     The subject  sat in a small soundproof cubicle (figure  3). He was instructed  to repro-
duce  the  6 digits as rapidly and as  accurately as  possible by pressing on the keys  of a
keyboard. Visual control of  keyboard operations was recommended in order  to minimize
motor errors.

                                         418

-------
         ACQUISITION
RETENTION
                          3700
          REPRODUCTION
1
1
-innnnr-

"
!4
T
u
I
I
I
I
1
5
j EXPECTAT.
! O*
I II
V 1 II
h j *
Ull U U 11 ll DEL.
-RESPONSE--*!
1
7400
                                         11880
Figure 2 Time structure of a complete acquisition reproduction cycle. The reproduction phase <1>5 can be
split into two subphases $3 and $4. In phase $2 noise would be produced only until the 6th response-key
press. This particular condition has not been studied in the present investigation.
     Twenty-one Ss participated in the experiment. A replicated latin square design was
used. Thus each S had to perform four 30-min. sessions, each corresponding to noise in one
of the four phases. Practice sessions took place in the morning and experimental sessions in
the afternoon. These were  separated by  pauses lasting 35 minutes during which Ss were
tested on a audiometer and  were required to fill in a questionnaire concerning the subjective
ratings of the task difficulty. A more detailed questionnaire concerning the subjective  noise
effects was carried out at the end of the experiment.
     All data were directly processed by an on-line PDP-8 computer.

Results:

     The main results are shown in figure 4, where $ 1 stands for noise during the acquisi-
tion phase, O2 during the retention phase, $5 during the reproduction phase and $4 during
the expectation phase.Because during the expectation phase session did not interfere with
any noise active mental processing the performances during that session were used as com-
parisons for the three other  phases.
     Performance  on  the accuracy scores, with  errors and  omissions grouped together,
deteriorated  significantly (p< .05) when noise was produced during the acquisition and
retention phases, but there was no difference when noise was  produced in the  response
phase.
     Speed scores were computed by a special procedure: the time between the beginning
and the end of each response list was divided by the total number of keys swept over by the
hand while responding. Thus these elementary time scores obtained from different digit lists
could easily be compared. There were only very small differences  between the four  obtained
means of speed scores. However, by combining accuracy and speed scores through a T-score

                                         419

-------



Figure 3 Inside of the Subject's cubicle. The stimulus sources are located in front of the S on a semi-
circular screen. The keys are arranged on the keyboard (all Ss were righthanded) in such a manner that they
can be easily reached through mere forearm movements.
procedure, a hierarchical effect can be shown: deterioration appeared to be the most impor-
tant when noise  was produced  in  the acquisition  phase and the least  important  in  the
expectancy phase.
     If now we examine the results concerning the subjective ratings (Figure 5) we observe
that the most  unpleasant and most difficult session was the session during which noise was
produced  during  the acquisition phase. In  that phase too, memorization  was judged  to be
the most  difficult,  whereas it was judged  as facilitated  when noise was produced  in  the
response phase.

Discussion:

     There seems to exist now enough  experimental evidence  for  the hypothesis that
memorizing implies the  translation by  the central  mechanisms of the visual message into an
auditory  message which is processed and stored by the brain. We  should  then expect noise

                                         420

-------
               ERRORS
                p
               10
               ms
               205
               200
                50
                45
 ACCURACY

       t   p(bn)
01 vsOt: 2,143 <05
01 vs05 2,235 4: 2,130 <.05
02 vs $5: 2,046 
-------
                 <)>1     $2     1     (J>2     5    $4

MOST  DIFFICULT TRIAL
            10
                                                                r
             MEMORIZATION MADE EASIER
  -MADE  MORE  DIFFICULT
                               SUBJECTIVE  RATINGS
Figure 5 Subjective Ratings, n = number of subjects judging at the end of the experiment which session
had been the most unpleasant, the most difficult, etc.
below 95  dB. Schonpflug and Schafer (1962) found that a  1000-Hz sound at 95 dB im-
proved memory relative to performance with the same sound but at only  55 dB and they
observed  that  the differences in  retention reflected differences in the  activation level.
Hermann  and Osterkamp (1966) confirmed the hypothesis that intermittent white noise at
95 dB has a harmful effect on both the level of retention and organization of material to be
memorized by interrupting logical and associative connections.
     Rabbitt (1968) found that when  Ss  tried  to remember lists of digits played to them
through pulsed white noise, the number of errors they made was greater than in normal
conditions. According to that author, the digit-recognizing process in a noisy condition may
preempt channel  capacity which is necessary  for efficient retention  in immediate memory
storage. The present experimental conditions were not quite the same as those of Rabbitt, as
the digit  lists in  that  author's experiment were spoken  through noise.  Thus no  visual-
                                         422

-------
auditory  translation was necessary. Nevertheless in the retention phase, noise interference
with rehearsal was clearly present.
    An interesting point to be mentioned is the sound fitting of subjective ratings to the
performance scores. Thus both efficiency and  comfort in memory-task performance are
liable  to  be  seriously impaired by noise produced while information is being taken in and
edited in  storage.

                                 BIBLIOGRAPHY

HORMANN, H., OSTERKAMP, U. (1966, a): Uber den Einfluss von kontinuierlichem Larm
    auf die Organisation von Gedachtnissinhalten. Z. exp. angew. Psychol., 13, 31-38.
HORMANN, H., OSTERKAMP, U. (1966, b):  Uber den Einfluss von diskontinuierlichem
    Larm auf  die Organisation von Gedachtnissinhalten.  Z. exp.  angew.  Psychol., 13,
    265-273.
MILLER, H. (1957): Effects  of high intensity noise on retention. J.  appl. Psychol., 41.
    370-372.
SCHONPFLUG,  W.,  SCHAFER,  M.  (1962): Retention und Aktivation bei  akustischer
    Zusatzreizung. Z. exp. angew. Psychol., 9, 452-464.
SLOBODA,  W., SMITH, E. E. (1968):  Disruption effects  in  human short-term memory:
    some negative findings. Percept, mot. Skills, 27, 575-582.
RABBITT, P. M. A (1968): Channel-capacity, intelligibility and immediate memory. Quart.
    J. exp. Psychol., 20, 241-248.
                                       423

-------
   THE EFFECT OF ANNOYING NOISE ON SOME PSYCHOLOGICAL FUNCTIONS
                                 DURING WORK

                                 Irena Franszczuk
                         Central Institute of Work Protection
                                  Warsaw, Poland


    Psychological investigations concerning the influence of noise on human performance
were carried out with the cooperation of our Acoustical Department. Engineers dealing with
technical acoustical problems often met, in practice, some psychological and physiological
questions, as follows:
     1)   How does noise act on human performance  and the feeling of comfort, and what
         is the psychological mechanism of this effect?
    2)   Which kinds of work are most disturbed by noise?
    3)   On which physical characteristics of noise does its disturbing influence on human
         comfort and performance depend?
    The solutions of these problems, as we know from the literature, are rather ambiguous.
    The causes of this ambiguity have been very well described in the report of Dr. Gulian.
In our investigations we tried to avoid  ambiguity or at least some of the factors influencing
it. All conditions were constant in each of our experiments, except for the noise. The noise
was generated by loudspeakers  driven by a  tape recorder during all experiments except the
control  experiments  in silence. We used various natural noises recorded  in factories and
different bands  of  white  noise  from  generator. The acoustical  variables  were level,
frequency, and bandwidth. The level and spectrum of noise used were measured in all our
experiments.
    We  used no noises greater than  90 dB  SPL. Experiments were  carried out under
laboratory conditions with  students performing for some hours a task requiring attention
and finger dexterity. The level of psychic  performance  during the work was measured by
means of different psychological tests. The first problem in these investigations was to find a
test sensitive enough to measure subtle small changes. All our results were statistically
tested. The annoyance caused by the noise was evaluated by each investigated person on a
six-level scale.
    In  this report  I would like  to summarize the  most important results of our four
experiments, all  of which are  published in the Quarterly Journal of our  Institute, "Prace
CIOP", (1, 2,3,4). They are shown in Figures 1 and 2.
     1.  A noise  band  of median frequency 4000 Hz prolongs simple reaction time. An
annoying noise in which sounds of frequencies near 4000 Hz predominate, even though the
level is not greater than 85 dB, produces an increase of simple reaction time to both light
and sound stimuli. Independent of differences in the individual sensitivity to noise, the
phenomenon of  prolongation of reaction time has been observed in each of our 24 test
persons  in all four of the experiments cited (Figures 1 and  2). This prolongation is very
small but statistically significant (p < 0,01).
     2.  Broad-band white noise has more effect on simple reaction time than narrow bands
of white noise with the median frequency 250 Hz (3). This dependence can be seen in

                                        425

-------
 21
Figure 1 Simple reaction time under different acoustical conditions (4). Median values of the reaction time
in hundredths of second obtained from 8640 measures  of 12 persons:  (a) to light stimuli (b) to sound
stimuli. Silence S; Noise A: octave band of white noise with median frequency 4000 Hz; Noise B: octave
band of white noise with median frequency 250 Hz.
Figure 2, in which we have the average results of 720 measures of simple reaction time* of 6
persons, in each of the experimental noises and in silence.
     3.  Under certain conditions, noise (the level of which does not exceed 85 dB, without
dominant components of high frequencies) may be an activating factor, and may shorten the
reaction time both to light and to auditory stimuli. We have obtained the shortest reaction
time with narrow bands of white noise with the median frequency 1000 Hz (ref. 3) (Figure
2). Perhaps  these results may be explained  by  some arousal hypothesis. The  similarity of
changes in simple reaction time to light and sound stimuli supports such a hypothesis.
     4.  The noise reduces perception efficiency.
     Using in our investigations the octavebands of filtered  white noise with the median
frequency 4000 Hz and level 80-85 dB, we found a statistically significant prolongation of
                                         426

-------
24
23
22
21
20
 19
 18
                            CM


i
       b)
21
20
19
18
17
16
-
v>

-

.0

!NJ

^
C\j
i
i

1
1


|

^

~,
^.

i
i

i

figure  2 Comparison  of simple reaction  time  under different acoustical  conditions and in silence
{Franaszczuk, 1968,1971). (In hundredths of seconds).
    Median value obtained from 720 measures, 6 persons:
    (a) to light stimuli;
    (b) to sound stimuli.
     Silence          —S
     White noise      -WN
     Tone            — T
     1/3 octave band  -1/3
     Octave-band      — 0
     250,1000,4000  - median frequency in c/s.
                                             427

-------
 average simple  reaction time  both to  light  and to  sound stimuli in comparison with the
 average reaction time in silence. The increment of the average reaction time to sound stimuli
 is greater than the increment  of reaction time to light stimuli (4). This difference was very
 little, 0.01 sec, but statistically significant.
     We have obtained similar differences in some other experiments too (2).
     The simple reaction process to the stimuli is composed of the phase of perception and
 of the  motor phase (pressing the key).  The motor phase is  identical in both kinds of
 reaction, but the  phase of perception is different,  because the sound stimuli are partially
 masked by the noise. This explains the greater increment of the reaction time by auditory
 stimuli in noise and suggests that the perception process is the most disturbed by the noise.
     5.  All experimental noise conditions causing a prolongation of simple reaction time
 were  evaluated  as more annoying than noises  not causing it. The subjective feeling of
 annoyance  is a  signal of decreasing psychic performance even when  these changes  are not
 measurable and not observed.
     Experiments  involving other psychological tests and tests of choice reaction time did
 not give us any differences between noise and silence conditions.
     The simple reaction time is the simplest measure, giving much data in a very short time.
 With this test it is not possible to compensate for a decrement  of performance by exerting
 additional effort,  as one can in  spite of tiredness, during other work. The simple reaction
 time measure is closely connected with the subjective feeling  of annoyance and may be
 considered an indicator  of general influence of noise.

                                     References

     1.  Franaszczuk, I., Examination of Changes in the Effectiveness of Mental Functions
 During Work Under Noisy Conditions. Prace CIOP Nr 49. 1966.
     2.  Franaszczuk, I., Investigation  on Fluctuations  of Reaction Time to Visual and
Auditory Stimulus During Work Under  Noise Conditions. Prace CIOP Nr 54,  1967.
     3.  Franaszczuk, I., Effect of Width of  Frequency Band of Noise Upon Reaction Time
of Man To Light And Sound Signals. Prace CIOP Nr 58, 1968.
     4.  Franaszczuk, I., Effect of Annoying Noise of Determined Frequencies on Simple
and Composed Reaction Times. Prace CIOP Nr 64, 1970.
     5.  Franaszczuk, I., The influence of disturbing acoustic stimuli on  some psychic
 functions. Prace CIOP 71,1971, 335-343.
                                        428

-------
                        SESSION 5
NON-AUDITORY PHYSIOLOGICAL AND PATHOLOGICAL REACTIONS

               Chairmen: E. Grand jean, Switzerland
                       S. Kubik, Czechoslovakia
                            429

-------
                       NON-AUDITORY EFFECTS OF NOISE
           PHYSIOLOGICAL AND PSYCHOLOGICAL REACTIONS IN MAN

                                    Gerd Jansen
                                Inst. Hyg. Arbeitsmed.
                           University of Bochum, Germany


     During the Conference on "Noise as a Public Health Hazard" in 1968, an earlier report
on "Effects of Noise on Physiological  State" was presented. It was shown  that a formula
exists, established by systematic research, that allows prediction of the vegetative reactions
by means of sound levels and bandwidths (Jansen, 1967),  Moreover,  it was possible to
establish the limits of normal vegetative reactions (Fig.  1).  These limits have been applied to
concrete noise situations,  especially for assessment of noise-induced disturbance of health
around airports. These limits were  confirmed by two investigations:
     1.   noise-exposed steelworkers showed more vegetative disturbances due  to noise
         than those workers from "quiet" factories (Jansen, 1959);
     2.   by means of vasodilative medications which produce a contrary reaction, it was
         shown that pathological reactions  during noise applications  began  beyond the
         limits mentioned above. The medication  does not work in healthy men; therefore,
         we saw no influence on noise-induced reaction during application of subcritical
         noise bursts, whereas during application of supracritical noise a significant  com-
         pensation occurred (Jansen, 1969).
     In Dec. 1969, the AAAS organized a symposium "Physiological  Effects of Noise"; in
the  course of this Symposium physiological  and  patho physiological  noise reactions  were
reported. The main problems described were: cardiovascular noise reactions, the influence of
noise on adaptation processes, resistance against disease, endocrine and metabolic functions,
biochemical and pharmacological  influences, effects on reproductivity and some neurologi-
cal and sleep disturbances (Welch and Welch, 1970).
 Investigation Concerning Activation

     The summary of all published results, especially those of Washington and Boston, is
 that noise acts as a stimulator for activation of a group of physiological functions designated
 as "arousal reactions".
     Arousal, therefore, means an elevation of excitation level of certain systems of the
 body. The activation-theory, as experimentally affirmed by Hebb (1955), Malmo (1959)
 and many others, differentiates between cortical, autonomic, motor, endocrine and affective
 arousal.  The last concept is closely linked to emotional stress. Acoustic stimuli are con-
 ducted to the cortex via  the ascending  reticular activation system (ARAS). The latter
 responds to both qualitative and quantitative changes in an ongoing stimulus; it is influenced
 by and at the same time influences the connections to the structures of cortex, subcortex,
 cerebellum, sensory neurons,  motor innervation and vegetative centers. It plays a specific
 role in regulating vegetative and affective behavior.

                                         431

-------
       us

       dB
       105


       95


       65


       75


       65


       55
Schadigung '
deutliche
vegetative Reaktion
mogliche
vegetative Reaktion
                                             •t-
                125    200   320    500    800    1250   2000   3200    5000   0000 Hz
             WO    160    250    400    640   1000    1600   2500   4000   6400
                        Grenzwerte  zur  Beurteilung
                   larmbedingter vegetativer Reaction
                                       Figure 1
Classification of Activation Reactions

     The activation reactions due to high and moderate sound levels include: inhibition of
gastro-intestinal peristaltic activity,  inhibition of secretion of gastric juices and saliva, dila-
tion  of the pupil,  temporary increase  of blood pressure, diminution of GSR, moderate
decrease of stroke volume of the heart, influence on the pulse rate (especially during sleep),
increased production of catecholamines  and steriods. a decrease in skin temperature, and an
increase of cortical blood volume.
     The activation reactions mentioned above comprise in large part the so-called "orient-
ing reflex" according to Sokoloff (1963) and Lynn (1966).  Orienting reactions are necessary
to increase  the readiness for action  to a high  level and to  guarantee the possibility of
immediate  reaction. Repetitive stimuli, however, produce more or less rapid habituation, so
that  in  the course of time  the reaction pattern diminishes or disappears. It  should  be
                                         432

-------
pointed  out  that the orienting reaction and its ability to habituate are necessary for the
human life.
     From experimental studies with noise-induced peripheral vasoconstriction in the skin,
the relations between intensity and frequency of stimuli can  be presented in  a severely
simplified, schematic figure (Fig. 2). This figure is only a basis for discussion and not a result
of research, but it may lead to further research using the concepts of orienting and defensive
reaction. By "defensive  reactions"  we mean  those noise-induced  reactions that do not
habituate with increasing intensity and/or with frequent stimulus presentation.
     In my own experiments (Jansen, 1973) it was possible to  demonstrate orienting reac-
tions that rapidly become defensive  reactions with stimulus repetition. If the sound level is
high enough, defensive reactions will appear immediately. To prove this we simultaneously
recorded peripheral blood volume at the extremities and at the  head. We found that orient-
ing reactions were characterized by constant or even augmented pulse amplitudes (vasodila-
tion) at the head, while pulse amplitudes at the fingers showed a  decrease (vasoconstriction).
          :0
                   keine Reaktion
                                                      TTT
                                                          w
                                                           m
HI: Defensivreaktion
  IffMtiOT!
                             Orientierungsreaktion
                                                        Intensitat
         Beziehung zwischen  Reizhdufigkeit und -intensitdt
                            (Gewohnungseffekte)
                      - schematise?)  nach  Sokoloff-
                                      Figure 2

                                       433

-------
 When the orienting reaction changes to a defensive reaction, then vasoconstriction occurs at
 the head as well as in the fingers.
     As you know,  peripheral blood volume begins to decrease at sound levels of 60 to 70
 db(A); another classical activation or orienting reaction is a change in the GSR (or skin
 resistance) and this begins at a lower level. The skin resistance indicates that physiologically,
 a certain level of activation exists; it is released by affective or emotional stress. Due to its
 high capability of habituation, GSR seems to be characteristic of orienting reactions. It is a
 matter of fact that females show stronger reactions in GSR recordings during noise applica-
 tion than men.                     .
     The indicator of activation so far discussed, the orienting response, indicates, of course,
 something about the severity of the noise exposure, and that there is an influence on certain
 functions, but says nothing about the harmfulness of the noise; this is more appropriately
 indicated by the defensive reactions, which, however, are to be expected first at very high
 sound levels.


 Problem-oriented Experiments

     In order to apply the results and knowledge of noise investigations to the assessment of
 noise in  different practical situations, research in certain  areas should be developed. One
question  in urgent need of an answer is  the question of noise influence during phases of
decreased activation due  to  biological  rhythms, especially circadian  rhythms. Another
important problem is how noise may affect those people who biologically need quiet (for
instance,  shift workers after a night shift). Finally, how noise affects ill persons or those in a
state of convalescence is also an important question.
     The enumeration of these problems leads to the central questions of noise research; the
enumerated results are only small parts  of the complexity of noise effects. Therefore,  it
seems necessary to introduce multivariate research methods to psychosomatic noise research
as is done in other scientific disciplines.  Concerning the question  of noise and biological
rhythms  Griefahn (1973) will report in this  Congress on the  different noise reactions
during  certain periods of the  ovarian cycle. In another project in our  laboratory,  Dams
(1972) studied some moderator-variables in noise situations. He tested the influence of age,
sex, ambient noise and medication with a vasodilative  substance in the modification of aural
and extraaural noise reactions.  He investigated  peripheral pulsations, the pulse rate and the
breathing rate. He thereby demonstrated that in young  female persons, stronger vegetative
reactions were caused than with young male subjects; the reactions in old male and female
persons were  the same but they were much weaker than those of the young. Concerning the
influence of ambient noise, Dams demonstrated that  the "depth of modulation" is impor-
tant for  the amount of the vegetative reaction; older females show stronger reactions in a
high ambient noise (Fig. 3). The latter fact may be caused by the more accentuated connec-
tions between neuro-vegetative functions and emotions in female subjects.
     In my own investigations (Jansen, 1970),  e.g. the results presented at Boston, I found
that there is on the average a negative relation between the vegetative reaction and the TTS.
This means that a subject showing a  high TTS has only  small vegetative reactions when
exposed to higher sound levels than the limits for normal vegetative reactions, and vice versa

                                        434

-------
      Q.

      O
3  50
    A3CD
                                                                   I
                                                                Inilialreaklion

                                                                durchscriniltlichf
                                                                Gfsamtreaktion
                                                                    5 mat 5 mm Ruhf und
                                                                    5 min BreitbondgfrOusch
                                                                    105 dBIA) im Wtchsel
                                                                   t • n : Jhtophyltin - nicol/nof
                                                                         Grundpegel

                                                                   600 mg I - n

                                                                   Gesehlecht dtr Vpn

                                                                   Alter dtr Vpn
                            Vegetative  Schallreaktionen
                Alter  Geschlecht  Theophyllin- nicotmat und Grundpegel
                                        Figure 3
(fig. 4). These experiments were repeated by Dams (1972), Bergmann (1973), Meier (1971)
and Hezel (1972). They found the same relation. Bergmann, in addition, demonstrated that
subjects with good hearing and those with slightly poorer hearing showed different behavior
concerning the relation  between pulse amplitude and  TTS: he concludes that the  quality
(elasticity) of the regulation of blood vessels is different in  the two groups. In particular, the
results showed that the better-hearing subjects showed a negative and the worse-hearing
subjects a positive  correlation between pulse amplitude and TTS (fig. 5).
     Summarizing, it may be concluded that noise stimuli beyond the critical curve limit for
normal vegetative  reaction is 99 dB(A) at maximum, and  that between 90 dB(A) and 100
dB(A), a general hazard to human health must be considered. These  possible disturbances
might be found or manifested in various manners, even in psychic behavior, as there  is no
function in the human  body exclusively affected by noise.

Psycho-physiological Investigations

     Similar to the correlation  between vegetative reactions and TTS,  there might be corre-
lations between vegetative and  psychic reactions  caused by noise.  In an  interdisciplinary
                                         435

-------
              •« 5
.

B

•



20
              E
              i
              i
              -
                                         — Ohr mil der grofleren  TTS
                                         --Mr with grater TTS
                                         y-. 32.i -0.15 FPA
                                         r--0.62
                         —Ohr mil der genngeren TTS
                         —ear with smtller TTS
                         y*27.7-0.19 FPA
                         T: -0.72
                          •10  0  -10 -20  -30 -tO -50  -60 -70 -80 -90  -700%
                           d Finger-Pull-Amplitude IFPA) 20 see ntch Ltrmbeginn
                       A Finger-Pulse- Amplitude (FPA) 20 tec past beginning of noise
            Beziehung  zwischen Finger-  Puts - Amplitude  und  TTS
              Relation  between  Finger -  Pulse - Amplitude and  TTS
                                        Figure 4
study. *•_• investigated this  possibility with  psychomotor test procedures, psychological
classifications and physiological reactions (Jansen, 1962). We tried to define psyc'iic dimen-
sions of the subjects and their reactions to annoying noise by using moderator variables such
as ncuroticism, test anxiety, social desirability, etc. Simultaneously with the noise stimula-
tion we  recorded the  pulse  amplitudes. We confirmed its dependence  or intensity  and
bandwidth. A factor analysis of all psychological and physiological parameters resulted in a
factor "pulse amplitude" (beside other factors) which was not correlated very highly with
other test  factors. We interpret this result as  an indication  that even without psychic con-
comitants-i.e., without a positive or negative  attitude to the noise source  the possibility of
harmful noise influence on the human body exists.
     When meaningful noise  stimuli were applied, it was found that physiologica' reactions
depended on the above-mentioned dimensions of personality (Fig.  6). First of all. extremely
labile subjects reacted  physiologically more strongly  if the noise wa; meaningful. Stable
persons reacted less. If the noise lost its meaningfulness. this difference vanished: there was
                                         436

-------
                       ohnt  t.-n
                                            mil t.-n.
                                                       '—— / --O.I09'
                                                     /	/ . 0.261 J
                                                                            30
                                                                            dB
                                                                            2L


                                                                            :.


                                                                            •i


                                                                            rfl
                                                                      •-
                                                                      \
           '-tO  -20   0   20   tO   60 V. 80  -to  -20   0   20   40  60 '/. 80
                  mittlert Reaction der Fingerputsamplitudc in emer 8 mm Larmphase
iorm 6 rrun Srf/rfiontfperouicft 705 3B(A}
t -n i 750 fig Ttwophytlin • nicofi^of 160 ffim  tor
                                                   atr Vrrloubuny
             6 Vpn. mitttere Ruhehorschwette  8,0 15.7 dB  »Hlr (t096 Hi / 5732 Mi/bnar  Otr»r\//t vp < Vrrtuchrl
                        Vegetative   Reaction  und  TTS2
                      bei  Versuchen   ohne  und  mit  t.-n.
                                         Figure 5
seen only the dependence on intensity and  bandwidth. In a practical sense, this means that
all new and unexpected noises may cause different psychological and physiological reactions
in men depending on the dimensions of personality, but that the reactions become homoge-
neous  if the noise has lost its meaningfulness by habituation, frequent repetition or positive
attitude.
     In additional experiments, Hoffman (Jansen. 1972) found out  that the overall sensa-
tion of annoyance due to a noise is aggravated at peak levels of 90 dB(A).
     We therefore propose establishing "limiting  ranges" instead of a single limit for total
noise exposure. Within this range it might be possible to establish a criterion as "representa-
tive" of expected noise effects.
     As human health is endangered by single noise events as well, it seems justifiable to
demand an assessment of noise not only by the calculated equivalent continuous noise level
l^q, but also by limits for single noise events which must  not  be exceeded even if  the Loq is
below the criteria fixed in standards or laws.
                                          437

-------
F/ngerpulsamp/iludp CM
(18 sec noch Larmbeginn)
^-1 <» <0 5
o o o o
informationslos
75 dB 95 dB
U




Signifikanzen
75ns-


zwischen vegetaliven Schollreaktionen
75 dB 95 dB Ind ^B 1 	 1



ns




ns
5
5
ns
5

S
ns
5



Neuroliker Nichtneuroliker
hohe geringe
Prufungsongsl Prufungsongsl
hemmende fordernde
Prufungsongst Prufungsongst
geringe Soz/ale hohe soziale
Se/bsleinschofzg Selbstemscholzg
Introverliertheil Extrover tiertheil
(weiblich) (monnlich)
in/ormationslos
weifles Rauschen
informationshaltig
Industriegerousch und Ton 3 200 Hz
s = signifikanl
ns = nichl significant
Personlichkeitsmerkmate und vegetative Reaction auf
informationslose und informationshaltige Gerausche
                                       Figure 6
Conclusions

     Reviewing the results described above, it is clear that the relations between high sound
levels and their psychophysiological influences  are quite unequivocal. There are reactions
that  may be judged to endanger human well-being and health. It is obvious, too, that single
noise events  whose  intensities exceed established limits are as important as  equivalent
continuous sound  levels. The  risk  of  hearing damage seems on the other hand to be de-
scribed fairly well  by  Leq, whereas this assessment does not correspond to the extraaural
reactions.
     Concerning the  middle range  of sound intensities, the  assessment should  be based on
psychomotor disturbances (and in  addition with physiological methods) and -  most impor-
tant  - by determining the degree of annoyance.  In the middle range, Leq and especially the
rate of presentation of stimuli are just as important as maximum level assessment.
     In regard to low noise levels, only psychological classifications (and perhaps GSR) may
be used for evaluation of the exposure. Leq here seems to be the preferable method of noise
assessment.

                                        438

-------
     It should be mentioned that a decreasing sound level does not lose its meaningfulness
at corresponding rate, so we have to consider, besides the acoustical characteristics, the
information coupled to the noise.
     Though  the pathophysiology and pathopsychology of noise exposure is not so well
developed as  its  psychophysiology, we know on the other hand some facts of noise-induced
disturbances of human well-being. One of these facts will be reported later when Rohrmann
presents an abbreviated report  on our German Research Society Airport Study. It will be
demonstrated that by means of epidemiological methods and concepts, the complex noise
reaction—especially in the middle range of intensities—can be described much better by
psycho-sociological than by physiological research alone.
     The existing standards and criteria in different countries are in most cases compromise
agreements. These agreements have sometimes been partly  verified;  this refers especially to
the 15 dB(A) difference between day and night  standards. It should be the aim of noise
research to verify the existing standards or to show how we should modify them. But this is
only possible with a coordinated cooperative attack  in many lands by scientists from many
disciplines. Only by cooperation and integration of results  can the complex effects of noise
be assessed adequately.
                                        439

-------
          INDUSTRIAL NOISE AND MEDICAL, ABSENCE, AND ACCIDENT
                     RECORD DATA ON EXPOSED WORKERS*

                                  Alexander Cohen
                 National Institute for Occupational Safety and Health
                                 Public Health Service
                   U.S. Department of Health, Education and Welfare
                                Cincinnati, Ohio 45202

                                    Introduction

     This  paper describes the  first findings  in a  project seeking to determine evidence
coupling severity of occupational noise exposures with occurrences of extra-auditory prob-
lems of consequence to worker health and safety. Specifically, comparisons are reported of
the frequency of medical disorders, absences, and job accidents entered in company records
of workers subjected to high and low noise levels at their workplaces. This evaluation was a
retrospective one, using data contained in worker files of two  manufacturing firms located
in the southeastern  United States. In each company, the entries of interest were extracted
and tallied for the 5-year period, 1966-1970, which was just prior to the establishment of a
hearing conservation program for those exposed  to the higher-level noise.

                                 Sources for the Data

     Approximately 90% of the worker records evaluated  in this study were drawn from a
plant complex which manufactured large pressure boilers.  This facility, referred to as Com-
plex A, consisted of four manufacturing buildings, each divided into numerous shop areas,
bays, and  offices. Key sources of noise within work areas were generated by machinery used
in vertical turning, boring and  facing  of large-diameter boiler sections.  Other high noise
emitting equipment included arc-air flame-cutting tools, air compressors, heavy presses, and
many automatic panel welding machines. Also typical  of the high noise producing opera-
tions in Complex A were chipping and grinding on large nuclear reactor vessels and compo-
nents weighing up to 1,000 tons. Many of these noise operations took place simultaneously,
radiating into many work areas within each building in this complex.
     A secondary source of  record data was a plant engaged in the production of electronic
missile and weapon parts. High noise levels in this plant, called Complex B, were generated
in the  operation of boring  machines, grinders, pneumatic  presses, air  compressors, and
riveting machines. Complex  B was less  than one-fifth the size of Complex A in manufactur-
ing area, and utilized production equipment and machinery far smaller in scale and massive-
ness. Noisy operations in Complex B were  also more localized to those areas where indi-
vidual tools were in  use.
*This paper is a condensation of a report prepared by the Raytheon Service Company (1972) which under-
took this records study via contract (HSM 099-71-6) with the National Institute for Occupational Safety
and Health. Mr. Robert Felbinger served as project director for the Raytheon Service Company.

                                         441

-------
     The  choice of these establishments took account of some considerations which are
mentioned in connection with the procedural aspects described below.

                                 Procedural Aspects

     Repeated noise measurements had been made in various work areas within each com-
plex spanning the same time period as the record data to be evaluated. These noise measure-
ments permitted the division of such work sites into a high noise classification, defined by
the presence of sustained or interrupted noise levels of 95 dBA or more, and  a low noise
classification,  defined  by levels  not exceeding 80 dBA, regardless of temporal pattern.
Division of work areas into more  than two noise classifications was precluded by incomplete
information relating actual exposure time to .specified noise levels, and the probable fluctua-
tion of such exposure conditions from year to year due to changing production schedules or
other factors. Noise surveys  were performed at the outset of this study to confirm previous
readings and to make final decisions regarding work sites to be assigned to the high and low
noise classes. Figure 1  describes  the average 'sound  levels in dBA (re 0.0002 microbar) for
continuous or intermittent noises found in the areas finally chosen. Each data point repre-
sents a work area of one or more exposed persons. The range  and distribution of sound
levels for  the high noise workplaces in Complex A reflected more intense noise conditions
than those shown for Complex B.
     Both plant complexes  included a sizable complement of employees who had worked
for many consecutive  years  in the areas defined as having highland low noise  levels. This
assured realization of a study goal which was to evaluate record data for a minimum of 500
workers with prolonged experience in noisy jobs and 500 workers with comparable experi-
ence in quieter ones. Medical,  accident and  attendance files on all personnel working in the
classified high and low noise level areas in each complex were made available after measures
were  taken  to insure their confidentiality. The record data of interest had  been logged by
medical and administration staffs in each establishment which had remained intact over the
years, lending consistency to the record-keeping process. Initial screening excluded those
persons who were not employed  in the high and low noise jobs for the entire 5-year period
of this  records study,  and/or whose pre-employment health examinations  indicated ear
trouble (hearing problems) or active health disorders. The remaining workers in each com-
plex were  then sampled to form groups of comparable size in high and low noise areas  that
were matched as much as possible in.age, experience at  the workplaces designated, work-
shift, etc.  Table I shows the, age-job experience makeup of the final groups constituting the
sources  of record  data for  the  noisy  and  quiet work locations in Complexes A and B.
Equivalence  in  these variables for the  high  and low noise groups is only  approximate.
Complex A workers were all males but a few females were included in both the high and low
noise  groups of Complex B. No blacks were entered in the  worker groups selected for
Complex B.  Less  than  10%  were blacks in  the Complex A samples. This small number of
females and blacks precluded any efforts to evaluate separately the effects on extra-auditory
 - •   '•-":"    •        "  —""   "     ,*-•'-,".         • - ,.  i -     1
problems of sex or race. Hourly and salaried persons covering a wide range of salary levels
were represente'd in the high and low noise groupings of each complex. Match-ups between
actual job  functions by workers in the high  and low noise sites were few. More will be  said

                                       442

-------
                                      -140-
                                      -130-


                                      -120-
                                      _J	
                   HIGH     «:..           I                    HIGH
                  NOISE    ji.       -10-              NO|SE
                            • • • ••          i              •
                 (95dBAOR ;  !'	
                  HIGHER)   H:       -100-

                            	
                                      -90  -
                                 ••    -  80  -
                                           1
                   LOW     •              I              .     LOW
                  NOISE    -        -  70  -              NOISE
                 (80dBA OR              	!	

                                      -60-
                                      -50-

                                      — i —
                                      -  40  -
                       PLANT COMPLEX         I         PLANT COMPLEX
                           A
 Figure 1 Sound levels in dBA for continuous or intermittent noise conditions observed in the work areas
'constituting the high and low noise groups in complexes A and B. Each data point represents a work area of
 one or more exposed persons.
about this point in the course of the paper. At least 45% of the workers in the high and low
noise groups of each complex worked the first or early shift of the workday.
     The medical, attendance and accident-files for all workers selected for this survey were
searched and  relevant  entries  collated  by research  assistants subject to additional close
review by other members of the investigating team. To protect against recorder bias, none of

                                          443

-------
                                       Table I
         COMPOSITION OF WORKERS BY AGE AND EXPERIENCE IN HIGH AND LOW
                        NOISE GROUPS IN COMPLEXES A AND B.
IGE IN
reARs
BELOW 26
26-35
36-45
46-55
56-65
TOTAL
SAMPLE
DUMBER OF WORKERS
COMPLEX A COMPLEX B
HIGH LOU HIGH LOW
NOISE NOISE NOISE NOISE
60 9 7 0
155 102 21 8
89 138 19 27
89 141 17 28
61 59 2 2
454 449 66 65
YEARS AT
PRESENT
JOB
5-10
11-15
16-20
21-25
26-30

COMPLEX A COMPLEX B
HIGH LOW HIGH LOW
NOISE NOISE NOISE NOISE
345 269 50 10
44 102 16 54
43 41 0 1
6 14 0 0
15 21 0 0

these persons was permitted knowledge as to whether a given set of files  belonged to a
worker in a high or low noise group. All record entries were accepted at face value. In regard
to occurrences  of- accidental injuries, there were notations of only the  type of injury and
body part  affected. Circumstances surrounding such reported accidents were typically not
detailed. Both minor accidents, necessitating dispensary treatment only, as well as  major
ones, involving lost time, were tallied.
     Record entries reflecting health problems included references to disorders diagnosed  by
the attending physician and symptomatic complaints reported by the worker. These entries
were classified into nine different categories of medical problems based on the nature  of the
diagnostic information or symptoms reported. These categories are shown in Tables  II and
III, to be discussed later. Absence data collected over the time period of this evaluation was
coded in two ways, namely, in total days and as the number of discrete events lasting one or
more days. Absences for reasons other than reported illness or injury were  not included—
that is,  only sickness-absenteeism  was considered. One cannot discount the  fact, however,
that a certain number of these reported absences may have been for  reasons other than
bonafide sickness.
                                        444

-------
                                  Results and Discussion
General Findings
     The  accident,  medical and  absenteeism data extracted from worker records for  the
5-year period, 1966-1970, were  evaluated  separately  for Complexes A and B. Figures  2-5
present such  data in  the  form of cumulative percent frequency distributions.1 There  are
plotted the percentages of workers in high and low noise job sites with specifiable numbers
of accidents (Figure 2), diagnosed disorders (Figure  3), and absences, both discrete (Figure
4) and total days (Figure  5), as logged for the 5-year period in each complex. For Complex
A, the distribution curves  for the high and low noise exposed workers show clear differences
for each of these problem indicators. Specifically, greater proportions of the worker group
exposed to high level noise reveal more accidents,  more health disturbances, and greater
amounts of absence than  in the comparison group  not so  exposed. As an illustration,  the
accident data in Figure 2 show  that fewer than  5% of the workers in the quieter areas of
Complex  A had  15  or more  accidents for the 5-year period of this records study. In
'In actuality, these distribution curves have been  plotted in an inverse way to display more clearly dif-
ferences in the  frequency of extra-auditory problems among workers in the high noise group relative to
those in the low noise group. Each point on a given curve represents the percentage of workers in the
specified group  having as many or more of the number of problem occurrences shown on the abscissa for
the 5-year period of record collation.
              PLANT  COMPLEX    A
                      HIGH NOISE GROUP (N-454)
                0	...    NOISE GROUP (N*449)
                -P-—O---6--.6-- -A..-A--.A.
                                       45  50
PLANT  COMPLEX   B
       H.  HIGH NOISE GROUP (N-66)
       -»  LOW NOISE GROUP (N*65)
    15  20  25  30  35 40 '45  50
                               NUMBER OF ACCIDENTS-5  YEAR TOTAL

 Figure 2  Cumulative percent frequency distribution of workers in high and low noise groups in complexes
 A and B with a specifiable number of accidents over the 5 year period, 1966-1970.
                                           445

-------
CO
             PLANT  COMPLEX   A
                      HIGH NOISE GROUP (N-454)
                          NOISE GROUP (N'449)
        PLANT  COMPLEX   B
           o	° HIGH NOISE GROUP (N-66)
           o	o LOW NOISE GROUP ( N-65)
          5  10  15 20  25 30  35  40  45 50       >l    5  10  15  20  25 30  35  40  45  50

                 NUMBER  OF DIAGNOSED  MEDICAL  PROBLEMS  -   5   YEAR  TOTALS

 Figure 3 Cumulative percent frequency distribution of.workers in high and low noise groups in complexes.
 A and B with a specifiable number of diagnosed disorders over a 5 year period, 1966-1970.
CO
              PLANT COMPLEX   A
                      HIGH NO|SE GROUP (N-454)
                      LOW NOISE GROUP (N-449)
          5  10  15 20  25  30 35  40  45 50
         PLANT  COMPLEX   B
             —» HIGH NOISE GROUP (N-
             ---o LOW  NOISE GROUPIN-
65)
2t  .5  10  IS, 20  25,30   35 40  45 50
                         NUMBER  OF DISCRETE ABSENCES  -  5  YEAR  TOTALS

Figure 4 Cumulative percent frequency distributions of workers in high and tow noise group in complexes
A and B with a specifiable number of total days absent over a 5 year period, 1966-1970.
                                            446

-------
CO
             PLANT  COMPLEX    A
             u       HIGH HOISE GROUP (N-454)
                 ----«>LOW NOISE GROUP (N-449)
          10  20 30  40  50 60  70 80  90
             HIGH NOISE GROUP (N*66)
             iow  NOISE GROUP (N« 65)
10  20  30  40 50  60 70 80  90
                         NUMBER OF ABSENCES  (DAYS) -5 YEAR  TOTAL

Figure 5 Cumulative percent frequency distributions of workers in high and low noise groups in complexes
A and B with a specifiable number of total days absent over a 5 year period, 1966-1970.
contrast, 35% of the workers at the noisier worksites had 15 or more accidents in this same
time span and 10% had as many as 40.
     The  medians of the distribution  curves shown in Figures 2-5 indicate that the 5-year
record entries for a typical worker in a noisy area of Complex A include 8-9 more accidents,
3-4 more diagnosed  medical  problems, 40 more days  of absence, and  25 more discrete
occasions of absence than that found for a counterpart worker in a less noisy area of the
same  facility. Statistical evaluation of these differences in medians  found them all to be
significant.
     Differences in  the cumulative frequency distributions of  the  accident-j health and
absence data recorded for workers in the high vs. low noise areas of Complex B, however,
were  riot1 as great as those seen in Complex A. Based on .-median values,  the 5-year record
data for a typical worker in the high noise area of Complex B, relative to a typical worker in
the quiet, show one  more accident, equal occurrences  of diagnosed medical problems, 2
more days  of  absence  and 4  more  discrete instances  of absence.  Though  slight, the
differences  in accidents and  discrete numbers of absence between the high and low noise
exposed workers attained statistical significance.
     The  median numbers of accidents, health disturbances and absences when computed
by individual years within the 5-year collection period in Complexes A and B yielded results
consistent with the overall totals as described above.
     Why Complex A showed greater  differences between high and low noise groups than
Complex  B in the frequency of extra-auditory problems  is conjectural. One possibility is
that the high noise  classification  for Complex A included areas  with  much higher sound
levels than those classified in the high noise group of Complex B (see Figure 1). Apart from
                                         447

-------
this noise factor, the differential risk of injury or illness specific to jobs and work areas
classified in the high versus low noise groups of Complex A might have been much greater
than that of Complex B.
Specific Evaluations

     Different comparisons of the record data  were made to define the influence of age,
length of job experience, work shift and other variables. Select results can be summarized as
follows:
     (1)   The number of accidents per worker was greatest for the younger persons in noisy
          jobs and/or those who had the least experience at such jobs in both complexes.
          This accident rate diminished with increasing age and job experience for workers
          in noisy workplaces, with similar though less obvious changes  noted for those
          located in  quieter ones. For the 5-year period, the youngest workers (25 years or
          below) with least experience (10 years or less) in noisy jobs showed typically 9-10
          more accident occurrences than their peers in quieter jobs, and 8 more than those
          found for the oldest (over  55  years), most experienced (greater than 25 years)
          workers in noise.  These  results agree with other findings in the literature which
          generally report more  accidents among younger, less experienced workers (Mann,
          1944; Hale and Hale,  1972, Freeman et al (undated)). That high  levels of noise
          may  act as an  additional potentiating factor in this context  seems plausible.
          Drawing such a conclusion, however,  presupposes the  same jobs or equally risky
          ones being performed by the subject workers in both the noisy  and non-noisy
          areas. Assurances of these conditions  were lacking for this study, as they seem to
          be for other research'concerned with more general effects of age and experience
          on accidents (Hale and Hale, 1972, Freeman et al (undated))."
     (2)   Younger workers in  both the high  and low  noise level groupings showed  the
          greatest number of diagnosed disorders entered in  their medical files for the
          5-year period. Differences revealing more frequent medical problems for workers
          in the high vs.  low noise jobs were only apparent in Complex A, and  became
          smaller with increasing age.  Variations  in these differences with job experience,
          apart from age, were uncertain.
     (3)   Sick-absences, either in terms of total days or discrete occurrences, were found to
          be greatest for the younger workers, especially those in the high noise level group.
          This amount of absenteeism tended to decrease in the middle age  groups only to
          increase again for the oldest workers. A similar U-type relationship was seen in the
          absence rates of workers in noisy areas  as a function of years of job experience.
          Absenteeism measures for the workers in the low noise group showed no change
          (Complex  A) or increased (Complex B) with advancing age or longer years of job
          experience. The higher rates of absenteeism among young workers in the noisier
         jobs can  be a natural consequence of the  increased  numbers of accidents  and
          health disturbances also noted  in the records of this group. Taken  together, these
          findings may depict the initial strain of coping with a work situation subject to
          intense noise and possibly other stressors as well. Older workers, though showing
          fewer  injuries  and  health  problems in their  files, may  be liable  to more

                                         448

-------
         absenteeism due to greater susceptibility to illness, not necessarily job connected,
         and some loss in recuperative capacity.
     (4)  No consistent differences emerged in comparing frequency differences in recorded
         accidents and medical  problems or amounts of  sick absence as a function of
         workshift for workers in the high and low noise groups in either complex.
     Additional evaluations were  performed to clarify certain aspects of the health data and
elaborate further on the overall results. For example, when sorted into diagnostic categories,
the medical entries filed for workers in both complexes revealed respiratory disturbances to
be most common, irrespective of workplace noise levels (see Tables II, III).
     For workers in the higher noise, however, more respiratory cases involved hoarseness,
laryngitis and sore throats. An undetermined  number of these  ailments could be attributed
to the shouting of workers in communicating in the noisy work sites. Other more frequently
noted disorders  for  workers in the high vs. low noise groupings  fell  into the allergenic,
                                       Table II
        NUMBER OF DIAGNOSED DISORDER BY MEDICAL CATEGORY FOR WORKERS
                            IN HIGH AND LOW NOISE GROUPS

                               COMPLEX A - 5 YEARS
CATEGORY
OF
DIAGNOSED DISORDERS
RESPIRATORY
ALLERGENIC
MLJSCULO/SKELETAL
CARDIOVASCULAR
DIGESTIVE
GLANDULAR
NEUROLOGICAL
UROLOGICAL
NUMBER
HIGH
NOISE
331
196
75
64
50
39
34
29
AFFLICTED
LOW
NOISE
146
86
31
37
21
10
11
14
NUMBER OF
HIGH
NOISE
2152
358
104
114
66
48
49
40
OCCURRENCES
LOW
NOISE
590
118
47
70
30
14
29
15
                                        449

-------
                                      Table 111
        NUMBER OF DIAGNOSED DISORDERS BY MEDICAL CATEGORY FOR WORKERS
                           IN HIGH AND LOW NOISE GROUPS
COMPLEX B - 5 YEARS
CATEGORY
OP
DIAGNOSED DISORDERS
RESPIRATORY
CARDIOVASCULAR
ALLERGENIC

MLJSCULO/SRELETAL
GLANDULAR
DIGESTIVE
UROLOGICAL
NEUROLOGICAL
NUMBER
HIGH
NOISE
56
27
13
2
2
1
1
0
AFFLICTED
LOW
NOISE
59
13
21
3
o
0
0
0
NUMBER OF
HIGH
NOISE
384
28
20
2
2
1
1
0
OCCURRENCES
LOW
NOISE
360
16
33
3
0
0
0
0
musculo-skeletaJ, cardiovascular and digestive categories, especially in Complex A. Symp-
toms and  diagnostic signs here were less specific in nature  or origin as related to noise. In
this regard,  health  examination surveys of workers in noisy  industries have also noted
increased incidence  of circulatory, allergenic and neurological problems of assorted descrip-
tions which have been ascribed to excessive occupational noise exposure (Jansen, 1961,
1969; Shatalov et al, 1962, Anticaglia and Cohen, 1970). At the same time, however, this
research has been  criticized for the inability to control other adverse workplace or job
factors,  apart  from noise, which  may  have influenced the results (Kryter, 1970, Miller,
1971).
    The present study can be similarly criticized since, as already noted, job situations for
workers in the high and low noise groups could not be matched on a one-to-one basis.
Partial equating was tried, using jobs with the same functional titles which were found in the
high and low groups of both Complexes A and B. Comparisons of the record data by select
                                        450

-------
job titles and noise levels yielded differences which were most often in directions showing
either more numerous accidents, medical problems or absences under the higher level noise
(see Table IV).  The magnitudes, and in a few instances, the directions of these differences
for the specified jobs were quite variable when compared to one another, and to the overall
differences based on the total group comparisons. This variation stresses the importance of
"the. job.factor and the need to better account for it in this type of research. On this latter
point, a  most effective approach would be to contrast the incidence of extra-auditory
problems in the same workers before and after noise reduction takes place, especially if such
controls did not materially change the  nature of the job operations or alter other non-noise
hazards attendent to the total work situation. There exists an opportunity to implement this
approach as  part of a follow-up effort  to this record study. Specifically, hearing conserva-
tion measures stressing the use of personal ear protectors have been in effect in Complex A
and B for the past two years, and it is  planned  to evaluate again the medical, accident, and
sick-absence entries of the subject workers subsequent to the establishment of this program.
Reduction in  individual worker noise exposures through ear protectors should diminish the
occurrences of medical, safety, and related sick-absenteeism problems if, in fact, noise was a
causal factor. Positive findings here would also indicate the extent to which efforts designed
to reduce noise hazards to hearing can also offset extra-auditory problems as well.
     This additional work is slated to be undertaken only at Complex A. This is to capitalize
on the large  group of workers available for study, and the  fact that their initial record
entries, as reviewed above, showed the  clearest indications of increased health and accident
problems among workers  in the high noise workplaces. Any conclusions regarding noise as a
major or contributing cause of these extra-auditory problems would be incautious at this
time, and should be deferred pending the outcome of the follow-up  study. Indeed, the data
available at present offer only circumstantial evidence.

                                       Summary

     Entries  in  medical,  attendance, and accident files for over 500 workers situated in
noisy plant areas (95  dBA or higher) were compared with 500 others in quieter workplaces
(80 dBA or less) gathered over a 5-year period in two plant complexes. Most of the record
data were taken from the larger  of the two  establishments which manufactured  boiler
equipment, and  which was  also  found to have generally more intense noise conditions.
Workers subjected to the high workplace noise  here showed greater numbers of diagnosed
medical  problems,  absences for illness,  and job related  accidents  than were  noted for
workers in the quieter areas of the same plant. Medical diagnostic categories showing signifi-
cant differences between high and low noise level jobs were respiratory (hoarseness owing to
shouting in noise) and  non-specific allergenic,  musculoskeletal, cardiovascular and gastro-
intestinal disturbances.  Differences between high and low noise  level groups showed wide
variation when sorted by job type, suggesting that the increased frequency of extra-auditory
problems can be greatly affected by this variable, regardless  of noise level.  Evidence for
increased medical, absence, and  accident problems in comparing the high and low noise
exposed groups in the second plant complex, which produced electronic missile and weapon
parts, was not as prominent as that noted in the first one. A follow-up study is planned to

                                         451

-------
                                     Table IV
      TYPICAL OCCURRENCES OF MEDICAL PROBLEMS,SICK-ABSENCE, AND ACCIDENTS
        CLASSIFIED BY JOB TITLES FOR WORKERS IN HIGH AND LOW NOISE GROUPS.
                          COMPLEX A - 5 YEAR TOTALS
JOB
TITLE
FOREMEN
PESTS &
[NSPECTS
ADMINIS-
TRATIVE
TOTAL
SAMPLE
NOISE
LEVEL
HIGH
LOW
"HIGH"
LOW
HIGH
LOW
HIGH
LOW
N
34
138
48
38
10
45
459
449
AVERAGE OCCURRENCE PER WORKER
MEDICAL
2.0
3.7
3.9
3.6
4.5
0.8
3.9
0.4
DISCRETE
ABSENCE
9.1
3.1
.6
5.7
74.7
4.7
30.3
4.2
TOTAL
ABSENCE
21.6
8.3
.5
18.1
107.1
9.4
49.8
8.8
JOB
ACCIDENTS
4.5
4.8

3.4
7.6
0.3
9.0
0.4
                          COMPLEX B - 5 YEAR TOTALS
JOB
TITLE
&CHINE
)PERATORS
ASSEMBLY
WORKERS
TOTAL
SAMPLE
NOISE
LEVEL
HIGH
LOW
HIGH
LOW
HIGH
LOW
N
16
16
46
38
66
65
AVERAGE OCCURRENCE PER WORKER
. MEDICAL
PROBLEMS
4.7
4.7
7.3
7.3
4.8
5.5
DISCRETE
ABSENCE
14.0
4.4
16.6
7.5
10.8
7.0
TOTAL
ABSENCE
17.5
9.9
26.3
38.2
18.2
16.5
JOB
ACCIDENTS
1.8
0.3
2.0
1.4
1.7
0.7
   NOTE:   VALUES  BASED ON CELL SIZES OF 10 OR MORE
evaluate entries in the records of the same workers over a period subsequent to the establish-
ment of an ear protection program in the first plant complex studied. Reduction in indi-
vidual worker noise exposure through  ear protectors should diminish the  occurrence of
medical, sick-absence, and accident problems if, in fact, excess noise was a causal factor.
                                   References

Anticaglia, J.  R., and Cohen, A. Extra-auditory effects of noise as a health hazard. Amer.
    Indust. Hyg.Assoc. J. 31, 277-281 (1970).
Freeman, F., Goshen, C. E. and King, B. The role of human factors in accident prevention.
    (Operations Research Inc. Contract Report SAph-73670) Public Health Service, Dept.
    Health, Educ., Welfare, Wash. D.C. (undated).
                                      452

-------
Hale,  A.  R. and Hale, M. A review of industrial accident literature. National Institute for
    Industrial Psychology, London, England (July 1972).
Jansen, G. Adverse effects of noise in iron and steel workers (in German). Stahl u. Eisen, 81,
    217-220, 1961.
Jansen, G.  Effects of noise on the physiological state. Proceedings of National Conference
    on Noise as a Public Health Hazard,  Amer. Speech and Hearing Assoc. Kept. No. 4,
    89-98, (Feb. 1969).
Kryter, K. D. Effects of Noise on Man (Academic Press, N.Y. 1970), p. 508.
Mann, J. Analysis of 1009 consecutive accidents at one ordnance depot. Indust. Med. 13,
    368-374, 1944.
Miller, J. D. Effects of noise on people.  Kept. No. NTID 300.7, Environmental Protection
    Agency, Wash. D.C. 20460 (Dec. 1971).
Raytheon Service Company. Industrial  noise and worker medical, absence,  and accident
    records. Contract No. HSM 099-71-6, Burlington, Massachusetts (1972).
Shatalov, N. N., Saitanov, A., and Glotova, K. V. On the state of the cardiovascular system
    under conditions of noise exposure. Labor Hyg.  Occup. Diseases, 6 (7), 10-14 (1962).
                                         453

-------
        FACTORS INCREASING AND DECREASING THE EFFECTS OF NOISE

                                   D. E. Broadbent
                  Medical Research Council, Applied Psychology Unit,
                                 Cambridge, England

     Human beings have a limit to the number of features of their surroundings which they
can perceive in any limited period  of time, and therefore anything which happens in the
environment has  to  compete  with  other events for their attention. Until about 1960, a
number of the effects of noise  could be explained simply by considering that intense sounds
have a tendency to win in such a competition. On this view, a man in  a noise would show
failures of perception  because important signals would fail to be analyzed while  he  was
being 'distracted' by the noise. Physiological changes  could then be explained as  due to
compensating mechanisms which attempt to combat this distracting effect.
     Since 1960,  however, it has become increasingly  clear that this analysis may confuse
cause and effect; it may be that exposure to noise produces a change in the state of the man
and that this changed state is reflected in failures of selective perception. The evidence for
this changed interpretation comes from a number of experiments, but  for those who have
specialized in other areas it will be sufficient to quote a result from Wilkinson (1963). In
this experiment, men were asked to perform a task with and without the presence of  100-dB
noise, each condition being met when the  men were in a normal  state  and when they  had
been deprived of sleep for 24 hours. Three main points appeared in the results. First, the
usual harmful effects of sleeplessness were reduced by the presence of the noise. Second, the
effects of noise itself on the task  were harmful if the men had slept normally, but  if they
were sleepless, noise  actually improved their performance.  These  two findings suggest that
noise  creates some general  state of arousal which reduces  the effects of sleeplessness,  and
which only impairs efficiency if the man is already as highly aroused as is desirable. Such a
conclusion is supported by a great deal of related evidence (Broadbent, 1971).
    Wilkinson's third finding  is similar  to that of many other experiments on noise: the
effects are greater when the task has been continued  for a prolonged  period in the noisy
conditions. There are two  possible explanations for this. One is that  the work  produces
some kind of change in the man, which we may call 'fatigue' if we can avoid defining that
word  too precisely, and that the noise affects the man more when he has been 'fatigued'.
The other possibility is that the noise gives rise to some cumulative effect, so that the longer
one stays in the noisy environment, the more incapable one becomes of performing even a
novel  task. Since  Wilkinson's experiment,  like most others, started the noise at the same
time that the man started  work,  it is impossible to  distinguish these possibilities from his
results. In this  paper, I am going to outline three recent  as yet  unpublished experiments
from  our laboratory which show that noise changes  the  general state of the perceptual
system rather than merely distracting it;  and the first is one which indicates that noise gives
a cumulative effect on the man which may persist even when the noise itself has ceased.
     Hartley has used the same task as Wilkinson, in noise and in quiet,  and in each case tor
a work period of 20 minutes.  The main interest of the experiment lay  in the condition to
which the man was exposed during the 20 minutes before the measured session; he might be

                                        455

-------
 reading, or he might be performing the task, and  in either case he might be in quiet
 conditions or else in 100-dB noise. There were therefore eight experimental conditions by
 which one can separate the different theoretical possibilities already mentioned.
     At first sight the results are complex, but this is only a superficial difficulty. Perform-
 ance is worse if the man has worked for a previous 20-min period, and also if he spent the
 previous 20 min in a noise environment. (This latter  fact in itself shows that the noise has
 changed the state of the man rather than simply acting as  a 'distractor', since it makes him
 inefficient even when he is subsequently working in quiet and there is no noise to distract
 him). Thirdly, performance is worse when noise is present  than when it is absent. The  key
 point however in distinguishing the different theoretical explanations  is whether the effect
 of noise is bigger when the man has previously worked than when he rested; and it is not.
 On the other hand, the effect of noise is greater when the man has previously been exposed
 to noise. The combination of all these findings looks complicated, but in fact the conclusion
 is simple; noise affects people at the end of a work-period because it produces a cumulative
 change in  their state, and not because  it affects them more  when they are 'fatigued' by
 work.
     The next experiment I wish to discuss considers whether the general state which noise
 produces is one which might change the function of the senses and perceptual mechanisms.
 McLeod has  devised a method  of measuring the integration time of the  eye, following
 techniques introduced by Allport (1968). The  basic method is to present a series of lines on
 a cathode-ray tube, one after another, each separated by 1 cm from the previous one. The
 man controls the number of lines presented before the equipment returns to the original  line
 and repaints  it. His task  is to set the number  of lines present  at such a  value that  the
 addition of one more would cause the whole display to appear to flicker. At this point the
 man is seeing simultaneously a number of lines which have all beea presented to the  eye
 successively, and this is therefore  a method of assessing temporal resolution in the visual
 system.
     As is well known, in low levels of illumination the integration time of the eye increases,
 which is obviously adaptive in extracting as much visual information as possible from a weak
 signal. McLeod's results show however that a similar change occurs in loud  noise, the two
 effects  interacting so that the  effect  of noise is  only  statistically significant  at 0.25
 foot-lambert and not at 40 foot-lamberts.
     We thus have evidence that noise produces a general change in men exposed to it,  and
 that this change affects the intake  of sensory information.  The last point I wish to make is
 that the perceptual changes are of such  a kind that they would resemble 'distraction'. In a
 series of studies by myself and  my wife, we  presented  visually mixtures of relevant  and
 irrelevant information, and found that noise impaired the ability to select the one from the
 other. In the most definitive trials, we used words interleaved so that the  odd-numbered
letters   came  from  one  word  and  the  even-numbered ones  from the  other,   e.g.
 LjEuAnDgEIRe; a  difference of colour between the two words was also introduced, so that
some men could be asked to identify  the black and word and  some the red one. If now the
exposure duration was increased until correct identification  of the word took place, noise of
about 100 dB had no harmful effect on threshold for the easy word (capital letters, black
print, common word), the threshold for the  difficult word (small letters,  red print, rare

                                        456

-------
word)  was increased by  about thirty per cent. Control experiments using the words sepa-
rately showed no such effect; thus the effect of the noise was to impair the functions which
would normally suppress  the large and conspicuous, but irrelevant, letters.
     To summarize, these experiments show changes which cannot be due to distraction by
noise, but which would have the effect of producing failures of perception;  noise may not
always distract,  but rather make men more distractible. This view fits well  with results
discussed elsewhere in  this meeting, and particularly with the complexity of the effects on
performance  described by Dr.  Gulian;  tasks in which the maintenance of attention is no
problem will  show no impairment by noise,  nor will tasks performed when other conditions
are unarousing. The disruption of performance in tasks which do require selective percep-
tion such as the Stroop test is however to be expected and so is the increased deterioration
which  will follow arousing conditions such as deprivation  of control  over the situation
(Glass and Singer, 1972). It is particularly worrying that Hartley's results, like those of Glass
and Singer, show a persistence of the effects  of noise after the stimulation has ceased.  One is
reminded  of  the finding of Jansen (1959)  that family disturbances are significantly more
common amongst those who work in noise, and of the higher rate of admissions to Spring-
field Hospital from streets with a high exposure  to aircraft noise (Abey-Wickrama  et al,
1969). In each  case, there may be factors other than  noise which might  be alternative
explanations  of the effects; but equally there is the  speculative possibility  that there is a
chronic effect of noise in distorting perceptual input, which disturbs personal relationships
as well as laboratory tasks. There is a need for further work  on chronic effects of noise, if
only to eliminate this possibility.
                                     References

Abey-Wickrama, I., a'Brook, M. F., Gattoni, F. E. G., and Herridge, C. F. Mental hospital
     admissions and aircraft noise. Lancet, 1275-77 (1969).
Allport, D.  A.  Phenomenal simultaneity and the perceptual moment hypothesis.  Brit. J.
     Psychol,. 59, 395-406(1968).
Broadbent, D. E., Decision and Stress, Academic Press: London (1971).
Glass, D. C., and Singer, J. E. Urban Stress: Experiments on Noise and Social Stressors.
     Academic  Press: New York and London. (1972).
Jansen,  G.,  Vegetative functional disturbance caused by noise. Archiv. fur Gewerbepatho-
     logie und Gewerbehygiene. J 7, 238-261 (1959).
Wilkinson,  R. T. Interaction of noise with  knowledge of results and sleep deprivation. J.
     Exp. Psychol., 66, 332-337. (1963).
                                         457

-------
           EXAMPLES OF NOISE-INDUCED REACTIONS OF AUTONOMIC
             NERVOUS SYSTEM DURING NORMAL OVARIAN CYCLE

                                 Barbara Griefahn
                               University of Bochum
                                    Essen, BRD

    The extent of reactions of the autonomic nervous system caused by ergotropic stimuli
are dependent on the vegetative status of the test person. People with trophotropic circula-
tion function, which means those with small pulse rate, small  cardiac output, and great
peripheral resistance, show greater vasoconstriction during noise exposure than people with
ergotropic circulation function (Jansen 1969, Jansen and Schulze  1964, Matthias and Jansen
1962, Oppligerand Grandjean 1959).
    Women with normal ovarian cycle show in the premenstruum a characteristic increase
of pulse rate, minute output  respiration frequency and basal  temperature; in the post-
menstrual period, they show a decrease of these values. Thus, the vegetative functions of a
fully-developed woman are characterized by a cyclic succession of the trophotropic follicle
phase  and the ergotropic corpus-luteum-phase (Brehm  1959, Doring 1948, 1953, Goodland
and Pommerenke 1953, Artner 1960).
    The question  concerned here is  whether these changes of vegetative status are great
enough to cause different responses to the same stimulation.

Method

    12 females were tested in  57 experiments, 2-4 days before and  after the beginning of
menstruation and 2-4 days before and after ovulation  for two complete cycles. The tested
persons were 17-39 years old, their cycle lasted between 27 and 32  days. They were
healthy; none of them had a hearing loss greater than 20 dB. None of the test persons took
hormone preparations or circulatory preparations.
    None of the test persons had measured basal temperature, so it was  necessary to find
other  parameters which point to  the existence of normal or anovulatory  cycles. In accord-
ance with the results of other authors (Doring and  Feustel 1953, Brehm  1959) we found the
values of pulse rate and respiration  frequency  in the  corpus-luteum phase significantly
greater than in the follicle phase. Therefore and because of the careful selection of the test
persons it is very probable that we tested only within normal ovarian cycle.
    During the experiments, the test persons sat in a comfortable chair in a sound-proofed
room.  After  a quiet period of 15 min a  white  noise of 95 dB(A)  with a duration of 2
minutes was presented 5 times, each time followed by a quiet period of 3 minutes.
    During the experiments we recorded finger-pulse amplitudes, pulse rate and respiration
frequency.

Results

    On each day of examination (Figure  1) we found a great initial  decrease of the finger-
pulse  amplitudes. In the first half of the cycle (the follicle phase; 2 - 13 days or, in Figure 1,
                                       459

-------
                 0123

                      Breitbandgerausch  95 dB(A)
                                                                     :
                                                                 2-£  Tag
                                                                des Cyclus

                                                                16 Versuche
                                                                 8 Vpn
                                                                    b
                                                                 11-13  Tag
                                                                des Cyclus

                                                                15 Versuche
                                                                11 Vpn
                                                                17-19 Tag
                                                                des Cyclus

                                                                IS Versuche
                                                                10 Vpn
                                                                    d
                                                                25-27 Tag
                                                                des Cyclus

                                                                11  Versuche
                                                                6  Vpn
                  Fingerpulsamplitude  bei  Larmbelastung
                      zu  unterschiedlichen Cycluszeiten
Figure 1  Relative amplitude of the expansion of the finger (vasoconstriction effect) associated with heart
beat in response to a 95-dB(A) white noise burst of 2 rnin duration (cross-hatched). The value from 0 to 0.5
min on the abscissa is taken as lOO^. The parameter is the number of days since menstruation.
                                         460

-------
curves a and b) this initial decrease is followed by a gradual increase, in the second half (c
and d: the corpus luteum phase) the finger-pulse amplitude remains near 85%. (The dif-
ferences between the  follicle and corpus-luteum phases of the cycle are significant until the
24th second of noise exposure).
     The results will  be much clearer after calculation of only one average value for each
day of examination  (Figure 2). Contrary to the values within one phase (a-b  or c-d) we
found that the values in the follicle phase are very significantly smaller than those of the
corpus-luteum-phase.
     These results are concordant to the results of some other authors, who found a greater
vasoconstriction in trophotropic  than in ergotropic  people (Jansen 1970, Jansen  and
Schulze 1964, Heinecker,  Zipf and  Losch 1960); and they are in agreement with the rule
found by Wilder (1931),  which  postulates that the reactions of the autonomic nervous
system will be smaller, the greater its excitation.
     After dividing the total reaction into the initial and the residual reaction (the latter
being the value at the end of the  noise burst) we found that the curve of the residual
reaction  follows  closely the total reaction; the values of the initial reaction show great
deviations, so we could suppose a psychic reason. But after calculation of the initial reaction
within the second cycle after beginning of the experiments (Figure 3) we found the same
curve. Psychic conditioned reactions are dependent on habituation, so that the extent of the
reaction  becomes  less after a lot of stimuli.  Therefore it is very  probable  that psychic
displeasure has only an insignificant influence on the results.
     We thought that the cause of the different reactions are the ovarial hormones. There-
fore,  with the use of the crosscorrelation function we calculated the delay  between the
curve of the reaction values and that of the hormonal level.
     Usually, crosscorrelation functions will be calculated  in  order to discover a periodic
event within an apparent  stochastic curve. In  this case,  with well-known periodicity, the
hormonal curve is displaced against the residual reaction until the best agreement is found,
as shown by a maximum  of the crosscorrelation function. The place of the maximum, in
this  case  the number of  days until temporal agreement, is therefore the degree for the
probability of a causal relation.

     The data used for crosscorrelation (Figure 4) are those of the remaining reaction, the
daily  urinary excretion of estrogens, after Brown, Klopper and Loraine (1958) and the daily
gestagen level in plasma, after Neill et al (1967).
     The maximum of the crosscorrelation function between estrogens and reaction is r = 4,
between  gestagens and  reaction r =  1  (Figure 5). That means that the alteration of the
reaction follows the increase of the estrogens after 4 days, the increase of the gestagen level,
however,  after only  1  day. Therefore  it seems that the lower reaction is  caused by the
appearance of the gestagens.
     If the extent of vasoconstriction caused by noise exposure is really an indicator for the
degree of excitation of the autonomic nervous system, the ergotropic situation in the second
half of cycle will be caused by the increase of the gestagens, the trophotropic situation in
the first half by the decrease of the gestagens and not by the increase  of the estrogens.

                                        461

-------
*/•} bei BB 95 dB(A
Ql
tn

o wi
                         cr
                                                           •A
es


«
                                                               sr.
                                                               f X
                        a/b    a/c
a/d    b/c
b/d
c/d
                   or 2 - 4  Tag  des Cyclus   c = 17-19 Tag des Cyclus
                   b:11.-13.  Tag  des Cyclus   d-25-27 Tag des Cyclus

                   • Haltewert    O Gesamtreaktion    A  Imtialreaktion
                Larmbedingte  periphere  Vasokonstriktion

                  an  verschiedenen   Tagen   des   Cyclus
Figure 2  Noise-dependent peripheral vasocon strict inn on different days of the menstrual cycle. The dotted

curve shows the change in finger volume immediately after onset of the 95-dB(A) noise, the solid line the

effect at the end of the 2-min noise burst. The dashed curve is the average change over the entire 2 min.
                                        462

-------
75
s
5 *
o» 70
CO
CD
S 65
•o
1 »
O
3
h.
&
c
so








/
/•
9 '
*
6


P"
/ ,
/ /
.//
'/

.U.
• *m
/N
f






^
\
a



Imtiolrtoktion
















t t 8 12 16 20 2t 28
Tag des Cyclus
Figure 3 Comparison of initial vasoconstriction effect during the first (solid curve) and second (dotted
curve) month of observation.
Respiration Frequency

     Ovarial  hormones have a stimulating  effect on respiration frequency (DOring  1953,
Doringet al,  1950, Wilbnuul ot al.. 1<)51>).
     According to Harmon (1933) and Stevens (1941). noise also effects a small increase of
metabolism and by  that a higher level of  carbonic acid concentration in blood, which is
accompanied by an  increase of respiration  frequency. Though effects of hormones and of
noise on metabolism and therefore on respiration rate are very small, it seems possible that
both together will effect a significant reaction.
     Two  to four days after the beginning of menstruation, when there is only a small
production of estrogens and  gestations, respiration rate shows, according to this theory, an
insignificant change (figure 6): 2 - 4 days before ovulation, when the first peak of estrogens
appears,  we found a significant increase. Two to four days after ovulation, when the estro-
gen level decreases, and at  the same time gestauen level  increases, the reaction is significant
too. Shortly  before  the beginning of the menstruation,  after the decrease of the estrogens
and gcstagens. the reaction is insignificant.
     The maximum of the crosscorrelation  function (Fig. 7) between estrogens and respira-
tion rate appears at T  = 27  or   1. This precocious agreement is explainable, because the
values of the estrogens are excretion values.
                                         463

-------
        -
        I









— Cestagene
— Oestrogene


...
r^m
^__
• •••


«





^_ . _
...









....
— • -

- - -




. _ -


_.-





_ . _

_ _ _




-•-





....

...




.._
^^^




• . —

— —


• . V


^_



- . *






- -
- - _>





*»••
1
— • —



....







1 •
_



—
...
_._


• V •*

— —


•
	

IV.


-^— —
_ _ «

r:.:


...
—
...




_ . -,



- --
•^^M
_ • —
	 1
Fingerpuls-Hattewert
Atmungsreaktion











^r



irrr

                             S
J2
20      2i       28
       Tag des Cyctus
          Werte  fur  die  Kreuzkorrelationsfunktion
Figure 4 Data used to calculate cross-correlation functions between hormone levels (gestagens, solid lines,
estrogens, dot-dash) and autonomic response to noise (residual vasoconstriction, dashed lines; change in
respiration rate, dotted).
                                       464

-------
• -
)
Relativwerte
1 o +



'



'


N,
--


\



\



\
N


^.
\


X
\
\
N
\
N

X
\
\
\

\

\
->.


S



\




/



7


.

4 8 12 16

.

**

/
'-
•^IMM
^•M «
/
—
	
.

'
-

t
9
•
•


•
•


'
•


^
'
•


'
•

\

•

N

H

\

i

\
\
•

\
.• •
A


•


'
'



•

'•
•
28 32 36 r
(Toge)
Kreuzkorrelationsfunktionen
                           Figure 5 Cross-correlation function between hormone levels and vasoconstriction cause by noise.

-------
        t>
        HI
        01
107,5

 *
105.0
        !T   102.5
         e
         c
            KW.O
         c
         •)
         o»
            97.5
            S5.0
                            .XI-
                            t *
                                                 16
                                             20
                                                                 2t       28
                                                                Tag des C/c/us
        0
        A
        o
—  Durchschnittlicrie Grsamtrtaktion alter Versuchspersonen
• -  Gesomtreaktion der Versuchspersonen Nr 2. Nr 3. Nr 4 und Nr 10
—  Gesamtreoktion der Versuchspersonen im  I  Cyclus der Versuchsreihe
    Gtsamtreaktion der Versuchspersonen im  II Cyclus der Versuchsreihe
                       Durchschnitttiche  Gesamtreaktion
                 der Atmungsfrequenz  und  Vergleichskurven
            bei  Larmbelastung zu unterschiedlichen Cyctuszeiten
Figure 6  Change  in respiration rate caused by 95-dB
-------
                                            Oestrogene/Atmungsrcoktion
                                                             •.
                                                                  36    T (Togel
                     Kreuzkorrelationsfunktion
Figure 7 Cross-correlation function between estrogen level and change in respiration rate induced by noise.
                                 REFERENCES

1.  ARTNHR. J.:  Vegetative  Ausg.mgskige uiul Cyclus.  Arch, (.ynaek.  /'/.?  (>0)
    379-392
2.  BRE11M.  U.:  Der  Kreislauf der Fran, scin Alternsuang. seine Anderung  with rend
    Puhertat,  Klinuikterium.  Zyklus. Sehw;inj:ersch;irt uiul  Wochenbett und seine  Heein-
    flussung durch  Zufuhr  von Ovarial-  und   tioiuuloiropen  HDnnoneii. Habil-Schrift.
    l-Yanklurt Main. 24. Sept. ll>5'>
3.  BROWN.  H. B.. A.  KLOPPI R und .1. A LORAIM  The Urinary l-xcrction olOestro-
    genes. Pregnaediol and (ionadotrophins During the Menstrual Cycle. J. Fmiocrin. /7
    (1958)401-410.
4.  DORINC. (i. K., H.  II. I  ()1 S( IIKI  und B. ()( H\V\I)I  Ueitere  rntcrMichiinpon iiber
    die Wirkung der Sexualhormone aui" die Atmung.  Plliigcrs Arch.  252. < I(»5(M 216-230
                                       467

-------
 5. DORING, G.  K. und  E.  FEUSTEL:  Uber Veranderungen der Pulsfrequenz im
    Rhythmus des Menstruationszyklus. Klin. Wschr. 31II, (1953) 1 000-1  002
 6. DORING, G.  K.:  Uber  rhythmische  Schwankungen von  Atmung und  Korper-
    temperatur im Menstruationscyclus. Pflugers Arch. 250, (1948) 694-703.
 7. DORING, G. K.: Uber Veranderungen der Atmung wahrend des Cyclus. Arch. Gyneak.
    ; 82 (1953) 746-751.
 8. GOODLAND, R. L. und W. T. POMMERENKE: Cyclic Fluctuations of the  Alveolar
    Carbon Dioxide  Tension During the Normal Menstrual Cycle.  Fertil. and Steril. 3,
    (1952)393-401
 9. HARMON, F. L.: The effects of noise upon certain psychological and physiological
    processes. Arch. Psychol. (N.Y.), 747(1933) 140-164.
10. HEINEKER, R., K.-E. ZIPF und H.-W. LOSCH: Uber den Einfluss kSrperlichen Train-
    ings auf Kreislauf und Atmung. II. Mitteilung: Trairiingseinflub auf die Reaktionsweise
    des Kreislaufs bei Anwendung von Kalte-, Flickerlicht und Larmbelastung. Z. Kreislauf-
    forsch. 49, (1960)924-935.
11. JANSEN, G.: Relation between Temporary Threshold Shift and Peripheral Circulatory
    Effects of Sound. Physiological Effects of Noise, 67-74 Plenum Publishing Corporation
    (Vortrag Boston) AAAS, 28.12.1969 Edited  by: Bruce L. Welch and Annemarie S.
    Welch.
12. JANSEN, G. und J. SCHULZE: Beispiele von Schlafstoningen durch Gerausche. Klin.
    Wschr. 42, (1964)132-134.
13. MATTHIAS, S.  und G. JANSEN:  Periphere Durchblutungsstorungen durch La'rm bei
    Kinderri. Int. Z. angew. Physiol. 19, (1962) 201-208.
14. NEILL, J. D., E. D. B. JOHANSSON, J. K.  DATTA und E. KNOBIL:  Relationship
    Between the Plasma Levels of Luteinizing Hormone and Progesterone During the Nor-
    mal Menstrual Cycle. J. Clin. Endocr. 27/II, (1967) 1167-1173.
15. OPPLIGER,  G. und E. GRANDJEAN: Vasomotorische  Reaktionen der Hand  auf
    La'rmreize. Helv. Physiol. Acta 17, (1959) 275-287
16. STEVENS, S. S.: The Effects of Noise and  Vibration on Psycho-Motor Efficiency.
    Harvard University, Psychological Laboratory,  OSRD Report 32 and 274 (1941)
17. WILBRAND, U., CH. PORATH, P.  MATHAES und R. JASTER: Der Einfluss der
    Ovarialsteroide auf die Funktion  des Atemzentrums. Arch. Gynaek. 191,  507-531
    (1959)
18. WILDER, J.: Das "Ausgangswert-Gesetz" - Ein unbeachtetes biologisches Gesetz; seine
    Bedeutung fur Forschung und Praxis. Klin. Wschr. 10, (1931) 1889-1893.
                                     468

-------
        THE INFLUENCE OF NOISE ON AUDITORY EVOKED POTENTIALS

                           J. Gruberova, S. Kubik, J. 2alcik
                          Institute of Occupational Hygiene
                             Bratislava, Czechoslovakia
     The auditory  evoked potential, which is defined as an electrical response of brain to
acoustical stimuli,  has recently attracted the attention  of audiologists for the purpose of
objective audiometry.
     The auditory evoked potential detected from the scalp is a nonspecific response widely
distributed over the scalp, with  a maximum in the vertex region.  A typical example of
auditory evoked potential is shown in Figure 1. The sequence of negative and positive waves
in characteristic.
     Seventeen healthy experimental persons were investigated before and after noise expo-
sure. The recording electrodes were placed, according to the ten - twenty  system (Jasper,
1958), in positions O^, P£, C^ andT3- The reference electrode was placed on the chin.
     Recorded  potentials were amplified by a Schwarzer EEC apparatus and added by a
multichannel analyzer NTA  512  of KFKI Budapest. One hundred  responses were always
summed up.
     A 500-msec acoustical stimulus of level about 90 dB with irregular pauses (from 0.5 to
5.0 sec) was used.  Auditory evoked potentials were investigated at three frequencies: 500,
1000 and 2000 Hz. Stimuli were delivered by earphone directly to the ears.
                       100
200
300
400
                     Figure 1 Typical average auditory evoked potential.
                                        469

-------
     Each experimental subject was investigated in four sessions. At the beginning of every
session the auditory evoked  potentials to the three frequencies were recorded and then a
white  noise of level about 90 dB was applied.  The period of white noise application was
changed at every session. Four periods of 0.5, 1,  1.5 and 2 hours were used. Immediately
after termination of the noise, the auditory evoked potentials at three frequencies were
again investigated.
     Ten of the subjects were also tested under the same conditions by classical audiometry.
In comparing the auditory evoked potentials recorded before and after noise, we  concen-
trated  our attention on  amplitude  differences of waves N2 —  ?2- The amplitude was
measured  peak to peak. The mean differences of the whole group were calculated. The
results are shown in Fig. 2.
     A statistically significant difference was found in parietal, central and temporal records
after 0.5, 1, and 1.5 hours of exposure. No statistically significant difference was found
after 2 hours of exposure.
     In Figure 3 can be seen the mean shifts of acoustical threshold of ten persons, obtained
by classical audiometry.
     In  the evaluation of results  obtained after  the noise,  it is  necessary to  take into
consideration the fact  that the white noise has,  besides an influence on the hearing appara-
tus, also an influence on the state of vigilance of the subject, and the amplitude of auditory
evoked potential does depend on vigilance.
                                                                        T,
Figure 2 Decrease in the peak-to-peak difference between N2 and P2 of the auditory evoked potential after
noise exposures of various durations (abscissa) for four different electrode placements (parameter). Stimu-
lus frequencies 500 Hz (thin line), 1000 Hz (dashed line), & 2000 Hz (thick-line).
                                         470

-------
                  dB

                  10 —
                              -I—
                               30
Figure 3 Changes in behaviorally-determined auditory sensitivity (TTS) after exposure of various durations
{abscissa) to 90 dB(A) noise, at 500 Hz (thin line), 1000 Hz (dashed line), and 2000 Hz (thick line).
     The changes of amplitude of auditory evoked potentials with changes of vigilance are
explained by Fruhstorfer and Bergstrom (1969) in terms of a decline in activity of certain
brain functions which are essential for the maintenance of vigilance.
     The Roumanian author Edith Gulian showed also that noise application produces a
decrease of vigilance of experimental persons. She found in her experiments that continuous
noise of 90 dB after 1.5 hours produces a clear decrease of vigilance. In EEC she found a
clear decrease of alpha index at the end of a 1.5-hr session.
     Jerison  (1959) observed the  mental performance of experimental persons exposed to
noise levels of 85 and 115 dB in four 0.5-hr intervals. He found the greatest decrease of
performance after  1.5 hour.  In the last half hour an  improvement in performance was
observed.
     Those  time  relations can  also be  seen  in  our experiments. While after 1.5 hr, the
decrease of amplitude is statistically significant, after 2 hr of noise exposure the decrease
was not significant.
     Amplitude of auditory evoked potentials is influenced, then, by the state of vigilance
on the one hand and changes of hearing produced by white noise exposure on the other.
     We can still compare the changes of auditory evoked  potentials and the changes of
auditory threshold measured by classical audiometry. From classical audiometry results we
can see the maximal decrease of hearing at the end of the first hour, while in  auditory
evoked responses there is a statistically  significant decrease of amplitude even at the end of
1.5 hr. This difference  can be explained in no other way than by changes of vigilance.
     The results obtained by the method of classical audiometry are responses mediated by
the specific auditory pathways, while the auditory evoked potentials are mediated by non-
specific pathways-that is, by the ascending reticular formation of the  brain stem and the
nonspecific thalamic system — and these structures are essential for the state of vigilance.
                                         471

-------
                                    Literature

Fruhstorfer, H., Bergstrom, R. M., Human vigilance and auditory evoked responses. Electro-
     enceph. din. Neurophysiol., 27, 346-355 (1969)
Gulian, E., Effects of noise on an auditory vigilance task, Rev. Roum. Sci. Social. Psychol,
     10,175-186(1966)
Gulian, E., Effects of noise on reaction time and induced muscular tension, Rev. Roum, Sci.
     Social Psychol, 13,33^5 (1967)..
Gulian, E., Effects of noise on arousal level in auditory vigilance, Acta Psychol (Amst.), 33,
     388-393, (1970)
Jasper, H.  H., The ten twenty electrode system of the  international federation, Electro-
     enceph. din. Neurophysiol, 10,371-375 (1958).
Jerison, H. J., Effects of noise on human performance, J.appl psychol, 43, 96-101 (1959).
                                        472

-------
 SOME DATA ON THE INFLUENCE OF NOISE ON NEUROHUMORAL SUBSTANCES
                          IN TISSUES AND BODY FLUIDS

                                  Lech Markiewicz
                        Department of Physiology and Hygiene
                           of the Central Research Institute
                               for Labour Protection
                                  Warsaw, Poland


     Noise, especially at high intensities, is a strong bionegative stimulus acting as a stressor
on the organism.  Changes in sympathetic nervous system reactivity and endocrine activity
constitute non-auditory effects of the noise. Yet  long-lasting  exposure to noise of high
intensity  has a considerable effect  on endocrine system reaction. This was pointed  out in
previous research  work (see references).
     The level of catecholamines can  be the indicator of stress magnitude as well as the
modified sympathetic  nervous  system reactivity.  Many authors have found an  increased
excretion of catecholamines in urine due to noise of high intensity, especially when it comes
unexpectedly and is short-lasting.
     The results of experiments given here prove that noise of high intensity and various
frequencies have some bearing on the catecholamine level in the organism.
     Experiments were carried out  on white rats by exposing them to noise for three hours
daily. The noise frequencies 50, 4,000 16,000 and  20,000 Hz. and intensities from  100 to
130 dB were used. The  catecholamines were estimated in tissues (brain, heart, suprarenal
gland), blood and  urine.  The urine for catecholamine determination was collected in
metabolic cages during 24 hours, beginning immediately  after the exposure to noise. Blood
and  tissues were collected on finishing the experiment. Determinations were performed after
1, 3, 6, 8 and 24 weeks of exposure. In this report some of the more interesting results are
presented.

1. Catecholamines in urine

     Stimuli of frequency 50 Hz cause an increased level of excreted noradrenaline (NA) in
urine only within the  first week of exposure. In the later  phases of the experiment the
excretion of NA remains at the level of  the control value (Fig.  1). On the other hand,
concentration of  adrenaline (A) is higher not only within the first week but also in the third
week of experiment.
     Similar results were  also obtained in the second series of experiments, when  a noise of
frequency 4,000  Hz and  of intensity 100 dB was applied. An increase of NA excretion was
evident within the  first  week only, becoming slightly lower than the control value in the
third week of experiment, but the  more intensified A excretion still  remained in the third
week of experiment. However, the influence of an acoustic stimulus of frequency 16,000 Hz
leads to a decrease of the excretion of both catecholamines; although there is  a marked
tendency for the return of noradrenaline to the control value within the sixth week of
experiment, adrenaline is still being excreted at a low level. Prolonged experiments with

                                         473

-------
  10
M9/24h
1.25-1
            50
   1.00-
   0.75-
   0.50-
   0.2 5i
-^
£
  0.3-
  0.2-
  0.1-
     0  1  3
frequency c/s 50
intensity dB 108
                                    4.000 16.000 20.000
                     NoradrenaLine
   111!     |ll  Lili
   0136     013   0136   01
                         0136
                    Adrenaline
                                         I
                                    01    6  24 (weeks)
                013   0136
                 4.000      16.000
                  100      128

                  Figure 1 Catecholamines in urine
                                    0 1
                                       20.000
                                        130
1
24 (weeks)

-------
20,000 Hz show a low level of catecholamines in the sixth week of experimentation, and
their recovery only after 24 weeks of applying  the acoustic stimulus. Such a long lasting
deviation from normal indicates a lesser degree of adaptability of the adrenergic system to
an acoustic stimulus of high frequencies than to a  stimulus of, shall we say, 4,000 Hz.

2. Catecholamines in blood

     Since the previous experiments have shown different catecholamines excretion in urine
when lower (50 and 4,000 Hz) and higher (from 16,000 to up 20,000 Hz) frequencies were
applied, the level of NA in blood was examined just after termination of a long-lasting series
of acoustic stimulus activity. Figure 2 shows the results, indicating an increased level of NA
in the third as well as  in the first week of the experiment, when noise of 4,000 Hz and  100
dB intensity was applied.
     On  the other hand, long-lasting  stimulus activity of  20,000 Hz did not lead  to an
increase, but on the contrary, to a slight decrease  of NA level in the blood. Nevertheless,  it is
interesting that the A level in blood was, in the initial phase of experiment, increased in
spite of the fact that the excreted amount of A in urine was insignificant.

3. Noradrenaline in brain

     Application of low or high frequencies leads to increased amount of NA in brain tissue,
which is illustrated in fig. 3.  This increase is observed with 50  Hz frequency in both the
third  as well as the first week and it returns to normal not sooner than after 8 weeks time. It
is worthwhile to emphasize  that  with the 4,000-Hz stimulus  the changes are the least
marked. However, in this case an increase of NA in, the brain is found.
     Stimuli of 20,000 Hz applied cause a two-phase reaction to appear; first there is an
increase of NA level in the brain, lasting till  the 6th week of experimentation. Only the
determinations done after 24  weeks, when the experiment has been stopped, have shown
outstanding decrease (nearly by half) of NA level  as compared with the control value.
     Experiments  concerning behavior  of  serotonin carried  out  together  with  Dr.
 Markowska have shown that an increase of this compound occurred in the brain as well as in
 blood with simultaneous excretion of its metabolite 5-HIAA (5-hydroxy-indolacetic acid) in
 urine.
 Comment

     The experimental results, as described above, have proved that metabolic disturbances
 of catecholamines and serotonin are produced under the influence of long-lasting acoustic
 stimuli. They point out the lack of adaptability to this harmful factor of human environ-
 ment.  Stress reaction at the initial phase is characterized by the increase of catecholamines
 in urine, with high level these compounds found in blood and brain at the same time.
     Longer-lasting stimuli, especially of high frequencies, lead to changes in synthesis and
 degradation  of biogenic amines.  Indeed, the influence of noise causes  reduction of

                                        475

-------
       ug/1
      20.0 -I


      15.0-
S     10.0H
NA
        5.0-
                                           03        013 (weeks)

  frequency c/s 4.000                                20.000

  intensity dB  100                          122              122
                                Figure 2 Catecholamines in blood

-------
                                                                 I  T
        10     20      50    100            500   1.000        4.000          20000

        /g
     1.00-


     0.75-
-^

     0.50 H
     0.251 M  •  •     M                M M  M      MM      mm



           0138                  013        06        4 (weeks)

     frequency c/s   5O                           4.000              20.000

     intensity dB  108                            100                 130
                                     Figure 3 Noradrenaline in brain

-------
monoamineoxidase activity with simultaneous intensified excretion of cortisol and thyroid
gland hormones.
     The negative stimuli of this kind, as applied in our experiments, become factors that are
able to cause serious changes in  functioning of the whole organism which, if repeated
habitually, may lead to disease.
     A particular part is played by acoustic stimuli which are on the border of audio- and
ultrasonic devices and are getting more' common. As Konarska and I have pointed out, even
destruction of Corti's organ does not prevent metabolic disturbances of catecholamines and
corticosteroids.

                                  REFERENCES

1.   Arguelles A. E. - Introduction to Clinical Endocrinology editor: E. Bajusz, Basel 1967
     p. 123.
2.   Hrubes V,Benes v. - Activities Nervosa Superios, 7 p. 165, 1965.
3.   Markiewicz L. Konarska M. - Prace Centralnego Instytutu Ochrony Pracy, 19 p. 233,
     1969.
4.   Markiewicz L., Markowska L., - Prace Centralnego Instytutu Ochrony Pracy 18, p.365,
     1968.
5.   Sackler A. M., Weltman A. S., Jurtshuk P. -Aerospace Med. 31, p. 749, 1960.
                                        478

-------
     STRESS AND DISEASE. IN RESPONSE TO EXPOSURE TO NOISE - A REVIEW

             Gosta Carlestam, *Claes-Goran Karlsson** and Lennart Levi**

     Noise has been defined as any unwanted sound, the most prevalent "waste products"
of pur age. Numerous authors claim to have shown that noise provokes physiological stress
reactions, not only as concomitants to the distress reactions implicited in the very definition
of noise, but also through reflex stimulation of the auditory nerves and on to the hypothala-
michypophyseal system. It is  occasionally claimed that exposure to  noise  can cause  a
number of diseases belonging to the  field  of psychiatry and internal medicine, either by
these or by some other mechanisms.
     The  purpose of this  paper is to examine critically the evidence in favor of these
hypotheses and to report,  in summary, a study  conducted  at the laboratory for Clinical
Stress Research.** At the  National Institute of Building Research*, David Wyon and his
associates are studying noise as a component in  the indoor environment.

Noise and physiological stress

     The  term "stress"  is used here in the sense  that Selye described it, namely, the non-
specific response of the  body to any demand made upon it; a stereotyped, phylogenetically
old adaptation  pattern primarily preparing the  organism for physical activity, e.g. fight or
flight.
     It is conceivable that in the dawn of the history of mankind, noise very often was a
signal of danger or else  of a situation requiring muscular activity. In order to survive, the
human organism  had to prepare itself for activity, inter alia by  the non-specific adaptive
reaction  pattern  defined as stress. More often than not, noise  in  today's  industrialized
societies  has a meaning very different from what  it had  during stone age. Yet, according to
one  hypothesis, our genetically determined psychobiological programming still makes us
react as if muscular activity would be an adequate  reaction  to any sudden, unexpected or
annoying noise stimulus. True, it can be argued that some authors have demonstrated not an
increase but rather no reaction or even a decrease in hormonal activity in response to noise
(Bugard,  1955; Sakamoto, 1959). One explanation for  this  controversy might be that the
measurements have been made  at varying intervals after noise exposure. Various endocrine
systems can react after various intervals or even in different directions, some of the reactions
being diphasic.  Accordingly, some reactions,  present immediately after the exposure, may
have disappeared or changed direction in some instances but not in others.
     As one may expect, the reaction pattern  to noise  is not entirely non-specific but is
partially  conditioned by the specific  characteristics of  the reacting  organism. One  man's
meat may be another man's poison. Comparing adrenal hormone reactions in response to
noise in  healthy controls  with those of patients  with cardiovascular diseases  or schizo-
phrenia, Arguelles et al. (1970) found increases in hormone excretion in all three groups, the
reactions  in the two patient groups, however, being significantly more  pronounced.

 •National Institute of Building Research Box 27 163, S-102 52 Stockholm
**Laboratory for Clinical Stress Research, Karolinska sjukhuset, Pack, S-104 01 Stockholm
                                        479

-------
     Horio et al. (1972) exposed rats to various noise levels, measuring corticosteroid levels
in the adrenal glands. They found rapid increases in concentration reaching a maximum
after  15 minutes of noise exposure. At moderate noise levels, the corticosteroids soon re-
turned  to  initial levels. At  higher noise levels, however, corticosteroid concentration re-
mained elevated over longer periods, interfering  with the circadian rhythm.
     Measuring 17-ketosteroid excretion in urine in response to meaningful and meaningless
noise of moderate intensity, Atherley et al..(1970) found that the meaningful but not the
meaningless variety did induce physiological stress reactions.
     In an experiment conducted at our laboratory, 22 young female IBM operators were
studied in  their usual work. In half of the group, the  noise level produced by their IBM
machines increased 6 dB from one day to the next during four consecutive days, the noise
levels being 76, 82,  88 and 94 dB-C, respectively. The other half were subjected  to the
same noise levels but in the opposite order (i.e., 94, 88, 82, and 76 dB-C, respectively). The
noise level normally  prevailing in the office was 76 dB—C. Every  working day started with
two hours  of rest without noise exposure, followed by three 2-hour work periods with noise
exposure as indicated.
     Contrary to what might be expected, the subjects reported only minor increases in
self-rated fatigue  (figure  1) and  "distress" (figure  2).  Although these ratings  increased
slightly with increasing noise,  the  rating differences between the highest and lowest noise
levels were conspicuously small. The corresponding epinephrine and norepinephrine excre-
tion levels (figures 3 and 4) were low or moderate and the changes from control to noise
periods and from low to high noise levels were usually non-significant. Thus,  not even the
objectively rather considerable noise levels used were particularly potent as stressors. This
may be due  to the familiarity of the noise and to the generally positive attitudes of these
subjects to the job per se and to the experiment. It is conceivable that such factors may have
counteracted the stressor effects of the noise. Briefly, then, noise may be a potent stressor
under some circumstances and in some individuals, but need not generally be so.

Noise and disease

     Sakamoto (1959) found  that more than 50%—i.e. a rather high proportion-of the
inhabitants living close to an airport complained of  various  types  of somatic distress,
possibly induced by the aircraft noise.
     In  epidemiological  studies,  several  authors (Mjasnikow,  1970; Andriukin,  1961;
Shatalov et al., 1962; Ratner et al., 1963) report an increased incidence of hypertension in
workers exposed to  high noise levels. According to Mjasnikov, this increase  in morbidity
manifests itself after 8 years of exposure, reaching a maximum after 13 years of exposure.
     Similarly, other authors (Jerkova and Kremarova, 1965; Andrukovich, 1965; Strakhov,
1966;  and Dumkina,  1970) found  an increased incidence of "nervous complaints" in
workers habitually exposed to higher noise levels. Living in areas close to a noisy airport was
accompanied  by increased number of admissions to psychiatric hospitals (Abey-Wickrama et
al., 1969 and 1970). However, the causal implications of this statistical relationship can be
seriously questioned (Chowns, 1970).
     Jensen and Rasmussen Jr.  (1970) inoculated mice with various infectuous agents,
before or after exposing them to noise.  It was found that those inoculated with stomatite
                                        480

-------
oc
                    Ext'eme 5 i
                  Notot all 1
                                       94 dB (C)
                                    A
 08-10  10-12 12-14 14-16
[: jFRIDAY
   ~~l TUESDAY
                                                                    FATIGUE
                                                               88dB(C)
                                                         82 dB (C)
                                                        J]
                                    i
08-10  10-12 12-14 14-16
[jUJTHURSDAY
[     [WEDNESDAY
                                 4
08-10 10-12 12-14 14-16
  [J WEDNESDAY
  ~~I THURSDAY
                                                                                                                  76dB(C)
                         i
                                                                                                                   t
                                                                                                           08-10 10-12  12-14  14-16
                                                                                                           (•   J TUESDAY
                                                                                                           [     1 FRIDAY
                       Figure 1. Self-rated "fatigue" under different noise conditions and during different times of the day.

-------
 EXTREME
                                       DISTRESS  SCORE
12

11

10
9
oo 8
K)
7






- T
A





'





011.1
•








































NdB(C) I24BIC] naBK.


J
fl1

NOT AT All 6' — •— •- *— • — f~* — ? • • I'
ot-w r-u u-u u-w ot-io w-i2

1



12 -U
I | rfiinAv [ 1 THURSDAY

1


J

U-K 01-10 10

;
.
u

\



»



u-u u-w


1



M-K)
| jWfliNESDAY |

i



.



i'




W-t2 12 -V U-W
| TUESDAY
^]TUEf.rAY [ | WEDNESDAY f^^] THURSDAY [ [FRIDAY
Figure 2. Self-rated "distress" under different noise conditions and during different times of the day.

-------
         Adrenaline
1
to

I
7
t
5

*
3
2


"3/nin





. 76dB-C 82







1


X


X
x
X
x
X
X
X
X
X
X
x








X
x
X
X
x
X
X



X
X

X
X
X
X
X
X
x
x
OI-M »-
X
X
X
X
X


L i


X
X
x
X
X
X
X
X

x
X
X
X
X
X
X
x
x
1






xT
>
X
X
X
x
x
X
X
x
X
X





dB-C 88c

^
X
X
x
X

x
x
x
x
x
x
x
x




X
X
X
X





X
X
X
X
x
X"
x
X
X
X
x


I ,
x
'
x
X
X
X
X
n n-n «-i*  io-u n-w w-







*

X
X
X
X
X
X

X
X
X






IB-C



p

X
X
X
X
X
x
X
X
X
X
It M-M /«•






x




'
X
X
X
X
X
X
X
It 11-






X
X



'
X
X
X
X
X
X
n w







X
X








?fdB-C



X
X
X
x
X
X
X
x
X
•H M-
TUESDAY WEDNESDAY THURSDAY







X
x




LX
X
X
X
X
X
X

ro ;o-
X



x
X
X
x
x
x
X
X
X
X

1 tt-
X


X
3
X
X
x
X
X
X
X
x
X
A* *-








» Tim*
FRIDAY
     II
         Noradrenaline
      t


X

x

X

x
x1
X

X

X

X

X

X
x
X
X
x
X
X
x
X
x

x^





















X

X"

x

X

X

X
X

x

x

X

x

X
X
X
X
X
X
x
x

^










X
X
X
XI








X5
X
X
X
X
X
X
X
X
X
X











X
X
X
x^







X
X
X

xl
X
X
X
X
X
X

                                     X
                                         x
X ~
X
X
X
x
X
x
x*
X
X
x
-X
X
.x
x
x
X
x












X
x

x

x

x




x1
X
X

X

X

X
x

x

X

X

••

x
x

x

X
x
X
X









x
X
X
X
x
X
x
x





X
X
X
X
X
x
xl
X
X
X
X
X
X
X
X

X

X

>

X

X
ff
X
X
X
x
X
x
x
X
X
X
X







X
X
X
x-






T
X
X
X

x

X
X
X
X
x
X
x
X
X
XI
X
X
X
X
X
X
X
X
X
x
           O*-M io-i»  a-*  H-I*    «-»  W'li a-*  «-<*    «-io  »-«  /»-*  *-/«   »-»  »-«  /a-»  «-« TfMl
             TUESDAY      WEDNESDAY      THURSDAY        FRIDAY

Figure 3.  Urinary excretion of adrenaline and noradrenaline during four consecutive days with increasing
noise level.
                                           483

-------
      Adrenalin*
s
7
6
5
3

4
1



I
• "ftdB-C 88dB-C


it
/
/
/
/
1 r4-1
J
X

X
x
X
X
X
X
X
T
>
X

X
X





s>

x
X
X

X
X
x






1
'lx
x*
x
X
X
X
X
x
X
X

X
X
X
X
X

X
X







X

x
X
X
X
X
x







/
x
X
x
x
X
x
x
x

82dB-C 76dB-C


X
X
X
X
X
X
X
X

pL.
x
X

X
x
X
X
X
X
X
>
^
X
x
X
x
_L






t
X
X
X
x
£
x
x
X
X
5
X
X
X
x
xK
X
X
X
^x





OHO w-12 ft-w NH» M-W n-B a-n n-it ot-w w-a n-n *H« «-/• w-a a-w «-« r/«f
TUESDAY WEDNESDAY THURSDAV FRIDAY
      Noradrcnallne
4"9/"
"1
It
n
i«
ii
15



n

•

.





J-

X
X
run



"
x
X
X
x
x
x
x
x








7
X
X


;
X
X
XI
X
x





1
xj
<
X

X
X
X
.Xl
X
X
X
X
X
X
v>\


/

s

'•

x1

X

X
X
s
X
X
X
X
X
X
X
x<
f^


• *



X
X
X
X
x
x
x
X
X
r^>1








*
r.
                                          x
                                            X
                                           -X
                                            X

                                          X

                                          X
        ono w-n n-H H-W
           TUESDAY
oi-io to-a  Q-H «-/*
 WEDNESDAY



X





X

X
X
X
X




xH LX
^ K
X
X
<^1 ixj
7
;
X
x
X
x
X
&d








x
X
x
X
1
X

-------
virus just before noise exposure were more susceptible, whereas those inoculated after the
exposure were less susceptible, than non-exposed controls.
     Reviewing studies  on noise and mental disease, Lader  (1971) concludes  that noise
exposure does not generally increase psychiatric morbidity but might be of some etiologic
significance in neurotic and anxious subjects.
     Briefly, then, some of the physiological reactions found  in response to noise exposure
seem  to be closely related to the non-specific  physiological reaction pattern  defined as
"stress". Stress has been hypothesized to be one of several pathogenetic mechanisms acting
by increasing  the "rate  of wear and tear" in the organism. Although some circumstantial
evidence has been presented, there is still no proof.
     Some epidemiological studies seem to indicate a higher occurrence of "psychosomatic"
and mental disorders in  subjects  exposed to prolonged and rather intense noise. However, it
should be kept in mind that such an exposure is often accompanied by exposure to a variety
of other potentially noxious stimuli. In addition, various segregational forces  may "sort
out"  particularly susceptible individuals to noisy, unpleasant and/or pathogenic environ-
ments.
     Accordingly,  we have  to conclude that the  evidence in favour of noise as a major
pathogenetic  environmental agent is rather shaky. To solve this controversy, future research
should focus  on controlled intervention studies with  an interdisciplinary and multifactorial
design.
                                     References

Abey-Wickrama, I., Brook, M. F., Gattoni,  F. E., and Herridge, C. F.: Mental - hospital
     admissions and aircraft noise. Lancet 1275-1277 (1969).
Abey-Wickrama, I., Brook, M. F., Gattoni,  F. E., and Herridge, C. F.: Mental - hospital
     admissions and aircraft noise. Lancet 467 (1970).
Andriukin, A. A.: Influence of sound  stimulation on the development of hypertension.
     Clinical and experimental results. Cor Vassa 3:285-293 (1961).
Andrukovich,  A.  L: Effect of industrial noise  in winding and weaving factories on the
     arterial pressure in operators of the machines. Gig.Tr.Zabol. 9:39-42 (1965).
Arguelles, A.  E., Martinez, M. A., Pucciarelli, E., and Disisto, M. V.: Endocrine and meta-
     bolic effects of noise in normal, hypertensive and psychotic subjects. In: Welch and
     Welch (Eds.): Physiological Effects of Noise, Plenum Press, New York-London, pp.
     43-56(1970).
Atherley, G.  R. C., Gibbons, S. L., and  Powell, J. A.: Moderate acoustic stimuli: the
     intervelation of subjective importance and certain physiological changes. Ergonomics,
     13:5:536-545(1970).
Bugard, P.: Presse med., Paris 63:24 (1955).
Carlestam, G., and Levi, L.: Urban Conglomerates as Psychosocial Human Stressors. Stock-
     holm, Royal Ministry for Foreign Affairs, (1971).
Carlestam, G., Noise - the Scourge of Modern Society Ambio Vol. No. 3 1972 Stockholm.
Chowns, R. H.: Mental-hospital admissions and aircraft noise. Lancet 467 (1970).

                                        485

-------
Dumkina, G. Z.: Some clinico-physiological investigations made in workers exposed to the
     effects of stable  noise.  In: Welch and Welch (Eds.): Physiological Effects of Noise,
     Plenum Press, New York-London, p. 346 (1970).
Horio, K., Sakamoto,  H., and Matsui, K.: Adrenocortical response to noise exposure. Re-
     print from: Joint Meeting of International Societies for Hygiene Preventive and Social
     Medicine, 29 Oct. -1 Nov (1972).
Jensen, M. M., and Rasmussen Jr., A. F.: Audiogenic Stress and susceptibility to infection.
     In: Welch and Welch (Eds.): Physiological Effects of Noise, Plenum Press, New York-
     London, pp. 7-20 (1970).
Jerkova,  H., and Kremarova, B.r Observation of the effect of noise on the general health of
     workers in large engineering  factories;  attempt at evaluation. Pracovni Lekarstui
     17:147-148(1965).
Lader, M. H.T Responses to repetetive stimulation. In: Levi, L. (Ed.): Society, Stress and
     Disease: The Psychological Environment and Psychosomatic Diseases, London,  New
     York, Toronto: Oxford Univ. Press, pp. 425429 (1971).
Levi, L.  (ed.): Society, Stress and Disease - The Psychosocial Environment and Psychoso-
     matic Diseases. Oxford University Press, London (1971).
Levi, L.  (ed.):  Stress  and Distress in Response to Psychosocial Stimuli. Pergamon Press,
     Oxford (1972).
Mjasnikow, A. L.: In:  The pathogenesis of essential hypertension. Proceedings of the Prague
     Symposium 153-162 (1970).
Ratner, M. V., Medved, R. A., Film, A. P., Skok, W. I., Rodenkow, W. F., und Makarenkow,
     N. A.: Thesen des Berichtes der allunionswissenschaftlichen Tagung iiber methodische
     Probleme  der Larmwirkung auf den Organismus. Institut fur Arbeitshygiene und
     Berufskrankheiten, AMW, UdSSR (1963).
Sakamoto, H.: Endocrine dysfunction in noisy environment. Report L Mie Medical Journal.
     9: 1:39-58(1959).
Sakamoto, H.:  Endocrine dysfunction in noisy environment.  Report  II. Mie  Medical
     Journal. 9: 1:59-74(1959).
Shatalov,  N. N., Saitanov. A. O., and Glatova, K. V.: On the State of the cardiovascular
     System  under conditions  of exposure-to continuous  noise.  Report T-411-R,
     N65-15577 Defense Research Board, Toronto, Canada (1962).
Strakhov, A. B.: Some questions of the mechanism of the action of noise on an organism.
     Report N 67-11646, Joint Publication Research Service, Washington, D.'C. (1966).
                                       486

-------
                   SOME LABORATORY TESTS OF HEART RATE
                          AND BLOOD VOLUME IN NOISE

                                   Karl D. Kryter
               Stanford Research Institute, Menlo Park, California 94025
Introduction
     Two soundproof rooms, decorated as residential living rooms, were arranged so that
two subjects placed in each room could be monitored continuously for EKG and peripheral
blood flow. The subjects in each room were separated by a drape so that they could not see
each other.  The subjects, seated in easy chairs, were asked to read  novels or magazines
during the test sessions. By means of hidden loudspeakers, noise could be introduced into
the test rooms.
     Each daily test session  lasted for two continuous hours (except for one 5-minute
break). The subjects were instructed to try to behave as  though they were resting and
reading in their own homes. They were also told tha,t they, would not be exposed to noises
at a  level any. greater than they might hear in a home and that there was to be no regular
pattern of noise, or no-noise, during any  session. Six adult housewives, age  25-45 years,
served as subjects in the two pilot studies.                                         .
     The data for  the blood  volume from a photoelectric plethysmograph, and the heart
rate, determined by means of an electrocardiograph, were each averaged for each successive
10-sec epoch during the test sessions except for a forced 10-sec period that occurred every 5
min to clear the storage register of the computer.

Results

     Study /-Repeated Sessions of Exposure to a Similar.Pattern of Quiet-Noise-Quiet.

     Figure  1  gives the  average heart rate in beats per second for  4 subjects who were
exposed to a "quiet" (35 dBA) ambient background with 1-4-min bursts of noise at a level
of 90 dBA interspersed at more or less random intervals. (See top section of Table 1 for
details of procedure.)
     It is seen that during the periods of quiet  for any one session, the average heart rate
was less  than during the noise interval. This would, of course, suggest that the bursts of
noise caused an increased stress-arousal reaction in the subjects as compared to their some-
what more relaxed physiological state during the periods of quiet. Of equal interest is that
the average heart rate for both the quiet and noise conditions showed a progressive decrease
from the first session to the third daily session.
     As seen in  Figure  2, the noise also caused an increase (although  of smaller relative
magnitude)  in  physiological  tension measured  by peripheral  blood flow. However, the
average blood volume during each of the different sessions was very similar. It is to be noted
that  the units for measuring blood volume were subtracted from the number "50" in order
that  an increase in the  score, as for heart rate,  would indicate an increase in "stress", i.e.
greater peripheral vasoconstriction with reduced peripheral blood volume.
                                        487

-------
           Q = 35 dBA Ambient (Quiet)
          N* - Bursts of Pink Noise at 90 dBA
               Bursts of Noise of 1 — 4 minute Duration
               Period Between Bursts, 7 — 10 minutes
81.0 	


80.5 	

/y.b 	
78.6
78.5 	
Q
o 78-°
o
UJ
uj -IT r
cc
HI
a.
77 n
<
"- 76.5 —
a:
*£ ir n
Uj /U.U
X
75.5 	
75.0 —
•JA n 	
73.5 —

Q

N*




AVERAGE

Q



76.7





—

Q


N*








AVERAGE


75.6


Q


	
a

N*
•
	 	 76.9

AVERAGE
1





Q
—
Q

N*




AVERAGE



Q
	


	
                                SESSIONS
   Average
of All Sessions
         TA-8755-5
          Figure 1 Average heart rate during pilot study.
Each of four  subjects in  three daily sessions of quiet  interrupted
by  bursts  of noise.
                              488

-------
                                   Table 1 Sequence of test sessions and acoustical conditions
                                                   for pilot studies
                                Study I:   Four Subjects (A, B,  C,  and D)
Session
1
2
3
Acoustic
35 dBA Pink 90
7-10 minute ' 1-4
duration
Same
Same
Condition
dBA Pink
Q
minute '
duration
as 1
as 1
                                                                               35 dBA Pink
                                                                               7-10 minute
                                                                                    duration
oo
Study  II:   Four Subjects  (A,  B,  E, and  F)
      1  Session
                 Acoustic Condition
1

2


3
l.-M -.._
1
: 4

1
:5

A & B
E & F
A & B
E & F
A & B
E & F
A & B
E & F
A & B
E & F
Q (35 dBA,
N (85 dBA,
NQN
QNQ
QQQ
«5 
-------
     Study II—Daily Sessions of Exposure to Different Patterns of Quiet and Noise.

     The basic data of this study for heart rate are summarized in Figure 3 and the results
for peripheral blood volume  in  Figure 4. (See lower section of Table  1 for details of
procedure.) We see in Figures 3 and 4 that:
     (1)  Unlike the findings  in  Study I, there is  no consistent  relationship  between the
         presence of noise and increased stress as revealed by increased heart rate or
         decreased peripheral blood volume; for example, bursts of quiet during ambient
         noise, or burst of noise during ambient quiet both caused some apparent decrease
         in stress.
     (2)  for the average overall sessions there was a somewhat greater heart rate when the
         ambient  was noise, whereas the blood volume indicated that the least amount of
         stress was present when the ambient was the noise.                    :
     We  feel, however, that the results cannot be interpreted  in a meaningful way with
respect to  the  effect of the  presence  or absence of noise  on these physiological stress
reactions, but rather that there was an interaction effect with the alternation from session to
session of  the  patterns of noise and  quiet that occurred within any  one session. This
possibility is revealed in Figure 5 where it is seen that during the second or (repeat) session
for a given acoustic sequence  there is a decided increase in stress as measured by heart rate.
This increase, as is seen oh  the right-hand portion of Figure 5, occurred for all of the
individual conditions  scored  separately; that  is, the heart rate as found during the two
ambient conditions and also between the two "burst" conditions.
     This result is contrary to the adaptation that occurred with repeated sessions of the
quiet-noise-quiet sequence in  Study I  (Figure 1) and must be related to the interposition
between  the repeat sessions, for a given acoustic sequence,  of sessions utilizing different
acoustic sequences or patterns of stimulation. Qearly, changing test conditions from session
to session had a decided effect upon the heart rate of the subjects, regardless of the ambient
and non-ambient acoustic conditions that were utilized within  any one session.
     On  the other hand, peripheral blood volume, as shown in Figure 6, showed a small
decrease  in  "stress" between  the .first  and second sessions of a given sequence of acoustic
conditions. However,  it is suggested that these peripheral blood volume data are probably
not as reliable an indicator of what is usually considered in the present context as "stress,"
as is the measure of rate of heart beat, for the, following reasons: (a) the changes in blood
volume, as measured, are insignificantly small; (b) there appeared to be no consistent trend
of the average  blood volume from session to  session during Study  I; and (c) the  rather
extreme degree  of vasodilation (as indicated by  a lower number on our blood volume
measure) during all sessions of Study  II (about a  25 percent lower score than found in
Study I)  was perhaps due  to  the fact that during the  period Study II was conducted the
outdoor ambient temperature was generally high, on some days exceeded  100°F.  Even
though the test  chamber was  air-conditioned, the temperature in the room was somewhat
higher during extremely hbt days and we suspect that the general vascular condition of the
subjects was influenced to some extent by these conditions.      .
     Accordingly, it is hypothesized that although  changes in peripheral blood volume may
be indicative of relative conditions of stress within rather short spans of time,  participation

                                        490

-------
   N*  = Bursts of Pink Noise at 90 dBA
    Q  = 35 dBA Ambient (Quiet)
         Bursts of Noise of  1 — 4 minute Duration
         Period  Between Bursts 7 — 10  minutes
   NOTE:  The larger  number of the  vertical ordinate the less the blood volume (the
           greater constriction  of  the peripheral blood vessels).
44.0

42.5 	
X 41 5 41-4
I
LU
O
>
Q
O
O
m 39.5 —
39.0 —
38.5 —
38.0 —
37.5 —



39.6
Q


•

42.9
N*

AVERAGE





\ AVERAGE




41.6
Q

1



V



—
—
—
—
—


41.5
Q


41.8

42.0
N*




41.9
Q





—
—
—
—
~*






41.0
Q

2
SESSIONS

;
41.4


41.7
N'

AVERAGE


41.4
Q


—
-
—
—
—
~~*



40.7
Q


41.5


42.2
N*






41.6
Q



	
—
—
—
—
~
3 : Average
of All Sessions

                                                                             T A-8 7 55-6
         Figure 2 A measure of peripheral blood volume during pilot study.
           Each of four subjects in three  daily sessions  of quiet  interrupted by
           bursts of  noise.
                                        491

-------
                      Four Subjects Each Session = 2 hours
                      N = Pink Noise at 85 dBA
                      Q = 35 dBA Ambient (Quiet)
                     N* - Bursts of Pink Noise at 85 dBA Ambient
                          Bursts of Noise of 1 - 4 minute Duration
                          Period Between Bursts,  7-10 minutes
   77.0
8
tr
LU
Q.
HI
I
76.5

76.0

75.5

75.0

74.5

74.0

73.5

73.0

72.5

72.0

71.5

71.0

70.5

70.0

69.5

69.0

68.5

68.0
                                                                          AVERAGE
                                        Averages for  Conditions,
                                                  Pilot Study I
                                                                  79.0-
                                                                  78.0
                                                                       77.0
                                                                  76.0-
                                                                  75.0-
                                                                  74.0-
                                                                              lf>
	 ; — — 72.9- . — . -

AVERAGE
7.1.6
	





^^
ffl
LU
cc
9-
Q

L 9 DURING 71.7



n
t-
O
a
Q







o
to
p.
LU
CC
a.
N

r><
CN
r~
0
z
£
o

«
<-
to
O
N
AVERAGE

Q N-Q.





	 /1 .4


n
695!
I*

jjE
r
r>
O'
O
2
CC
Q
N-
AVERAGE
l-
V)
O
a.
Q
CQ



CO
*
a
6

en
oi
Z
i
b
i



AVcnAljc 	 —





CO
8
o
•
2
O
              AMBIENT
               QUtET
              35  dBA
                           AMBIENT
                            85 dBA
                          PINK NOISE
                         WITH  BURSTS
                          OF QUIET
   AMBIENT
    QUIET
     WITH
   BURSTS
OF PINK NOISE
    AVERAGE
      OVER
PRE-DURING-POST
                                                                          TA-8755-7
                   Figure 3  Average heart rate during pilot study.
                                       492

-------
     N - Pink Noise at 85 dBA Ambient
     Q = 35 dBA Ambient  (Quiet)
     N* =  Bursts of  Pink Noise at 85 dBA Ambient
           1—4 minute Duration
     Q* =  Intervals of Quiet 1—4 minute Duration
     Period Between N* and Q* Bursts 7-10 minutes
     Each Session =  2 hours

     NOTE:  The larger the number on the vertical ordinate  the less the blood volume,
             the greater the constriction of the peripheral blood vessels.
*M ft
T3 5
*>o n . . ,i ,
*V7 ^ 	

X
i
s 31-&
UJ
31.0
_l
O r^^
> JU.t>
Q 30.1
o 	
O 30.0 	








•^

















AVERAGE

— i -wi n










Average for Conditions,
Pilot Study I









30.5
AVERAGE













42.0 -
41.5 	
41.0-
--IS
UJ
-40.0- £
—
DURING 42.2 J
41.5 AVERAGE

CD
CO
o
Q.
Q N* Q





	 l AVERAGE

30.2





AVERAGE


m
   29.0

   28.5

   28.0



   27.0

30.1

	 —



]m^m
8
UJ
DC
Q.
Q
AVERAGE
| O DURING 30.1
R
1-
O
a.
Q
30.0






in
8
UJ
CC
a.
N
AVERAGE

| 9 DURING 29.5
a
H
w
O
a.
N






UJ
CC
a.
Q
| 2 DURING 30.4 |
— ^
cp
1-
in
O
a.
Q

30.2








o"
n
a
9
a

p
z
9
z

in
8
C3
*
Z
a
AVERAGE







             AMBIENT        AMBIENT      AMBIENT
               QUIET          85 dBA          QUIET
              35  dBA       PINK NOISE   WITH BURSTS
                           WITH  BURSTS OF PINK NOISE
                             OF QUIET
                                                                          TA-8755-8
          Figure 4 A measure of peripheral blood volume during pilot study.
                                        493

-------
  80
   78
   76
ui
D
u
UJ
I
   72
   70
   68
   66
O  Q"N"Q
X  QQQ
•  N( )N

O  QUO

X  "Q"

O  "N"
                                                          _L
         FIRST SESSION
            FOR A
          CONDITION
                SECOND     MIDDLE
                SESSION     SESSION
                 FOR A    (QQQ ONLY)
               CONDITION
FIRST SESSION
    FOR A
  CONDITION
SECOND SESSION
    FOR  A
  CONDITION

       TA 8755 2
                    Figure 5 Average heart rate in study for various test sessions
 of the peripheral vascular system in homestatic functioning of the body may make interpre-
 tation of the relation of blood volume to stress rather difficult.

 Summary and Conclusions

      (1)  It would appear that heart rate and peripheral blood flow are possibly correlated
          measures of physiological stress reactions, but that peripheral blood flow, as
          detected and measured in  these studies, is probably a somewhat less sensitive and
          reliable measure of physiological reaction to "stress" than heart rate.
      (2)  When tested in successive sessions with a quiet ambient, subjects showed increased
          heart rate and a small decrease in peripheral blood flow when occasional bursts of
          noise (90 dBA) were presented.
      (3)  When the ambient conditions were varied between test sessions, either bursts of
          "quiet," in  a noise-ambient, or  bursts of "noise," in a quiet ambient, resulted in
          an apparent slight decrease hi stress as measured by decreased heart beat rate or
          increased peripheral  blood volume.  However, the change  was  larger  with the
          bursts of "quiet" than with the bursts of "noise."
                                          494

-------
   34
   33
   32
X
 i
8
-  31
tu
5
D
O

Q
O
g
03
30
   29
   28
   27
                                     T
                          •  N"Q"N
                          O  Q"N"Q
                          X  QOQ
             I
                        I
                                           I

                                        • N( )N
                                        O Q{ )Q
                                        X "Q"
                                        D "N"
                                           I
       FIRST SESSION
          FOR A
        CONDITION
                     SECOND
                     SESSION
                      FOR A
                    CONDITION
QOQ ONLY
                    FIRST SESSION
                       FOR A
                     CONDITION
SECOND SESSION
     FOR A
  CONDITION
  Parameter  is acoustic  condition  (see Table  1 for  specifics).
           Figure 6 Average of a measure of blood volume in study II for various test sessions.
      (4)  The average heart rate and freedom of peripheral blood flow of the subjects were
           as much, if not more, related to the experimental design of the test sessions as to
           the presence of quiet  and noise per se. For example, average heart rate of the
           subjects increased  with continued testing day after day when (Study II) quiet-
           noise-quiet sessions were alternated on successive days with so-called noise-quiet-
           noise  sessions, or  a completely quiet session; whereas, when the subjects were
           exposed on consecutive days only to  the quiet-noise-quiet pattern (Study  I), a
           significant progressive decrease was evident in this stress response (i.e. adaptation
           or  habituation occurred). In that  regard, it could be  suggested that in Study I
           (with  only the quiet-noise-quiet sequence being used) the subjects in addition to
         ,  being  apprehensive about the general experimental situation viewed the noise as
           the obvious change in the environment that was potentially stressful and to which
           they might be expected to respond. Whereas in Study II, it is conceivable that the
           subjects, from session to session, became somewhat confused as to what were the
                                          495

-------
         true experimental variables that were being manipulated. This presumably could
         cause a general increase in tension as the sessions progressed without the adapta-
         tion or habituation that occurred in Study I.
    (5)  These findings must be considered as tentative in view of the small number of
         subjects involved, the exploratory nature of the studies, and the lack of consist-
         ency between decreased peripheral blood volume and increased  heart  rate as
         measures of "stress."

Acknowledgment

    This paper was prepared under Grant NIMH 18161-02  from the National Institute of
Mental Health, Department of Health, Education and Welfare.
                                 REFERENCES

 1.  BERGMANN, F.  J., Untersuchungen zur Frage der Kompensation vegetativer Schall-
     reaktionen durch eine vasoaktive Substanz (unpublished Dissertation, Essen).
 2.  DAMS, E. H., Kompensationserfolge larmbedingter Funktionsstorungen durch Theo-
     phyllin - nicotinat  in  Abhangigkeit  vom  Umgebrunspegel so'wie  von  Alter und
     Geschlecht. Dissertation Essen (1972).
 3.  GRIEFAHN, B.,  Examples of Noise induced  Reactions of the Autonomic Nervous
     System during Normal Ovarian Cycle. (Int. Congress on  Noise as  a Public Health
     Problem. Dubrovnik, 1973).
 4.  HEBB, D. O., Drives and the C.N.S. (Conceptual Nervous System) Psychol.  Rev.  62
     (1955)
 5.  HEZEL, J. S. Zur vasokonstriktorischen Reaktion bei Impulsschallexposition (110  dB
     (A))  und unterschiedlicher zeitlicher  Energieverteilung  (unpublished Dissertation,
     Essen).
 6.  JANSEN, G. Zur Entstehung vegetativer Funktionsstorungen durch Larmeinwirkung.
     Arch. Gewerbepath. u. Gewebehyg. 17, 238-261 (1959).
 7.  JANSEN, G. Zur  nervosen Belastung durch La'rm. Beihefte z. Zentralbl. f. Arbeitsmed.
     und Arbeitsschutz, 9 (1967) Dr. Dietrich Steinkopff Verlag
 8.  JANSEN, G. Relation between Temporary Threshold Shift and Peripherical Circula-
     tory Effects of Sound.  Physiological Effects  of Noise, 67-74 (1969) Plenum  Press
     Publishing Corporation,  New York.  Edited by: Bruce L. Welch and  Annemarie S.
     Welch.
 9.  JANSEN, G., Grenz- und Richtwerte in der Larmbekampfung und ihr psychophysio-
     logjscher Aussagewert. Tagung: "Der La'rm als Gefahr fur den Menschen in der zu
     planenden und gebauten Umwelt" DAL, Stuttgart 10. -11.10. 1972.
10.  JANSEN, G., Untersuchungen Uberdie psychophysiologische Wirkungvon Gerauschen
     mit unterschiedlichem Bedeutungsgehalt. Forschungsbericht an das Bundesministerium
     deslnnern(1973)

                                       496

-------
11.  JANSEN,  G., Fluglarm.  Eine interdisziplinare Untersuchung iiber die Auswirkungen
    des Fluglarms auf den Menschen.-Arbeitsphysiologische Sektion.- (unpublished).
12.  JANSEN,  G., W.  KLOSTERKOTTER und  R.  REINEKE:  Experimentelle Unter-
    suchungen   zur   Kompensation   larmbedingter   Gefassreaktionen.  Schriftenreihe:
    Arbeitsmed. Sozialmed. Arbeitshyg. 29, 303-330 (1960).
13.  LYNN, R., Attention, Arousal, and the Orientation Reaction. Pergamon Press (1966).
14.  MALMO,  R. B., Activation:  a Neuropsychological  Dimension.  Psychol. Rev.  66
    (1959).
15.  MEIER,  F. J.,  Untersuchungen  viber  die Wirkung von Schallreizen mit gleicher
    Energiesumme bei unterschiedlichem Zeitgang auf Gehor und peripheren Kreislauf-
    widerstand. Dissertation, Essen (1971).
16.  SOKOLOFF, N., Perception and  the Conditioned Reflex. Pergamon Press, Oxford
    (1963).
17.  WELCH, B. L.,  and A.  S. WELCH, "Physiological Effects of Noise".  Plenum  Press
    Publishing Corporation, New York (1970)
                                       497

-------
             SESSION 6

SLEEP AND ITS DISTURBANCE BY NOISE

       Chairmen: B. Metz, France
                M. Levi, Yugoslavia
                499

-------
                     EFFECTS OF NOISE ON SLEEP: A REVIEW

                                 Harold L. Williams
                               University of Minnesota

     The effects  of acoustic stimulation on behavioral and physiological phenomena of
human sleep depend on several factors, including: (a) the stimuli: their physical parameters,
qualitative aspects and scheduling; (b) stage of sleep and accumulated sleep time; (c) instruc-
tions to the subject and his psychophysiological and motivational state; and (d) individual
differences on such variables as age, sex, physical condition and psychopathology. Stimulus-
response relations vary, not only for different physiological systems and  behavioral events,
but also with subtle differences in measurement criteria for specific physiological systems or
behavior. Furthermore,  habituation and adaptation to acoustic stimulation during sleep
depend on both  the programming of stimuli and  the response system under study.  This
paper reviews a number of recent investigations which indicate that the measurement of
responsiveness during sleep is a multivariate problem of considerable  complexity.

                                  Sensory Processes

     In general, with neutral and brief acoustic stimuli such as clicks or tones, the intensity
required to awaken a sleeping subject is somewhat higher than his sensation threshold during
wakefulness.  This well-known fact has been taken as evidence that in  sleep, either the
thresholds of sensory transducers are  raised or conduction in afferent pathways is dimin-
ished.  However, several investigators, recording various behavioral and physiological re-
sponses to  sensory  input, have  concluded that neither sensation thresholds  nor the
information-handling capacity of sensory systems is impaired during sleep, at least in man.
For example, Davis et al. (1939) reported that during stages 1 and  2  the K-complex of the
sleep electroencephalogram (EEC) was regularly evoked by rather faint tones, about 20 dB
above the noise level of the bedroom,  Williams et  al. (1964) found that in sleep stages  1, 2,
and 3 (see Rechtschaffen and Kales, 1968), simple acoustic stimuli no more than 5 dB above
waking sensation  threshold  could evoke statistically reliable EEC, autonomic and behavioral
responses, while  slightly higher  intensities  were usually  required for stages 4 and REM.
Keefe  and  his  co-workers  (1971), delivering increasingly intense  tones to young males,
reported that the  average threshold for nighttime awakening was about 25 dB (range 5 to 35
dB)  above the sensation threshold established during waking,  or about  60 dB referred to
.0002 dynes/cm2. Awakening thresholds for daytime sleepers were about  15 dB higher  than
those for night sleepers,  and thresholds (at night) were slightly lower in stage 2 than in REM
or high-voltage slow-wave (delta) sleep. In that study, as in the investigation by Williams et
al. (1964), statistically reliable EEC and cardiac responses occurred to stimuli only 5 to 10
dB above waking sensation threshold.  Thiessen's  studies of recorded truck noise found
significant shifts  toward  EEC awakening in 10%  of truck sounds  at 40 dB(A) and some
subjects awakened more than half the time at peak noise levels no  greater than 50 dB(A).
(See also Thiessen's report in this Symposium.) We can conclude that the frequent failure to
awaken the sleeping human with sensory stimuli whose intensity is above waking sensation

                                        501

-------
levels is  not generally due to raised thresholds at the periphery or to impaired sensory
transmission. (See also, Rechtschaffen et a!.,  1966; and Watson and Rechtschaffen, 1969).
                              Properties of the Stimulus

Brief Sounds.

     The likelihood and magnitude of EEC, autonomic and behavioral responses to brief
(msec  to sec) neutral acoustic stimuli delivered during sleep is a monotonic function  of
stimulus intensity (e.g., Williams et al> 1962; Watson and Rechtschaffen et al. 1969; Keefe et
al} 1971; Anch, 1972). However, different response systems show differential sensitivity to
stimulation. For example, in the Keefe et al, 1971, study, EEC responses were most sensi-
tive to 1,000 Hz tones, followed by cardiovascular variables. Electrodermal responses were
found  only in the presence of full arousal, and respiratory periodocity was not altered by
stimuli sufficiently intense to cause EEG or behavioral awakening. Despite variations among
response measures, each of these studies showed that humans are capable  of perceiving
graded auditory stimuli and of responding proportionately to their intensity in all stages of
sleep, throughout the night. (See also Derbyshire and McDermott, 1958, for a simjlar con-
clusion.)
     Responsiveness during sleep also varies with other physical parameters of simple audi-
tory stimuli. Vetter and Horvath (1962)  found that with brief acoustic stimuli, relatively
low frequencies (100 Hz) and fast rise times (1 msec) were most effective for eliciting the
EEG K-complex in  stage 2, and unpublished pilot studies in our laboratory confirmed these
same effects for frequency of responding on a microswitch taped to the hand. Similarly, Hutt
et al. (1968) reported that square-wave tones were more potent than  sinewave tones for
eliciting electromyographic responses in human neonates, especially square-wave tones with
low-frequency fundamentals (i.e., 100 Hz).
     Several investigators have observed that the arousal value of a stimulus depends on its
quality as well as its strength (See Koella, 1969). Weak stimuli that are novel or unexpected
(e.g., offset of a continuous sound), relevant to biological drives (e.g., the smell of meat to a
hungry animal) or that possess acquired significance (e.g.,  one's own name, or the whimper
of one's child) can cause immediate awakening. In some stages of sleep, humans can analyze
and respond differentially to such complex sounds as spoken names (Oswald et al., 1960),
sentences  (Lehmann and Koukkou,  1971), or complex instructions (Evans et al, 1966).
Furthermore, differentiated responses acquired during waking to specific acoustic stimuli,
persist  during sleep in both  animals and humans (e.g., Buendia  et al., 1964; Granda and
Hammack,  1961; Zung and Wilson, 1961; Williams et al., 1966; Schicht et al., 1968). These
results indicate that during some stages of sleep, sensory analyzer mechanisms in  the brain
remain operative  so that incoming signals  can  be  encoded and  categorized. Moreover,
instructions or information  acquired prior to sleep  must remain in long-term  memory,
available to the analyzer mechanism.

                                        502

-------
Fluctuating and Continuous Sounds.

     Crescendos of white noise rising over a period of seconds, sounds of airplane flyovers,
and  fluctuating sounds of automobile traffic can cause gross alterations in sleep, including
inhibition  of  delta sleep  (stages 3 and 4), increased body movements,  wakefulness,  and
delayed onset of sleep (Lukas and Kryter,  1970; Schieber et al, 1968; Pearsons et al [with
Globus and co-workers] 1973). The  results of Schieber and his colleagues (1968) in Metz's
laboratory with recorded traffic noise are particularly interesting.  They found that low-
density traffic sounds averaging 61  dB were more disruptive of sleep than high-density
traffic averaging about 70 dB. These data suggest that relatively infrequent sounds (one or
two per minute) which exceed background noise levels may cause more general disturbance
of sleep then relatively frequent sounds with higher average intensity. Perhaps in the latter
case where the surprisal value of the stimuli  is less, adaptation is more likely to occur.
     In Thiessen's experiments, where  sleeping subjects were exposed to recorded noise
from a passing truck at selected sound levels, there was a 5% probability of behavioral
awakening at 40 dB(A) and a  30% probability at 70  dB(A),  but with wide individual
differences. Some subjects awakened more than half the time at 50 dB(A) whereas others
almost never awakened, even at 75 dB(A).
     In a study by Scott (1972), continuous high-intensity noise (95  dB) that was turned on
at bedtime caused a loss of stage REM,  but had no substantial effect on non-REM states or
on  other  measures of sleep  disruption. By the  second  night of stimulation, stage REM
percent was returning toward baseline control level. Scott estimated that adaptation would
have been complete after a few nights of noise exposure. Other evidence concerning adapta-
tion and habituation will be discussed in another section of this review.

                              Properties of the Response

     As was  mentioned, different response systems are differentially sensitive to neutral
acoustic stimuli of moderate intensity. Whereas, during waking, EEC, autonomic and motor
responses occur simultaneously to tones adjusted to sensation threshold, during sleep these
responses  show a consistent hierarchy. The EEC is most sensitive, followed by the cardio-
vascular system (i.e., heart rate and peripheral vasoconstriction), followed by electrodermal
activity, respiration and motor behavior. Further, this hierarchy of sensitivity is consistent
over the several stages of sleep. Factors such as  accumulated sleep time, time of day or
night, or the presence of phasic  activity such as rapid eye movements during stage REM
apparently do not alter the ranking or responsiveness (Keefe et al., 1971).
     As would be expected, thresholds for the awakening response depend on its definition.
Full  EEC arousal often  occurs  without a  specified motor response, particularly during
high-voltage sleep stages 3 and 4 (Keefe  et al., 1971). Investigators disagree, however, as to
whether the converse can occur during  sleep—that is,  a designated or conditioned motor
response without associated EEC arousal. Keefe et al. (1971) found no such events, whereas
Williams et al. (1966) reported  that many  motor responses to  specific auditory signals
occurred without prior signs of EEC awakening. It is agreed that the  obtained threshold for
behavioral awakening increases with  the  complexity of the required motor response.  For

                                         503

-------
 neutral stimuli, pressing a microswtich taped in the hand occurs systematically to noise or
 tones at about 25-35 dB above waking sensation threshold (Williams et al., 1964; Keefe et
 al.,  1971). The threshold for obtaining complex verbal responses signifying recognition of
 specific properties of the stimulus appears to be about 65 dB above background noise levels
 (Rechtschaffen  et al., 1966), as does  that for more complex motor responses  such as
 reaching for and pressing a button on the headboard of  the bed (Lukas and Kryter, 1970).
 (See Miller, 1971, for a review of these and other aspects of the problem.)

                      Stage of Sleep and Accumulated Sleep Time

 Sleep Stage.

     The EEC stages of sleep 1 through 4 in the Dement  and Kleitman (1957) classification
 were labelled  for their order of occurrence  after sleep onset, and for their apparent ordinal
 relation to threshold for arousal. Stage 1, characterized by loss of the alpha rhythm of quiet
 waking,  and stages 3 and 4 with their slow high voltage delta waves were classified as the
 "lightest" and deepest" states of sleep respectively. The  discovery  by Aserinsky and Kleit-
 man (1953) of the periodic state of REM sleep, associated with a low-voltage stage 1 EEC,
 complicated the situation because arousal  thresholds in  this stage were higher than in the
 stage 1 episodes found at the onset of sleep. In general,  however,  the likelihood of behav-
 ioral responding to neutral stimuli is a decreasing function of the amplitude and period of
 the  background EEC rhythms (Zund and Wilson, 1961; Williams et  al., 1964; Rechtschaffen
 et al.,  1966; Keefe et al., 1971). Although simple acoustic stimuli of moderate intensity can
 elicit specific  physiological responses in any stage of sleep, during  high-voltage delta sleep,
 the  likelihood of a specified, easily executed motor response is low. For example, Williams
 et al. (1966) found instrumental responding on a microswitch taped in the hand to tones at
 35 dB above sensation threshold only during low-voltage EEC stages 1, 2, and REM. More-
 over, Evans et al. (1966) were able to elicit relatively complex motor responses to verbal
 instructions only in stage REM. Although the reasons for this are not entirely understood,
 we can no longer conclude that  the stages of sleep,  1 through 4, define a universal con-
 tinuum either of depth of sleep or thresholds of arousal. When awakening is defined as full
 EEC arousal, the awakening thresholds in stages 2 through 4 and REM are apparently nearly
 identical (Keefe et al, 1971). The relative  loss of designated motor responses in the high-
 voltage stages  of sleep is not due to raised sensory thresholds, and probably not to failure of
 the signal analyzer system. Schicht et al (1968) found that discriminated classically condi-
 tioned cardiovascular responses  acquired during wakefulness  could be elicited regularly
 during extinction trials in stage 4 sleep. This finding, if confirmed, is evidence that signal
 analysis is still possible during stage 4 and that previous failures to observe discriminated
responding during that stage occurred only because stage 4  is relatively incompatible  with
the organization and execution of motor responses. Thus, Keefe and his colleagues (1971)
reached the tentative conclusion that impaired motor responding during high-voltage slow-
wave sleep was due to the disorientation and confusion which accompany sudden awakening
from that state rather than  to raised response thresholds. Broadly speaking, the human is
neither deafferented nor de-efferented during sleep.

                                        504

-------
Sleep Time.

     Several investigators have found that thresholds for awakening decreased as time asleep
increased  (e.g., Williams,  1966; Rechtschaffen et al., 1966; Watson and Rechtschaffen,
1969; Morgan and Rice, 1970; Keefe et al.,  1971). However, in all of these studies amount
of accumulated  sleep was confounded with chronological time. Thus, it is  not known
whether  the amount of  accumulated sleep or a circadian biological rhythm, relatively
independent of time asleep, is the principal factor in this effect. Williams (1966) proposed
the latter explanation on grounds that the curve of behavioral responsiveness over the night
is  reminiscent of the circadian  curve of body temperature  described  by Kleitman and
Ramsaroop  (1948).  However, Keefe et al.  (1971) reported  the same temporal trend in
awakening thresholds for both day and night sleepers. But the day sleepers had been on the
reversed sleep-waking schedule for at least seven days so that circadian biological cycles were
probably also reversed.
                            Motivation and Pre-Sleep State

     Studies summarized earlier in this review provide clear evidence that motivational and
incentive factors can influence the  probability of either physiological  or behavioral re-
sponses to noise. Sounds which are relevant to survival, or which, through conditioning, or
instructions, acquire signal properties  are more likely to arouse  the sleeper  than neutral
sounds. As Miller (1971) suggested, for weak stimuli, the effects of motivation depend on
the stage of sleep.  For example, Williams  et al. (1966)  found that as the motivation to
respond to  designated 35-dB tone stimuli  was enhanced by instructions and contingent
punishment, instrumental responding on a microswitch increased about  five-fold in low-
voltage stages  2 and REM, but very little  in high-voltage delta sleep. On the other hand,
Zung and Wilson (1961) showed that for moderately intense stimuli, instructions and finan-
cial incentives induced a marked increase in frequency of EEC arousal and  waking responses
in all stages of sleep.
     As in wakefulness, small differences in instructions given prior to sleep can have sub-
stantial effects on behavioral  and  physiological responding. For example,  the FAA (CAMI)
group in Oklahoma  City*, employing simulated sonic booms of 1.0 psf, did not label sounds
as sonic booms.  The  subjects  were told that  the investigators  were  interested  in sleep
behavior, moods, and performance; that noises might occur during the night (including the
presence of experimenters in the test room);  but  that the subject's task was to ignore
disturbances of any kind and get  the best night's sleep possible. The frequency of full EEC
arousal to hourly booms in this investigation was considerably less than that found by Lukas
and Kryter  (1970), even in  elderly subjects. The  latter investigators were more explicit
about the nature of the stimuli, and the response requirement of pressing a button attached
to the headboard of the bed.
*Personal communication from Dr. Wm. Collins. See also Collins' report in this Symposium.

                                        505

-------
     Other aspects  of the  pre-sleep  state alter the sleep EEC profile, and possibly the
subject's responsiveness to disturbing stimuli. For example, Lester et al (1967) reported that
a moderate increase of daytime stress, such as that occasioned by a college examination, was
associated with increased spontaneous arousal and inhibition of delta sleep. Jansen (1969)
cites evidence that emotional factors, stress and  neuroticism influence responsiveness to
noise in waking subjects. It is reasonable to predict similar positive relationships between
disturbed emotional states  and responsiveness to  noise during sleep. Indirect evidence for
such a relationship comes from studies showing that 64 hr of sleep deprivation caused a
systematic reduction in behavioral and physiological responsiveness to noise stimuli in all
stages of sleep (Williams et al, 1964). Keefe and his colleagues (1971) suggest that the higher
awakening thresholds found in their daytime sleepers may also have resulted from chronic
loss of sleep.

                                Individual Differences

     As  mentioned .earlier in this review,  responsiveness to noise during sleep varies in
relation to the age of the subject, sex, psychopathology and physical condition. The series
of studies by -Lukas and his co-workers used simulated sonic booms ranging in "outdoor"
intensities from  .06 to 5.0 psf, and recordings of subsonic jet flyovers, ranging in "outdoor"
intensity from  101-119 PNdB. They found  that children 5-8  years old  were  relatively
undisturbed by  either type oftooise,  whereas elderly men were much more disturbed than
younger subjects (Lukas and Kryter,  1970a  and 1970b). In general, this age effect was
confirmed by Collins' group, using simulated sonic booms with "outdoor" intensities of 1.0
psf. However, the average magnitude of boom effects was considerably less in Collins et al's
investigation than in  the studies by  Lukas et al.  (See Collins' report in this symposium.)
Possible reasons for this difference include differences in instructions, scheduling of subjects
and variation of the boom intensity parameter. Steinicke (1957)  reported that both the
elderly and people under thirty were more readily  awakened by noise than the middle-aged,
and that manual workers were  more susceptible  to noise awakening  than intellectual
workers. He concluded, incidentally, that the  noise in bedrooms should not exceed 35
dB(A).
     Although the sleep of small  children and normal infants (e.g., Gadeke et al, 1969) is
less disturbable  by acoustic  stimuli  that that of adults, babies subjected to gestational
difficulty or birth trauma may be hyperresponsive. Murphy (1969) on the basis of clinical
observation suggested that the short gestation, anoxic or brain-injured infant, in particular,
displays  exceptional responsiveness to sounds. Bench and Parker (1971), however,  in an
interesting application of signal detection theory, failed  to confirm this assertion. In fact,
their short-gestation babies tended to have higher awakening thresholds than full-term in-
fants.
    For neutral auditory  stimuli delivered during sleep, the threshold  for EEC  arousal
responses is lower in women than men (Steinicke, 1957; Wilson and Zung, 1966). Lukas and
Dobbs  (1972) found  similar greater sensitivity in middle-aged women to the sounds of
subsonic jet aircraft flyovers and simulated sonic booms. The women were particularly
responsive to  the sound of aircraft  flyovers. Wilson, and Zung (1966)  suggest that this

                                        506

-------
tendency toward hyperacuity in women may have adaptive significance for the mothering
role.
     There is evidence that EEC arousal thresholds differ for different types of psycho-
pathology.  For example, Kodman and Sparks (1963) reported that schizophrenic patients
showed a marked elevation of auditory sleep thresholds, whereas Zung et al. (1964) found
markedly reduced EEC arousal  thresholds in the depressive disorders. In fact the auditory
sensitivity during sleep in depressed males was greater on average than that found by this
same group in normal middle-aged females (Wilson and Zung,  1966). As had been men-
tioned, it is probable also that the sleep-disturbing effects of acoustic stimuli increase with
neuroticism. Monroe (1967) found that the sleep of neurotic subjects was grossly disturbed,
even in a quiet environment, and Jansen and Hoffmann (1965) reported that subjects high
in neuroticism were generally more sensitive to and disturbed by noise than normals.

                  Short-Term Habituation and  Long-Term Adaptation

     Whether short-term  habituation or long-term adaptation can occur during sleep is a
subject of  debate. For the EEC and autonomic responses which comprise the orienting
reflex,  Johnson's group in San Diego has found  no evidence of habituation over a few trials
or adaptation over many  nights  (e.g., Johnson and Lubin, 1967; Townsend et al, in press).
Similar findings are reported by Lukas and  Kryter  (1970) and Collins' group (this Sym-
posium) for simulated sonic booms,  and by Hutt et al. (1968) for EMG responses in the
human neonate.  Firth (1973) did find some habituation trends for autonomic and EEC
responses to short  runs of closely spaced 1000-Hz (70 dB) tones. However, this result, if
replicated, is more significant for theories of brain functioning during sleep than for applica-
tion.
     Anecdotal evidence suggests that the frequency and duration of behavioral awakening
or gross disturbances  of  sleep should show long-term adaptation. We are all familiar with
accounts of soldiers sleeping undisturbed in the presence  of artillery fire, or city  people
sleeping in the presence  of high levels of urban noise. Yet, so far, neither laboratory nor
field studies have produced unequivocal evidence of long-term adaptation. Lukas and his
associates  did  report some adaptation in college students, but only  to sonic booms of low
intensity (about 0.7 psf, "outside") and only in stage 2 sleep (Lukas,  1969). Townsend et al.
(in press) suggest that the relatively small effects on sleep found in their study of young men
exposed day and night to "pings" may be due to pre-sleep adaptation. However, as will be
reported by Dr.  Friedmann in this Symposium, the sleep of middle-aged couples who had
resided in  the vicinity of Los  Angeles International Airport for more than 5 years was
considerably disrupted by jet flyovers.

                              Summary and Conclusions

     During wakefulness, the presence of  raised  thresholds  for psychophysiological or
behavioral responding often permits an inference about the status of the sensorium.  During
sleep, however, interpretation of, the same finding requires more complex analysis.  Raised
thresholds  could be due  to alterations in sensory analyzer systems, in systems which link

                                        507

-------
sensory and motor processes, or in mechanisms which mediate the selection and execution of
psychophysiological or motor behaviors. Taken together, recent studies of responsiveness to
acoustic stimulation indicate that the states of sleep are not accompanied either by in-
creased thresholds for sensory transduction, diminished conduction in afferent pathways,
gross impairment of sensory analyzer systems or loss of sensory-motor links. Differentiated
and systematic EEC and cardiovascular responding are found during sleep either to condi-
tioned  or  biologically relevant stimuli whose intensities are  very  near waking sensation
thresholds. Thus, the sleeping human can encode, categorize and respond differentially to
near-threshold simple and complex acoustic stimuli. The increased thresholds for behavioral
responding often reported for sleeping subjects are probably due to the fact that some states
of sleep are generally not compatible with the selection and  execution  of certain motor
responses.
      Although the likelihood of behavioral responses to neutral acoustic stimuli is lower in
high-voltage than in low-voltage states of sleep, the notion that the stages of sleep, 2 through
4, represent a universal continuum of depth of sleep is no longer tenable. When awakening is
defined as full EEC arousal, stages 2 through 4 and REM are nearly identical. Moreover, the
relative impairment of motor responding found during high-voltage stage 4 sleep may be due
to the disorientation and confusion which accompany awakening from that state rather than
to raised response thresholds per se.
      The interpretation of stage-of-sleep data is further complicated by the fact that respon-
siveness varies with chronological time. Whether this is a function of amount of accumulated
sleep or the phase of  a circadian biological rhythm  is not known.  Finally, as in  other
psychophysiological studies, responsiveness during sleep is altered by subject variables such
as age, sex, instructions, motivation,  medical illness and psychopathology. Thus, it is not
surprising that investigators in the field have been unable to recommend uniform guidelines
for the regulation of noise in the environment.
                                    Bibliography

Anch. M., The auditory evoked brain response during adult human sleep. (Abstract) Paper
     delivered at the  11th annual meeting of The Association for the Psychophysiological
     Study of Sleep (1972).
Aserinsky, E., Kleitman, N., Regularly occurring periods of eye motility, and concomitant
     phenomena during sleep. Science, 118: 272-274(1953)
Bench, J. and Parker, A., Hyper-responsivity to sounds in the short-gestation baby. Develop.
     Med. Clin. NeuroL,  13: 15-19 (1971)
Buendia, N., Sierra, G., Goode, M. and Segundo, J. P., Conditioned and discriminatory
     responses in wakeful and sleeping cats. Electroencephal. Clin. NeurophysioL, Suppl.
     24,199(1964)
Davis, H., Davis, P. A., Loomis, A. L., Harvey, E. N. and Hobart, G. Electrical reactions of
     the human  brain to auditory stimulation during sleep. /. NeurophysioL,  2: 500-514
     (1939)

                                        508

-------
Dement, W. C, Kleitman, N., Cyclic variations in EEC during sleep and their relation to eye
    movements, body motility and  dreaming. Electroencephal. Clin. Neurophysiol,  9:
    673-690(1957)
Derbyshire, A. J. and McDermott, M., Further contributions to the EEC method of evalua-
    ting auditory  function. Laryngoscope International Conference on Audiology, 68:
    558-570(1958)
Evans, F. J., Gustafson, L. A., O'Connell, D. N., Orne, M. T., Shor, R. E., Response during
    sleep with intervening waking amnesia. Science, 152: 666-667 (1966)
Firth, H., Habituation during sleep. Psychophysiology, 10: 43-51 (1973)
Gadeke, R.,  Doling, F. K. and Vogel, A., The noise level in a children's hospital and the
    wake-up threshold in infants. A eta Paediat. Scand. 58:  164-170 (1969)
Granda, A. M. and Hammack, J. T., Operant behavior during sleep. Science, 133: 1485-1486
    (1961)
Hurt, C., von Bernuth, H.,  Lenard, H. G., Hutt,  S. J.,  Prechtl, H. F. R.,  Habituation  in
    relation to state in the human neonate. Nature, 220: 618-620 (1968)
Jansen, G., Effects of noise on physiological state. ASH A Reports No. 4 (1969)
Jansen, G. and Hoffmann, H., Larmbedingte Anderungen der Feinmotorik und Lastigkeit-
    sempfindungen in Abhangigkeit von bestimmten Personlichkeitsdimensionen. Z. Exper.
    Angew. Psychol,  12: 594-613 (1965)
Johnson, L.  C. and Lubin,  A., The orienting reflex  during waking and sleeping. Electro-
    encephal. Clin. Neurophysiol, 22: 11-21 (1967)
Keefe, F. B.,  Johnson, L. C. and Hunter, E. J., EEC and autonomic response pattern during
    waking and sleep stages. Psychophysiology,  8: 198-212(1971)
Kleitman, N.  and Ramsaroop,  A.,  Periodicity  in  body temperature and heart  rate.
    Endocrinology, 43:  1-20(1948)
Kodman, F.  and Sparks, C., The auditory sleep threshold of the catatonic schizophrenic./.
    Clin. Psychol, 19: 405 (1963)
Koella, W. P., The central nervous control of sleep.  In  W. Haymaker, E. Anderson, W. J.
    Nanta (Eds.) The Hypothalamus,  Springfield: Charles C. Thomas. Pp. 622-644
Lehmann, D. and Koukkou, M. Das EEC des Menschon beim lernen von neuem und bekann-
    tem material. Arch. Psychiat. Nervenkr,, 215:  22-32  (1971)
Lester, B. K., Burch, N. R. and Dossett, R. C., Nocturnal EEG-GSR profiles and influence  of
    presleep states. Psychophysiology, 3: 238-248 (1967)
Loomis, A. L., Harvey,  E. N., Hobart, G. A.,  Cerebral states  during sleep as studied by
    human brain potentials. /. Exp. Psychol., 21: 127-144 (1937)
Lukas,  J. S.,  The effects of simulated sonic booms and jet flyover noise on human sleep.
    Proceedings of the  Sixth Congress on Environmental Health, Chicago, A.M.A., April
    (1969)
Lukas, J. S., and Kryter, K. D., Awakening effects of simulated sonic booms and subsonic
    aircraft noise. In  Physiological Effects of Noise, B. L. Welch and A. S. Welch (Eds.)
    Plenum Press, New York, pp 283-293 (1970a)
Lukas, J. S. and Kryter, K. D., Awakening effects of simulated sonic aircraft noise on six
    subjects  7 to 72 years of age. NASA  Report No. CR-1599, Washington, D.C. (1970b)
                                       509

-------
Lukas, J.  S. and Dobbs, M. E., Effects of aircraft noises on the sleep of women. NASA
     Contractor Report CR-2041 (1972)
Miller,  J.  D.,  Effects  of Noise on  People. U. S.  Environmental Protection  Agency,
     NT1D300.7(1971)
Monroe, L. J. Psychological and physiological differences between good and poor sleepers.
     /. Abnorm. PsychoL 72: 255 (1967)
Morgan, P. A.  and Rice, C. G., Behavioral awakening in response to indoor sonic booms.
     Institute of Sound and Vibration Research,  University of Southampton, Technical
     Report No. 41(1970)
Murphy, K. P., Differential diagnosis of impaired hearing in Children. Develop. Med.  Child.
     NeuroL, 11:561(1969)
Oswald, I., Taylor, A. M. and Triesman, M., Discriminative responses to stimulation during
     human sleep. Brain, 83: 440-553 (1960)
Pearsons,  K. S., Fidell, S., Globus, G., Friediiiann, J. and Cohen, H. The effects of aircraft
     noise on sleep electrophysiology as recorded hi the home. Bolt, Beranek and Newman
     Report 2422 (1973)
Rechteschaffen, A., Hauri, P. and Zeitlin, M., Auditory awakening thresholds in REM and
     NREM sleep stages. Perceptual and Motor Skills, 22: 927-942 (1966)
Rechtschaffen,  A. and Kale, A. (Eds.): A Manual of Standardized Terminology, Techniques
     and Scoring System for Sleep Stages of Human Subjects.  Public Health Series, U.S.
     Government Printing Office, Washington, D.C. (1968)
Schicht, W. W., McDonald, D. C.  and Shallenberger,  H. D., Progression of conditioned
     discrimination from waking to sleeping. (Abstract) Psychophysiology, 4: 508 (1968)
Schieber,  J. P., Mery, J. and Muzet, A., "Etude Analytique en  Laboratoire de Plnfluence
     due Bruit  sur Le Sommiel," (Rept. of Centre d'Etudes Bioclimatique du CNRS, Stras-
     bourg, France), (1968)
Scott, T. D., The effects of continuous, high intensity white noise on the human sleep cycle.
     Psychophysiology, 9: 227-232(1972)
Steinicke, G., Die Wirkungen von Larm auf den Schlaf des Menschen. Forschungsberichte
     des Wirtschaftes- u. Verkehrsministerium Nordrhein-Westfalen No. 416 (1957)
Steinicke, G., Cited in Urban Traffic Noise:  Status of Research and Legislation in Different
     Countries.   Draft report on  the Consultative  Group on  Transportation Research,
     DAS/C51/68.47 Revised; Organization  for Economic Cooperation  and Development,
     Paris, France, March (1969)
Thiessen, G. J., Effects of noise during sleep. In Physiological Effects of Noise. B. L. Welch
     and A. S. Welch (eds.) New York: Plenum Press, pp. 271-275 (1970)
Thiessen, G. J.  and Olson, N. Community noise-surface transportation. Sound and  Vibra-
     tion,  2:10-16 (1968)
Townsend, R. E., Johnson, L. C. and Muzet, A. Effects of long term exposure to tone pulse
     noise on human sleep. Psychophysiology. (In press)
Vetter, K.  and  Horvath, S. M. Effect of audiometric parameters on K-complex of electro-
     encephalogram. Psychiat. NeuroL, Basel, 144: 103-109(1962)
Watson, R. and Rechtschaffen, A. Auditory awakening thresholds and dream  recall in
     NREM sleep. Perceptual and Motor Skills, 29:635-644 (1969)

                                       510

-------
Williams, H. L., Hammack, J. R., Daly, R. L., Dement, W. C. and Lubin, A. Responses to
     auditory stimulation, sleep loss and the EEC stages of sleep. Electroencephal.  din.
     Neurophysiol,  16: 269-279(1964)
Williams, H. L. The problem of defining depth of sleep. In S. S. Kety, E. V. Evarts, H. L.
     Williams (Eds.), Sleep and Altered States of Consciousness. Baltimore, Md.: Williams
     and Wilkins, pp. 277-287 (1966)
Williams, H. L., Morlock, H.  C. and Morlock, J.  J. Instrumental  behavior during sleep.
     Psychophysiology, 2:208-215 (1966)
Wilson, W. P. and Zung, W. W. K. Attention, discrimination and arousal during sleep. Arch.
     gen. Psychiat., 15:523-528(1966)
Zung, W. W. K. and Wilson, W. P. Response to auditory stimulation during sleep. Arch. gen.
     Psychiat., 4:548-552 (1961)
Zung, W. W. K., Wilson, W. P. and Dodson, W. E. Effect of depressive  disorders on sleep
     EEC  responses. Arch. gen. Psychiat., 10:439-445  (1964)
                                        511

-------
              PREDICTING THE RESPONSE TO NOISE DURING SLEEP
                                  Jerome S. Lukas
                             Stanford Research Institute
                            Menlo Park, California  94025

                                  INTRODUCTION

     Recently, auditory stimuli  with very different spectra  have  been used to study  the
effects of noise on human sleep. In most of these studies the physical characteristics of the
stimuli were  described only  partially,  making questionable  direct comparisons of results
obtained in the  several laboratories. To provide a technique for comparing stimuli and to
estimate  relative sensitivities  of  subjects, a burst of pink noise was recommended (Rice,
1972) for use in laboratories, and tape recordings of that noise were distributed (Lukas) to
some.
     This paper  describes  a study  that used the recommended pink noise burst as one of
three different stimuli, and that correlated several physical  descriptors of the noise, with
different measures of response to those noises.

I     METHOD
     A. Procedure
     The SRI sleep laboratory consists of two identical, acoustically isolated rooms in which
four subjects, divided into two pairs, are tested simultaneously. A test period for one pair of
subjects was considered a control period for the other pair. Typically, test periods alternate
with control periods in each room. In any given room, stimuli are presented randomly with
respect to sequence, intensity, and interval between stimuli, but any two stimuli are  not
presented at  intervals of less than twenty minutes. On the average, stimuli occurred once
every forty minutes.
     The first stimulus on any test night was presented only after both subjects in the room
are in sleep stage 2, at least, or about one hour after the subjects went to bed.
     For any subject the procedure included, first, three accommodation nights in  the
laboratory, next two nights at home, then fourteen consecutive  nights in the laboratory.
The  first two nights of the fourteen, as well as nights 9, 10, and 14, were considered control
nights, during which the subjects were permitted undisturbed sleep. Stimuli were presented
during the remaining nine test nights.
     The subjects were instructed  to sleep as normally  as possible, but to push a button
attached to the headboard of each bed (the "awake switch") if they should awaken for  any
reason. They were never told when stimuli would  be presented  or how many had been
presented.
     Both electroencephalographic (EEC)  and  behavioral responses  to  the noises were
scored. The behavioral response was reserved exclusively for  the use of an "awake" switch
attached to the  headboard of the bed, while the EEC (central, C3, with reference to  the
contralateral  mastoid, A2) responses were scored on the basis of the criteria presented in
Table 1.

                                         513

-------
                                      Table 1

          CRITERIA FOR SCORING THE ELECTROENCEPHALOGRAMS VISUALLY
  Score
                          Response Required
    0
No change in EEC. This category also includes "K complexes," brief
bursts of Alpha (about 10 Hz activity), spindles, and eye movements, as
appropriate for the subject's sleep stage.*

 Sleep stage change of one or two steps, but without arousal. The change
 must occur within 30 s of stimulation and continue for at least an addi-
 tional 40 s.

 Arousal of at least 10 s duration, but without use of the "awake" switch.
 Typically such a record shows brief bursts of Alpha, 10 or more s  of
 low-amplitude Beta (20-40 Hz) activity, and gross body movements.

 Awake response, in which the subject, after arousal, will move about and
 use the "awake" switch. Usually the response occurs within one minute of
 stimulus termination.
*"K complexes," Alpha, spindles, and eye movements occur normally in the EEC in some
sleep stages. If such activity were scored as a response, the subjects in those stages would
appear to be overly sensitive to stimulation as compared to stages in which the activity does
not normally occur.
     B. Stimuli
     The three stimuli were (1) landing noise from a DC-8 without acoustically treated
engine nacelles, (2) landing noise from a DC-8 with acoustically treated nacelles (Langdon et
al., 1970), and (3) a burst of pink noise. The aircraft noises were originally recorded out of
doors, but for the purposes of the study were shaped to simulate noises as they would  be
heard indoors. The time-courses of the stimuli heard by the subjects are illustrated in Fig. 1,
and various physical descriptions of the noises are presented in Table 2.
     It is important to note,  in Table 2, that although the stimuli had nearly identical
nominal intensities (79 or 61 dBA maximum), as progressively more information about their
physical characteristics was added into the descriptors, the stimuli became relatively more or
less severe ("noisy"). For example, adding the tone correction to EPNdB (compare columns
EPNdB and EPNdBT) makes the noise from the jet without treated nacelles about 2 dB
more "noisy" than the jet with treated nacelles, and both these noises are at least 4 dB more
noisy than the pink noise, although most (about 3.5 dB) of this 4 dB difference is due to the
                                       ;514

-------
      <•)  DC-8 WITH UNTREATED ENGINE NACELLES LANDING AT  159.1 m I522 ft)  ALTITUDE
       Ibl  DC-8 WITH TREATED ENGINE NACELLES LANDING AT 154.4 m (507 ft) ALTITUDE
                                    (cl  PINK NOISE  BURST
Figure 1:   Time histories of the three test stimuli measured in a test room near the subject's ears.
                                             5I5

-------
                                                           Table 2

                                          PHYSICAL DESCRIPTIONS OF THE STIMULI



Stimulus
DC-8 with
untreated
nacelles
DC-8 with
treated
nacelles

Pink noise
Nominal
Overall
Duration
(s)

|30
30

(30
1 30
1 4
\
1 4

Rise Time
to Peak
(s)

16.5
16.0

19.0
18.0
1.0
1.0
Duration
to 10 dB
Down Points
from Max dBA

7.5
7.5

9.0
10.5
3.5
3.25

Nominal
Level
dBA

79
>- 61

79
61
79
61


Max
dBA

78.9
61.1

78.4
60.4
78.0
59.9


Peak
dBA

79.8
62.0

79.5
61.5
78.3
60.3


Max
dBD2

83.8
66.2

82.6
64.6
81.9
63.7


Max
PNdB

91.5
73.8

90.0
71.5
89.7
71,9


EPNdB
(1)

85.0
66.7

84.8
66.3
81.3
63.5


EPNdBT
(2)

87.5
69.3

85.6
67.1
81.7
63.8


EPNdBTM
(3)

88.6
70.2

87.2
68.7
83.1
65.4


EPNdBTM-ic
(4)

88.6
70.2

87.2
68.7
95.1
70.4
ON
    Definitions  and techniques for calculating the various physical units may be found in Kryter (19708, 1970b).

    (1)  Calculated using 15 s as a reference duration.
    (2)  Tone correction recommended by Kryter (1970a, 1970b).
    (3)  Modified to account for the critical bandwidth of the ear at frequencies below 355 Hz (Kryter, 1970a, 1970b)
    (4)  An impulse correction applied to Col. 5 to account for a rise of some 32 dB above background noise level
        (about 35 dBA) in the first 0.5 s of the "high" intensity pink noise burst, and a rise erf some 14 dB above
        background level in 0.5 s by the "low" intensity pink noise burst.

-------
differences in duration of the stimuli (see the columns labeled EPNdB and Duration to 10
dB ... dBA). The pink noise typically is less noisy than are the two jet noises; however, if
account (Lukas, Peeler, and Dobbs, 1973) is taken of its impulse characteristics (see Column
EPNdBTM-ic), the pink noise becomes 6 to 8 dB more noisy than the nominally high level
jet noises, and approximately equivalent to the jet noises at nominally low levels (61 dBA).

     C. Subjects
     Four middle-aged (46 to 58 years) males were studied. Three of the four had been
subjects in a previous study; thus they had slept in the laboratory and were familiar with the
procedures  used. All thought themselves  to  be reasonably normal sleepers, without any
particular bias for or against aircraft noise, and none of the subjects lived near or in the
flight paths to the local airports. Audiometry showed their hearing was within normal limits.
H     RESULTS
     Typically, during the control trials the subjects did not show any response (Response
0), although a few spontaneous changes in sleep stage (Response 1) occurred. As shown in
Table 3, in only one case was an arousal (Response 2) observed, and the awake switch was
not used during the  control trials.  It may be concluded that,  in the main, responses  to
stimuli rather than spontaneous changes during sleep are discussed below.

                                       Table 3

                  RESPONSE FREQUENCIES DURING CONTROL TRIALS
                          (Numbers in parentheses are percentages)
Test
Room
Number
1

2

Number of
Control
Trials
158

162



0
152
(96.2)
158
(97.5)


1
5
(3.2)
4
(2.5)


2
1
(.6)
0



3
0

0

Number
of Test
Trials
166*

159

     *During some test trials the control subjects may have been still awake from their previous test
     trial or were moving before and during stimulus occurrence. Such instances were not counted as
     control trials. Hence, the numbers of test and control trials are not equal.

     An earlier study (Lukas, Dobbs, and Kryter, 1971) suggested that response frequencies
 to  a noise during sleep are normally  distributed, particularly when  about eight or more
 subjects are studied. Since three of the four subjects were studied previously, it is reasonable
 to  assume that their responses also are within those normal limits. Consequently, the re-
 sponse frequencies of the four subjects are combined in the data presented below.
                                         517

-------
     A. Effects of Stimulus Intensity
     For each of the three stimuli a change of 18 dBA in stimulus intensity decreased the
frequency of 0 responses and increased the frequencies of 2 and 3 responses; but, as shown
in Table 4, the  magnitude  of these  changes was not  similar for the three stimuli. For
example, with the pink noise an increase of 18 dBA resulted in an increase of approximately
44 percentage points in the frequency of 3 responses, while in response to the treated jet
noises an increase  of only  7 percentage points  was observed.  Somewhat  less disparate
changes were observed with  respect  to the other responses for nominally equivalent
variations in stimulus intensity.
     In Table 5, the data are reorganized  to facilitate comparison  of the response fre-
quencies to the  three stimuli when they are of nominally equivalent intensity.  It should be
noted that, at 61 dBA, pink noise resulted in the lowest frequency  (6.3  percent) of be-
havioral awakenings (Response 3), while the untreated jet noise had the highest (about  24
percent), and no great differences were observed  in the frequency of 0 responses. At the
higher stimulus  intensity, however, the frequencies of behavioral awakening were about
                                        Table 4

                       RESPONSE FREQUENCIES TO THREE STIMULI
                               EACH AT TWO INTENSITIES
                            (Numbers in parentheses are percentages)
Stimulus
Untreated Jet
Treated Jet
Pink Noise
Nominal
Intensity
(dBA)
79
61
79
61
79
61
Responses
0
16
(29.1)
30
(65.2)
15
(44.0)
51
(70.8)
11
(15.7)
30
(62.5)
1
3
(5.5)
4
(8.7)
6
(17-6)
6
(8.3)
7
(10.0)
6
(12.5)
2
9
(16.4)
1
(2.2)
4
01.7)
3
(4.2)
17
(24.3)
9
(18.7)
3
27
(49.0)
11
(23.9)
9
(26.5)
12
(16.7)
35
(50.0)
3
(63)
X2
16.37*
7.5 6t
f
35.36*
       *3 df (degrees of freedom); p < 0.001.
       t3 df, 0.10 > p > 0.05, not significant.
       *3df,p<  .001.
                                         518

-------
                                         Tables

                   RESPONSE FREQUENCIES TO STIMULI OF NOMINALLY
                                 EQUIVALENT INTENSITY
                           (Numbers in parentheses are percentages)
Nominal
Intensity
(dBA)
79
61
Stimulus
Untreated Jet
Treated Jet
Pink Noise
Untreated Jet
Treated Jet
Pink Noise
Response
0
16
(29.1)
15
(44.0)
11
(15.7)
30
(65.2)
51
(70.8)
30
(62.5)
1
3
(5.5)
6
(17.6)
7
(10.0)
4
(8.7)
6
(8.3)
,6
(12.5)
2
9
(16.4)
4
(11.4)
17
(24.3)
1
(2.2)
3
(4.2)
9
(18.7)
3
27
(49.0)
9
(26.5)
35
(50.0)
11
(23.9)
12
(16.7)
3
(6.3)
X^
15.71*
16.08t
      *6 df, 0.02 >p> 0.01.
      t6 df, 0.02 >p> 0.01.
equal (about 50 percent) for the pink and untreated jet noises, and the pink noise resulted
in fewer (about 14 percentage points) 0 responses than did the untreated jet noise.
     These data indicate quite clearly that the jet noises  emanating from aircraft with
nacelle treatment are less disturbing than those from jets without the nacelle treatment. At
high noise levels, for example,  treated jet noise was associat