A STUDY OF THE EFFECTS OF LOW LEVELS
         OF CARBON MONOXIDE
UPON  HUMANS PERFORMING DRIVING TASKS

-------
A STUDY OF THE EFFECTS OF LOW LEVELS OF CARBON

MONOXIDE UPON HUMANS PERFORMING DRIVING TASKS
         CRC-APRAC Contract CAPM-9-69(2-70)

          June 15,  1970 - September 15, 1972


                     Final Report
 Prepared for the Coordinating Research Council, Inc. ,
 30 Rockefeller Plaza, New York,  New York, and the
 Environmental Protection Agency, Durham, North
 Carolina.
               Ross A. McFarland,  Ph.D.
               William H. Forbes, M. D.
               Howard W. Stoudt, Ph. D.
               John D. Dougherty, M. D.
               Thomas J.  Crowley,  M.S.
               Roland G. Moore, Ph. D.
               Theodore J. Nalwalk
    Guggenheim Center for Aerospace Health and Safety
              Harvard School of Public Health
                  665 Huntington Avenue
               Boston,  Massachusetts 02115

-------
                               FOREWORD
    This project has been sponsored by the Coordinating Research Council,
30 Rockefeller Plaza,  New York, N. Y. and the Environmental Protection
Agency,  Durham,  North Carolina.  The Air Pollution Research Advisory
Committee of the CRC appointed a Technical Advisory Group in cooperation
with EPA to assist in carrying out the research program.

    First of all, our thanks are due to Mr. Milton K. McLeod, Secretary
and General Manager of the Coordinating  Research Council, and to
Mr.  Alan E.  Zengel, Project Manager of the  Council.  Other members
of the advisory group who aided us greatly in  this study included Dr. John
H. Knelson and Dr.  D. W. C.  Nelson of the Environmental Protection Agency
(formerly the National Air Pollution Control Administration,  Mr. Fletcher
N. Platt of the Ford Motor Company,  Mr. Joseph M.  Collucci of the
General Motors Corporation,  and Dr. Wayne  Stewart of the Sun Oil Company.
On four different occasions this group spent from one to two days at the
Guggenheim Center to review the progress beugmade and to make constructive
criticisms aboutthe research.  These suggestions were of  special value
because of the wide experience and freedom from bias of the  advisors.
At no time was there any attempt to influence the outcome or  results of the
investigation.

    In the early stages of the project we were fortunate in having  the
assistance of Mr.  C. W. Dietrich,  of Bolt, Beranek and Newman, Inc.
Mr. Dietrich had previously carried out a study with an instrumented car,
using the Visual Interruption Apparatus.  He was able to assist in making
the appropriate installations of the equipment, and in advising our engineers
in designing certain improvements.   Later, we had the valuable advice of
Dr.  Thomas Triggs, also of Bolt Beranek, and Newman,  during the road-
test phase of our project.

    One of the most difficult phases  of the project related to having the subjects
inhale the desired percentages of carbon  monoxide mixtures  so as to reach and
maintain the desired carboxyhemoglobin levels over fairly  prolonged periods of
time.  Also, it was necessary to make frequent and accurate determinations of
the levels of carbon monoxide in the  blood.  We were fortunate in having the
technical assistance of the late Professor F. J. W.  Roughton of Cambridge
University, England, an international authority in the field of hemoglobin.  In
collaboration with Dr. W.H.  Forbes the biochemical aspects of the project
were performed with great care and accuracy.

-------
    This project was carried out chiefly during the period June 15, 1970 -
September 15, 1972.  The early stages of the work related primarily to
standardizing the laboratory tests and procedures, and instrumenting the
car.  Actually the final testing program continued until January 1,  1973.
The analysis of the data was completed during the next four months.

    During the first year, two-thirds of the funding for the project was
furnished by the Coordinating Research Council, and one-third by  the
National Air Pollution Control Administration of HEW - now the
Environmental Protection Agency.   For  the second year of the research,
funding was reduced by one third, limited to that available from the
Coordinating Research Council alone. This was because contractual negotiations
for the  second year between Harvard University and the Environmental
Protection Agency  reached  an  impasse over certain items in the contract
with the new agency.   The  most important objection was a "technical direction"
clause which was unacceptable to the University.  This reduction in the budget
for the  second year resulted in our being able to test considerably fewer
subjects than originally planned in both the laboratory and over-the-road
testing  programs.

    Special acknowledgement should be made to  Anthony  J.  Morandl, M.  A.
for his  assistance in helping to develop some of the laboratory tests in the
early phases prior  to his leaving the project to take a  position at Stanford
University.

    The recruiting and scheduling of subjects who had  passed the physical
examination and who could give enough time to the rigorous scheduling was a
difficult task. This was ably done by Bonnie Myers, Sharon Greene, and Toula
Coulas.  Mrs. Myers and Mrs. Greene also assisted in carrying out parts
of the testing program in a  competent way.  Miss Coules is deserving of
special credit for typing the manuscript and tables so  capably.  Mr.  Richard
Nardone provided essential technical assistance in regard to equipment and
and testing procedures, especially with the automobile.

    Finally, it is of interest to report that subsequently a study was undertaken
on the effects of marihuana on driving, under the direction of Dr.  John D.  Dougherty,
of the Guggenheim  Center through support from the National Institute of Mental
Health of HEW.  In  this investigation,  the effects of marihuana were studied
separately,  as well as in combination with small amounts of carbon monoxide or alcohol.
Although the same  laboratory tests were given,  a different procedure was used in
the over-the-road driving tests.  Where carbon monoxide alone was used,
significant effects  on the Complex Coordination  Test were obtained with somewhat
smaller amounts of CO than reported in the original study, although generally
the results were similar.
May,  15, 1973                                     Ross A.  McFarland
                                                     Project Director

-------
                       TABLE OF CONTENTS
Summary

  I.   Introduction                                                       \

          A.  Implications from Biostatistical Studies of Accidents         1
          B.  Representative Exposures of Drivers to the Carbon           2
               Monoxide of City Traffic
          C.  The Effects of Small Amounts of Carbon Monoxide on         7
               Human Performance
          D.  Measurement of the Performance of Drivers on the Road     9

 II.   Subject Selection and Medical Examination                          12

 III.   Carbon Monoxide Administration and Carboxyhemoglobin            13
      Monitoring

          A.  General Procedures                                       13
          B.  Equipment and Techniques                                 15
          C.  Alveolar CO Determinations                               18

IV.    Laboratory Tests of the Effects of Carbon Monoxide                 28

          A.  Introduction                                               28
          B.  General Procedures                                       28
          C.  Complex Central-Peripheral Reaction Test                 29

                 1.  Test Procedures                                    29
                 2.  Results                                             30

          D.  Dark Adaptation and Glare Recovery                        50

                 1.  Test Procedures                                    50
                 2.  Results                                             52

          E.  Peripheral Vision                                         57

                 1.  Test Procedures                                    57
                 2.  Results                                             57

          F.  Depth Perception                                          59

                 1.  Test Procedures                                    59
                 2.  Results                                             59

-------
                                                                     Page

V.   Driving Experiments                                              60

         A.  Introduction                                              60
         B.  Selection of Subjects                                      60
         C.  Roadways                                                61
         D*  The Test Vehicle                                         61
         E.  Experimental Procedures                                 65

                1. Training Sessions                                  66
                2.  Experimental Sessions                             67

         F.  Results                                                  68

                1.  Analysis  of Visual Occlusion Data                    68
                2.  Analysis  of Steering Wheel Reversal Data            73

Appendix A.  Percent COHb versus CO Pressure                        76

Appendix B.  Medical Forms                                           77

Appendix C.  Consent Form                                            80

Appendix D.  Data on Uptake of Carbon Monoxide(CO)                    81

Appendix E.  Data on Decline of COHb while Breathing Air and while
             Breathing a Mixture of 99% Q£ and 1%  COg                83

Appendix F.  Vehicle Specifications                                     85

Appendix G.  Recorded Instructions for Subjects                         86

Bibliography                                                          87

-------
                            List of Figures


Figure No.                                                             page

      1.      COHb levels during four hours of driving in                    3
             London traffic.

      2.      Effect of CO from smoking on light sensitivity,                 5
             compared to effect of high altitude.

      3.      Percentage COHb vs.  atmospheric CO, in                     6
             relation to duration of exposure.

      4.      Effects of hypoxia and CO on visual contrast                   8
             thresholds.

      5.      Theoretically expected effects of CO on                       10
             "response blocking".

      6.      Subject breathing mixture from gasometer.                   14

      7.      Subject in process of  " washing out" CO.                       16

      8.      Equipment for administration and monitoring                 17
             of breathing mixtures.

      9.      Relation of percent COHb and alveolar CO.                    19

    10.      Uptake and elimination of CO in a typical subject.             20

    11.      Rate of uptake of CO in one hour, at 710 ppm CO.             21

    12.      Elimination of CO, breathing air.                             26

    13.      Elimination of CO, breathing 99% O_/l% CO.                 27

    14.      Apparatus for Complex Central-Peripheral                   31
             Reaction Test, with subject positioned for
             test.

    15.      Distribution of omitted peripheral responses                  43
             in relation to  stimulus time and 17% COHb
             vs.  control.

    16.      Frequency distributions  of peripheral reaction                46
             times at stimulus presentation time 1.1, under
             control and 11% COHb.

-------
Figure No.                                                             Page

      17.     Frequency distributions of peripheral reaction               47
             times with simultaneous onset of central and
             peripheral stimuli,  for control vs.  11% COHb.

     18.      Visual discriminometer, with subject in                     51
             position for test.

     19.      Average curves of dark adaptation,  pretest vs.               54
             17% COHb.

     20.      Average curves of dark adaptation,  pretest vs.               55
             11% COHb.

     21.      Average curves of dark adaptation,  pretest vs.               56
             control.

     22.      Apparatus for peripheral vision test, with                    58
             subject in position for test.

     23.      The Visual Interruption Apparatus,  with                     62
             visor raised for full road vision.

     24.      The Visual Interruption Apparatus,  with                     63
             visor lowered to occlude vision.

     25.      Overall view of Visual Interruption Apparatus                64
             and control and recording  equipment.

     26.      Sample of record  of visual occlusions.                        69

     27.      Sample of record  of steering wheel movements               74
             and reversals.

-------
                         LIST OF TABLES


Table No.                                                            page

   1.   Central Responses - Control vs. 11$ COHb Tests(21 Subjects)     33
        Central Responses - Control vs. 17$ COHb Tests(22 Subjects)     33

   2.   Change in Average  Numbers of Central Responses in Subjects
          Grouped by Order of Test Days                                31;

   3.   Average Numbers of Central Responses by Day of Test            35

   i|.   Summary of Response Times to Central Stimuli                   36

   5.   Summary - Numbers of Peripheral Responses                      37

   6.   Changes in Average Numbers of Peripheral Responses in
          Subjects Grouped by Order of Test Days                       38

   7.   Average Number of Incorrect Peripheral Responses by
          Stimulus Location                                            39

   8.   Average Number of Omitted Peripheral Responses by
          Stimulus Location                                            39

   9.   Comparison of Numbers of Subjects  with One  or  More Instances
          of Response Blocking in Control  and CO Tests                 1^1

  10.   Prolonged Reaction Times to Peripheral Stimuli as Percentage
          of Responses which Exceed Selected Times  at  Stimulus  Pre-
          sentation Times .1; Sec.  Before to .1; Sec.  Following Central
          Stimulus Time                                                ^9

  11.   Daily Timetable for- Road Testing                                68

  12.   Average Percent Occlusion Time as  a Function of Gas Intake
          and Vehicle Speed.                                            71

  13.   Changes in the  Cumulative  Distance  Parameter as a Function
          of Gas  Exposure.                                              72

-------
                               Summary
      The purpose of this research was to determine the effects of low levels of
 carbon monoxide on human performance in driving-related laboratory tasks and
 in over-the-road vehicle driving.  Twenty-seven subjects ranging in age from
 20 to 50 years participated in these experiments under conditions  of 17% COHb,
 11% COHb, and "Control",  or no-administered CO.  The laboratory tests
 measured:  (1)  Complex psychomotor reactions involving simultaneous perform-
 ance of a primary and secondary task;   (2) dark adaptation and glare recovery;
 (3) peripheral vision; and (4) depth perception.  The driving task was designed to
 evaluate driver visual information needs and the steering wheel movements
 required to keep a vehicle properly positioned within the driving lane at different
 speeds.

      The  CO administration and COHb  monitoring phase of the study demonstrated
 considerable intersubject variability both in CO retention and in the rate of CO
 uptake.  The latter may be  inversely related to age. Rates of CO elimination in
 air and in oxygen also showed considerable variability between subjects.

      Results of the laboratory tests are as follows: For the central and peri-
 pheral complex task the subject responded to red or green lights presented in his
 central field of vision by pressing foot switches, and concurrently responded to
 any one of six lights which might come on in his peripheral field by pressing
 appropriate finger buttons.   When the numbers  of correct,  incorrect and omitted
 responses on the central task were compared, no significant differences were
 found between CO and control conditions, except for a suggestion of more
 incorrect responses with CO with minimal prior test experience.  Thus it  is pos-
 sible that a deleterious CO  effect would  be most apparent during the learning
 period, or during unfamiliar situations.  COHb levels of 11% and 17% showed no
 effect whatever  on central task reaction times.

      When the peripheral task is considered, the overall numbers of correct,
 incorrect and omitted responses  showed no CO-related differences, though there
 was  a finding of greater variability with CO at one level. The relationship of
 the location of the peripheral stimuli, i. e.,  15°,  30° or 45° from  center to
 numbers of incorrect or missing responses showed  no consistent pattern suggest-
 ing a general effect from CO, but there  was an  increase in omitted responses
 at 30° out from  center on one side at both levels.

      Response blocks, or attentional lapses or gaps in  performance, appear to
 have a CO-related effect of marginal statistical significance.  It does appear
 that more subjects  showed response blocking at both CO levels than under the
 control condition.

      If the possible effects of CO on interactions between the central and peri-
pheral tasks are considered, we  find that peripheral responses are more
frequently omitted at 17% COHb,  but less markedly  at 11%, if their stimuli

-------
 appear close in time to a central stimulus.  This may be a normal tendency
 which is enhanced by exposure to CO.  When response times to peripheral
 stimuli are considered,  there appears to be a tendency to longer reaction
 times when the stimuli are close to a central stimulus, but here a  close CO
 effect is not apparent.  Nor did the frequency of excessively long response
 times appear to be related significantly to CO.

      When final dark adaptation threshold values obtained on different days
 from the  17% COHb,  11% COHb and Control sessions are compared, no sig-
 nificant differences were found.  However, when intraday pretest (before
 gassing) results were compared with post-gassing 17% COHb and control
 results, both showed statistically significant differences in the direction  of
 more light needed during the test than the pretest.  The direction of change
 was the same for  11% COHb test-pretest,  though the difference was not
 significant.  Comparisons at other time points on the dark adaptation curve,
 i. e., at 1,  4,  and 10 minutes showed a generally similar pattern of pretest-
 test differences.  However, since  the control (or placebo) session  produces
 results which parallel those of the CO sessions, there is a strong suggestion
 that the gassing procedure itself, and not the resulting COHb,  is producing
 these effects.  Glare recovery time  showed no significant CO-related  differences.

      On a test of peripheral vision subjects missed significantly more targets
 presented at 20° from their central fixation point with 17% COHb than  under
 Control conditions.  They also missed more at 11% though the  difference was
 not statistically significant.  There does appear to be a CO-related decrement
 in peripheral vision here, though the extent of the data does not permit firm
 conclusions in  this regard.

      No differences whatever were found between measures of depth perception
 with a standard Verhoeff apparatus at COHb levels of 11%,  17%, and Control.

      For the driving phase of the  study a test vehicle was equipped with  a
 Visual Interruption Apparatus for evaluating driver attentional demands and
 information processing,  as well as with a potentiometer for measuring steering
 wheel movements.  With the visual interruption apparatus the driver's vision
 was normally occluded by a visor until he  pressed a foot switch giving him  a
 0.5 second "look".  When occluded distances travelled on standardized runs
 were compared for CO (17% COHb) and no-CO conditions, no  statistically
 significant differences were found at either 30 or 50 mph, i.e. , drivers were
 demanding no more,  nor less,  visual input with CO than without it. It was also
 observed that at 50 mph the percentage of occlusion time was always less
 (i.e. ,  the driver needed more visual information) than at 30  mph.  This held
 true for both CO and control conditions, but with CO the'subjects required
 relatively more roadway viewing at the higher speed than they did without CO.
 This CO-related effect was statistically significant.

      Analysis  of steering wheel reversals standardized for roadway distance
and for time indicated only negligible changes associated with the administration
of CO, intra- and intersubject variability being the most significant factor.
 Thus 17% COHb was found to have no significant effect on steering wheel reversals.

-------
 I.  Introduction:  Statement of the Problem

       The purpose of the present investigation was to study the effects of small
 amounts of carbon monoxide on human performance in driving-related tasks  in
 the laboratory, and in vehicle driving over the road.  The first phase of the
 study involved the utilization of laboratory tests.  The procedures developed
 and employed in this program are those believed most likely to influence the
 visual reactions and control responses  of the driver.

       One of the major problems  in studying the effects of carbon monoxide on
 driving performance relates  to the development of sensitive tests.  Many of the
 studies previously carried out by others have been handicapped by the fact  that
 the effects of CO on the subject could have been concealed by exerting greater
 effort, or by failure  to control for practice effects or learning.  In our  studies
 of the effects of altitude, carbon monoxide and of other agents in the environ-
 ment, we have developed techniques which have proved to be reliable and
 effective in measuring threshold values.  These have included sensory tests  of
 visual perception and tests of mental performance and information processing.
 During the first year of the study, considerable effort was expended on  the
 design of the laboratory experiments  and the fabrication of the equipment.
 After all of the  experimental designs  were finalized and all laboratory systems
 were operational, data were obtained on twenty-seven subjects in the laboratory
 phase of the project.

      The second phase of the study involved over-the-road driving experiments
 with an automobile especially designed for this experiment at the Harvard
 School of Public Health with assistance  from its subcontractor, Bolt, Beranek
 and Newman,  Inc. , of  Cambridge, Massachusetts.   Experimental protocols
 utilizing a Visual Interruption Apparatus were developed.  Each phase of the
 research program, including CO administration and monitoring, laboratory
 testing, and driving experiments, are  described in detail below.

      A.  Implications  from Biostatistical Studies  of Accidents

      It has been brought out repeatedly that  automobile accidents are non-
 repetitive in nature,  with non-specific causation and usually resulting from
 multiple causes.  It is  well known from  biostatistical studies that youthful age,
 lack of training  and experience, changes in skill with advanced age, and use of
 alcohol or  drugs are  factors which frequently enter into the interactions
 between the driver, the vehicle, and the environment.  It also has been  observed
 frequently  that the amount of illumination is an important factor and that high
fatality rates tend to  occur  during twilight or at night.  There may also  be
 greater consumption  of alcohol during these times of day.  (McFarland and
associates,  1972)

      Although careful  experimentation has not established the possible  ways
by which small amounts of carbon monoxide may influence driving perform-
ance, it might be assumed, however,  that the effects of small amounts existing

-------
 in city streets from automobiles might have a cumulative or synergistic effect
 in combination with other variables.  For example,  the role of carbon monoxide
 in impairing the uptake and delivery of oxygen to the tissues is very pronounced.
 In fact at body temperature, the hemoglobin molecule has an affinity for carbon
 monoxide about Z10 times as great as that for oxygen.   If a person has anemia,
 or has carbon monoxide in his blood from cigarette  smoking,  the small amounts
 from city streets might have more serious effects.  It is also well known that
 the effects of alcohol impairs the transport of oxygen to the tissues.  Also,
 there  are many commonly used drugs which have similar effects.   It would be
 unrealistic to assume,  therefore,  that drivers will be  influenced by any  one of
 the above possibilities, but more probably a number of them acting in combina-
 tion.   (McFarland, 1963,  1970)

       B.  Representative Exposures of Drivers to the Carbon Monoxide of
          City Traffic

       It is very important to consider the amounts of carbon monoxide which
 might  be found in city streets.  Obviously this would be a function of atmos-
 pheric conditions and winds, as well as the back-up of traffic on congested
 streets or in peak traffic periods.  Also,  it may be some time before  the
 improvement from the newer models with better control of engine exhaust will
 be apparent.  Only a few  examples can be reported here of the amount of carbon
 monoxide prevailing at present.

      During April 1971 one of us had an opportunity to visit the Medical
 Research Council Air Pollution Unit at St. Bartholomew's Hospital Medical
 College in London.  During the conference with the Director,  Dr.  P. J. Lawther,
 it was  possible to obtain information in regard to their studies on carbon
 monoxide in the City of London.  The variation in carbon monoxide in  the blood
 of four subjects driving for four hours in heavy London traffic is shown in
 Figure 1.  The average amount of  carbon monoxide in the streets was  reported
 to be 35 ppm.  It is interesting to note the striking difference between  non-
 smokers and smokers.  (Lawther, 1970)  Dr. Lawther reported that the  more
 recent studies have been  similar.  Also, it is interesting to note that no  impair-
 ment has been observed in a large number of psychophysical tests administered
 under  conditions simulating the amounts of CO in London city streets.

      One of our colleagues at the Harvard School of Public Health,
 Dr. Benjamin G.  Ferris, Jr. , has been carrying out a study  relating to the
 amounts of carbon monoxide found in Boston policemen.  In eighty subjects
 studied thus far, the highest value of COHb,  14.4%,  was found in a policeman
 who inhaled cigar smoke over long periods of time.  The next highest value
 was 12. 6%,  followed by a third at 10. 5%.  Almost all of the non-smokers had
 values of COHb  below 4. 0%,  and most of the smokers below 6. 0%.

     Any study of the effects of carbon monoxide on driving performance should
take into account the fact  that chronic cigarette smokers tend to have from 4 to
8% carboxyhemoglobin in  their blood.  The importance of this fact has  been

-------
                      SMOKER
  10
8
EQUILIBRATION  SATURATION
        NON  SMOKERS
                                  1
                                  2
                                 Hours
          Figure 1.  Variations in COHb % in four subjects during four hours
             driving in heavy London traffic;  mean street CO approximated
             35 ppm.  The uppermost curve represents a smoker starting the
             experiment after smoking 17  cigarettes;  the lower curve at the
             top represents a non-smoker  "gassed" to the level shown at the
             start.  (Lawther and Commins,  1970)

-------
 brought out in relation to the combined effects of carbon monoxide and altitude
 in pilots.   The true physiological altitude of an airman would be related to the
 decreased amount of oxygen while in flight.  Nomograms have been developed
 so as to show these combined relationships (McFarland,  et al.,  1944).  These
 data were developed on the  basis of the diminished ability to see targets at low
 levels of illumination with the dark adapted eye, along with the amount of
 carbon monoxide in the blood.

       The effects of small amounts of  carbon  monoxide from smoking are
 demonstrated in Figure 2.  Each subject studied was first brought in a chamber
 simulating high altitude to approximately 7, 000 feet or  higher, and the dimin-
 ished ability to see was determined.  After returning to sea level conditions,
 the subjects were then asked to inhale the smoke of three cigarettes.  As shown
 in the top part of Figure 2,  the effect of the saturation of the blood with CO
 (4. 1% COHb) was equal to that of about 7, 500  feet altitude.  Therefore, the
 subject was at a physiological altitude  of 7, 500 feet while still at sea level
 (McFarland. 1946)

       In recent years an interesting problem has arisen in the building of high-
 way tunnels at high altitude  in the Rocky Mountains.  For example, what would
 be the effects on drivers and passengers in the tunnels at 11, 000 feet? Three
 possible sources  of hypoxia  would result as follows:  (1) decreased partial
 pressures of oxygen due to elevation;  (2) inhalation of CO from cigarette smoke;
 and (3) inhalation of CO discharged in motor vehicle exhaust.   By means of the
 nomograms mentioned above and  other data, the combined effects of altitude,
 smoking and tunnel  CO were determined,  and  it was recommended that the CO
 concentration in the tunnel be maintained below 25 ppm,  among a series of
 other recommendations  in regard to smoking and the use of oxygen for engineer-
 ing purposes.  (Miranda, et al.,  1967)  It should be kept in mind that many parts
 of the United States have highways and cities at moderately high altitudes where
 the effects of carbon monoxide for non-acclimatized persons,  both smokers and
 non-smokers,  might be  accentuated.

      As indicated above, the amount of carbon monoxide resulting from auto-
 mobile exhaust is of importance in relation to its possible effects on driving
 performance.  In  Figure 3 the percent  of COHb in the blood is shown in relation
 to the percent and parts per  million of  CO in the atmosphere.   (Forbes,  1970)
 It is obvious from this figure that the length of the exposure is of great
 importance in relation to the concentrations in the blood.  Other variables  relate
 to the amount of activity, as well  as  general physical condition of the subject.
 It can be determined from this figure that the amounts of carboxyhemoglobin
 which might accumulate in the blood of the average  driver during short periods
 of time would be very small. Concentrations as high as 50 ppm have been
 observed only infrequently in city streets of the United States.  A study of the
 amounts of carboxyhemoglobin in  various segments of the population has
 recently been reported by Stewart (1972).  To aid the reader in interpreting
the various values, the relationships between the amount of CO in the atmosphere
and the percent of CO in the blood (COHb) are shown in Appendix A.

-------
                     EFFECT OF CO  FROM  INHALED SMOKE OF CIGARETTE
   36
      ROOM AIR
                15 7 X 02
               7.SOO FEET
               192 X 0;Hti)
                   100 X02
                          CIGARETTE
                             ROOM MR

                          CIGARETTE     CIGARETTE
7XC02* 91\02
                                                                                       230
 o
•a
o
ROOM AIR
  30
II09X 02

15,400 FEET
|76X02Hb)
                      IOOX 0,
    10      30      50
                                     CIGARETTE
                           05% COIIt
                       _^	I.J.I	L
                    70     90 MO     I JO      150
                                TIME  (MINUTES)
                                                                  190
                                                                          210
                                                                                 230
      Figure 2.  The effect of carbon monoxide from smoking on the light
         sensitivity of the eye is shown in comparison with that of high
         altitude for two subjects.   (McFarland,  19&3)

-------
                     PARTS  PER MILLION IN  ATMOSPHERE

                     200     400    600    800    1000
                                           QUITE UNCOMFORTABLE
                                         LIGHT WORK POSSIBLE
                                             I        '
1200
               0     002    0.04    OD6    0.08    0.10    0.12

                        PER  CENT  CO IN ATMOSPHERE
Figure3.  Percent COHb in blood vs. atmospheric CO at one, two,
   and four hours, and at infinite time.  The effects of various
   percentages of COHb are given on the curves of percentage
   saturation.  These are the effects expected if the individual is
   raised rapidly to the percentage COHb in question, and then
   maintained at that level.  There are considerable variations,  in
   persons in apparently good health,  and the symptoms also vary
   from person to person,  with headache the most common
   (Forbes,  1970)

-------
      C.  The Effects of Small Amounts of Carbon Monoxide on Human
          Performance

      In our laboratory many studies have been carried out on the effects of
lowered oxygen tension on the central nervous system whether produced by
ascents to high altitude or by agents which impair the delivery of oxygen to the
tissues, such as  carbon monoxide.  To obtain precise results relating to
impairment,  the  tests should possess a high degree of sensitivity and be inde-
pendent of conscious effort or learning on the part of  the subjects.  We have
found that psychophysical tests of sensory functions,  such as light sensitivity,
have many of the desirable characteristics.  Also it has been possible to devise
other tests that detect and scale increased effort.  In addition,  measures of
speed and accuracy can be combined into measures of rate of information pro-
cessing and space or reserve capacity.  (Me Far land,  1970)

      In this  report it will be possible to review only  a few of the studies which
have been concerned with impaired oxidation  in the nervous tissue.  The visual
tests selected were as follows:  (1) visual acuity at low  levels of illumination;
(2) dark adaptation; and (3) differential brightness sensitivity.  In studying the
effects of high altitude it was found that sensitivity of the dark adapted eye was
significantly  impaired as low as 4, 000-6, 000 feet altitude. It was also observed
that the effects of oxygen want were much greater at low levels of illumination.
(Me Far land,  et al. ,  1941) Additional studies were carried out on the effects of
small amounts of carbon monoxide.   The above tests also proved to be very
sensitive.  Analysis showed  a comparison of  the effects of carbon monoxide and
the reduced partial pressure of oxygen on the .brightness thresholds.  It is
relevant to note that the increase  in threshold or poorer performance was
apparent at 5% COHb,  or 25  ppm of carbon monoxide.  In Figure 4,  similar
comparisons have been made in respect to the effects on visual contrast thresh-
olds with various background luminances. This proved to be one of the most
sensitive tests, and  significant impairment resulted with highly trained subjects.
(McFarland,  et al. ,  1941, 1944)

      Studies with other psychophysiological tests have  shown that larger amounts
of carbon monoxide or higher simulated altitudes were required to produce
significant impairment.  Neuromuscular coordination and pursuit tasks were
influenced only in more advanced  stages of oxygen want, with auditory acuity
being most resistant to hypoxia.   Certain cognitive functions  such as immediate
memory proved to be more vulnerable,  the effects beginning  at about 7, 000
feet simulated altitude, or 5-8% COHb.  It is interesting to note that one of the
most sensitive tests of the effects of aging relates to  diminished visual acuity
at low levels of illumination, as well as loss  of capacity for short-term memory.
These results might be expected since it has  been fairly well established that
there is lowered  oxidation in the nervous tissue in one form or another during
the processes of  aging.  (McFarland,  1963)  Other studies have been reported
of the impairment which may result from small amounts of CO in selected
cognitive functions.  (Schulte,  1963)

-------
        Figure l\..  Effects of hypoxia and carbon monoxide on contrast thresholds for targets with various

           background luminances.  (Data from McFarland, Evans,  and Halperin, 1941)
CO
HI
S
                                                                            LOG
                                                                            O« -3.O7IHL
      0.1-1
      0.0
                                                             LUM.
                                                                                               a
                                                                                               a.
                                                                                               O
                                                                                              U
                                                                                              ••*
                    95
9O
85
8O
75
7O
65     ART. O2  SAT.  (%)
                                     SO  70
                        125
                      I7O
                     2OO
                             CO (ppm)
                                         10
                      15
                      20
                      25
                      30     COHb (%)
                                                                                                                 00

-------
       One of the most significant influences of reduced oxygen tension relates to
 the subject's capacity to receive and interpret information in the visual field
 such as in driving.  There is a tendency for the  central nervous system to
 respond in a less flexible manner.   Ideas tend to persist, or perseverate, and
 there may even be a blocking of responses.  These tendencies have been fre-
 quently observed in studies of oxygen want.  One of the earlier ones brought
 out the important principle of "blocking" during  hypoxia and mental fatigue.
 (Bills, 1937)  This may be an important way of interpreting the accident potential
 relating to driving performance as Teichner  has recently pointed out.  Indeed,
 response blocking may be one of the most  characteristic reactions which may
 occur during continuous, high-information processing tasks.  It may be  invol-
 untary and subject to all of the factors which affect driving skill thus related to
 accident probabilities.  (Teichner, 1968)  In Figure 5 the effect of carbon
 monoxide on frequency of response blocking is shown as postulated from studies
 relating to impaired oxidation, whether from high altitude or carbon monoxide.
 Similar effects might be expected from loss of sleep, excessive alcohol,  or
 various drugs.

       D.  Measurement of the Performance of Drivers on the Road

       Experimental attempts to measure the performance of drivers while
 operating vehicles on the roadway have been quite limited in their extent.  It is
 difficult to carry out controlled experiments with the driver's  being aware of
 the measurement procedures and being influenced by them.  Furthermore, the
 demands on  the ability of the driver to process the  incoming information vary
 greatly. At times the demands are considerably below his full capacity; at
 others this may be exceeded.  Some of the needs for information relate to tasks
 in the immediate and  ongoing control of his vehicle, for example,  in relation to
 lane holding, maintaining distance from a leading car,  steering, braking, or
 accelerating in relation to other traffic or  roadway hazards.  In addition the
 monitoring of dashboard instruments  and of roadside directional and warning
 signs  is necessary.  Thus, one approach to the measurement of performance
 may be based on the information processing inherent in driving,  utilizing the
 techniques of information theory.

      In the context of automobile driving,  one method has been concerned with
 measuring information processing as reflected in changes in performance on a
 subsidiary task such as mental arithmetic carried out while operating a vehicle
 under  various conditions of traffic and roadway complexity.  This method has
 been useful  in demonstrating  reductions in  "spare mental capacity" as attentional
 and information processing increases with  driving complexity.

      Another approach utilizes a technique of intermittent visual time sampling.
 The  field of view of the driver is interrupted systematically as he operates a
 vehicle on the roadway. This is accomplished by a helmet-mounted visor which
 alternately drops to occlude the driver's vision at controlled frequencies and
 duration.  Initial experiments have  been conducted on closed road systems and
unopened sections of interstate highways.   It was found that sufficient information

-------
   200
LU
CO
<
LU
a:
o
— 100-
LU
o
a:
LU
Q_
                 •NUMBER OF BLOCKS
                 •DURATION  OF  BLOCKS
               95
                                                                                 ART. 02 SAT. 1%)
          20.  18
     16
                                          12.
                                     10.50
         159
132
109
                                               90
                       25
                50  70
 P02 in 02-N mix.


P02 in Breathing  Air,mm
                                                125
                                        170
                                                                 200
                                                    CO 
-------
                                                                          11
for error-free lane holding operating was obtained if drivers viewed the road
for a period of one-half second at two-second intervals at a speed of 60 mph
at every four seconds at 25 mph, and every nine seconds at 5 mph.   The
records were  obtained under conditions where (1) the occlusion rate was
controlled by the experimenter and vehicle speed by the subject, and (2) where
vehicle speed  was controlled and the driver actuated the occlusion device as
needed.  The results were used to develop a mathematical model which related
the driver's informational content of the roadway in "bits" per mile, the  speed
of the vehicle,  and the driver's  estimates of his  own uncertainties.   (Senders,
et al., 1967)  This work was subsequently extended to include the information-
processing demands of car-following and passing,  and driving under normal
traffic conditions.

       A second application of information theory relates to the recognition of
signs and traffic  control devices.  Objective measures  of recognizing a variety
of roadside signs by subjects of different age and socio-economic status were
obtained in experiments  carried out on drivers wearing the vision interruption
device described above.   These data were related to the probabilities provided
by signal detection theory in a way which resulted in an index of relative
recognizability.  Thus far no studies  have been carried out with these tech-
niques while drivers were being influenced by carbon monoxide.  (Senders, et al. ,
1969)

-------
                                                                           12
II.  Subject Selection and Medical Examination

      The subjects for the laboratory phase of the present research ranged in
age from 20 to 50 years, with the majority in their twenties.  Both smokers
and non-smokers were included,  and although the sample was not stratified on
the basis of smoking behavior, smoking histories were obtained on each subject
for use in the data analysis.  All subjects  were paid $60 for participation  in
three laboratory sessions of about 4-5 hours duration each.  Subjects were
obtained through posted notices at the Harvard School of Public Health and
Northeastern University,  through the Massachusetts  Employment Service, and
through advertisements in the three major Boston newspapers.   The total
number of  laboratory subjects was  27.  Subjects for the driving phase were
selected from those  who have satisfactorily completed all of the laboratory work.

      Medical selection of the subjects  was carried out prior to the experiments.
Before selection, questionnaires were used to obtain subjects who had no  contra-
indications to the experimental hypoxia.  (See Appendix B) This procedure ruled
out individuals with  a history of disease involving the neurological, cardiovascular,
renal, hematopoetic or pulmonary  systems.   Subjects with a history of an allergic
or toxic response to any drug or chemical were not accepted.   Individuals with a
history of any emotional disturbance which required  treatment  were also  excluded.
All subjects  were initially tested for visual requirements  with a Titmus Vision
Tester in the laboratory. The medical examination included neurological and
pulmonary function tests, chest x-ray and complete blood count and urinalysis,
and an electrocardiographic examination,  as  well as a double Master's 2-step test
for subjects  over 39 years.

      Normal standards were defined as equivalent to U.S. Air Force standards
for entry into flight  training except as follows:  (1) Any subject with uncorrected
visual acuity due to  myopia or hyperopia which corrects to 20/20 was acceptable.
(2) Hemoglobin must be greater than 14. 5 gm/100 ml. (3) Vital capacity values
must be 80% of predicted for height and age,  and 1-second forced expiratory
volume 70% of the total.  (4) Sitting pulse  rate must be less than 85 beats per
minute.  After 20 hops the pulse rate must not be greater than  120 beats  per
minute, and 2 minutes after exercise not greater than 100 beats per minute.

      During the testing program no subject was allowed to continue exposure to
CO if his  symptoms warrant termination.  In addition, CO exposure was  term-
inated in the event of faintness, dizziness or apparent loss of gross mental
function.   At the end of their participation, all subjects underwent a second phys-
ical examination identical to the procedure used for  initial selection.   (See
Appendix  C for copy of consent form signed by all subjects. )

-------
                                                                         13
 III.  Carbon Monoxide Administration and Carboxyhemoglobin Monitoring

      A.  General Procedures

      The procedures for the administration of carbon monoxide (CO) and for the
 monitoring  of the Carboxyhemoglobin (COHb) levels of each subject participating
 in the experiments were as follows:  Immediately upon arriving at the laboratory
 the subject  had both a blood sample and an alveolar (expired air) sample taken
 to determine his initial COHb levels.  He was then seated comfortably,  a nose
 clamp applied, and he breathed one of two gas mixtures through a  mouthpiece
 (see  Figure 6).  Initially a face mask was  used to administer the mixtures, but
 leakage around the face seal on many subjects necessitated a change to the
 mouthpiece, which though less comfortable and convenient, eliminated the
 problem of  leaks.

      Each  subject breathed either normal room air for a control session in
 which his CO exposure was called "zero, "(that  is, no CO was added), or an
 air mixture containing 720 ppm of CO.   In either case the subjects breathed
 from a 500  liter  Tissot gasometer and exhaled into a second one,  so that by
 measuring the  CO in  a small sample from the second  Tissot and knowing the
 content in the first Tissot, we could monitor the amount retained by the subject.
 There were differences in rate of uptake of around 15% or more between sub-
 jects, and variations  within the same subject of as much as 10%, for reasons
 we believe were  probably due to differences in "restlessness" during the period
 of exposure to  the CO.  The Tissot gasometer system is  shown in Figure 6.

      During the  first set of experiments on 9 subjects,  four different levels of
 exposure  to CO were  used,  zero and enough to produce 6% COHb,  11% COHb,
 and 17% COHb.   In practice it was difficult to achieve  these figures precisely.
 Those of our subjects who were smokers came in with 2% to 6% COHb, and the
 non-smokers from 0. 5% to 2%, so that anything under 4% COHb was used as our
 "zero" exposure. The varying rates of uptake even in the same individual also
 made it difficult to attain the 6, 11 and 17% goals so that we were  sometimes .
 over  or under the desired saturation by  1% and occasionally 2% COHb, that is,
 in aiming at 17%  we might be as high as  19% or  as low as 15%.  The exposure
 lasted 80  minutes since we were using only 720  ppm CO in the inspired air.
 However,  if the subjects were low, they were given a  short additional CO
 exposure  to  raise their level to about 17%.  If high,  the tests  were run at this
 slightly higher  level,  as  the subjects fell from 19% to  17% after 20 minutes of
 breathing air during the  testing.

      After  the initial 80-minute exposure, blood and alveolar samples were
 obtained and COHb levels determined.  The subject then completed one portion
 of the experimental tasks.  At this time  an alveolar  sample was again taken, and
 the subject was re-exposed to the mixtures (room air  or CO mixture) for a short
period to regain the desired COHb level.  Alveolar and blood  samples were
again taken  and the subject completed the remainder of the experimental tasks.
After another alveolar sample the subject was "washed out,"  i. e. ,  had his COHb

-------
                                                                14
Figure 6.  Subject breathing mixture from gasometer.

-------
                                                                          15
level reduced by breathing a mixture of 99% oxygen and 1% carbon dioxide for
approximately 80 minutes.  (See Figure 7.)  No subject was released until his
COHb level was below 6%.  Final alveolar and blood samples were obtained to
determine this level.

      B.  Equipment and Techniques

      COHb levels were determined with a Model 182 CO-Oximeter (Instrumenta-
tion Laboratories).  A 3 cc venous  sample of blood was obtained:  (a) upon arrival
at the laboratory;  (b) after inhalation of the CO mixture;  (c) after re-exposure to
the CO mixture, and (d) at the end of the session. Determinations were made
twice with same sample which gave hemoglobin in grams/ml, oxyhemoglobin in
grams percentage /ml, and carboxyhemoglobin in grams percentage/ml.  The
second determination was within plus or minus 0. 2% of the first determination.
Correlative curves have been made with COHb levels and alveolar (ppm) levels
(see below).

      The CO-Oximeter has proven to be  reasonably reliable in determining the
COHb levels used in this study.  Scale linearity is well maintained throughout,
except below 3% where the findings are more questionable.  Maintenance of the
instrument has presented problems, however,  and it has been necessary to obtain
three replacement-instruments (under warranty) due to major component failures.
In each case the instrument had to be returned to the manufacturer for repair and
re calibration.  The last of the three instruments was thoroughly satisfactory.
A considerable amount of time was also expended establishing a set of techniques
for the operation of the instrument that would provide the required level of accuracy.

      A Beckman 315 AL Infrared CO Analyzer was used to determine the CO
concentrations in air and alveolar samples, and in the room air.  This instrument
has three ranges of CO sensitivity, 0-100 ppm, 0-500 ppm, and 0-1000 ppm.
This wide range of sensitivity is needed because of the diversity of CO concentra-
tions in the air samples that must be  tested by the equipment.  These samples
are:  (1) the CO breathing mixture at  720 ppm CO;  (2) air from the inspired air
line,  also at 720 ppm;  (3) air from the expired air line at about 370 ppm;
(4) total expired air, also about 370 ppm;  (5) Douglas bag samples at 720 ppm;
(6) alveolar samples, which may range from almost zero to 140 ppm; and,
finally,  (7) room air which may vary  between 0-10 ppm.

      The CO  gas mixture was made up in a 600 liter gasometer,  utilizing 380 cc
of CO with 530 liters of air.  This gave a  standard mixture of 720 ppm which was
used for all the COHb levels, with varying exposure times to the mixture used to
obtain the different levels.  It should  be noted that all subjects were on the
breathing apparatus  for the same amount of time for all sessions.  However, for
the lower levels of COHb, the subjects breathed only room air for a portion  of the
time - a fact of which they were unaware.  Figure 8 depicts an overall view  of the
gas apparatus  including gasometers and Beckman Analyzer.

-------
                                                                    16
Figure 7.  Subject "washing out" CO with 99% O / 1% CC>2 mixture.

-------
                                                                    17
Figure 8.  Gassing equipment including gasometer (rear),  infrared
analyzer, and control panel (right).

-------
                                                                             18


      C.  Alveolar CO Determinations

      The accuracy of COHb determinations has improved with experience in
 operating the Instrumentation Laboratory's CO-Oximeter.  The probable error
 in COHb is about _+ 0. 2%.  At values below 3% COHb the accuracy appears to be
 decreased somewhat.  In the alveolar samples the critical problem is to obtain
 an accurate sample of air from the subject.   With experienced  subjects  the
 variations usually are not over  5 ppm which corresponds to 0. 7% COHb.  (See
 Figure  9. )

      Figure  10 is  typical of the rate of uptake and elimination  of CO that we
 have obtained on our subjects.  On these  curves solid circles represent the
 values of percent COHb obtained by analysis of the blood.  Open circles repre-
 sent the values obtained by analysis of  the alveolar air translated into percent
 saturation from a standard curve derived from the formula

                        % COHb  =  Z30pCO
                        %O2Hb      arterial pO2

 and the assumptions that the arterial blood saturation is 98% and the arterial
 pO2 is 98 mm.  The values derived from the  blood samples have been used in
 drawing the curves since they are more consistent than the alveolar samples,
 but the  average discrepancy is under 0. 7% saturation.

      The line of small dots on the left  of the higher curve in Figure 10 is the  rate
 of uptake obtained in another experiment  on the same subject at 720 ppm,  the
 same exposure as in the solid curve. Although the elimination of CO in air and
 in oxygen actually follow logarithmic curves, as noted below, they are drawn
 here as straight lines, since intermediate measurements were not made.

      The subjects  showed considerable variation in the proportion  of the  inspired
 CO which they retained.   This variation was apparent not only from subject to
 subject but also within the same subject on different exposures, and even  within
 the  same exposure, fluctuations were observed.   The expired air was run con-
 tinuously through the Beckman Analyzer,  and the contents were thus observable,
 not breath by  breath, but over  15 to 20  second periods, since the analyzer
 responded primarily to the air that had passed through it during the time period.
 Consequently  a running integration was  available of the values of the last 3 to 4
 breaths.

      The variations in these values were larger than expected.  The retention
was occasionally as low as 35% of the inhaled CO or as high as  55%. The average
was 45%. The variations  were  quite irregular and were presumably due to the
 restlessness or boredom of the  subjects.   However, these variations were
 infrequent, and most of the time the subject retained from 42 to 48% of the
 inspired CO.  These irregularities tend to even out over time so that the uptake
 appears to be regular and  also linear within the limits of experimental error for
 periods of about one hour, or until about  one-third of the final equilibrium

-------
               % COHB VS ALVEOLAR CO
                                                 19
7
40      60      80      100

 ALVEOLAR CO IN PPM
120
Figure 9..  Percent COHb vs. alveolar CO in subjects from present
study.  Each dot represents one determination on a subject.

-------
                                                        May 13 6%	Way 20  11%—
o
o
o
o
                                                                    • BLOOD SAMPLE


                                                                    O ALVEOLAR SAMPLE
         9:20    9:40  1O:00  10:2O  10:40   11:OO   11;


                                           TIME
2O   ll:4O  12:OO  12:2O  12:40   l:OO
             Figure 10. Uptake and elimination of CO in one typical subject in present study.
                                                                                                      #0

-------
                                                                         21
 saturation value is reached.  With a concentration of 720 ppm in the inspired
 air the equilibrium value is 57% so that one would expect to be unable to
 measure any deviation from linearity  until 19 or 20% COHb is  reached.

      A wide range in the rates of uptake in our 28 subjects was found, as is
 indicated  in accompanying Figure 11.  In these graphs both the 11% and  17%
 experiments are shown and the rates calculated from each.  Ih both experiments
 they were exposed to the same 720 ppm of CO and the only difference was the
 time of exposure.   In 8 of the 9 subjects the discrepancy between the two runs
 is well beyond the experimental error of the measurements and is  almost
 certainly  due  to differences in  subject activity (i. e. ,  restlessness  or dozing).
 There was no  consistency as to which run was higher, that is,  sometimes it was
 the 11% run (which was the first one in all but two subjects) and sometimes the
 17% run. The values for each subject are  shown  in the Appendix D.

      The age of the subjects is shown on the graphs as  is their cigarette use.
 The black dots indicate cigarette smokers, all of whom  inhaled,  and the  half
 black dots designate light smokers.  One man of 23, WM, who did  not smoke
 cigarettes was, however, a light pipe  smoker (2 or 3 pipefuls per day).   He did
 not inhale and so was classed with the non-smokers.  If the subjects are divided
 into three groups,  the 8 who took up CO  most rapidly, the middle 9, and the
 slowest 9, only two cigarette smokers are found in the first group, and four in
 each of the other groups.   (The pipe smoker was in the first 8. )  No conclusion
 should be  drawn from these observations.

      Our subjects ranged in age from 20 to 50,  but only 5 were over 30.  Their
 ages were 31, 33,  38, 41, 50.  The  41-year  old subject was also a cigarette
 smoker and was in the fastest group in CO uptake; the other 4 were in the slowest
 group.  This is suggestive, but certainly not conclusive, evidence  that age may
 tend to lower the rate of uptake of CO.  It may be  that the older subjects  were
 less restless.

      The great variability between subjects  resulted in missing  the desired
 percentage of  COHb by an average of 16% of its value in the first experiment on
 a subject.  The misses were equally divided between high and low.   In our
 second experiment on the same man, the average  miss was 7% also equally
 divided between high and low.   The worst misses on the first experiments were
 +28% and -25%, and the worst on the second experiment were +14% and -14%.
 There  was no  demonstrable improvement in our ability to hit the target percent
 saturation in the first experiment on a subject as a result of experience  with
 other subjects.

      The elimination of CO in  air was observed for slightly more  than an hour
 in our  first subjects.   This was done by blood and alveolar samples just before
and just after  the series  of tests.  During this time the  loss of COHb was about
 17% of its value.   In the later work an effort was made to keep the  COHb  more
 constant during the tests by giving a "refresher" dose of CO half way through
 the testing period,  that is, 30 to 35 minutes after  the start of the tests.   This

-------
                                                                      I    I    I   I
                                                    O NOT CIGARETTE  SMOKER
                                                       UNDER 15 CIGARETTES  PER DAY
                                                       OVER 15  CIGARETTES  PER DAY
                   1 Hour,  WM
                       I/  I
              Hour. GF
Figure 11, Rates of uptake of CO at 710 ppm: 2 determinations (23 subjects), 1 determination (3 subjects).

-------
Figure 11 (continued}
                                                                                                          INJ

-------
Figure  11 (continued^
                                                                                                            INJ

-------
                                                                          25
 meant that the fall in COHb was between 1 and 2 percent COHb - too small a
 change to be measured accurately, and consequently the figures from the earlier
 work which are presumably about twice as accurate are  given here in Figure 12.

      The rate of elimination of CO when breathing room air may vary  in dif-
 ferent subjects  from about 11% to 19% per hour with an average value of 15%,
 the figure generally found in the  literature for men in sedentary occupations
 (elimination increases with the respiratory ventilation).  Figure 12 shows the
 rate of elimination of CO of subjects in the present study when breathing room
 air.  Breathing 99% O^ increases the rate of elimination by a factor of  3. 5 to 4,
 again with a considerable variation among subjects.  On the average, 55 or 60%
 of the CO is eliminated in the first hour, and 60% of the  remainder is eliminated
 in the next hour.  (See Figure 13. )

      In contrast to the uptake of CO, which is linear or virtually so, the
 elimination of CO is logarithmic, and also,  in air, is slower than the uptake,
 being about 15% of whatever the current value is per hour.  However, under
 the conditions of our experiments in which we wished to  avoid keeping the sub-
 jects at high levels (17% COHb) for any longer than necessary, small errors in
 measurement could make considerable errors in our estimation of the rate of
 elimination.  Since our subjects  were put on O  as soon  as the tests were finished,
 we usually had only about an hour, often less, TEo observe the rate of  fall of COHb
 from, say 18% to 15. 3%.   Thus,  an error of 0. 3% in either figure would make
 over 10% error in our estimation of the rate of elimination.  This prevented any
 accurate determination of the rates  of elimination in these cases.

      The rates of elimination in oxygen, however, could be,  and were,
 measured with much greater accuracy.   Again there were considerable differ-
 ences between the various subjects.  There were also complicating factors such
 as the degree of activity of the subjects and the fit of the masks through which
 the O2 was delivered.  The elimination of CO was  usually between 3 and 4 times
 as fast in O.  as it had been in air.  It was not as great as might be expected
 from the fact that the O-^ pressure theoretically  was  6 times as great.  In
 practice, however,  the dead space of the mask plus the possibility of leakage of
 air into it lowered the rate of elimination.  Approximately half (sometimes a
 little more) of the COHb in the blood was eliminated each hour while breathing
 oxygen at rest.

      The numerical values for the  rate of  elimination of CO for each subject while
 breathing air and also while breathing oxygen, are shown in the Appendix E.  As these
observations were not a primary  objective of this study,  but purely incidental to it,
the design of the experiments was not optimal for obtaining great accuracy in these
measurements.  However, the average values probably give a reasonably accurate
picture  of the rates of elimination of  COHb in air and in a mixture of 99% oxygen and
1% carbon dioxide.  The considerable  amount of individual variability or  scatter of  the
results  is due in part to (1) individual differences and (2)  the  measurements being
made over a  short period of time. The protocol of the experiment required that each
subject could not be released until the COHb level had reached 6% or lower. It was
for this reason that the values on rate of elimination of CO were obtained.

-------
                                                                           26
Figure 12.  Elimination of CO of subjects in present study when breathing
room air.

-------
o
o
                                                                                27
       Figure 13. Elimination of CO of subjects in present study when breathing
       99% O  - 1% CO  mixture.

            M         O

-------
                                                                           28
IV.   Laboratory Tests of the Effects  of Carbon Monoxide

      A.  Introduction

      The laboratory tests used in this study were designed to detect the effects
of relatively small amounts of carbon monoxide on human performance on tasks
related to vehicle driving.  More  specifically we wished to utilize objective
methods  for appraising the  effects of carbon monoxide on  the oxidative mech-
anisms of the  central nervous system. The alteration of certain visual functions
by hypoxia andother  physiological stresses has yielded results of considerable
theoretical interest and practical  significance.  The study of vision under  such
conditions is important because visual functions are believed to reflect changes
in the central nervous system, of which the retina of the eye is  essentially a part.
In addition,  vision is one of the primary functions or requirements of the driving
task.

      The accuracy of such studies has, in the  past, often been  limited by the
nature of the specific test or physiological function used as  an index of impair-
ment.  To obtain satisfactory results a test should possess  certain features
including:  (a) a high degree of sensitivity; (b) precision of the physical measure-
ments involved; (c) independence of the results from the degree of conscious or
unconscious effort which may be exerted;  and (d)  stability of the function during
control experiments.

      Tests of certain visual functions, particularly light sensitivity,  possess
most of these desirable qualities.  They provide a useful  took with which to
measure the effects  of hypoxia and related stimuli.  The changes manifested by
the visual mechanism when its oxidative processes are disturbed are  of consider-
able magnitude.  The physical measurements of light intensity involved in these
tests can be made very accurately.   Moreover, the control of such experiments
is simplified by the  fact that the  subject is not aware of changes in his own visual
sensitivity or changes in the physical intensity of the stimulus,  since,  at the
threshold level,  the stimulus  always has the same appearance.  The subject,
therefore,  cannot mask the impairment by exerting additional effort.   The
specific  tests used in this study,  which contained most of the above features,
were: (1) central-peripheral complex reaction task; (2) dark adaptation and
glare recovery; (3) peripheral target recognition, and (4) depth perception.

      B.  General Procedures

      In  our initial experimental design,  laboratory testing was carried out under
four different levels of exposure: 0% COHb (in practice 4% or less),  6%,  11%
and 17%. Each subject was exposed to one of these  levels on each of four dif-
ferent days.  In all cases,  however,  the general  procedures followed were
similar.  At no time during the course of the laboratory tests did any subject know
to which of the four levels of  carbon monoxide  he had been  exposed.

      All subjects were paid $20 per  session.  In the first phases of the laboratory

-------
                                                                           29


 work,  subjects were tested first under the control condition, followed by a
 randomized sequence of the three COHb levels.  Later the levels were random-
 ized.   In addition, for each subject the same fixed order of presentation of
 laboratory tests was followed for all of his four sessions.

       The overall experimental procedure on each day,  regardless of which of
 the four gas mixtures was used, was as follows:  As soon as the subject arrived
 at the laboratory, usually about 9:00 a. m. ,  blood and alveolar  samples were
 taken to determine COHb levels  at that time.  Immediately following he was
 given pretests on the glare recovery-discriminometer test, and the central and
 peripheral complex task (see below).  When this was completed, usually in  about
 45 minutes time,  the subject  was placed on his gas mixture for 80 minutes, at
 the conclusion of which both blood and alveolar samples were taken to determine
 his COHb level at the  beginning of the testing period.  After about  35  minutes,
 when half of the tests  were completed,  the subject was again exposed to his gas
 mixture for that day to restore his COHb to  the intended level.  Alveolar  samples
 were taken immediately before and after  this "refresher."  The remainder  of the
 tests were  then carried out, whereupon both blood and alveolar samples were
 again  taken.  The subject was then "washed  out" with the O^-CO, mixture for
 about  60 to 90 minutes, depending on his  COHb level,  whereupon his final blood
 and alveolar samples were taken.  He was then released for the day.

       C.  Complex Central-Peripheral Reaction Test

           1.  Test Procedures

      A method devised during recent years  to measure the effects on human
 performance of moderate stresses or small amounts of  various noxious agents
 has involved the simultaneous performance of two separate tasks.   It  has  been
 shown in various experiments that under such circumstances performance on a
 main task often may show no measurable  decrement.  However, the ability to
 carry  out a subsidiary concurrent task may be impaired, as measured by
 various objective criteria, with the effect of the agent under study  interpreted as
 reduced "reserve  capacity."

      The Complex Central-Peripheral Reaction Test used in the present
 experiment was adapted from  procedures previously developed  in the  Guggen-
 heim Center at Harvard for measuring effects of environmental variables  on
 performance, especially in studies relating  to altitude.  The test task shares
 with driving the need to make  accurate and rapid responses to visual events
 occurring simultaneously or sequentially in different parts of the field of view.

      The primary task was to respond to red or green lights displayed in the
 central field of vision by pressing foot switches, right for  green and left for red.
 These  stimuli, 440 in all, were presented in random order at regular 2-second
 intervals throughout the 15-minute test.  The stimulus lights were  extinguished
by a subject's making a response, or in the case of no response, they auto-
matically went off at 1. 7 seconds.

-------
                                                                          30
      The stimuli for the secondary task were small white lights appearing in
 random order at any of six locations in the peripheral field.  These were at
 15°,  30°,  or 45° from center, both right and left.   The subject responded to a
 peripheral light by pressing the appropriate finger button from 3 on the left
 and 3 on the right which corresponded to the location of the light.  There were
 96 peripheral stimuli in each test, 16 at each location, separated by intervals
 ranging from 4 to 11 seconds. Each stimulus had a duration of . 6 second.
 Time of presentation was randomized across the 20 one-tenth second points
 within the  two-second interval between central stimuli.

      The  display-response apparatus with a subject in position is shown in
 Figure  14.  A Varatek PDP-8/S Computer programmed for randomized stimulus
 presentation and linked with a teletype console for  instructions and automatic
 recording was located in an adjacent room.  This arrangement provided a
 continuous time-based record (to  . 01 second) on punched paper tape of all
 stimulus and response events throughout the 15-minute test.  These data were
 then transferred to magnetic tape for computer processing.

      Test performance on the complex reaction test was measured on different
 days  under three conditions:  (1) a "control" session,  after subjects went through
 the "gassing" procedure,  but were exposed only to room air;  (2) when CO was
 administered and a COHb level of 11% was attained, and (3) similarly, at  17%
 COHb.  Subjects were given a period of practice on the apparatus on a day
 previous to any testing, and on test days a 5-minute warm-up "run" on the
 apparatus immediately preceded the experimental test.

      Order of test sessions was randomized with respect to  the control and 11%
 COHb days.  Scheduling problems and some equipment malfunctions,  however,
 resulted in the majority of subjects performing the 17% COHb test  at the last of
 the three sessions.

      Valid records for analysis were obtained on twenty-three subjects, with
 twenty under all three conditions.   For two subjects data were available only for
 control and 17%  COHb tests; one subject performed only under control and 11%
 COHb.

          2.  Results

      The basic  information in the individual test records included whether or
not a  particular  central or peripheral stimulus was followed by a response,
whether a response was correct or incorrect, and when each response was made.
 In addition, the continuous real-time recording  of all events in a test  provided
 the basis for an analysis of interactions between central and peripheral stimuli
and responses in relation to the sequences of events and their time relationships.

      The results are presented in terms of the effects of CO on overall measures
of performance on both central and peripheral tasks, and in relation to interaction
 effects observed in the data.   Specifically, the control-test comparisons were

-------
                                                                       31
Figure  14.  Central and peripheral complex task.  Subject reacts with
foot switches to red or green lights displayed in the central panel.
Simultaneously he responds with the  appropriate finger-activated buttons
as the six lights on the perimeter are illuminated.

-------
                                                                          32
made in regard to the total numbers of correct, incorrect, and omitted
responses, response times,  incorrect and omitted responses in relation to
peripheral stimulus location,  frequency of "response blocking, " and the
influence of stimulus proximity on peripheral responses and  response times.

          a.  Effects of CO on Overall Performance - Central Task

                1.  Numbers of Correct,  Incorrect, and Omitted Responses

      The summary data relating to correct, incorrect, and  omitted responses
in the central task are shown in Table 1 for the control vs. 11% COHb and
control vs. 17% COHb tests separately.  The differences in Table 1 between the
mean values  under  control and CO conditions are obviously small,  subject
variability was considerable,  and the pattern of changes in the distribution of
means among the three categories of response is not consistent across the two
groupings. While mean changes between control and 11% COHb appeared to be
in the direction of "poorer" performance under CO, none were of sufficient
magnitude to reach the . 05 level of significance (paired difference "t" tests).
Even in the case of correct responses, where the numbers per subject were
relatively large,  the probability corresponding to the "t" was between . 08 and
.09.

      Comparison of the means for control and 17% COHb does not  suggest any
real change.  The pattern of very slight numerical differences observed would
seem to be a reversal of that in the other  set of data, but none of the "t" tests
of the differences were significant, or were even borderline  in that respect.

      Since the subjects were generally more familiar with the test at the time
of performance under 17% COHb, the above results raised the possibility of a
confounding of possible effects of CO and test familiarity or practice.  Thus,
the control vs.  11% COHb data were divided according to whether the CO test
was taken on a day  subsequent to the control test (experienced group) or on a  day
prior to the control test (inexperienced group).  A comparable division could  not
be made of the  17% data.

      The mean changes in numbers of correct, incorrect and omitted responses
in control vs. 11% COHb tests (i. e. ,  COHb minus control) are given in Table 2.

-------
                                                 Table 1
                         Central Responses - Control vs. 11% COHb Tests (21 Subjects)
Test
   Number of
Correct Responses
    Number of
Incorrect Responses
   Number of
Omitted Responses
Control
11% COHb
Control
17% COHb
Range
348-435
343-434
348-435
563-430
Mean
413. 3
407. 4
Central
413.8
415. 9
S. D.
17. 48
21. 29
Responses
17.25
18.56
Range
1-31
3-57
- Control vs.
1-31
0-55
Mean
12.
14.
17%
13.
12.
95
76
COHb
23
50
S.D.
7. 72
11.47
Tests (22
8.46
12. 24
Range
0-56
0-42
Subjects)
0-56
0-46
Mean
10.62
14. 19
10. 09
9.32
S.
12
13
12
10
D.
.41
. 57
. 39
.33
                                                                                                            U)
                                                                                                            UJ

-------
                                                                        34
                                Table 2

        Change in Average Numbers of Central Responses in Subjects
                     Grouped by Order of Test Days*

                           Experienced Group     Inexperienced Group
   Central Responses       	(N-10)	          (N= 11)

      Correct                     -1.18                  -11.1

      Incorrect                    -  . 27                  +3.9

      Omitted                     + 1.64                  +5.4

   #
     +indicates increase in CO test over control;  - indicates decrease in
     CO test from control.
      None of the differences between means in the "Experienced" column were
statistically significant.  In regard to those in the "Inexperienced Group", the
decrease in correct responses and the increase in incorrect responses under CO
were  significant at P<. 057, >. 02.  The mean increase in omitted responses
did not reach significance (  P approximately . 15).  Additional "t" tests were also
carried out for the row values across the columns.  The difference for "correct"
responses approached significance (P = . 08); those for "incorrect" and "omitted"
were well within chance expectations.

      The data were also analyzed for effect of day, with control and 11% COHb
data treated separately.  The average numbers of correct,  incorrect, and
omitted responses obtained are shown in Table 3.

-------
                                                                          35
                               Table 3
           Average Numbers of Central Responses by Day of Test
 A.

 Responses

 Correct

 Incorrect

 Omitted
Control Test
on First Day
   (N-ll)

   410.4

    16.5

     9.7
Control Test on
  Second Day
    (N=10)

    416.5

       9. 1

     11.6
Differences Between
       Means
  (Day 2 -Day 1)

       +6. 1

       -7.4

       + 1.9
 B.
 Correct

 Incorrect

 Omitted
11% Test on
 First Day
  (N=10)

  405.4 .

   13.2

   17.3
                                11% Test on
                                Second Day
    409.3

     16.2

     11.4
       +4. 1

       +3.0

       -5.9
      There is a suggestion of "better" performance with practice in the increase
in mean numbers of "correct" responses in day 2 tests over day 1, and also in
the corresponding decrease in the totals of incorrect and omitted responses.
However, the only change reaching statistical significance is the smaller number
of incorrect responses in the control test performance of those taking the test on
the second day, as  compared to those having control first.

      It thus appears there are some factors of test experience or order of test
day operating in the data which might obscure possible effects of CO.  However,
significant impairment from CO,  in terms of decreased numbers of correct
central responses,  and an increase in incorrect responses can be demonstrated
only in the case of minimal prior  experience in the test.   Thus, the foregoing
analysis suggests that a deleterious effect from COHb in this amount may be
most apparent during the learning period of a task, with the  implications for
driving relating to performance in unfamiliar or novel situations.

              2. • Reaction Times, Central  Task
      The average reaction times per subject to central stimuli ranged from

-------
                                                                         36
. 38 sec. to . 85 sec. in the control tests, . 39 to 1. 14 under 11% COHb, and . 38
to . 87 with 17% COHb.  The means of the individual averages are shown in the
table  below, with times for correct and incorrect responses shown separately.
Both simple means and means adjusted to take account of differences in the
numbers of response times available per individual are given, along with the
adjusted means for correct and incorrect responses combined.
                                 Table 4
               Summary of Response Times to Central Stimuli
                Simple Means of
            Average Response Times*
Test
Condition

A.
Control

11% COHb
 (N=21)

B.
Control

17% COHb
 (N=22)
 Correct
Responses


   .661

   .668
  .673
   .666
Incorrect
Responses


   .633

   .640
  .547

  .588
     Adjusted Means of Average
	Response Times	
                         Combined
                         Correct and
 Correct     Incorrect     Incorrect
Responses    Responses    Responses
  .661
   666
   673
   665
Incorrect
Responses


   .605

   .650
  .615
  .653
.660
.660
.672
.664
   #In seconds.
      In general,  individuals tended to be quite consistent in the different test
conditions, i. e. ,  relatively slow or fast in all.  When the mean times for
individuals varied between  control and CO conditions, small increases or
decreases appeared equally often and seldom exceeded a few hundredths of a
second.  It is interesting to observe the slightly shorter response times for
incorrect compared to correct responses in  the Table. However, a differential
effect related to CO in this  regard could not  be demonstrated.  The summary
data thus do not indicate any demonstrable effect of CO at these levels on
reaction  times in the central task.

-------
                                                                          37
          b.  Effects of CO on Overall Performance - Secondary Task

               1. Numbers of Correct,  Incorrect, and Omitted Responses

      The average total numbers of correct,  incorrect, and omitted responses
 are given in Table 5.  It should be noted that omitted responses to stimuli at
 one of the peripheral locations (15° to the right) have been deleted from the
 totals, since an equipment malfunction had resulted in a spurious count of
 omitted responses to this light.
Condition
A.
Control
 11% COHb
 (N=21)

B.
Control

 17% COHb
 (N=22)
                                 Table 5

                Summary - Numbers of Peripheral Responses

             Correct Responses    Incorrect Responses     Omitted Responses
 Mean    Standard     Mean   Standard
Number   Deviation    Number  Deviation
  73.3

  72.9



  72.4

  72.9
12.89

16.62



13.53

13.74
 8.9

10.2



 6.7

 6.7
7. 02

9.80



6.72

6.66
                                  Mean    Standard
                                Number   Deviation
 9.2

10.7



11. 1

13.1
 8. 00

12.0.4



 9.58

10.29
      The differences between the mean values above are obviously small.
Numerically, a small increase appears in both average numbers of incorrect
and omitted peripheral responses in the 11% COHb tests, and in omitted
responses in the 17% tests.  However, individual variability was great and none
of the differences proved to be  significant with these numbers of subjects.
There is also a suggestion of greater  variability under CO in the data for control
vs. 11% COHb.  Here the  F ratio of the variables was significant at the . 05 level
in the case of the omitted  responses.

     An analysis in reference  to experience with the test was  also performed on
these data.  The changes between mean numbers of the peripheral responses
between control and 11% COHb  tests for the "experienced" and "inexperienced"
subjects (as previously defined) are given below.

-------
                                                                          38
                                 Table 6
             Changes in Average Numbers of Peripheral Responses
                  in Subjects Grouped by Order of Test Days


     Peripheral              Experienced Group    Inexperienced Group
     Responses                  (N=10)	    	(N=ll)	

     Correct                     -1.09                  + .50

     Incorrect                     -2.82                 +4.90

     Omitted                      +3. 18                 -2.40



      None of the mean changes between control and CO values are significant
 in either the "experienced" or "inexperienced" group.  When the comparison
 is made across rows, the  reversal in the case of incorrect responses reaches
 borderline significance (P = . 06),  again suggesting relatively greater inaccuracy
 under  CO if the task is being learned.

               2-  Numbers of Incorrect and Omitted Responses  in Relation
                  to Location of Stimuli in the Peripheral Field  of View

      Previous studies have suggested that stimuli in the periphery of vision,
 as compared to central location, are  more likely to be "missed"  after exposure
 to CO.  In the present data, this question involved a comparison  between control
 tests and those under CO in regard to the frequency of incorrect  and omitted
 responses to peripheral stimuli at each of the six locations,  i. e.  , 15°, 30°, and
 45   from center, to both right and left.  The average numbers of incorrect
 peripheral responses, by stimulus location, are shown  in Table 7,  and the cor-
 responding data for omitted responses in Table 8.

     A consistent pattern  suggesting an effect from CO is not apparent in regard
to the incorrect responses. Numerically, there were larger numbers of incorrect
responses under both CO levels at the 45  left position,  but the difference was not
 statistically significant in either case. The suggestion of slight increases under
 11%  COHb at the two 15° positions is  reversed in the 17% COHb data.  Overall,
five  of the means were numerically larger under CO, as against seven which
decreased under CO.   The  difference  between means was significant (P <. 05)
only in one case, i.e., there were significantly fewer incorrect responses at 30°
left under 17% COHb as compared to control.

-------
                                                                   39
                           Table 7

     Average Number of Incorrect Peripheral Responses by
Control
11% COHb
(N=21)
Control
17% COHb
(N=22)
Stimulus
Left
45U 30°
1.24 2.33
1.48 2.05

1.09 2.50
1.27 1. 18

Location
Center
15U 15°
1.62 .67
1.95 .71

1.45 0.59
.77 0.18

Right
30° 45°
1.76 1.61
1.00 1.76

1.64 1.40
1.91 1.36

Table 8
Averag

Control
11% COHb
(N=21)
Control
17% COHb
(N=22)
e Number of Omitted Peripheral
Stimulus
Left
45° 30°
2.28 1.28
2.05 1.62

2.68 1.59
3.32 1.59

Location
Center
15° *J_5°_
1.71
2.09

2. 04
2.23

Responses by
Right
30° 45°
1.52 2.24
2.43 2.24

1.95 2.77
3.14 2.27

*As previously noted, "omitted response" data at this location were
 found to be invalid.

-------
                                                                         40
      In regard to Table 8, small numerical increases in mean numbers of
omitted responses under CO appear in three of the five comparisons at the  11%
level and four of the five at 17%.  However, in only two of the ten did any of the
differences reach statistical significance.  At 30° right the  increases in mean
number of omitted responses over control at both levels of CO were significant
at P<. 05.

      Obviously,  the foregoing does not provide adequate answers concerning
the  influence of stimulus peripherally on response,  or a differential effect of
CO  in this regard.  One problem was that with only 16 stimuli at each location
per test the numbers of incorrect and omitted responses per location per
subject were very small.   Also, variability in the data was  found to be affected
by two other factors.  In general,  relative propinquity in time of peripheral
and central stimuli had an influence on the frequency of omitted responses.  In
addition, frequencies of stimuli at the different presentation times did not turn
out  to be sufficiently uniform for making precise comparisons.

              3.  Effect of CO on Frequency of Instances of Response
                  Blocking

      The phenomenon  of "response blocking" was first recognized and studied
systematically by Bills (1937) who noted that subjects performing long  series
of repetitive operations tended to show occasional gaps in performance during
which  the desired responses were not made even though the stimuli were
present. Generally, such gaps have  been interpreted as  reflecting attentional
lapses in high speed, continuous,  decision-making tasks.  In more recent
years  "blocking" has been considered in the framework of the human as a com-
munications system, along with other human time lags,  but still with an emphasis
on an attentional mechanism.   (Broadbent,  1958;  Teichner, 1968) In general,
"blocks" have been found to occur periodically throughout a response series,
and may include from 2-6 consecutive responses.  They tend to increase in
frequency and duration with increasing complexity of the task and/or rate of
information flow. Prolonged  time at the  task, sleep loss, and environmental
stresses, such as anoxia, have also been shown to increase blocking.  Driving
is a type of task  similar to many of those in which "blocking"  occurs, and
certainly is one in which failure to respond to external cues could under some
situations prove very hazardous.

      In the present experiment, an "instance" of response blocking was defined
as failure to respond to a peripheral stimulus in conjunction with failure also
to respond to the central stimulus next preceding and/or next  following the peri-
stimulus.  A few individuals showed  one  or more instances of such response
blocking under both control and experimental conditions: about half did not show
any instances under any of the conditions.  Table 9 below presents an analysis
of the numbers of subjects in whose records instances of response blocking were
found  under the different test conditions,  together with the  corresponding X
and P values.

-------
                                                                          41
                                Table 9

      Comparison of Numbers of Subjects with One or More Instances
               of Response Blocking in Control and CO Tests

                                  Number of Subjects
      Test                        With One or More
      Condition                    Response Blocks        X2, P

      Control                             5            3.634, P =  . 07

      ll%COHb                          11
      (N=21)
      Control                             4            3.804, P = . 052

      17% COHb                          11
       (N=22)
      Slightly more than twice as many subjects showed one or more instances
of blocking in the tests under CO as in the control tests.  With the small
numbers of subjects involved, the significance of the X^'s is only borderline.
The P value in the second comparison is just about at the . 05 level, while that
in the first is at . 07.  Thus, while there is a strong suggestion that some
persons may be more susceptible than others to response blocking after
exposure to CO at these levels,  the present finding requires further verification.

      Analysis of the data in regard to numbers of response blocks was made
on the basis of the total numbers of "blocks" for all subjects under a given test
condition in relation to the total numbers  of omitted responses.  The results
were  in the direction of an increase in the numbers of instances of response
blocking in the total group under both CO  levels as  compared to control.  How-
ever, in neither case did the X^ value quite reach a statistically significant
magnitude, with P approximately . 09 and . 07 for the 11% and  17% comparisons
respectively.

      The "response blocks" included some which involved failure to respond to
a peripheral  stimulus and to both the preceding and following central stimuli,
i. e.,  "blocks" of considerable duration, lasting somewhat longer than  2 seconds.
The numbers of subjects showing instances of this more protracted blocking were
too few for adequate statistical testing; however,  2 of 21 subjects showed such
blocks under control, vs. 5 at 11% COHb,  and  3 of  22 under control vs. 10 at
17% COHb.

      The test also at times involved relatively long sequences of central stimuli

-------
                                                                          42


 without an intervening peripheral stimulus.  Hence, the data were also scanned
 for instances of failure to respond to two or more consecutive central stimuli
 in such series.  The number of subjects showing this type of "block" was very
 small, and in view of the large numbers of central stimuli per test, these blocks
 were quite rare.  One or more were found in the records of 4 of 21 subjects
 under control vs. 8 under 11% COHb, and for 3 of 22 under control vs. 7 at 17%
 COHb.  A phenomenon which did not appear in  any of the control records, but
 was found for two subjects in both 11% and 17% tests, for 1 at 11%, and  1 at 17%,
 involved complete failure  to respond for periods  exceeding  4 seconds.  One
 subject failed to respond to 28 consecutive stimuli at one point in his 11% test;
 another did not respond to a series of 18 during the 17% COHb test; another
 missed 9 in a row.  These lapses extended from  16 to over 40 seconds,  and it is
 possible that  these subjects may actually have  fallen asleep for  brief periods.
 However,  the only evidence that this may have  been related to CO is the lack of
 such episodes in the control test data, and a definite relationship cannot be
 established on the basis of the present data.

      Previous  work has indicated that "blocks" are characteristically preceded
 by increased  reaction times just prior to  their  appearance.  It would thus be
 desirable in relation to the present data to examine the  reaction times just
 preceding  the occurrence of isolated omitted responses for  additional evidence
 of response blocking,  and  for making comparison in regard to possible effects of
 CO.

           c-  Effects of CO in Regard to Interactions Between Central and
              Peripheral Task

               1.  Frequency of Omitted and Peripheral Responses as a
                  Function of Stimulus Presentation Times

      Originally, a classification system of response  categories based on stimulus
 response sequences had been  devised for the analysis of interaction effects.
 Inspection  of the data in this form suggested that  responses  to peripheral stimuli
 tended to be omitted more  frequently when presented at times  close to the
 onset time of  a central stimulus.  However, when the attempt was made  to
 analyze the data tabulated into categories, variability between subjects and  bet-
 ween tests  was great,  and consistent trends related to CO could not be demon-
 strated. A considerable source of the variation was the  differences in numbers
 of stimuli appearing at the 20 separate presentation times.  In a given individual
 test, there was  a range in  stimulus  frequencies of from 0 or 1 to 10 or 12 at the
 different times, and the patterns of  stimulus frequencies varied between sub-
jects, and for the same individual between tests.

      For an alternative approach,  distributions were prepared of the total
numbers for all subjects of peripheral stimuli appearing at each of the presenta-
tion times separately for the three test conditions.  Corresponding distributions
of the total numbers of correct, incorrect, and  omitted responses were also
prepared and plotted as proportions  of the total  numbers  of stimuli presented at

-------
                                OMITTED  PERIPHERAL  RESPONSES
§
LU
    .50
    .40
LJ
    .30
"i
_iiiJ


feo:
1-0-

O

2
O
    .20
     .10
                     CONTROL CONDITION

                     17% CoHb

                     (22 Subjects)
      1.1   1.2   1.3  1.4   1.5  1.6  1.7   1.8   1.9  0.0   ,1    .2   .3   ,4   .5    .6    .7   .8   ,9
                                                2.0

       PRESENTATION TIMES OF PERIPHERAL STIMULI •' (CENTRAL  STIMULI  AT 0 AND  2 SECS.)
                                                                                                1.0
            Figure 15 •  Frequency of omitted peripheral responses as a function of  stimulus
                       presentation time,  relative to onset  time of central stimuli.

-------
                                                                            44
 each of the 20 time points.  Inspection of these graphs suggested that incorrect
 responses tended to be fairly evenly or randomly distributed over the 20 presenta-
 tion points under both control and CO conditions.  Omitted responses,  however,
 were relatively concentrated in a range of presentation times from shortly
 previous  to shortly after a central stimulus.  This pattern was  characteristic
 for all  three test conditions, with 30-50 percent of the stimuli at particular
 points close to central stimuli receiving no response.  Figure  15 illustrates
 this relative concentration and  shows the comparative distributions from control
 and 17% COHb tests.

      Statistical comparison of the distributions was  made through the calculation
 of Chi Squares for the interval  .4 sec. prior to .5 sec. following central stimulus
 time, and in regard to the frequency data at the separate points in this interval.
 In the control-11% comparison  none of the X2 values  reached the . 05 level of
 significance.   The distribution  of the  17% data, however, was significantly
 different  from control (P < . 046,  >. 025) indicating a  greater frequency of omitted
 responses in this area after CO than to be expected by chance.  Also, the numbers
 of omitted responses at several of the individual  time points were also significantly
 greater,  notably at . 1 sec. preceding, and . 1 sec. and . 3 sec.  following the central
 stimulus  point.

      The suggestion here is that there may be a "normal" tendency for relatively
 increased likelihood of omitting a response to a secondary stimulus  when it
 competes closely in time with a primary one,  and this tendency may  become
 enhanced after exposure to CO which eventuates in 17% COHb.   Obviously the
 above suggestion requires confirmation, especially through procedures which
 would involve greater numbers  of stimuli in the critical time areas and permit-
 ting more precise manipulation of effects  of individual variability in the statistical
 treatment.  The tentative finding does suggest an important direction for further
 research, not only in regard to effects of CO, but also for stimulus  response
 relationships in the case of temporally and spatially competing stimuli.

               2.  Effect of CO on Peripheral Response Times

      In the original computerized classification into  ten stimulus-response
 sequence  categories, the mean  reaction times varied considerably within cat-
 egories and between categories  for individuals and test conditions.  Observed
 changes between control and CO data were inconsistent in direction and
magnitude.  It  became apparent that category frequencies and consequently
numbers and variability of response times per individual per category were
affected by variations in the numbers of peripheral stimuli at the different pre-
sentation  times.  In this regard the stimulus randomization program,  in this
number of subjects, did not result in the desired uniformity.  Also,  whether a
given subject tended to be generally "fast" or "slow"  influenced the relative
allocation of responses between certain categories.

      A further programming for  computer analysis to provide comparisons of
response  times as based specifically on the stimuli appearing at each of the 20
successive presentation times would have been desirable but was beyond the

-------
                                                                           45
financial resources of the project.  It was, however,  possible to make a rough
approximation of the influence of central-peripheral interactions on peripheral
response times and of possible effects of CO in this regard.  Frequency dis-
tributions  were prepared from the categorized data of all the response times
for all subjects in respect to the stimuli at each presentation time and under the
three test  conditions.  These were then plotted for each presentation time as
proportions of the distribution of reaction times at that point falling into suc-
cessive reaction time intervals of . 10 seconds.  Visual comparisons were made
for changes across the 2-second span of presentation  times and for differences
between control and CO conditions.  In instances where divergencies appeared,
the statistical test for significance of differences in proportions was applied.
In the inspection process, emphasis was placed on overall shape of the distri-
butions, variations in central tendency, and relative frequency of prolonged
response times.

      Qualitatively,  the plots  ranged from a fairly uniform pattern quite
characteristic  from about . 6 to 1. 5, in terms of the presentation time points.
Figure  16  illustrates the  typical pattern for this range and also  shows the
comparative control and  11% COHb  distributions at presentation time 1. 1.  In
the range where peripheral and central stimuli were close together, or coincided,
in time, the patterns were more irregular, more platykurtic, and with less
clearly defined central tendencies.  Figure 17 shows the plots at the point of
simultaneous onset of stimuli for Zl subjects in the control and  11% COHb tests.

      As a rough index of central tendency, the plots were inspected for location,
and changes in location,  of the modal intervals,  i. e. ,  the reaction-time interval
containing  the largest proportion of the response times.  In the  control data for
21 subjects the modal interval was the one including reaction times of . 60-. 69
sec., at 16 of the 20 presentation points, from .3-1.8 successively.  At 1. 9
time, the largest proportion was  in . 70-. 79 sec. At 0. 0 (stimulus simultaneity)
the plots suggested possible bimodality in peaks at . 60-. 69 and  . 90-. 99 sec.
(See Figure 17)  These observations suggest that when central and peripheral
stimuli appear at the same time,  or in close  temporal proximity,  peripheral
response times may tend to be slightly longer,  on the average, or more variable,
compared to responses when the stimuli are more remote in time.

      In general, only minor variations between the distributions under control
and 11% COHb, and control and 17% COHb were  observed,  and the patterns in the
plots were  very similar.   In regard to the 11% level, the location of modal
intervals coincided closely with those  in the control plots,  and there was a clear
difference  in this regard  at only one point.  At time . 1, the modal interval was
. 90-. 99 sec. with CO, compared to . 70-. 79 sec. in control.  The variability in
both control and CO plots at 0. 0 time prevents a clear comparison of the modal
intervals.

      The distributions of response  times under control and 17% COHb followed
the same patterns described above quite closely, except that,  in the middle range
of stimulus times, the 17% COHb modal interval in several instances shifted to

-------
   .35
u_
o
   .30
CO
UJ
CO

1  .25
CO
UJ
o:
   .20
g  .15
h-
tr
   .10
   .05
          I     I    I     I     I     I    I     i     i     I    I     I     I
        DISTRIBUTION OF REACTION TIMES OF PERIPHERAL RESPONSES
        AT STIMULUS ONSET  TIME 1.1 SEC-
                                            CONTROL (88 RESPONSES)

                                      x—x  11% COHb (95 RESPONSES)
   Pi—--,
         .40-  .50-   .60-  .70-  .80-  -90-  iJOO-  1.10-  1.20-  1.30-   1.40-  1.50-
         .49   .59   .69   .79   .89  .99   1.09  1.19   1.29  1.39   1.49  1.59
                   REACTION TIME INTERVALS (SECONDS)
                                                                  1.60-  2.00
                                                                  THRU  AND
                                                                  1.99  OVER

-------
  .30-
to .25
UJ
co

O
Q_
CO

£.20
£
•—1	1	1	1	1	1	1	1	1	1     I    I    I    '

 DISTRIBUTION OF REACTION TIMES OF PERIPHERAL RESPONSES _

 AT STIMULUS ONSET TIME 0,0 SEC-
                            o—o  CONTROL (58 RESPONSES)  -

                            x—x  11 % COHb (73RESPONSES)
u_ .15
O
   .05
                  REACTION TIME INTERVALS (SECONDS)


                           Figure 17.
                                                        1.50-  1.60-  2.00
                                                        1.59  THRU  AND
                                                             1.99  OVER

-------
                                                                          48
the next interval lower than in the corresponding control plot,  i. e. , from . 60-. 69
sec. to . 50-. 59 sec.  This may reflect a previously mentioned influence of
familiarity with the test, or practice,  resulting in shorter response times in the
"neutral"  areas of the test and possibly obscuring any effects of CO m more
critical ones.

      The distributions  were also examined for the possibility, under CO, of an
increase in the frequency of excessively long response times.  This possibility
was suggested by the preliminary findings in a concurrent study at Harvard on
the effects on performance of marihuana and alcohol in which the  Complex
Reaction Test was also  used.  Accordingly,  tables were set up of the  proportions
of the distributions of response times  at each time point which, respectively,
reached or exceeded reaction times of 1. 00,  1.20, 1.30, 1.40, 1.50, 1.60, and
2  00 seconds, for both  the control and CO data.  Under both test conditions it
was apparent that instances where 10% or  more of the reaction times  were longer
than 1 second were concentrated in stimulus times near the time of simultaneous
onset.  The data  for control and 11% COHb at time points 1. 6  to . 4 are  shown
in Table 10.

      A consistent pattern of increased frequency of the longer reaction times
under  CO is  not apparent.   Only the increases at 0. 0 time were statistically
significant (at . 05 or less) in the cases of responses 1. 0 sec., 1. 3 sec. , and
1.4 sec. or longer.   None of the differences in the direction of decreased fre-
quency of the longer  reaction times under CO reached significance.  However, a
satisfactory assessment of the  significance  of differences in this  table is difficult
because of the small absolute numbers of prolonged  responses and variations in
the totals of the distribution being compared.

       The data relating to control and 17% COHb were analyzed in a similar
manner.  In this case,  significant increases under CO were observed at stimulus
time 1. 8  in regard to the reaction times exceeding 1. 0, 1. 1,  1. 3, and  1. 4 sec.
and with increase of borderline significance at 1. 2 and 1. 5.  Borderline signifi-
cance was also indicated for an increase in  times longer than 1. 0 sec.  at
 stimulus  time . 3.  Again, none of the differences  in the direction of fewer
prolonged reaction times were significant.  Also,  the rare instances of exceed-
ingly long reaction times,  i. e. , 2. 0 or greater, seem to be scattered through
the stimulus time range and not to differ in  frequency between CO and control
 conditions.
       A further treatment to provide larger cell frequencies by combining the
 data for two or more adjacent presentation times would have been desirable in
 order to permit more precise comparisons and estimates of significance by
 Chi-Square techniques.  This,  however, could not be carried out within the
 resources  of the project, and verification  of the slight  suggestion that prolonged
 peripheral response times may be more frequent  under certain stimulus condi-
 tions at these levels of COHb awaits further investigation.

-------
                                         Table 10




Prolonged Reaction Times to Peripheral Stimuli as Percentage of Responses Which Exceed Selected
Times
Stimulus
Time
1.6
1. 7
1.8
1.9
0.0
0. 1
0.2
0.3
0.4
at Stimulus
Test
Condition
Control
11% COHb
Control
11% COHb
Control
11% COHb
Control
11% COHb
Control
11% COHb
Control
11% COHb
Control
11% COHb
Control
11% COHb
Control
11% COHb
Presentation

1.00
Sec.
13.2
8.0
17.9
19.2
20.2
28. 6
33.2
32.6
25.6
42.0
27.1
25.9
13.2
24.0
25.6
17. 1
12. 0
11. 1
Times

1. 10
Sec.
8.3
8.0
13.4
13.5
18.8
24.9
20.0
27.3
18.8
28.4
19.6
13.6
6.5
12.8
15.8
11.8
9.6
9.9
. 4 Sec. Before
Reaction
1.20
Sec.
6.4
6.0
13.4
10. 1
14.6
19.9
16.0
23. 1
18.8
23.0
10.5
13.6
3.2
8.0
10.9
9.2
7.2
7.4
to .4
Time
1.30
Sec.
4.5
4.0
10. 0
9.0
8.4
12.4
12.0
16.8
8.5
21.7
6.0
7.5
3.2
6.4
8.5
7.9
6.0
6.2
Sec. Following
Central
Stimulus
Time
Reached or Exceeded
1.40
Sec.
3.6
2.0
8.9
5.6
7.0
8.7
8.0
10.5
5. 1
17.6
4.5
6.0
1.6
4.8
4.8
3.9
4.8
3.7
1.50
Sec.
3.6
1.0
6.7
2.2
5.6
6.2
8. 0
8.4
5. 1
14.9
3.0
4.5
1.6
3.2
2.4
2.6
2.4
1.2
1.60
Sec.
0.9
1.0
4.5
2.2
4.2
3.7
6. 0
6.3
5. 1
9.5
1.5
4.5
0
3.2
2.4
1.3
1. 2
1.2
2.00
Sec.
0.9
0
1. 1
1. 1
1.4
0
2. 0
0
1.7
1.3
0
1.5
0
1.6
0
0
0
0

-------
                                                                           50
           D.  Dark Adaptation and Glare Recovery

                 1.  Test Procedures

      Among the critical visual tasks which drivers encounter is the ability to
 see after exposure to high levels  of illumination at night, a condition that arises
 primarily from the headlights of oncoming vehicles.  The effects of such glare
 tend to disrupt the visual photochemical system.   Although  glare has been
 studied under a variety of conditions by investigators at this School,  little has
 been done to analyze the amount of time necessary to achieve the pre-existing
 levels of dark  adaptation.  In order to study this effect,  two test procedures
 were available.

      In the first of these, a visual discriminometer was used to investigate
 recovery from light shock (see Figure 18).  The subject was seated in a light-
 tight compartment and used his left eye to locate a red dot in the center of  the
 visual field.  He was then presented  with a pre-exposure light source of 2000 ml
 for 3 minutes,  after which the light is turned off,  and the subject attempted to
 relocate the  red dot.  While fixating  on this dot, a 1° test field, located 7° to
 the left of the fovea,  was illuminated with light flashes of 1/25 of a second.
 Threshold values were obtained by increasing the  intensity  of these light flashes
 until the test field was just seen by the subjects.   The threshold values were
 recorded in log luminance-millilamberts, and were taken first at less than one
 minute after the exposure to the light source,  then at approximately one minute
 intervals until  final thresholds were  obtained at around 12 and 13 minutes.  The
 resulting values were recorded in tabular form and plotted  as curves  of the
 dark adaptation of each subject under each test condition.

      After these final threshold dark adaptation values were obtained with the
 visual discriminometer,  each subject then had introduced into his visual field
 a bright flash of light of one second duration.  His 1° test field was then
 illuminated with 1/25 of a second flashes of light in rapid succession.  The
 elapsed time for the subject to first see this test field illuminated by the  same
 light value as that of his final dark adaptation threshold was then recorded as
 his glare recovery time.

      In  the second test for glare  recovery,  a Biometrics glare testing device
was utilized.  Each subject was first tested for his glare threshold by his ability
to identify Landolt rings (circles with gaps at the top or bottom, right or left)
under progressively lower levels  of illumination.  Once this threshold was
obtained, a glare source was  turned on in the subject's field of view, immediately
to the right of the testing  device.  After looking at this glare for 15 seconds,  the
 subject then again attempted to identify the gaps in the Landolt rings until three
 successive correct determinations were made.  The time required to return to
his original threshold was then recorded.

-------
                                                                        51
Figure 18. Visual discriminometer and subject.   Testing is carried out
in darkness while the  subject is in the experimental enclosure.

-------
                                                                          52
          2.  Results

      Data for dark adaptation threshold values are available for 25 subjects
over the time range from approximately 30 seconds to 13 minutes at one-
minute intervals after exposure to the bright light source. For the first 9 of
these subjects, data are available from each of four test sessions, those at
17%,  11% and 6% COHb, and control-or less than 4% COHb. After these 9
subjects were  run, two changes were made in the experimental design. First,
the 6% COHb level was eliminated since preliminary results clearly indicated
an inability to  differentiate between this level and the  control condition.
Secondly a pretest was added to each of the remaining three sessions,  or 11%
and 17% COHb, and control.   The pretest was always  given prior to the gassing
procedures.

      There are thus two different sets of results for  comparison.  First,  the
inter day comparisons of the  results from each of the  three sessions, or: 17%
COHb vs. control, 11% COHb vs.  control; and 17% COHb vs. 11% COHb.
These data are available for all 25 subjects.  The second set of results contains
the intraday comparisons of  the pretest-test results.  Here the pretest (or
before gassing) results are compared with those of the same test administered
after  the gassing procedures.  The comparisons here are: 17% COHb, pretest
vs. test;  11%  COHb, pretest vs.  test; and control, pretest vs. test.  These
data are available for 16 of the subjects.

      The results for the inter day comparisons of the effects of CO on dark
adaptation as measured by the discriminometer are as follows:  For the final
dark adaptation threshold values obtained at about 13-14 minutes after bright
light exposure, there were no  statistically significant differences between the
results of the  three  CO exposure groups when evaluated by means of the paired
"t11 test.  This  holds true for 17%  COHb  vs. control;   11% COHb vs. control;
and 17% COHb  vs. 11% COHb.  Overall mean differences were slight, and
changes in the  direction of the difference between subjects were common.  (The
highest level of significance found was P <. 5, as opposed to the commonly
accepted minimum level for statistical  significance of P< . 05).

      The intraday pretest-test comparisons were then made.   Here as noted,
we are comparing the results of the dark adaptation threshold values of subjects
as measured before and after gassing on the same day.  For the final dark
adaptation threshold values,  a statistically significant difference (paired "t"
test, P< . 02) was observed between 17% COHb pretest-test results.  The 11%
COHb pretest-test results  showed no statistically significant results, though
the control pretest-test comparisons did show such a  difference.

      Since the pretest-test comparisons clearly demonstrated a greater likeli-
hood of establishing significant differences between no-CO and CO conditions,
it was  decided  to further evaluate  the dark adaptation  data at <1,  4,  and 10-
minute intervals after light exposure. The less-than-one minute interval was
selected since  this was the first determination made after light exposure, while

-------
                                                                           53
the four-minute interval represents the time at about which the curve of the
cone function of the eye for dark adaptation is beginning to flatten out.  The
10-minute interval was selected as roughly around the midpoint of the rod
function for dark adaptation.  The results were as follows:

      At the   1-minute interval, neither 11% COHb pretest-test comparisons,
nor the 17% pretest-test comparisons showed statistically significant differences
(paired "t" test) for dark adaptation.  However,  the control pretest-test dif-
ference did just attain statistical significance (P £. 05).  It is interesting to note
here that the direction of the difference shows a higher light value for the pretest
condition,  a  reversal of the pattern that holds fairly consistently elsewhere.

      At the  4-minute intervals there were no statistically significant differences
between any  of the  pretest-test conditions, though the 17% COHb level came
closest at P< . 1.   At the 10-minute time interval neither control nor 11% COHb
pretest-test  comparisons showed statistically significant differences, but the
17% COHb pretest-test comparison was significant (P < . 05).

      Figures 19,  20 and 21 demonstrate graphically the dark adaptation curves
and the values  in log luminance-millilamberts at each time interval from
approximately  30 seconds to 14 minutes after lighr exposure.  These curves are
based on the  averaged values  for all  subjects for which pretest-test data are
available.  The general tendency towards  higher values, i. e., more light needed,
can be noted for both 17% COHb and  11% COHb pretest-test conditions.  This
distinction is less marked for the control  condition.

      To summarize, the results of the dark adaptation values obtained on the
same subjects  under different CO conditions  on different days do not seem to
yield significant differences.  This could be due to the fact that interday variability,
for whatever reason, is so  great that it overrides any effect of the CO admin-
istered.   This, of course, would only be true if the effect of such variables were
reasonably randomly distributed in this small sample - which may or may not
be the case.  If there is a CO effect operating here,  it may be cancelled out by
these other variables.

      Alternatively, it is possible that the test is not a sufficiently sensitive one
to be able to distinguish between the  effects of such relatively low levels of CO.
This would not appear to be the case for two reasons: (a) the test has, in the
past, been successfully used to differentiate between the effects of lower levels
of CO and no CO,  and (b) as the result described above demonstrate, the test
does indicate some sensitivity on an intraday pretest-test basis.

      Here,  interday variability within and between subjects is  removed but,
although statistically significant differences do appear with 17% COHb pretest-
test conditions, they are also found under the control pretest-test conditions.
It seems, therefore,  that there may be some factor or factors in the gassing
procedure itself that influences dark adaptation in the subjects.  This could be
fatigue, motivation, altitude,  boredom, discomfort,  time of day, blood-sugar

-------
                    Dark Adaptation
    54
Log Luminance— Mill ilamberts
M OJI roi — i c












\
u
A
_M






























\







































\
^v
\







































\
-X






































s^
>»j







































X
\







































s

V-
^






































s

\






































\

*\






































s

%






































s

>






































v







































^^

'•s






































^
*^






































•»







































































































































































































































— p
	 M


















































































































































































































re-exposure
r% COHb








































































































*

































                   10                20
                Time  in Dark-Minutes
30
Figure 19.  Dark adaptation curves for 15 subjects before
exposure  to CO and after exposure with 17# COHb.  The first
part of the curve depicts the cone function,  the second
part the  rod function.

-------
                      Dark Adaptation
55
Log Luminance — Millilamberts
w oil roi —i c





\

\






































\
\
i







































\
\
\





































L
\
\



























1











s
V







































\
^
\






































k
\
\







































s
\







































\
\







































s
t
V


































•




s
V
\






































\
\







































*s
^.






































Sy








































k
»



































































































































































































































































































































































































	 Pre-exposure
	 ll%COHb






























































































































































































 0                 10                 20                30

                 Time  in  Dark-Minutes


Figure 20. Dark adaptation curves for 11* subjects before
exposure  to  CO and after exposure with lljg COHb.  The first
part of the  curve depicts the cone function,  the second
part the  rod function.

-------
                     Dark Adaptation
    56
Log Luminance— Mi Ililamberts
W wl roi —i c



1
\
•

\






































1
\
0
\







































^
»
^







































V
^
-^







































s







































t
Vv
>







































\
\







































s.
V







































V
^ *
s







































s
\







































\
^1







































k
<
^







































s
N







































V
<






































«*
"s
































































































































































— — — '













































Pr
C(




































'6-
DH




































•e
tr




































XF
ol









































































































losure












































































































































































                   10                20
                 Time  in  Dark-Minutes
30
Figure 21.  Dark adaptation curves for 15  subjects before
and after exposure to a CO placebo (Control).  The first
part of the curve depicts the cone function, the second
part the rod function.

-------
                                                                         57
levels, etc. - any of a variety of possible variables.  (It should be remem-
bered that, of necessity,  the test is always given about one and one-half hours
after the pretest. )  In addition, while the subjects were not highly trained, or
specially motivated, they were fully familiar with, and practiced in,  the task
expected of them,  and they appeared reasonably we 11-motivated.  It should
also be observed that since pretest  results generally and fairly consistently
required less  light than the test results,  the differences  could  hardly be
attributed to training or practice.

      In other words under these testing  conditions, and with these subjects,
there is little  indication that COHb levels of up to 17%  COHb have  any significant
effect on dark adaptation.

      For glare recovery, i. e. , the time,  in seconds, required to return to the
previous final dark adaptation level after exposure to a one-second bright flash
of light, the results are as follows: When interday comparisons were made,
i.e.,  17% COHb day vs.  control day,  11% vs.  control, or 17%  vs.  11%,  none
of the results  approach statistical significance (P <. 6 by paired "t" test,  the
highest association).

      When pretest-test values were evaluated,  the results  approached (but did
not achieve) statistical significance more closely as follows:  17% COHb,  P<. 3;
11% COHb, P< . 2; control, P< . 1. Clearly there is no suggestion at all here
of a CO-related effect on glare recovery.

      E.  Peripheral Vision

           1.   Test Procedures

      Since the driving task involves the  ability to assimilate a variety of visual
stimuli emanating from different parts of the environment,  it is clearly important
to investigate the ability to discriminate  or recognize  stimuli in the central and
peripheral visual fields.   In this test the subject was seated with his eyes
positioned 20  inches from a translucent screen in a darkened experimental
chamber.  Using his left eye he fixated on a lighted dot in the center of the screen,
his other eye  covered by a patch.  With a Kodak Ektagraphic Carousal Projector,
60 slides  were presented to the subject.   On each slide there were four  dots:
half of the slides have the dots arranged in a square, half in a diamond.   In 20
of the slides the dots are presented so that they fall 10° from the  subject's point
of central fixation.  Twenty slides have the dots  at 20°,  and a  third group of 20
at 30°.  The slides are mixed and presented in random order,   each for  0. 0125
second.   The  subject's task was to  report the number  of dots seen of the four
dots that were always presented. Figure 22 shows the test apparatus.

           2.  Results

      Twenty-two subjects completed this test under  conditions of 17% COHb,  11%
COHb,  and control, each test taken on a different day. No  statistically significant

-------
                                                                          58
Figure 22.  Subject in position for peripheral vision test.
are projected on screen to left in darkened booth.
Dot patterns

-------
                                                                          59
differences were found between any of the CO levels for the slides with dots
presented at 10°.  Almost all subjects saw most of the dots most of the time,
regardless of CO level.  Similarly, no statistically significant differences were
found between any of the CO levels and the 30° slides.  Here, most of the sub-
jects missed most of the dots most of the time - regardless of COHb level.

      For the target patterns presented at 20 ,  however, the results were
different.  There was an increase in  mean number of dots missed for the 22
subjects from control to 11% to 17% COHb.  The 11% vs. control comparisons
were not statistically significant,  but the 17% vs. control results were.  When
total numbers of dots missed were compared (with the  paired "t" test),  the
significance level was P< . 05.  If the total numbers of slides presented in
which any dots were missed were compared, the level  increased to P < . 01.

      There thus does appear to be some sort of a CO-related decrement here
in near peripheral vision,  though the extent  of the data does not  permit any
firm conclusions in this regard. The 10° test was too  easy, and the 30° test
too difficult.   Future tests in this area should vary the intensity of illumination
of the presented targets,  as well as their size,  exposure time and position,
until a more  critical and sensitive  set of conditions are found.  The more
important area is likely to be that farther out to the periphery of the field of
vision.

      F. Depth Perception

          1.   Test Procedures

      Depth perception, which could be of considerable importance to the driving
task was evaluated by means of a standard Verhoeff apparatus.   In this equipment
a 1/2 by 2 inch illuminated opening contained three vertical bars which varied
in width and in depth in eight different presentations.  The  subject was asked to
determine which of the bars were closest or most distant.

         2.  Results

      The data from the depth perception test utilizing  the Verhoeff apparatus can
be summarized as follows: Each subject was presented with 8 successive
determinations, his  score expressed as  a fraction of those correct, e.g., 8/8,
all correct;  4/8, 4 incorrect.   Of the first 10 subjects run, 8 scored 8/8 on
control and at all COHb levels,  while 2 subjects, obviously with poor  depth
perception,  had equally poor scores, i.e., 1/8, 2/8, 3/8,  etc.,  both at the
control condition and at all three COHb levels.  (A similar pattern of results
were found on subjects run with and without  4 ounces of 80° alcohol.) While the
test  is apparently a good measure of  a subject's binocular stereopsis,  it
obviously was not able to  differentiate the effects of relatively low levels of
carbon monoxide (up to 17% COHb) -  at least under these conditions with this
equipment.  The test was, therefore,  eliminated.

-------
                                                                            6L


 V.  Driving Experiments

      A.   Introduction

      The purpose of the driving experiments was to attempt to detect and
 measure certain changes in actual driving performance that may take place after
 exposure to carbon monoxide.  These experiments were conducted on an unopened,
 divided highway utilizing a visual interruption apparatus as a measuring tool in
 a 1971 Plymouth Fury test vehicle.

      Specifically, the experiments were designed to measure any decrement in
 the driver's performance in a task which is fundamental to the driving process:
 the processing of visual information required for positioning of the vehicle on the
 roadway.  By standardization of car and road, and specially designed apparatus
 to control the amount of visual attention the driver could devote to the task, these
 experiments involved a driving situation that was both  realistic  and safe and
 provided objective measures of driver performance.

      Experimental attempts to measure the performance of drivers while
 operating vehicles on the roadway have so far been limited in their results.  It
 is difficult to carry out controlled experiments without the driver's being aware
 of the measurement procedures and being influenced by them.  Furthermore,
 the demands on the ability of the driver to process the  incoming information
 vary  greatly. At  times the demands are considerably below his full capacity; at
 other times his capacity may be exceeded.   Some of the needs for information
 relate to tasks in  the immediate and ongoing control of his vehicle, for example,
 in lane holding,  maintaining distance from a leading car,  steering, braking,  or
 accelerating in relation to other traffic or roadway hazards.  Thus, one approach
 to the measurement of performance may be  based on the information processing
 inherent in driving, utilizing the techniques  of information theory.

      In the present project a technique of intermittent visual time sampling was
used.  The field of view of the driver is interrupted systematically as  he operates
a vehicle  on the roadway.  This was accomplished by a helmet-ircunted visor
which alternately dropped to occlude the driver's vision at controlled frequencies
and duration (the Visual Interruption Apparatus,  or VIA).  In previous experiment
conducted on closed road systems and unopened sections of interstate highways,
it was found that sufficient information for error-free lane holding operation was
obtained if drivers viewed the road for a period of one-half second at 2-second
intervals at a speed of 60 mph,  at.every 4 seconds at 25 mph,  and every 9 seconds
at 5 mph.  The records were obtained under conditions where (1) the occlusion
rate was controlled by the  experimenter and vehicle speed by the subject, and
(2) where vehicle speed was controlled and the driver actuated the occlusion
device as needed.   The results were used to develop a mathematical model which
related the driver's informational content of the roadway in "bits" per mile, the
speed of the vehicle,  and the driver's estimates of his own uncertainties.  This
work was subsequently extended to  include the information-processing demands
of car-following and passing, and driving under normal traffic conditions.

-------
                                                                           61
      The Visual Interruption Apparatus itself consists of a helmet with an
attached face shield which can be moved up and down by means of a gas-operated
piston mounted on the helmet.  The piston can be operated by a solenoid valve
whose switch is accessible to the driver's left foot, or alternatively by an auto-
matic timing sequence.  The shield is designed to be non-transparent, but
translucent.  During testing,  the normal position of the shield is down (the
occluded position).  Depressing the solenoid switch drives the helmet into the up
(seeing) position for a fixed viewing time before the visor closes again.  (See
Figures 23, 24 and 25. ) Based on  previous experience, a viewing time of
approximately 0. 5 sec. has been found to be suitable. (For a more detailed
description of the  VIA see Senders, et al. , 1967, 1969.)

      B.  Selection of Subjects

      All subjects used in the driving experiments were drawn from the pool of
those who have already participated in the laboratory phase of the  experiments.
In addition to the health requirements  previously described (see Section II),
additional screening for the driving phase  ensured that all subjects:  (a) had a
current driver's license; (b) had  at least 50, 000 miles driving experience;
(c) had no more than two re portable accidents for six years; (d) had no more than
three moving violations for two years,  (e) Finally,  no one who disliked driving
was  selected as a  subject.

      C.  Roadways

      Two different roadways were used in the driving experiments.  The first
was  a closed loop, two-lane winding roadway in the  Arnold Arboretum in  Boston.
Initial experiments establishing testing procedures with the visual interruption
apparatus were carried out on this  roadway from which all other vehicular traffic
was  excluded.  The winding nature  of the road,  however, plus the  occasional
presence of pedestrians and bicycles,  made it  undesirable for the  basic driving
experiments.  For this purpose we used a completed but unopened segment of
Interstate 95,  a limited access eight-lane  divided highway in Danvers, Massa-
chusetts,  about 20 miles from the School of Public Health.   The test course was
1. 9 miles in length and consisted of the two middle lanes of the northbound
segment of this highway.  During the testing no other vehicles or persons were
present on the highway.

      D.  The Test Vehicle
      The test vehicle,  a 1971 Plymouth Fury sedan was equipped with a number
of options, some primarily for safety purposes,  and some to facilitate carrying
out  the experimental conditions.  In all cases the intent was  to keep the vehicle
as "ordinary" as possible insofar as the subjects were concerned.  The special
equipment on the test vehicle is listed by description and by  manufacturer's code
in Appendix P.   Certain other equipment was necessary toaccommodate the
experimental apparatus and experimental protocol.  For example, the electrical
system of the vehicle had to supply additional power for the timers and recorders,

-------
62
       Figure 23.   Visual Interruption Apparatus:  visor raised for full road
       vision.

-------
                                                                          63
Figure 24.  Visual Interruption Apparatus: visor lowered for occluded
vision.

-------
Figure 25.  Overall view of Visual Interruption Apparatus and control
and recording equipment (lower right).  Plastic tubing supplies and
vents the  compressed CO  which operates the helmet visor.

-------
                                                                          65
and these requirements were met by a larger 70-amp-hour battery and 60-amp
max. capacity alternator, monitored by a 60-amp calibrated ammeter.  A
magnetic tape recorder was used for voice recording.  The vehicle power plant
had to produce very low amounts of CO when operated at the speeds  expected
during the experiments,  since the car was to some extent exposed to its own
exhaust.

      Other special features included a seat back from a 1970 Plymouth bucket
seat, with a non-integral head restraint, which was installed on the  driver's seat
to allow shorter test subjects to sit normally without interference from the
actuating cylinder on the VIA.  Inertia reel shoulder harnesses were also
installed. Aside from improved convenience,  the inertia reels provided a feeling
of freedom of motion similar to that afforded by not wearing a shoulder harness.
Since most of the driving population  choose not to wear seat belts, let alone
shoulder belts, this freedom of motion  is important in making the subjects feel
comfortable in the experimental vehicle.  A dry chemical fire extinguisher was
installed in the passenger compartment.

      In addition to the data obtained by the VIA apparatus,  steering-wheel move-
ments and measures of vehicle position in lane were recorded.  Steering-wheel
position was measured by means of a potentiometer mechanically connected to the
steering column.  The position of the vehicle in the lane was monitored contin-
uously with scanning photocells contained in each of two units mounted on the
vehicle directly in front of the two front wheels.  These sensors  detected each
crossing of a white lane marker.  A Rustrak event recorder was used for
recording helmet usage (number and time interval), and lane-holding errors.
Steering wheel movements were traced on a Rustrak potentiometric  recorder.
The  cassette equipment in the vehicle was used to play back a standard
instruction tape to all subjects.

      E. Experimental Procedures

      Each subject in the driving experiments participated in three separate
sessions, the  first of which was a training session of several hours  duration
devoted to familiarization with the driving task, and the use of the visual inter-
ruption apparatus.  The second and third sessions lasted approximately eight to
nine hours each,  from the time the subject first appeared in the morning for
gassing until he was dismissed after having his COHb level reduced  to 6% or less.

      During both the training and the actual driving runs on the test course,
the subject was required always to maintain a path within the normal 12-foot
marked lanes  on the highway.  To accomplish this, he had to select  a viewing
frequency to match the fixed speed conditions of the vehicle. In essense, the
driver was required to stay in lane exactly as if he were able to view the road
continuously.  Occlusion data were obtained in runs at two speed levels, 30 and
50 mph,  and each subject drove at both speeds  on  each day.  These two speed
levels permitted interaction effects between CO and speed to be assessed.

-------
                                                                            66
       In previous experimental programs,  experience was gained with both fixed-
 vehicle-speed,  subject-controlled occlusion conditions, and with subject-controlled
 vehicle speeds and fixed occlusion rates.  Subjectively, it was observed that the
 variable-occlusion condition provided the driver with more immediate control
 over his perceptual environment.  The response time of the visor to the depression
 of the foot switch is virtually instantaneous.  On the other hand,  the vehicle can
 be considered as relatively slow to respond to minor changes of acceleration  or
 brake pressure.  There is evidence to show that the driver even  perceives
 straight and level road as possessing a variable information  rate.  Consequently,
 the fixed-vehicle-speed, subject-controlled-occlusion condition is likely to be the
 more desirable of the two experimental modes.  (Senders, et al., 1967, 1969)

       !• Training Sessions.  Subjects were trained in pairs.  After reporting to
 the Harvard School of Public Health in the morning, the subjects  were driven in
 the test vehicle to the Danvers course.  During the one-half hour trip, the subjects
 listened to a pre-recorded tape which described the experiment they were about
 to participate in (the text of this tape is presented in Appendix Q).  After this
 introduction the experimenter answered any questions and then explained in more
 detail the experimental task,  the operation of the Visual Interruption Apparatus,
 and all other experimental instrumentation in the vehicle.

      After arriving at the test  site, the experimenter first drove over the course
 to familiarize  the subjects with it, noting the starting and finishing points, and
 explaining  the operation of the vehicle's automatic speed control.  The exper-
 imental equipment was then checked, and the lane  sensors mounted on the vehicle.
 The  experimenter donned the  VIA and drove a test run to demonstrate the
 operation of the helmet. The first subject then took his place in the driver's
 seat, adjusted it to his normal driving position, adjusted his  lap belt and upper
 torso restraint, and positioned the left foot switch which raised the visor of the
 helmet.  He started driving and locked in the automatic speed control at 30 mph.
 The  experimenter activated a switch which lowered the visor over the subject's
 eyes. From this point on the visor remained down until the subject actuated his
 foot  switch which raised the visor for 0. 5 sec. of full vision.

      Throughout the training session the subject was instructed to take as many
 looks as needed, but no more, in order to keep the vehicle at all  times within
 the  12-foot width between the lane markers. Errors in lane holding were indicated
 by the sensors which activated a light on the instrument panel.  The subjects were
forewarned that during the later tests crossing the lane markers  would cause  the
 run to be aborted and restarted from the beginning.  The subject  then made as
many runs  over the  1. 9 mile course as necessary to make his driving largely,
 though not  necessarily completely, error-free.  At this point he began a series of
 8 runs, 4 north and 4 south over the course at 30 mph.  These runs duplicated
those that he would undertake in the two subsequent "data" sessions (during  the
training sessions all data were recorded, but were not necessarily intended for
use in the analyses).

      The second subject then underwent the same  training procedures for his

-------
                                                                           67
30 mph runs.  Following a lunch break, the first subject began a series of eight
runs at 50 mph,  after which the second subject did the same.  At this point most
subjects were found to be essentially error-free in their driving performance,
that is, they were able to keep their vehicle within the lane markers with a
minimum number of "looks".  Where this was not the case, additional runs were
made until the experimenter was  satisfied with the subject's performance.  No
training sessions, or experimental runs,  were ever carried out during rain,
snow,  wet roads, high wind,  poor visibility, or any other conditions that might
adversely affect driving performance or provide significantly different driving
conditions from  one day to the next.

      2.  Experimental Sessions.  Once the subjects were trained, the data-
gathering  runs were scheduled within the next several days.  The  same  two
subjects participated in pairs for these sessions.  For the first day one of the
subjects was randomly selected to breathe a carbon monoxide mixture (700 ppm)
which raised his COHb level to 17%.  The second subject concurrently underwent
an identical breathing procedure, but was exposed to only room air.  On the
second day the exposures were reversed,  i. e.,  the  subject who had CO now
received only air and vice versa.  The  experiment was double blind in that neither
subject knew his exposure on either day, and the experimenters were  equally
unaware as to which subjects received CO on which  days, until  the conclusion of
all data runs.

      The general schedule for the road testing  is outlined in  Table 2.  Initial
procedures at the School were the same as those previously described for the
laboratory sessions.  However, upon arrival at the  test track,  a "refresher" is
given to restore the subject's  COHb level to 17%.  For this purpose six 200 litre
Douglas bags were used,  3 containing a CO mixture of 700 ppm for the CO sub-
ject, and 3 containing plain air for the non-CO subject.  Breathing times on the
Douglas bags were based on curves plotted from previous laboratory data for the
rate of uptake of CO for the subject exposed to CO.  The other subject was on
plain air from a Douglas bag for an exactly equal length of time.  Five minutes
after the conclusion of the gassing, an  alveolar sample was taken from each
subject, immediately after which the first subject began his series of driving
runs at 30 mph.   Again accuracy in lane holding and error-free operation was
stressed,  consistent with the fewest number of  "looks" with the helmet needed to
achieve this goal.   Errors in lane holding or any departures from normal in the
operation of the  equipment (failure of the  speed control was the most common)
caused the run to be recommenced.

      A few minutes before the first subject finished his first run, the  second
subject was given his second CO "refresher", to again raise his COHb level to
17%.  Similarly, just before this  subject completed  his 30 mph  runs, the first
subject had his second refresher and then began his 50 mph session.  Finally,
the second subject had his third refresher immediately before beginning his
50 mph runs.  In all cases alveolar samples were taken 5 minutes after each CO
refresher.  Each subject began his driving run  within about 5-10 minutes after
breathing the CO mixture and was, therefore, very  close to his intended 17%

-------
                                                                           68
COHb level at that time.  Each driving session at 30 mph and 50 mph consisted
of eight runs of 1. 5 miles each, 4 in a northerly and 4 in a southerly direction.
                                  Table 11

                       Daily Timetable for Road Testing
      8:00 - 8:10       Arrive at School of Public Health, first alveolar
                         sample taken
      8:10 - 9:30       Subjects  gassed
      9:30 - 9:45       Blood and alveolar samples taken
      9:45 - 10:30      Drive to  test course
     10:30 - 11:00      Set up testing equipment; subjects regassed
     11:00- 11:35      1st subject,  1st session, 30 mph
     11:40 - 12:15      2nd subject,  1st session, 30 mph
     12:20 - 12:45      1st subject,  2nd session, 50 mph
     12:50- 1:15       2nd subject,  2nd session,  50 mph
      1:15 - 2:45       Pack-up; lunch; return to School of Public Health
      2:45 - 4:15       Subjects  degassed on O^
      4:15 - 4:30       Final blood and alveolar samples taken
      4:30             Subjects  released
      Upon completion of the last run, the subjects returned to the laboratory,
where alveolar and blood samples were taken.  They were then placed on a
breathing mixture of 99% O^ and 1% COg for approximately 80 to 90 minutes
until the COHb level of the CO subject for  that day was determined to be under
6%.  At that time both subjects were dismissed.

      F.  Results
         1.  Analysis of Visual Occlusion Data

      The time records of visor position provided the basis for measures of driver
uncertainty in carrying out the task.  Figure 26 shows a representative sample
from the record of one subject during a driving run,  with arrows indicating por-
tions of the tracing representing periods of occluded and full vision.  Viewing
times were fixed at . 5 second duration each.  The numbers shown in the figure
above the occlusion period portions of the tracing represent the time (in milli-
meters of tape) the subject could drive without vision of the roadway before
having to activate the visor to obtain  more information.

      a. Number of Occlusions: The first analysis of the Visual Interruption
Apparatus  data provided a distribution of the numbers of occluded periods under

-------
Cf

4.1



' V,
1
IW
w
4>
" ^ 	 -J
^r
i»
*
CM
O>
-4
•
F V.
3.9
<




4



H




.3





r



14
k
h



.5
pa
'Pk
w
4>
rn
M
-*
al
w
o>
-<
•

.1
)CC




lusi
^



14
on



.5
per




iod
r>



13
»



—
M
W
*

10
M
»
a
a>
-4
•

.6





L
«
j



13
-fix



.8
ed


i i_

viev
<


j i_
14
vine
>


t
—
10
jtir
CM
4>
M
ISJ
14.
ie
o
Ol
•
Figure 26.  Sample of time record of Visual Interruption Apparatus
(VIA) taken during one driving test.  The fixed viewing times are
0.5 second in duration.  Occluded periods vary in length according
to the Judgement of the driver.

-------
                                                                           70
 the various experimental independent conditions.  It became apparent that some
 of the observed variability in the total number of occluded time periods between
 and among subjects could be accounted for on the basis of variations  in "locking-
 in" the automatic speed control at the exact experimental speed prescribed.
 Hence, for the fixed distance of the experimental route,  there were unequal
 traversal  times.  Thus the total number of occluded time periods varied in part
 because of the time differences in test runs.  (The speed control device was  a
 production item designed for general use in highway cruising and proved to be
 only marginally useful for original experimental requirements.)

      b.  Percentage of Occlusion Time;  For comparative analyses,  it would
 be more useful to examine  total occlusion time as percent of actual running time
 as a measure of performance.  The occlusion time percentages were averaged
 for each subject's eight test runs under the various experimental conditions and
 the values are presented in Table  12.

      c.  Distance Traveled During Occlusion:  A physical interpretation of
 average occlusion time percentage is possible by transforming the percentages
 into equivalent road distance measures.  That is, the transformed data will
 represent  the cumulative total road distance that the vehicle travelled while the
 subject's vision was obscured.  In this transformation, the  data were normalized
 in termsof an arbitrary one-mile base.  These cumulative occluded distance
 (COD) results are presented in Table 13.

      In general differences are  apparent in the average  values between the CO
 and no-CO runs.  These differences, however,  are small in magnitude and exhibit
 several reversals in the direction of change between the  test and control conditions
 for both the 30 and 50 mph speeds.  A paired t-test for a comparison of treatments
 (carbon monoxide vs.  placebo) showed that there was no  statistically significant
 difference  as a result of the treatments for either the 30 or  50 mph speeds.  How-
 ever, although the COD values fell within a narrow range across subjects,
 variability in the individual subject's performance between days was apparent,
 complicating the determination of any effects of carbon monoxide on the perform-
 ance measure in the above comparisons.

      Having no previous experimental data regarding directional change in the
 arbitrary performance measure as a function of CO level, a pilot study was
 undertaken using the VIA procedures in which laboratory personnel participated,
 and in which alcohol was used as the experimental agent.  In this study, subjects
with a blood alcohol level of 0. 10% consistently showed a decrease in  the per-
 centage of  occlusion time compared with their own control results.  It may be
noted that these results were obtained from tests that were performed on the
 same day and,  therefore, were not affected by the inter-day variability that seems
to be prevalent in the CO data.

      The randomization in the experimental design across days, and treatments,
and with the 30 mph test always preceding the 50 mph test, made it possible to
analyze the data for effects  of CO while controlling inter-day variation.  It was

-------
                                                             71
Average Percent Occlusion Time as a Function of Gas
Intake and Vehicle Speed
CO Condition
Subject
G. F.



W. F.



J. L.



H.A.



0. R. - I



O.K. - II



P.L.



J. D.



A. C.



M. H.



and Speed
No-CO -
CO -
No-CO -
CO -
No-CO -
CO -
No-CO -
CO -
No-CO -
CO -
No-CO -
CO-
No-CO -
CO -
No-CO -
CO -
No-CO -
CO -
No-CO -
CO-
No-CO -
co-
No-CO -
CO -
No-CO -
CO-
No-CO -
CO -
No-CO -
CO -
No-CO -
CO-
No-CO -
CO-
No-CO -
CO -
No-CO -
CO -
No-CO -
CO -
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
30
30
50
50
Percent
Occlusion Time
83.5
84. 4
82.9
83.2
89.1
87.4
86.7
83. 1
87.7
88.4
85.3
85.1
87.5
88.4
87.0
88.0
86.1
85.5
84.8
84.9
87.6
87.5
85.7
84.8
87.0
88.8
85.9
86.4
87.5
88.6
86.3
86.1
87.5
89.6
86.9
86.0
88.6
87.4
88.0
88.0

-------
                     Table 13                                 72




Changes in the Cumulative Distance Parameter as a


Function of Gas Exposure
CO Condition Cumulative
Subject
G.F.



W. F.



J. L.



H.A.



O.K. - I



O.R. - II



P. L.



J. D.



A. C.



M. H.



and Speed
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
No-CO - 30
No-CO - 50
CO - 30
CO - 50
Distance (ft)
4409
4377
4456
4393
4704
4578
4615
4388
4631
4504
4668
4493
4620
4594
4668
4646
4546
4477
4514
4483
4625
4525
4620
4477
4594
4536
4689
4562
4620
4557
4678
4546
4620
4588
4731
4541
4678
4646
4615
4604

Difference (ft)
30 vs. 50
32

63

126

227

127

175

26

22

69

31

100

143

58

127

63

132

32

190

32

11



Test Day
1

2

2

1

1

2

2

1

2

1

1

2

1

2

2

1

2

1

1

2


-------
                                                                           73
observed that,  without exception,  the percentage of occlusion time at 50 mph was
less than that obtained at 30 mph on the same test day.  This difference was found
to be highly significant (P <. 001)  in a "t" test of paired differences.   Jt would
appear then that the subjects found the 50 mph speed to be the more difficult,
i. e., they needed to view more of the road, hence  they did not accumulate as
much resultant occluded distance at the faster speed.   The suggestion is that
subjects adapted  first to the requirement of the 30  mph runs, then attempted at
50 mph to replicate their performance at the lower speed driven earlier on the
same day,  the test itself, of course, remaining the same.

      The greater difficulty of the task at the  50 mph speed relative to 30  mph is
reflected in the COD measures as presented in Table  13 on the basis of data
from tests carried out on the  same day,  with  the 30 mph values serving as base-
line.  A two-way analysis of variance was then carried out on these data to
determine any differential effect of CO.  The  results indicated a significant
difference between the control and CO conditions which was  significant at  the . 05
level.  Thus the findings  indicate that, under  all conditions, the  subjects required
more viewing of the roadway to maintain accurate lane performance  than at 30 mph.
In addition,  at 17% COHb subjects required even more roadway viewing at the
higher speed relative to  the lower speed, as compared to the control condition.

          2. Analysis of Steering Wheel Reversal Data

      Records of steering wheel movements were taken concurrently with those
for visual occlusions, registered on a Rustrak Series  400 Recorder.   Input to the
recorder was from a potentiometer, with independent power supply,  coupled by
means of a friction wheel to the steering wheel hub.  The chart speed was
calibrated at 1.69 mm/sec.,  thus permitting  identification of reversals occur ing
at minimum intervals of 250 milliseconds.  The data were recorded as an analog
trace on a continuous graph calibrated at 1 mv intervals.  The scale  ranged from
-50 mv,  and a 1 mv deflection on the graph represented a 2  degrees turn of the
wheel.  Figure 27 shows portions  of a record for one  subject in runs at 30 and
50 mph,  under the CO condition.

      In this analysis, steering wheel reversals were  operationally defined as
deflections of the steering wheel of 2-9 degrees in  magnitude, and 10 degrees
or greater,  when accompanied by  a change in directional input to the wheel.
The choice of these values was determined partially on the basis of previous
research (Platt,  1963) and in  consideration of vehicle  and occlusion test
constraints.  Deflections less than 2 degrees were ignored,  since these smaller
deflections represented changes within the "slack"  of  the steering mechanism,
and would not cause directional changes of the vehicle.  Reversals of 10 degrees
or greater were included to represent relatively large corrective changes in
vehicle direction.  Because of the "error free" performance demanded by the
driving test design,  reversals of 20 to 25 degrees represented the practical
maximum upper limit of deflections, since such deflections  would cause the
vehicle to cross the lane line  and end a particular run.

-------
                                                                     74
         Subject: WP

         Dec.  8, 1972
CO: 17* COHb
S ^"\ ^> ^\ f\ f~\ ^*\ ^*} f
-L.


Rui
Sp

a:
sed

3
l:

Noi
30

•th
MPH
N


"\ '
	 _


p c
•' ^_
n

5 C
4

3 '


-\ /•
— ^c


> '
i — <



^ '


-) r



^ ^
— <
J

? c
> —


? c


i

c


) <;
^
1
J=

I
         Subject: WP

         Dee. 8, 1972
CO: 17% COHb
         o 00000090009  Q 99  9 999<
       Run:  3 Worth
       Speed* 50-MPH
Chart  Speed: 1.69 mm/»ec.   Reiolutloni 250 mllliieoonde by 2 degrees.
Figure  2?•  Sample  records of  steering wheel  reversals  of one
subject driving with 17% COHb  at  30 and 50 miles per hour.

-------
                                                                          75
      The data were treated statistically by analysis of variance.  These
 analyses  were made independently for frequencies of reversals of the two mag-
 nitude ranges for both 30 and 50 mph, with data standardized for time of run
 and distance of roadway travelled for CO vs.  control conditions,  as below:
          Reversals

    1.  _2°  10° reversals

    2.      _10° reversals

    3.  _2°  10° reversals

    4.      _10° reversals

    5.   2°  10° reversals
   6.
10  reversals
  Condition

CO vs. No-CO

CO vs. No-CO

CO vs. No-CO

CO vs. No-CO

CO vs. No-CO

CO vs. No-CO
Standardization

 distance

 distance

 distance/time

 distance /time

 time

 time
      Data were standardized for distance of roadway by computing the length of
the record to be analyzed as a function of a known chart speed and known vehicle
velocities over a fixed length of roadway.  As a result, all distance analyses
indicate frequencies of reversals for 1. 12 mile segments of roadway.  Data
standardized for time were computed to reflect frequencies of reversals ac-
cumulated at 30 mph during the  same period of time as those at 50 mph.  These
analyses, standardized for time, indicate frequencies of reversals for 1. 33
minutes of run time.  The analyses were  based on the data for individual subjects
for each run, rather than means, to take  advantage of the greater statistical
power of this approach.

      The analyses of variance indicated highly significant intra- and inter-subject
variability (P {. 01) and negligible changes in variability associated with CO
administration.  In only two of these analyses was an order effect apparent.  In
these the significance level was  P^ 05. The interpretation of the analyses  is that
carbon monoxide at the level of  17% COHb does not differentially affect the  fre-
quency of steering wheel reversals.

-------
                Percent COHb versus CO Pressure (APPENDIX A)
         71
      The following table shows the percentage saturation of the hemoglobin
with CO at various alveolar pressures of CO, calculated from the Haldane
formula % COHb = 23Qf(pCO) .  The figure 230 has been used rather than
        %0 Hb     pO
Haldane1 s original value of 210 because recent work indicates that 230 is more
nearly correct. The alveolar O   pressure is assumed to be 98 mm Hg,
which is the  generally accepted figure for sea level.
%COHb
0.87
1.73
3.45
5.05
6.63
8.16
9.63
11.08
12.46
13.80
15.11
16.37
17.60
18.78
19.95
21.05
22.15
23.23
24.26
25.24
26.22
PPM
5
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
170
180
190
200
                                           % Atmosphere
                                                0.0005
                                                0.001
                                                0.002

                                                0.003
                                                0.004
                                                0.005

                                                0.006
                                                0.007
                                                0.008

                                                0.009
                                                0.010
                                                0.011

                                                0.012
                                                0.013
                                               .0.014

                                                0.015
                                                0.016
                                                0.017

                                                0.018
                                                0.019
                                                0.020
mm: Hg
0.0038
0.0076
0.0152
0
0
0
0248
0304
0380
0.0456
0.0532
0.0608

0. 0684
0.0760
0.0836

0.0912
0.0988
0.1064

0. 1140
0.1216
0.1291

0.1368
0.1442
0.1520

-------
                                                APPENDIX B
77
                          Medical Questionnaires  Used in Subject Selection
Standard Form 89
(Rrx MABCH IWi)
BUREAU OF THE BUDGET
CIRCULAR A-32
REPORT OF MEDICAL HISTORY
n-iM-oi
THIS INFORMATION IS FOR OFFICIAL USE ONIY AND WILL NOT BE RELEASED TO UNAUTHORIZED PERSONS in •»>.«.( H. ID-IISI
1 US! IAME-FIKT HAME-NIOOLE RAM
4 HOME ADDIESS (Number. Uriel ar RFD. city or
J SE1
1! OATEOFIIITH
s uu

num. Stall and ZIP
Cadi)
i TOTAL tuts GOVEIHMIIT smicf
MiiiiAii
13 PLACE OF IIITH
(IVI1IU

IS EMNIIIIG FACILITY 01 EUMTHEI, WD AODItSS
J MtDE MD COMfOKK! 01 HHITIOI
5 wmSE OF EUMIMIIOI
10 AGEHCI
1 IDEITIFICAIIOI 10 MID SSA*
6 DATE OF EUMIMTHM
11 OKAIIZATIOH IMI1
14 HAME. niATIOHSHIP, AID UDIESS 01 HOT OF II*
11 OTHEI IIFOIMATIOI
17 STATEMEHT OF EMIIIItE'S PIESOII KAITH IN OWH WOIDS (Follow h description of fail bularv. if complaint nuts)
11 FAMILI HISIOI1
IELATIOI
FMHEI
MOTHEI
SPOUSE
IIOMEIS
AHD
SIS1EIS
(HILDIEH
.


•CE












STME OF HEALTH












IF DE«D. UUSE OF DEATH












AGE »!
DEATH












If H»i Ml HOOD IEIATIOH (Part nl. bnlbir. niter, other)
01 HUSIIID 01 WIFE
tES












DO












(Chick tach Him)
NtD TUKICOLOSIS
HtD STPHILIS
HID DIIIETES
HID CAICEt
HAD (IDIIEI IIOUHE
HAD HUH TtOUILE
HAD STOMACH HOUILE
HAD IHEUMAIISII (Arthritis)
HAD ASTHMA. HAI FEVEI. HIVES
HAD EPIIEPSI (Fill)
(OIUinED SUICIDE
IEEI IHANE
KIAIIOII(S)












  HAVE TOO EVEI HAD 01 HAVE VOtl «0» fPlaci chick at lift of tach Him)
tis


















•0


















(Chick tach Him)
SCAIIEI FEVEI. BVSIKUS
DIPH1HOIA
IKEIUUm FEVEI
SWOILEH 01 PAIIFUl MIITS
VUIIK
(OLOi ium«ss
FIIOUEH1 01 SEVEIE HEAMCHE
OI1IUESS 01 flllllM SKILS
EVE IIOIIItE
EAI. HOSE 01 TNIOIT TIOUILE
IIMHIHG EAIS
HEAIIIS IOSS
(HIOHIC 01 FIEOUEHT COLDS
SEVEIE TOOTH 01 GUI TMMIIIE
SKOSIFIS
HW FEVEI
HISTOIV OF HUD MJUn
SIIH DISEASES
TES


















m


















(Chick tach mm)
COITEI
TUIEICUIOSIS
SOAriHG SWEATS (Night I u rats)
tSTHW
IMHTIESS OF HEATH
Mil 01 riESSUIE 11 CHEST
CHIDIIC COUGH
PALNIATIOII 01 POUniK HEAI1
HIGH 01 LOW HOOD PIESSUIE
CHNPS II TOIII IEGS
FIEQUMI IHOIGESTIOII
STWAffl, UVEI 01 IITESTINAl TIOUILE
GALL IIAODB TMUILE 01 GALL SIOIES
uuma
All IEUTIOI TO SEIOM. OIUG 01
•EDICIIE
IISTOII DF IIOIEI IOIES


71 HAVE nil EVll (Chick tach item)










WOM GUSSES-COITACT IEIIS
WOM 1H AITIFKIAI ETE
WON HEAIIIt AIDS
SIlinEIED 01 STAMNEIEO
won i IUCE 01 uci SUPPOIT
23 HOW KMT JOIS HAVE TOD HID 11 THE
PAST IHIEE TEAIS7










ATTUFTEO SUICIDE
KEH A SLEEP WAKEI
LIVED WITH AIVOIE WHO HtD
TUIEICUIOSIS
COUOHEO UF ILOOD
NED EKESSIVEll AFIEI IIIUII 01
TOOTI EITIACTIOI
14 WHAT IS THE 10MEST PEIIOD TOU
HEU All OF THESE MIS!
MOUTHS
IIS


















10


















(Check each item)
TUMI, CIOWTH CIST. CAICEI
IUPTUIE/HEIIIA
APPEHOKITIS
PILES 01 IKTAL DISEASE
FIEOUEIT 01 UMFOl UIIHIIIOI
IIDHEI STOIE 01 ILOOD II UIIIE
SUGtl 01 ALIUWI II UIIIE
MILS
VD-SVPHILIS. GOIOIIHEA. ETC
IECEIT GAW 01 LOSS OF WEIGHT
AIFHIITIS 01 IHEUIATISII
IOHE, JOIIT, 01 OTHH DEFOUITI
UMEIESS
LOSS OF AM. LEG. FIHGEI. 01 TOE
PAIIFUl 01 IIIM" SHDULDEI M ELIOW
IAO TIOOIIE OF AIT HID


2? FEWLES Olll A HAVE 100 EYEI-










IEEH PIEGIAII
HAD A VACUA! DISCHAKE
IEEH TIEATEO Kl A FEMALE DISOIDEI
HAD PAIIFUl KEISTIUATIOI
HAD IIIEGUIAI MEISTIUATIOI
IES


















m


















(Check tach item)
TIIC( 01 IOCIED MEE
FOOT TIOOILE
IEUIITIS
PAIAirSIS (lac infantile)
EPILEPSI 01 FITS
CAI, IIAII. SEA, 01 All SICHESS
FIEOOEI1 IIOUILE 1LEEPIIG
FIEOUEHT 01 TEIIIFIIIG HIGHTMAIES
DEPIESSIDI 01 EiaSSIVE WONT
LOSS OF MEMOIT 01 AMIESIA
IED WETTIIG
HEIVOUS TIOUILE OF All SOU
All DUG 01 HAKOIIC HAIIT
EXCESSIVE DIIIIIIG HAIIT
HOMOSEXUAL IEIDEICIES
PEIIODS OF UKOHSCIOUSIESS


1 COMPLETE THE FOILOWIHG




AGE AT OISET OF MEHSTIUATIOI
IHTEIVAl IETWEEI PEIIOOS
DUUTIOI OF PEIIODS
DATE OF LAST PEIIOO
OUAITITI D MIMAL D EX(ES!I« Q SCAITV
!S WHAT IS TOUI USUAL OCCUPATION
26 AIE TOU (Check one)
Q IIGHT HAIDEO Q LEFT HAIDEO

-------
Appendix  B (cont.)
78
     ns
                                                       cuta EACH ITU IK N u  EVEN inn CHKOD  m MUSI IE NUT EVUUW u luu vtu <* nun
                   v HIVE ion «u RHIJEO EWLOMIM 01 IEEII
                      10 HOLD * (01 KCMItE OF
                      1  SEIUIWIK ra (HEMIUU. OIHI. SUIIIGH1. E1(
                      I  IMIIlin TO KIFOU Cflttll MTIMS
                      C IHIIllll 10 UVHt CfltAIH nKIIICMS
                      D  OTHK MiDIUI msOHS (If jn. gii'f reasons)
                   II HIVE TOU EVEI WOMED WIIM UDIOtCllVE SUUTUKE?
                   W  HAH (OU Hit KM OEIIID UK IISUIMCP  (If lit
                      Hale reason and fife details i
                   31  HIYt TOU HU. 01 HIVE IOU IEEN ADVISED 10 HIVE,
                      Mr OTUIinS'  (Ifyts dncribe jad girt
                      age at whnh occurrtd)
                      mi TOD EVEI IE» * PtlllNF  fiammilltd or
                      mluntary) III WITtl HOWTU 01 StNllOIIUH'
                      llfyts. spmf) u-btH. u-hirt uhy and
                      auiai a{doctor, jnd tvinpleit aJitreit of
                      htafttal or fitaic )
                   » HAVE fOll HU KU HI IlllESS   (If v». gift complnr
                     attdrtti oft/ottor, ftaipital, itmif, and
                     iletaillt	
                  )5 mu rou iiuim VOIIHIIF fit IIIIESSES OMII mu
                     IMOI C0in>  (If yet. uthah illnmn)
                  16 HIVE IOU (VH lEEK IIIECIEO FOI MIUIilY IEIVICE
                     IKIIISE OF PH1SIHL. Hum, 01 ITHEI I(«SO«S'
                     (If }rt, gtl e tlate and rtaion /or re/ec-
                  II HIVE roil EVEI IIB HUHMBEO FIOIl DIEirilT SEHICE
                      IIUIM OF nrsicu, »(«w, ai OIKEI iuw»i'
                      < If yes. gilt dale, reason, and lyfe a/
                      discharge   ubelber  konarakle.  other
                      thau  boaorab/f. far unfilneii at un-
                  II HIVE TOU EVEI lECEIVEO, IS THEIE niOIIII. 01 NtVE
                     IOU imiEB FOI PEIUOC 01 (OHIEIUIIOI IN EIIS1
                     HC OISMIIIIT'  (Ifyei. ifeaf) u-hal kind.
                     granted bf wham, and U'hal amount.
                     when, u by I
                   11  010 rOU HIVE OlFIKULtt WITH SCHOOL SIUOIIS 01
                      11ICHEK) (Ifytt girt details /
  WIBIIC  I HIS! 01 OISMMSI MSVEI 10 MT OF IHE OUBflOB 01 THIS (Oil WT H lUMiHEC II FIX 01 II
                                                                                          (IIII it  INI)

  I aillFI INITI HAVE KVIENEO fHE KiHWIlt IMOIUIIM SUmiEO IT W UD Ml II IS mil MO COMPLETE 10 THE US! OF NT HOWLEOtt
  I IU1HOIIZI MT Of IHI OOCFOIS. HOSPIIILS. 01 (UIICS WITIOIED MOUt TO rilMISH IHE MVtlmlEIIT I COWIEIEIIUSCIIPT 01 H MIDIUI WOO FOI HIKHU 01 PMOSilUC « tmiUIIM FOI THIS EUTIOTMIIT 01SEIVIO.
  ITFED 01 PIWTED HUU OF EUDIIUI
                                                                                  SIHIIUK
  »  PHTSKIU S SUMMIT ANO ELAIOUTICW OF All PEHIKIT Din  (Phytnian shall commtnt en all pantile ansutn in Hem 20 thru 38)
  ITKD 01 nillED KUI OF NIUCIM 01DUAIIEI
                                                             DUE
                                                                                  ilUATUIE
                                                                                                                                            MMIEI OF imtKtO
                                                                                                                                              WEEK
                                                                                                                     •US GOMRMHtNT NINriltC OFriCl  I9U OT—71O-MI

-------
Appendix B (cont. )
                                    7/9
Confidential
             Guggenheim Center for Aerospace Health and Safety
                       Harvard School of Public Health
                           665 Huntington Avenue
                        Boston, Massachusetts  02115
   SMOKING:
        1.  Do you smoke ?          Yes        No
        2.  If yes,   Cigarettes	  Cigars	  Pipe	
        3.  If cigarettes, how many packs do you smoke per day?
   ALCOHOL:
       1.  Do you drink alcoholic beverages ?   Yes
       2.  How much do you drink?
                 No	
               per day

               per week

               per month
   DRUGS AND MEDICATION:
       1.  Do you take any pills,  drugs, or medication frequently or
          regularly?    Yes	  No	
       2.  If yes, what kind?
                                        Signature
                                        Name (please print)
                                        Street Address
  Experimenter's Signature
City
State
Zip Code
  Date
                                        Tel. No.

-------
                                APPENDIX 5                                80
  Confidential

               Guggenheim Center for Aerospace Health and Safety
                         Harvard School of Public Health
                             665 Huntington Avenue
                          Boston,  Massachusetts 02115


              1 CONSENT FORM: CARBON  MONOXIDE STUDIES

  1.     I agree to participate in the study "The Effects of Low Levels of
        Carbon Monoxide on Driver RelatedVisual Tasks."  For this study
        I will be  present for six experimental sessions and receive
        $150.  00 for my participation.

 2.     I understand that  three of the sessions will be conducted in the
        laboratory at the  Harvard School of Public.Health, and that
        three  of the sessions will be  conducted while driving an auto-
        mobile on a test track.  During these sessions small amounts
        of blood will be withdrawn from me periodically.

 3.     I further  understand that I will inhale gas mixtures which will
        give carboxyhemoglobin levels of 6%, 11% and 17%,  and agree
        to remain with the experimenters for a period of time determined
        by the experimenters after each session.

 4.      I acknowledge that I have a valid driver's license and agree
        to have it in my possession at all times  during the study.

 5.      I further acknowledge that at  any time I may withdraw from
        this study upon notification to the experimenters.  I also ac-
        knowledge the right of the experimenters to terminate my
        participation at any time.
                                    Signature
                                   Name(please print)
                                   Street Address
Experimenter 's Signature          City          State      Zip Code


Date       ~              ~~      Tel. No.    ~

-------
            APPENDIX D



DATA ON UPTAKE OF CARBON MONOXIDE (CO)
                                                   81
NAME
L.D.
K.G.

D.H.
P. L.
B.M.
W.M.
M.M.
T.T.
J.T.
J.D.
R. F.
O.K.
DATE
May 6
May 11
May 19
May 4
May 6
May 13
May 7
May 12
May 17
May 18
May 26
-. June 1
May 18
May 26
June 1
May 11
May 20
May 27
May 12
May 24
May 13
May 20
May 27
May 6
May 17
May 20
Sept. 13
Sept. 15
Sept. 27
Sept. 15
Sept. 21
Inspired
CO (ppm)
300
715
715
700
300
720
300
715
720
510
715
690
510
720
715
715
510
715
715
720
315
720
720
300
720
720
720
720
720
720
720
Exposure
(mm. )
73
65
78
50
45
85
55
69
74
58
52
82
44
60
69
85
39
62
78
86
65
55
78
55
82
60
106
33
78
55
40
Final %
COIIb
5.2
9 S
16.5
8 7
5.3
15.7
6.1
14.6
16.2
7.7
10.6
16.2
7.2
13.3
17.3
13.3
8.2
15. 3
12. 1
18.6
6.4
12.9
19.9
6.8
16.9
12.8
9.9
11.6
15.6
14.9
11.0
Initial %
COHb
2.0
1.3
2.2
4.6
4.7
6.4
3.4
2.2
2. 1
1. 1
1.7
o.o
1.8
1.7
0.8
2.2
1. 3
0 0
1.3
1.2
2.0
3. 1
2.2
3. 0
1.4
1.3
0.7
5. 1
1.5
4.8
4.0
^ COHb
3.2
8.3
14.3
4. 1
0.6
9.3
2.7
12.4
14. 1
6.6
8.9
16.2
5.4
11.5
16.5
11. 1
6.9
15.3
10.8
17.4
4.4
9.8
17.7
13.8
15.5
11. 5
9.2
6.5
14. 1
10. 1
7.0
COHb/60
min
2.7
7.6
11.2
4.9
0.7
6.4
2.4
10.8
11.9
6.9
10.2
12.3
7.2
11.5
14.4
7.8
10.7
4.8
8.4
13. 1
4. 1
10.6
13.2
4. 1
11.3
11.5
6.4
12.0
10.9
11.0
10.5

-------
Page 2 - Data on Uptake of Carbon Monoxide (CO) APPENDIX D (continued)        8Z
                                                                       A
NAME
M.H.


J.C.
A.V.


W.F.

S.P.

A.C.

J.B.

H.A.

D. W.

J.L.

R.Y.

R. L.
W.F.

B.C.

J.D.

R. F.

DATE
Sept. 14*
Sept. 14
Sept. 17
Sept. 20
Oct. 1
Oct. 6
Oct. 13
Oct. 14
Oct. 21
Oct. 14
Nov. 5
Oct. 15
Oct. 19
Oct. 15
Oct. 19
Oct. 20
Oct. 26
Oct. 22
Oct. 26
Ocb. 29
Nov. 2
Oct. 29
Nov. 2
Oct. 28
Nov. 4
Nov. 12
Nov. 3
Nov. 9
Nov. 16
Nov. 18
Nov. 8
Nov. 15
Inspired
CO (ppm)
720
720*
720*
720
720
720
720
720
720
720
720
720
720
720
. 720
720
720
720
720
720
720
720
720
720
720
720
720
720
720
720
720
720
Exposure
(min. )
54
54
63
77
63
74 '
96
49
59
42
78
24
64
37
74
37
75
50
93
48
75
89
94
43
54
85
48
102
47
105
45
75
Final %
COHb
13.4
14.5
16.1
17.3
12.0
13.5
18.0
14.0
14.6
9.5
17.5
9.2
15.8
9.6
16.0
8.6
15.0
9.6
18.6
11.3
19.2
14. 1
17.7
11.5
11.7
16 9
8.4
17.0
7.9
17.7
11.8
20.2
Initial <5
COHb
0.7
0.7
0.8
2.9
0.3
1.3
1.3
0.9
1.4
0.8
0.4
3.0
2.5
3.5
3.0
0.2
0.3
0.8
1.0
0.6
0.6
0.4
0.2
3.0
0. 7
0.8
0.5
1.0
0.7
1.0
3.6
3.6
'o COHb
12.7
13.8
15.3
14.4
11.7
12.2
17.7
13. 1
13.2
8.7
17. 1
6.2
13.3
6.1
13.0
8.4
14.7
8.8
17.6
10.7
18.6
• 13.7
17.5
8.5
11.0
16.1
7.9
10.0
7.2
16.7
8.2
16.6
COHb/ 60
min
13.9
15.6
14.5
li.3
11.2
10.0
ID.,?
16. 1
13.4
12.7
13.3
14.3
12.5
9.8
10.7
13.4
11.7
10.6
11.5
13.2
14.9
9.2
11. 1
12.4
12.2
11.8
P. 7
9.5
9. 1
9.6
11. 1
12.6
              *Average of 2 slopes

-------
                                                     APPENDIX E

          Data on the decline  of COHb while breathing air and while breathing a mixture of 99% 0  and 1% CO-
                                  AIR
                                                                      OXYGEN + 1% CO,
Name

 L.DiB.


K.G.
same day
D.H.blood
    alv


P.L.
B.M.



W.M.



M.M.


T.T.



J.T.


J.D.D.
same day

B.F.


O.R.
% COHb
start
10.4
15.7
8.7
( 5.0
( 3.6
15.7
14.6
13.7
16.2
7.6
11.5
16.1
7.3
13.2
17.0
12.4
8.2
16.3
11.2
18.7
6.4
12.9
19.9
16.7
12.8
( 9.9
( 8.2
11.4

end
9.4
14.9
7.4
3.6
1.8
13.3
11.4
12.2
12.6
6.6
10.7
12.9
6.8
10.8
15.7
10.8
6.3
14.7
9.7
16.0
5.7
10.6
15.6
15.2
10.8
8.2
6.5
10.4

time
35 min
34
40
70
115
58
65
65
63
47
58
87
28
77
36
85
85
56
60
55
75
90
72
38
75
58
217
47
11.5
10.9
31

,A
1.0
0.8
1.3
1.4
1.8
2.4
3.2
1.5
3.6
1.0
0.8
3.2
0.5
2.4
1.3
1.6
1.9
1.6
1.5
2.7
0.7
2.3
4.3
1.5
2.0
1.7
1.7
1.0

0.6
% COHb
% start start
9.6%
5.0%
14 . 9%
28.0%
50%
15.2%
21.9%
10.9%
22.2%
13.1%
6.9%
19.8%
6.8%
18.1%
7.6%
12.9%
23.1%
9.8%
13.3%
14.4%
10.9%
17.8%
21.6%
8.9%
15.6%
17.1%
20.7%
8.7%

5.2%
9.4
14.9
7.4


13.3
11.4

12.6
7.0
10.7
12.9
6.8
10.8
15.7
10.8
6.3
15.8
9.7
16.0
5.7
10.6
15.6
15.2
10.8


10.4
15.6
10.9

end
6.1
7.6
6.0


8.3
6.9

6.7
4.4
5.5
5.9
4.3
5.7
5.8
7.5
4.0
5.6
7.0
10.7
3.6
4.6
3.5
7.8
5.4


6.0
6.5
4.5

time
35 min.
67
90


45
55

50
53
72
68
48
57
74
40
38
88
30
70
35
65
85
63
60


56
59
72
                                                                                       3.3
                                                                                       7.3
                                                                                       1.4
                                                                                       5.0

                                                                                       4.5

                                                                                       5.9
                                                                                       2.6
                                                                                       5.2
                                                                                       7.0
                                                                                       2.5
                                                                                       5.1
                                                                                       9.9

                                                                                       3.3
                                                                                       2.3
                                                                                      10.2
                                                                                       2.7
                                                                                       5.3
                                                                                       2.1
                                                                                       6.0
                                                                                      12.1

                                                                                       7.4
                                                                                       5.4
4.4
9.1

6.4
                                                                                    % start

                                                                                     35.1%
                                                                                     48.9%
                                                                                     18.9%
                                                                                     37.5%
                                                                                     39.4%

                                                                                     46.8%

                                                                                     37.1%
                                                                                     48.5%
                                                                                     54.2%

                                                                                     36.7%
                                                                                     47.2%
                                                                                     63.0%

                                                                                     30.5%
                                                                                     36.5%
                                                                                     64.5%
                                                                                     27.8%
                                                                                     33.1%
                                                                                     36.8%
                                                                                     56.6%
                                                                                     77.5%
                                                                                     48.6%
                                                                                       50%
42.3%
58.3%

58.7%
                                                                                                       oo
                                                                                                       UJ

-------
                                               APPENDIX E(continued)

          Data on the decline of COHb while breathing air and while breathing a mixture of 99% 0?  and  1%  CO.
Name

M.H


J.C.

A.V.
G.F.blood
    alv
    alv

S.P.alv.
    alv.

A.C.alv.
    alv.

J.B.
H.A.alv.

D.W.


J.L.


R.Y.
    alv.
W.F.


R.L.

B.C.


J.D.
    1

R.F.
blood
%
start
14.6
16.7
16.5
12.0
19.0
14.0
13.5
17.2
12.9
19.0
11.2
12.1
17.7
20.4
12.4
19.2
15.3
20.0
11.2
12.7
17.9
12.0
17.7
12.6
20.2
COHb
end
12.0
15.8
15.4
11.0
16.7
10.2
11.6
15.7
11.3
16.8
10.6
9.6
15.6
16.8
10.4
15.0
12.5
15.7
10.6
11.7
16.7
11.4
15.3
9.9
16.3
AIR
time
97 min
28
32 <•
42
60
82
82
35
38
24
27
67
45
67
42
65
60
50
24
34
32
33
67
60
60
                                                                                    OXYGEN + 1% CO,
                                        2.6
                                        0.9

                                        1.1

                                        1.0

                                        2.3
                                        3.8
                                        1.9
                                        1.5

                                        1.6
                                        2.2

                                        0.6


                                        2.5


                                        2.1


                                        3.6

                                        2.0
                                        4.2

                                        2.8
                                        4.3

                                        0.6
1.0
1.2
0.6
2.4

2.7
2.9
% start
17.8%
5.3%
6.6%
8.3%
12.1%
27.1%
14.0%
8.7%
12.4%
11.5%
5.3%
20.6%
11.8%
17.6%
16.1%
21.8%
18.3%
21.5%
5.3%

7.8%
6.7%
5.0%
13.5%
21.4%
14.3%
% COHb
start
12.0
15.8
15.4
11.5
15.9
16.7
10.2
11.6
15.7
11.3
16.8
10.6
12.8
9.6
15.0
15.6
8.9
16.8
11.7
15.0
12.5
15.7
10.6
16.8
12.8
11.7
16.7
11.4
15.3
9.9
16.3
end
4.6
5.7
6.3
6.0
4.5
4.3
4.4
4.7
5.4
5.1
5.0
4.3
4.9
3.7
8.1
3.6
3.6
5.8
5.0
5.0
5.3
5.1
5.4
6.9
5.0
5.0
5.1
5.1
4.8
3.9
5.5
time
80
85
76
65
87
110
67
67
78
54
80
73
100
60
55
100
70
70
70
77
65
80
57
70
87
67
108
75
100
84
65
 7.4
10.1
 9.1
 5.5
11.4
12.4
 5.8
 6.9
10.3
 6.2
11.8
 6.3
 7.9
 5.9
 6.9
12.0
 5.3
11.0
 6.7
10.0
 7.2
10.6
 5.2
 9.9
 7.8
 6.7
11.6
 6.3
10.5
 6.0
10.8
% start

 61.6%
 63.9%

 59.0%

 47.8%
 71.6%
 7 It.2%

 56.8%
 59.4%
 65.6%

 54.8%
 70.2%

 59.4%
 61.7%
 61.4%
 46.0%

 76.9%
 59.5%
 65.4%
 57.2%
 66.6%

 57.6%
 67.5%

 49.0%
 58.9%

 60.9%
 57.2%
 69.4%

 55.2%
 68.6%

 60.6%
 66.2%
                                                                                                                 00

-------
                                                                    85
                        APPENDIX F
                    Vehicle Specifications
Basic Vehicle Code
     PH43
Paint
     EW1
Fury HI
White
Options
     A01
     A 04
     B42
     C14
     C92
     D34
     D91
     E61
     F13
     F25
     F95
     G21
     G31
     H31
     H51
     J21
     J24
     J55
     N25
     N51
     N88
     P25
     P41
     R26
     R32
     R33
     S16
     S25
     S62
     W25
Light Package
Basic Group
Power Disc Brakes,  Special
Shoulder Belts, Rear
Accessory Floor Mats, Color Keyed
Torqueflite
Sure Grip Differential
383 Cu. In. 8Cyl. 2 BBL
60 Amp.  Alternator
70 Amp.  Battery
Calibrated Speedometer
Clear Glass
Manual Mirror, Right
Rear Window Defogger
Air Conditioning
Electric  Clock
Headlights Washer & Wiper
Undercoating W/Hood Insulator Pad
Engine Block Heater
Max.  Cooling Pkg.
Automatic Speed Control
Power Bucket or 50/50 Bench Seat, Left
Power Door Locks
Radio-AM W/Remote Stereo Tape
Dual Rear Seat Speakers
Microphone
HD Suspension
HD Shocks
Tilt Steering Wheel W/Rim Blow
HD Wheels

-------
                                                                          86
                             APPENDIX G

     Preliminary Recorded Instructions for Subjects in Driving Tests
      Following is a copy of the exact wording of the cassette recorded
instructions given to all subjects in the driving experiments before their first
training session.  This is followed by more detailed explanations and demon-
strations by the experimeter:

      "As you know, you will be participating in a driving study where we are
interested in your performance on an interstate type highway,  and whether
the small amounts of carbon monoxide which you breathe have any effect.
Certainly,  large effects will not be expected to occur.  However,  under all
circumstances the primary rule will be safety first. Always follow our
directives when behind the wheel.  There will be three driving  sessions, each
on a different day.

      During all driving runs,  you will be  required  to drive to the same level
of accuracy.   In normal highway  driving, you maintain the car  within the
marked lanes,  only touching the lines when changing lanes.  During the exper-
iment, we will ask you to always remain within the  designated lane, never
touching the edges.  In fact, when  such a crossing occurs, we will  recommence
that particular run.

      Now on all the experimental runs,  you will be wearing this  experimental
helmet which will be used to obscure your view of the road (periodically).  You
can raise  the visor by pressing the footswitch under your  left foot.  The
observer has a safety  switch which will raise the visor at any  sign  of trouble
and also you can raise the visor at any time with one hand.  In  normal operation,
when you depress the footswitch the visor will be raised for a half-second, and
we know that this will  give you enough time to get a  good view of the road ahead.
In order to get a good  glimpse of the road, you should maintain your vision
looking  ahead and as far as possible focussed at a long distance even when the
visor is closed.  Do not look just ahead of the car,  or to either side.   Use two
hands on the steering wheel in your normal driving  position.

      When you are driving with the visor  closed, you will begin to be  more and
more uncertain of the  car's position.  You should develop the technique of only
allowing uncertainty to a level where the car can be kept within the marked lane.
You can then "refresh" your view of the road by taking another  glimpse. Dif-
ferent speeds  should require different rates of viewing so you should change
your viewing rate to match the conditions of the particular run.  Try to avoid
sudden movements of the steering wheel in order to stay within the lane by
setting up a smooth rhythm of performance.  Remember you are  in no sense
competing with other subjects; we want you to do as well as  you can and yet
yield consistent, non-erratic performance.

      Any questions ?"

-------
                               Bibliography
Bills, A. G. , 1937.  Blocking in mental fatigue and anoxemia compared.
   Amer. J.  Psychol. ,  20:  437-452.

Broadbent, D. E. ,  1958.  Perception and Communication.  London: Pergamon
   Press.

Forbes,  W. H. , 1970.  Carbon  monoxide uptake via the lungs.  Ann. N. Y. Acad.
   Sci..  174(1):  72-75.

Halperm, M. H. ,  McFarland, R. A. , Niven,  J.I. and Roughton, F. J.W.,  1959.
   The time course of the effects of carbon monoxide on visual thresholds.
   J.  Physiol. , 146(3):  583-593.

Lawther,  P. J.  and Commings, B.T.,  1970.   Cigarette smoking and exposure
   to  carbon monoxide.  Ann. N. Y.  Acad.  Sci. ,  174(1): 135-147.

McFarland, R. A. , 1946.  Carbon Monoxide and Other Noxious Gases.  In:
   Human Factors in Air Transport Design.  New York: McGraw-Hill.
   Ch. 6, pp.  209-251.

McFarland,  R.A. ,  1963.  Experimental evidence of the relationship between
   ageing and oxygen want: in search of a theory of ageing.  Ergonomics,  6(4):
   339-366.

McFarland, R.A. , 1970.  The  effects  of exposure to small quantities  of carbon
   monoxide on vision.  Ann. N. Y. Acad.  Sci. ,  174(1):  301-312.

McFarland, R.A.  and Associates. ,  1972.  Publications in the Field of Highway
   Safety,  1950-1971.  Harvard School of Public Health, Boston, Mass.

McFarland, R.A. , Evans, J. N. and Halperin, M. H. ,  1941.  Ophthalmic
   aspects of acute oxygen deficiency.   Arch. Ophthal. , Chicago, 26:  886-913.

McFarland, R. A. , Roughton, F. J. W., Halperin, M. H. and Niven, J. I. ,  1944.
   The effects of carbon monoxide and altitude on visual thresholds.   J. Aviat.
   Med. . 15(6): 381-394.

Miranda, J. M. , Konopinski, V. J.  and Larsen, R. I. , 1967.  Carbon monoxide
   control in a high highway  tunnel. Arch, environ.  Hlth,  15:  16-25.

Platt, F. N. , 1963.  A new method of measuring the effects of continued driving
   performance.   In: Driver Characteristics,  Night Visibility, and Driving
   Simulation.  Highway Research Record #25, Highway Research Board,
   Washington, D. C.

Schulte,  J. H. ,  1963. Effects of mild  carbon monoxide intoxication.  Arch.
   environ. Hlth,  7(5):  524-530.

-------
                                                                          88
Senders, J. W.  and Dietrich, C. W. ,  1969.  Additional studies of driver
   information  processing.  Final Report.  Federal Highway Administration,
   U.S. Dept. of Transportation.

Senders, J. W. , Levison, W. H. and Dietrich, C.W., 1967.  The attentional
   demands of automobile driving.  Highway Research Record No.  195,
   National Academy of Sciences,  Washington, D. C.

Stewart, R. W. , 1972.   Determination of carboxyhemoglobin in various segments
   of the population.  Project CAPM-8-68,  Medical College of Wisconsin,
   Milwaukee, Wisconsin.

Teichner, W.H. ,  1968.  Response blocking: a necessary performance criterion.
   Pp. 166-180.  In:  Pre-crash Factors  in Traffic  Safety,  12th Annual
   Symposium,  G. G.  Snively, (Ed.), Sacramento,  Calif., American Association
   for Automotive Medicine.

-------
                                                                                         89
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1  REPORT NO.
   EPA-650/1-74-006
                                                            3. RECIPIENT'S ACCESSION-NO.
 4. TITLE AND SUBTITLE
   A study of the effects of  low-levels of carbon
   monoxide upon humans performing driving tasks
              6. PERFORMING ORGANIZATION CODE
 7 AUTHOR(S)
   R.  McFarland, W. Forbes, H.  Stoudt, J. Dougherty,
   T.  Crowley, R. Moore, T. Nalwalk
                                                            8. PERFORMING ORGANIZATION REPORT NO
 9. PERFORMING ORGANIZATION NAME AND ADORES
   Guggenheim Center for Aerospace  Health & Safety
   Harvard School of Public Health
   665  Huntington Avenue
   Boston. Massachusetts  02115
              10. PROGRAM ELEMENT NO.

                1AA005
              11. CONTRACT/GRANT NO.

               CPA 70-134
  2 SPONSORING AGENCY NAME AND ADDRESS
  Coordinating Research Council,  Inc.
  New  York,  N.Y.   10020
  U. S.  Environmental Protection  Agency
  Research Triangle Park. N.C.  27711
              13..
                                         = RED
   and
              14. SPONSORING AGENCY CODE
 15. SUPPLEMENTARY NOTES
                                                •
       Subjects  were exposed to  low  levels  (700 ppm) of carbon monoxide (CO) until
  carboxyhemoglobin (COHb) levels of 6%,  11%,  and 17% were reached,  and they were then
  tested  as  to their ability to  perform both selected driving-related  laboratory tests
  of visual  response and control reactions  and over-the-road vehicle driving.   These
  test results were then compared with  those on the same subjects  taken under control
  conditions without exposure to CO.

       The overall  pattern of results indicates that 6% COHb level had  no effect on
  driving ability,  and that COHb levels of  11% and 17% did not appear to seriously
  affect the ability to drive motor  vehicles,  as measured by the tests  administered
  in this study.  However, certain statistically significant differences  were  found
  in some of the  tests and suggest some decrement in performance as  a result of CO
  exposure.
 7.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b IDENTIFIERS/OPEN ENDED TERMS C  COSATI Held/Group
       Carbon Monoxide
       Carboxyhemoglobin
       Driving
   Carbon  Monoxide
   Carboxyhemoglobin
   Driving
 8 DISTRIBUTION STATEMENT
       general
       unlimited
19 SECURITY CLASS (This Report/
  unclassified
21 NO OF PAGES
   104
                                              20 SECURITY CLA.SS/TVliJ pare)
                                                unclassified
                           22 PRICE
EPA Form 2220-1 (9-73)

-------