BEHAVIORAL TOXICOLOGY:
AN EMERGING DISCIPLINE
EPA-600/9-77-042
December 1077
I"
I -
Health Effects Research Laboratory
Office of Research and Deueiopment
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
-------
EPA-600/9-77-042
December 1977
BEHAVIORAL TOXICOLOGY:
AN EMERGING DISCIPLINE
Proceedings of the Southwest
Psychological Association Annual Meeting
April 30, 1976
Albuquerque, New Mexico
Editors:
Harold Zenick
Department of Psychology
New Mexico Highlands University
Las Vegas, New Mexico 87701
and
Lawrence W. Reiter
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Health Effects Research Laboratory
Research Triangle Park, North Carolina 27711
-------
This report has been reviewed by the Health Effect Research Labora-
tory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
FOREWORD
Cerebration is among those characteristics that most decisively distinguish
humans from other animals. The study of higher thought processes has in-
trigued researchers for centuries, but conceptual, as well as methodological
problems, have made progress frustratingly slow. Recent years, however, have
seen a variety of breakthroughs in areas such as neurochemistry as well as the
application of information theory to artificial intelligence. There has also
emerged a growing concern for the application of behavioral research to socie-
tal problems. It seems natural then, that one of our recent, most pressing socie-
tal problems—the reconciliation of man's interaction with his environment-
has become an important driving force in neurobehavioral research. In effect,
the title of this symposium, "Behavioral Toxicology: An Emerging Discipline"
underscores the somewhat embryonic nature of this area of research.
Much environmental toxicology (appropriately) focuses on relatively gross
changes in biochemistry, morphology, or physiology. It is equally appropriate,
however, for the behavioral toxicologists to concern themselves with the subtle
alterations environmental agents may exert on man's capacity to cope with
his...environment.
Environmental health effects research has many inherent difficulties. Behav-
ioral research has many inherent difficulties. The juxtaposition of these two
disciplines results inescapably in a multiplying effect of the research problems
encountered. I have encouraged and supported my colleagues in environmental
neurobehavioral research over the years—but from a distance. Thus, it is per-
sonally gratifying to me to be able to recognize in a more explicit fashion the
progress they are making. This symposium represents one more important step
forward in the consolidation, coordination and promotion of the emerging
discipline of behavioral toxicology.
John H. Knelson, M.D.
Director
Health Effects Research Laboratory
Research Triangle Park, N.C. 27711
-------
PREFACE
This book contains the proceedings of a symposium entitled "Behavioral Toxi-
cology: An Emerging Discipline" held in conjunction with the Southwest
Psychological Association meetings, April 30, 1976, in Albuquerque, New
Mexico. Authors of formal presentations later reviewed and enlarged their
contributions. In emphasizing the interdisciplinary nature of this area, the
participants were drawn from a number of different disciplines including toxi-
cology, pharmacology, psychology, physiology, and veterinary medicine. The
audience that this publication addresses is not only contemporaries in the field
but also those researchers in behavioral toxicology. Thus the intent of this
book is to provide a "state of the art" overview of the area. We have attempted
to do this by incorporating chapters that discuss some of the most fundamental
aspects such as sources of funding (Chapter 1), and basic methodological and
experimental considerations (Chapters 2, 3, 4). In addition, there are chapters
that address specific areas of investigation including the utilization of tech-
niques to assess the interaction of the animal with its environment (Chapters 5,
6, 7), considerations ins sensory toxicology (Chapter 8), and employment of
electrophysiological techniques (Chapters 9, 10).
IV
-------
CONTENTS
1. John A. Santo/ucito
Overview of Federal Funding Avenues for Support of Neurotoxicology
2. W. B. Buck, D. L Hopper, W. L. Cunningham, and G. G. Karas
Current Experimental Considerations and Future Perspectives in Behav-
ioral Toxicology
3. Harold Zenick
A Review of the Developmental Models Employed in Behavioral
Toxicology
4. Robert L. Bornschein and I. Arthur Michaelson
Methodological Problems Associated with the Exposure of Neonatal
Rodents to Lead
5. Stata Norton
Observational Techniques in Behavioral Toxicology
6. Lawrence Reiter, George Anderson, Miriam Ash, and L. Earl Gray
Locomotor Activity Measurements in Behavioral Toxicology:Effects of
Lead Administration on Residential Maze Behavior
7. Phyllis Mullenix
Altered Behavioral Patterning in Rats Postnatally Exposed to Lead: The
Use of Time-Lapse Photographic Analysis
8. Hugh L. Evans
Behavioral Assessment in Sensory Toxicology
9. Dorothy E. Woolley
Electrophysiological Techniques in Toxicology
10. David Otto
Neurobehavioral Toxicology: Problems and Methods in Human Research
-------
OVERVIEW OF FEDERAL FUNDING
AVENUES FOR SUPPORT OF
NEUROTOXICOLOGY
JOHN A. SANTOLUCITO
Environmental Monitoring and
Support Laboratory, Las Vegas
U.S. Environmental Protection Agency
Although this symposium is entitled "Behavioral Toxicology," I believe it is
acceptable, if not desirable, for me to address the general subject of neuro-
toxicology. Simply defined, neurotoxicology is the systematic study of
unwanted effects of toxic agents on the nervous system. Recent interest in
neurotoxicology is largely due to the growing concern over the number of
toxic agents to which we are exposed in our highly industrialized society. The
number of synthetic materials being produced, as well as the large amounts of
industrial and municipal waste products being introduced into the environ-
ment, have all added significantly to the urgency of neurotoxicologic research.
In order to understand adequately the biological effects of environmental
insults, it is necessary to consider the potential of the insult to impair func-
tioning of the nervous system by direct or indirect alteration of neuronal
physico-chemical states. Such impairment may appear, for example, as changes
in the mental ability of an individual to react, remember, or learn; as changes
in emotional states, such as anxiety or depression; or as changes in neuroendo-
crine function. These changes may result in decreased capability to cope with
everyday living conditions, with a subsequent reduction in the "quality" of
life. Health effects such as these, which fall short of morbidity and mortality,
become increasingly important as the requirements for our society become
more complex and demanding.
It is not necessary to stress further the importance of neurotoxicology in the
overall assessment of responses to toxicologic agents. However, as a discipline,
it is in a relatively early developmental stage, having received little formal
recognition apart from that given to specific occupational risks prior to 1969
(Weiss and Laties, 1969).
1-1
-------
The recent establishment of neurotoxicology as a discipline is attracting the
interest of a new cadre of investigators seeking assurance that the Federal
Government is interested in and supportive of the discipline. A review of in-
house research and extramural support for the past decade provides this
assurance. All Federal agencies having responsibility in the field of toxicology
have recognized neurotoxicology as an important investigatory discipline
since the middle 1960's. While the interest initially focused on the direct
effects of toxic agents on the central nervous system (CNS), it later encom-
passed the use of behavioral and/or neurophysiologic indicators of toxicity in
which the CNS may not be the primary target (Colleti and Krueger, 197.0).
The present diversity of interest includes physical contaminants (e.g., noise and
microwave) as well as chemical contaminants in the environment and psycho-
logic as well as physiologic effects (Berglund eta/., 1974).
The problem faced by Federal agencies in evaluating the neurotoxicologic
effects of numerous insults, given that the CNS integrates and is reactive to the
status of the whole organism, is the lack of definitive equations relating alter-
ations in neurophysiologic indices, including behavior, to life span, health, pro-
ductivity, or well-being. For this reason, available funds for support of neuro-
toxicologic research are modest.despite the recognition of its importance. It is
also worth noting that in times of a tighter economy many agencies tend to
shift extramural support disproportionately in favor of contracts over grants.
That is to say, agencies use available funds to solve specific problems rather
than to expand a general information base having no immediate payoff. This
tendency should cause no great concern to those seeking research support.
It simply necessitates formulating specific objectives which meet the specific
research needs of the agencies. Grant monies are not unavailable. Rather, they
may be relatively more difficult to obtain.
Contracts may be competitive or noncompetitive, solicited or unsolicited.
Federal Procurement Regulation states, "All purchases and contracts shall be
made on a competive basis to the maximum possible extent." Because indivi-
dual agencies' needs will vary, so will their interpretation of maximum possible
extent. All solicited competitive contracts are initiated by an agency in the
form of a request for proposal (RFP) published in the Commerce Business
Daily* along with a closing date by which proposals must be received and the
name and address of the requesting agency. Unsolicited contracts are consi-
dered when they represent original effort by the offerer in the form of new
and unique ideas.
It is therefore apparent that the research needs of an agency and the research
capabilities of an investigator must be brought together. Research needs and
information about funding mechanisms can be obtained from each agency's
office or division responsible for extramural research or grants and contracts.
For example, a publication entitled "How to Do Business With EPA: A Guide
to Requirements for Grants and Contracts," is available from Environews, Inc.,
'Available from the Superintendent of Documents, Government Printing Office, Wash.,
D. C. 20402.
1-2
-------
This symposium discusses some of the testing methodologies currently used in
neurotoxicology, including behavioral, neurochemical, and electrophysiologic
techniques. One of the primary goals of neurotoxicology is to provide data
for use in the evaluation of potential health hazards associated with exposure
to various environmental pollutants. A thorough understanding of the methods
employed and of the properties being measured is critical to achieving this
objective. Sources of variability, control measures, and validation procedures
associated with the testing methodologies must be known. Baseline perfor-
mance or response of the test animal used and confirmation of the composi-
tion or purity of the toxic agent being tested are equally important. Ob-
viously, the more reliable and reproducible the test system, the more we
understand the biological effects of toxins and the better equipped we are to
define the risks involved in exposure. Quality assurance in biological research
is a matter of prime concern to Federal agencies.
Neurotoxicology is a young and vigorous discipline. While in its infancy, it is
being challenged to provide answers to complex and multifaceted real-world
problems. This symposium was convened to address some of these problems
and to assess current progress in several areas of neurotoxicology related to
environmental pollutants. The planning and implementation of this sympo-
sium, the contents of its timely papers, and the expertise in attendance afford
confidence that neurotoxicology will meet these challenges.
ADDENDUM
I. National Institutes of Health
Office of Extramural Research Support
Bethesda, Maryland 20014
Some of the Institutes of NIH will be more likely to have interest in neuro-
toxicology than others. For example: National Institute of Environmental
Health Sciences; National Institute of Mental Health; National Institute of
Child Health and Human Development. If it seems probable that a research
capability would be of interest to one of the Institutes in particular, it would
be advisable to so indicate.
II. National Institute of Occupational Safety and Health
Office of Extramural Research Support
5600 Fishers Lane
Rockville, Maryland 20852
III. Environmental Protection Agency
In EPA there are two Health Effects Research Laboratories with interest in
behavioral toxicology. They receive unsolicited proposals through the Grants
Administration Division, Environmental Protection Agency, Washington,
D.C. 20460.
IV. Food and Drug Administration
1-3
-------
Each of the units receives proposals directly as follows:
For environmental pollutants
For food additives/contaminants
For therapeutic drugs
For diagnostic agents
For veterinary products
Assistant Director for
Scientific Coordination
National Center for Toxicological
Research
Jefferson, Arkansas 72079
Director, Bureau of Foods
200 C St., S.W.
Washington, D. C. 20204
Director, Bureau of Drugs
200 C St., S.W.
Washington, D. C. 20204
Director, Bureau of Biologies
200 C St., S.W.
Washington, D. C. 20204
Director, Bureau of
Veterinary Medicine
200 C St., S.W.
Washington, D. C. 20204
V. National Aeronautics and Space Administration
Director, Office of University Affairs (Code P)
or
Director, Office of Life Sciences (Code SB)
NASA Headquarters
Washington, D. C. 20547
References
Berglund, B., Berglund, U., and Lindvall, T. (1974). Scaling of annoyance in
epidemiologic studies. In Recent Advances in the Assessment of the Health
Effects of Environmental Pollution: International Symposium Proceedings.
Commission of the European Communities, Luxembourg.
Colleti, R. B. and Krueger, K. K. (1970). Proceedings of the Workshop on
Behavioral Assays in Uremia. U.S. DHEW Publication No. (NIH) 72-37.
Weiss, B. and Laties, V. G. (1969). Behavioral pharmacology and toxicology.
Annu. Rev. Pharmacol. 9, 297-326.
1-4
-------
2. CURRENT EXPERIMENTAL
CONSIDERATIONS AND FUTURE
PERSPECTIVES IN BEHAVIORAL
TOXICOLOGY
W. B. BUCK, D. L HOPPER, W. L CUNNINGHAM, AND
G. G. KARAS
Behavioral Toxicology Laboratory
Veterinary Diagnostic Laboratory
College of Veterinary Medicine
Iowa State University
Introductory Comments
Any discussion of behavioral toxicology generally begins with an attempt to
put into words a concept which, as yet, has not been fully developed. This
"definition" problem manifests itself as an "identity" problem, not only
among ourselves but also in the eyes of our colleagues in other disciplines. The
urge to define comes from a feeling that we are well past puberty and repre-
sent a mature field of endeavor. However, considering that our evolutionary
process has only recently begun (we did not actually agree on what to call
ourselves until the late 1960"s), a final definition may be premature.
Rather than attempt a definition, we propose to discuss the scope and multi-
disciplinary aspects of behavioral toxicology. Whether one considers behavioral
toxicology as another approach for the study of the effects of toxicants on an
organism or as a means of studying animal and human behavior as affected by
neurotoxicants, we can surely agree that it is a field of endeavor that requires
the expertise of many disciplines.
It is unlikely that a single researcher can be found who possesses the many and
varied skills needed for the conduct of behavioral toxicology studies. There-
fore, studies using the behavioral toxicologic approach usually involve a multi-
disciplinary team effort. The various disciplines that are represented on this
*This work was supported by Environmental Protection Agency Contract 68-02-2288 to
Dr. William B. Buck, Director, Behavioral Toxicology Laboratory, Iowa State University,
Ames, Iowa 50010.
2-1
-------
team will be determined by the availability of individuals with a specific exper-
tise and the objectives of the research. Listed below are some of the disciplines
that may be found in a behavioral toxicology research program:
1. psychology
2. toxicology
3. veterinary and/or human medicine
4. chemistry
5. physiology
6. pathology
7. biophysics
8. biomedical engineering
9. computer science
10. statistics
The behavioral toxicologist strives to devise sensitive techniques to evaluate the
possible consequences of exposure to new and untested chemicals as measured
by the total integrated function of the organism being tested. Those toxicants
having direct effects on the nervous system, whether they be sensory, motor,
or cognitive, are more likely to be studied in behavioral toxicology. However,
toxicants which have their primary effect on other organ systems, such as the
liver and kidney, may also be sensitively assessed using the multidisciplinary
behavioral toxicologic approach. Perhaps one can summarize the goals of the
behavioral toxicologist as: (1) the utilization of behavior for the assessment of
the effects of toxicants, (2) the correlation of behavioral changes with concom-
itant physiopathologic alterations (overt clinical signs, pathologic changes,
electrophysiologic and biochemical alteration, and death), and (3) the assess-
ment of the effects of toxicants on behavior (sensory, motor, and cognitive).
Behavior as an Index of Toxic Effects
One need only make a cursory review of the literature over the past 10 years
or scan the references of some of the more recent behavioral toxicology review
papers (Weiss and Laties, 1975; Bignami, 1976) to appreciate the amount of
scientific research that has accepted behavior as a viable index for under-
standing the effects of toxic substances. The increased use of behavior should
give gratification to those early researchers who felt that this was a missing
dimension in experimental environmental toxicology (Puffin, 1963;
Brimblecombe, I968).
A majority of behavioral toxicologic research has dealt, for the most part,
specifically and directly with behaviorally measuring the effects of exposure
to toxic compounds. This "direct effect" type of research (the examination of
the effects on behavior only) has been and remains important for the following
reasons. First, it. has helped in understanding the kind and the magnitude of
behavioral effects that may result from exposure to toxic substances, effects
that would not be apparent from an examination of only neuroenzyme or
structural changes. Investigations from our laboratory over the past 10 years
which have involved the study of the effects of exposure to lead (Van Gelder,
et a/., 1973; Carson, et a/., 1973; Carson, et al., 1974), chlorinated hydrocar-
bons (Sandier, et al,, 1969; Smith, et al., 1976) and organophosphates (Reischl,
2-2
-------
Van Gelder, and Karas, 1975; Van Gelder, 1975), provide examples of this
type of research. Within the confines of these experiments, we have been able
to examine the behavioral functioning of an organism as it has been affected
by a specific toxicant.
Second, this type of investigation has helped to provide a research foundation
which gives the discipline credibility. By demonstrating that behavior can be
affected by toxic substances and that it can be as sensitive as other toxicologic
indices, we have established not only the credibility of behavior as a measure-
ment tool, but also the viability and reliability of the discipline as a competent
information provider.
However, it will probably be through an "indirect" research approach that
behavior will have its greatest utility and value as a research tool. That is, when
we examine behavior in conjunction with, rather than in lieu of other toxi-
cologic methods, we will be moving towards a better understanding of the
effects of toxic compounds. That is not to lessen our roles either as behavior-
ists or toxicologists. Just as one cannot hope to get a complete image of the
effects of a toxic substance or the overall functioning of an organism by
looking at enzyme levels, tissue levels, or an EEG alone, similarly, one cannot
hope to get a complete picture by looking solely at behavior. The multi-
disciplinary approach, by combining people and ideas from different disci-
plines, encourages attempts at correlating behavior with other indices of
toxicosis in order to obtain a more complete picture of the total functioning of
an organism. The behavioral toxicologist, or behavioral toxicology research
team, then, has an important place in terms of the overall understanding of
environmental toxicology.
Clinical Observations as a Source of Research Hypotheses
Clinical observations are perhaps the most direct and the oldest source of
information available to the toxicologist concerning the effects of a toxic
agent on an organism. Such observations can provide not only valuable infor-
mation as to the source and treatment of a toxicosis, but also important
clues in the continuing effort to delineate toxic effects, and they are, there-
fore, important to the formulation of viable research hypotheses.
One of the most apparent and readily determined signs of illness of toxicosis
is a change in the reflexive responses of the organism. Alterations in eye blink,
startle response, clasping, righting, or withdrawal, for example, give good
indications as to the source, extent, and severity of toxicoses. Knowledge of
reflexive response changes can aid in isolating toxic effects to various body
systems (CNS, vasculature, sensory organs, motor system, and musculature).
Many toxic agents have effects which are not associated with overt clinical
signs of which are manifested in situations or in organisms that do not lend
themselves to controlled experimentation. In such situations, the effects of
toxic agents on naturally occuring responses (although of interest in their
own right) can be a particularly useful resource for research hypotheses. By
"natural responses" we mean behavioral outputs characteristic of an organism
which require no experimenter training and which occur with a minimum of
2-3
-------
learning or experience on the part of the organism. Examples of such natural
responses include herding behavior, many of the stereotypic mating responses,
characteristic social behaviors, dominance hierarchies, orientation (homing,
spawning), and territorial behaviors. All of these behaviors are readily observ-
able and subject to toxicologic insult. The potential behavioral hypotheses are
innumerable.
Finally, a distinction must be made between clinical "signs" and "symptoms."
Clinical signs, such as changes in blood pressure or respiration, may appear with
a toxicosis in both humans and animals. However, symptoms are available to
the toxicologist only from human subjects and are merely the subjective report
given to the toxicologist by the subject. By extrapolation, symptoms reported
by human subjects may provide important clues to toxicoses and their effects
in animals; whereas, the reverse is obviously not available as a resource for
research hypotheses from animals.
Research Paradigms and Measures
Much of the research in behavioral toxicology has been paradigm oriented,
resulting directly from an effort to find behavioral paradigms which are both
sensitive to toxicologic insult and able to provide answers to questions of
specific behavioral interest.
Activity Tests
A favorite measurement of the psychopharmacologist and behavioral toxicolo-
gist has been the general activity of the organism. The term "activity" has been
most often used in reference to general motor activity. The method used to
obtain the measure is important in the use of activity measures. General activ-
ity can be measured in revolving wheels, stabilimeters, or tilt cages, by photo-
electric means, by ultrasonic or resonant circuits, or by direct observation.
Each of these methods of obtaining activity levels is subject to a number of
variables and may be differentially affected by each. The interpretation of
treatment effects due to a toxic insult can be most meaningfully understood if
stable pretreatment baselines are established, if the animal is properly habit-
uated to the activity apparatus, and if extraneous variables such as handling,
light, noise, health, age, sex, motivation, diet, population density, strain, and
species are controlled. For discussion of this topic see Finger (1972).
Although many activity apparatuses have the advantage of simplicity, they
cannot provide the wealth of information available through direct observation.
This method, when used properly by taking precautions against investigator
bias, avoiding subject-experimenter interaction, and insuring standards of be-
havioral classification, can provide activity information directly relevant to a
number of potential research hypotheses. This is not possible with activity
measuring devices. Norton (1973) has recently introduced a procedure, which
has subsequently been modified and automated in our laboratory, whereby
some of the sources of error and difficulties associated with employing the
method of direct observation can be reduced. This method of activity deter-
mination is called video pattern recognition. It employs a computer algorithm
2-4
-------
to determine the occurrence of experimenter-defined behavioral acts. Obser-
vation is via a movie camera input to a computer system. This methodology
has great potential both in objectifying a classically used observation method
and in providing a means of carefully examining aspects of a variety of behav-
ioral activities which have been previously difficult or impossible to obtain.
This method is described in greater detail by Norton in Chapter 5 of this study,
while application of a macroanalytic technique is discussed by Reiter (Chapter
6) and microsampling by Mullenix (Chapter 7).
Developmental Testing
Evidence that the immature organism is more likely to be affected by a number
of toxic agents is fairly well accepted among toxicologists. This research has
established that changes seen in an organism may be specific to one period of
the life span, thus making the time of exposure a critical variable in delineating
a toxic effect. In Chapter 3, Zenick reviews various developmental models that
have been employed in behavioral toxicology. The nature of a toxic insult
which occurs at one or more periods in the development of the organism and
manifests itself in another can best be detected by the use of a longitudinal
experimental design. The long-term assessment of the effects of toxicants on
animals or humans has only recently begun to be exploited, and we need a
great deal of research in this area if we are to understand the interaction of the
developmental processes with other variables of interest to the behavioral
toxicologist. A recent example of this research approach is discussed by Spyker
(1975).
Appetitive and Aversive Conditioning
Behavioral toxicologists have employed a broad range of classical, instrumental,
and operant paradigms in their efforts to elucidate the effects of toxic agents
on behavior. It is beyond the scope of this paper to discuss the nuances of the
many paradigms which have been employed; however, we feel that some com-
ments abouts the selection and use of behavioral paradigms are worthwhile. Of
primary importance to the selection of a behavioral paradigm are the realiza-
tions that behavior is not a single, simple entity and that failure to find a toxi-
cologic effect on behavior as measured under one set of circumstances with one
paradigm does not preclude effects with other paradigms under similar or even
dissimilar circumstances. The rule of behavior is its complexity and variability;
it is a function of antecedent conditions, current stimulus circumstances, the
properties of the behavior act itself, and the consequences of the act.
The constraints on the selection of a behavioral paradigm are many: the hypo-
theses to be tested, the subjects utilized, the effects sought (whether chronic
or acute), the nature of the toxicant, and the availability of equipment and
funds are only a few of the important factors in the selection of a paradigm.
The best tool available to the toxicologist to aid in the selection process is a
thorough evaluation of the research literature. It is hard to imagine a greater
frustration than discovering that the paradigm one has selected and in which
one has invested many hours and thousands of dollars has been shown by
colleagues to be inappropriate on the basis of a number of criteria.
2-5
-------
We have found in our laboratory that when studying the chronic effects of
various toxicants, it is strongly advisable to begin the experiment with as many
subjects as reasonably possible. For example, the need for a large number of
subjects is particularly true for between-group designs employing highly toxic
agents with potentially cumulative effects. A final admonition in the selection
and utilization of a behavioral paradigm is to expect the unexpected. Valuable
insights, or even factors in the delineation of a toxic effect, can be lost by a
mind that has an a priori set of expectations which precludes an insightful
examination of the data obtained from any paradigm which might be utilized.
Sensory Processes
The sensory systems are the means by which our information-processing brains
interact with the environment, and as such these systems place important func-
tional constraints on behavior. The study of sensory systems is one of the
oldest research areas in experimental psychology, and the methods employed
in this study are well developed and have stood the test of time. We refer of
course to the traditional psychophysical methods and their contemporary
improvements. By utilizing the method of limits, the method of average error,
or the method of constant stimulus differences, we can determine important
information concerning the effects of a toxicant on the absolute threshold,
terminal threshold, difference thresholds, and the discriminability of stimuli.
More frequently than is desirable, behavioral tests are conducted with com-
plete disregard for the effects of toxicants on important sensory systems.
Deficiencies ascribed to disruptions in short-term retention in a visual delayed
response task could well occur because of an alteration in the visual system.
We can apply this same logic to a number of the behavioral paradigms we have
employed in assessing toxic effects. Recent efforts in our laboratory (Reischl
et at., 1975) and by others (Hanson, 1975) have shown significant alterations
in sensory systems due to toxic insult. In Chapter 8, Evans examines the behav-
ioral principles likely to aid in the precise definition of sensory impairment
caused by toxins.
Cognitive Processes
In the last decade, there has been a great expansion of research on cognitive
processes in humans and, more recently (the past five years), with nonhuman
primates. The growth of cognitive research, particularly with nonhuman pri-
mates in the areas of short- and long-term retention, forgetting, and language
acquisition, can provide the toxicologist with promising behavioral tools. With
the upsurge of research on human cognitive processes, there has been a corres-
ponding development in cognitive theory. We feel that within this area the
behavioral toxicologist might find unifying concepts relating findings with
animal subjects to those with humans. It has been our experience that the more
complex testing paradigms have been more sensitive to toxicologic insult,
particularly during the period of acquisition. Of special interest is the fact that
some cognitive processes appear to be fairly localized functionally within the
CNS, for example, short-term memory and the hippocampus, language and
speech centers in the left cortical hemisphere, and the prefrontal cortex and
delayed response deficit.
2-6
-------
We are hopeful that the behavioral toxicologist will utilize the research para-
digms available to study cognitive processes in both humans and animals so as
to delineate subtle behavioral changes resulting from long-term chronic expo-
sures to a number of central nervous system toxicants.
Variables of Critical Interest
Interacting with all of the experimental approaches which we have just men-
tioned are a number of specific variables, which, if not controllable or mani-
pulable in some way, must at least be recognizable as contributing their effects
to our data.
One of these variables could be simply the species used within a given experi-
ment. One should always assume that his data is species specific until inter-
species experiments prove otherwise. Furthermore, one could have intraspecies
effects, such as one strain of experimental subjects displaying behaviors which
are slightly different from another strain of subjects comparably exposed to a
given chemical. In either case, the species and strain issues raise a question
about extrapolation of our research data.
A second variable which has shown itself to be of considerable interest is the
dose-response relationship. High-dose, acute-exposure experiments have consi-
derable importance in that they suggest tasks which might have potential for
being sensitive to much lower chronic exposure levels. For example, early in
our experimental program, we found that the conditioned avoidance response
task did not appear to be sensitive to dieldrin exposure in sheep even at levels
that were sufficient to produce convulsions and other overt clinical signs of
toxicosis. Most behavioral toxicologic research, however, will probably deal in
a chronic exposure framework at levels approximating human exposure or at
least at levels found within the environment. One needs to be aware that the
lower the level of exposure, the more likely it is that individual differences in
metabolism or other physiologic factors may bias the behavioral data.
A third variable is synergism, the effects between toxicants to which the test
subjects are either experimentally or environmentally exposed. Currently,
this problem can be conservatively and practically handled on a small scale
with a factorial experiment using two or three toxicants. However, the syner-
gism problem becomes more difficult or bewildering when considering the pos-
sible interactions between toxicants, inter-intra species specific effects, and the
dose-response curve.
Although many physiologic experiments may not attempt to account for these
variables, the very nature of behavioral toxicologic research necessitates their
not being taken lightly. Each of these factors may be highly correlated not
only with one another, but also with the total functioning of the organism
both in and out of the experimental environment.
A fourth variable of considerable interest concerns the critical period of expo-
sure. Regardless of when an organism is exposed to a toxicant, the effects
observed are the result of many interacting variables, some of which are deter-
mined by the state of development of the organism. As earlier emphasized,
2-7
-------
alterations following a toxicosis at one point in the life span of the organism
may be specific to that period only or may appear in other developmental
stages. It is important, therefore, to examine carefully experimental data in
light of developmental variables prior to (1) extrapolating findings from one
period of development to another or (2) assessing the hazard of a toxicant.
Finally, there are insufficient data even to suggest what effects environmental
factors such as bedding, frequency'of cage cleaning, temperature and climatic
variations, amount of handling, conditioning prior to experimentation, and
housing might have on the outcome of an experiment. Obviously these are fac-
tors which, for the most part, we take for granted and place with our experi-
mental error. However, when these factors are combined with those discussed
above, they have serious implications for the data we produce from our experi-
mentation and the methods we utilize in obtaining the data. Recent reviews
(Fouts, 1976; Vesell et at., 1976; Lang and Vesell, 1976) document the signifi-
cance of these variables. A further discussion of methodological problems is
provided by Bornschein and Michaelson in Chapter 4.
Predictions
Having attempted to describe some of the disciplines involved in the field of
behavioral toxicology, some of the methods employed by these disciplines, and
the variables which are potentially critical in the utilization of these methods,
it seems appropriate to ascertain what directions research may take in the
future. As our field of investigation matures, we may expect some fairly dra-
matic shifts in areas of study, changes in methodology, and a better delineation
of the concepts involved in behavioral toxicology. Therefore, we feel confident
that the future will see:
1. Continuation and refinement in the standardization of testing proce-
dures with a goal of establishing a battery of sensitive behavioral tests
for use by regulatory agencies in evaluating hazards of potential neu-
rotoxicants. The emphasis here will probably be to develop the bat-
tery so that the tests provide information on a number of psycholog-
ical capabilities, any one of which may be affected by the toxicant.
2. Increased use and improvement of automated and nonautomated
behavioral testing techniques. Toxicologic research will more fre-
quently use behavior as an index as techniques are improved and then
will attempt to examine behavioral functioning from both individual
and social perspectives.
3. Increased studies on the synergistic effects of neurotoxicants, not
only with respect to other neurotoxicants, but also in relation to
species effects and specific environmental or laboratory influences
on behavior.
4. More studies of the critical times of exposure during the develop-
mental stages of the CNS. Of particular interest will be studies linking
specific psychologic deficits with specific periods of exposure.
2-8
-------
5. Increased emphasis upon correlating behavioral changes with concomi-
tant physiopathologic changes. Researchers in our field will someday
be able to detect physiopathologic changes and reliably predict their
behavioral correlates, and vice versa. The capability lies in the future,
but the seeds are currently being planted.
6. Greater emphasis on the importance of theory in determining the
direction of research. We are only beginning to abandon the "knee-
jerk" approach to our research whereby we conduct experiments
without regard for what has gone before both behaviorally and phy-
siologically. With the advent of sound theory development generating
testable hypotheses, behavioral toxicologic research will assume direc-
tion and, therefore, long-range goals.
7. More emphasis on the knowledge of the target area of toxic effect.
8. More attention to interpretive fidelity, to establishing stronger cri-
teria for reliance upon behavioral toxicologic data before regulatory
action is undertaken. The end result will be an elimination of the
"straw-man" which may incorrectly support or deter the use of a
toxicant.
Although these predictions are not all-inclusive, we feel they do represent
developing interests among the growing group of people engaged in behavioral
toxicologic research. Progress in behavioral toxicology will undoubtedly come
from the strength of a diversity of disciplines and their contributions to meth-
odology and theory. It is a well intergrated interdisciplinary approach which
will ultimately provide the answers to the problems facing behavioral toxi-
cology.
References
Bignami, G. (1976). Behavioral pharmacology and toxicology. Annu. Rev.
Pharmacol. Toxlcol. 16,329-366.
Brimblecombe, R. W. (1968). Behavioural studies. In Modern Trends in Toxi-
cology (E. Boyland and R. Goulding, eds.), pp. 149-174. Appleton Cen-
tury Crofts, New York.
Carson, T. L, Van Gelder, G. A., Buck, W. B., Hoffman, L J., Mick, D. L,
and Long, K. R. (1973). Effects of low level lead ingestion in sheep.
Clin. Toxic-ol. 6, 389-403.
Carson, T. L, Van Gelder, G. A., Karas, G. G., and Buck, W. B. (1974). Devel-
opment of behavioral tests for the assessment of neurologic effects of lead
in sheep. Environ. Health Perspect. 7, 233-237.
Finger, F. W. (1972). Measuring behavioral activity. In Methods in Psycho-
biology (R. D. Myers, ed.), Vol. 2, pp. 1-19. Academic Press, New York.
Fouts, J. R. (1976). Overview of the field: Environmental factors affecting
chemical or drug effects in animals. Fed. Proc. 35(5). 1162-1165.
2-9
-------
Hanson, H. M. (1975). Psychophysical evaluation of toxic effects on sensory
systems. Fed. Proc. 34(9), 1852-1857.
Lang, C. M. and Vesell, E. S. (1976). Environmental and genetic factors affec-
ting laboratory animals: Impact on biomedical research. Fed. Proc. (5),
1123-1124.
Norton, S. (1973) Amphetamine as a model for hyperactivity in the rat.
Physiol. Behav. 11, 181-186.
Reischl, P., Van Gelder, G. A., and Karas, G. G. (1975). Auditory detection
behavior in parathion-treated squirrel monkeys (Saimiri sciureus). Toxi-
col. Appl. Pharmacol. 34, 88-101.
Ruff in, J. B. (1963). Functional testing for behavioral toxicity: A missing
dimension in experimental environmental toxicology. J. Occup. Med.
5(3), 117-121.
Sandier, B. E., Van Gelder, G. A., Elsberry, D. D., Karas, G. G., and Buck,
W. B. (1969). Dieldrin exposure and vigilance behavior in sheep. Psychon.
Sci. 15, 261.
Smith, R. M., Cunningham, W. L, Jr., Van Gelder, G. A., and Karas, G. G.
(1976). Dieldrin toxicity and successive discrimination reversal in squirrel
monkeys (Saimirisciureus). J. Toxicol. Environ. Health 1, 1-11.
Spyker, J. M. (1975). Assessing the impact of low level chemicals on develop-
ment: Behavioral and latent effects. Fed. Proc. 34(9), 1835-1844.
iVan Gelder, G. A. (1975). Behavioral toxicologic studies oT dieldrin, DDT,
and ruelene in sheep. In Behavioral Toxicology (B. Weiss and V. G. Laties,
eds.), pp. 217-239. Plenum Press, New York.
Van Gelder, G. A., Carson, T. L, Smith, R. M., and Buck, W. B. (1973).
Behavioral toxicologic assessment of the neurologic effect of lead in sheep.
Clin. Toxiol. 6, 405-4I8.
Vesell, E. S., Lang, C. M., White, W. J., Passananti, G. T., Hill, R. N., Clemens,
T. L., Liu, D. K., and Johnson, W. D. (1976). Environmental and genetic
factors affecting the response of laboratory animals to drugs. Fed. Proc.
35(5), 1125-1132.
Weiss, B. and Laties, V. G. (1975). Behavioral Toxicology. Plenum Press, New
York.
2-10
-------
3. A REVIEW OF THE DEVELOPMENTAL
MODELS EMPLOYED IN
BEHAVIORAL TOXICOLOGY
HAROLD ZENICK
Department of Psychology
New Mexico Highlands University
Introduction
In assessing the effects of a potential toxicant on the developing offspring, one
of the first questions to confront the investigator is "When should exposure
occur?" Few colleagues have addressed this question to any extent (Spyker,
1975); and although administration of toxicants has occurred at almost every
developmental period, it has rarely been done in any factorial manner within
a study. Furthermore, although authors have discussed the methodological
considerations in this general area of developmental psychopharmacology
(Young, 1967; Brimblecombe, 1968; Kornetsky, 1970), they have essentially
avoided this issue.
In many studies the primary concern of the experimenter is to demonstrate or
reveal an "affected" organism, and it is often nor clear to the reader why a
particular period of exposure was selected for study. Although this omission
does not reduce the significance of the findings, an equally important issue
should be the delineation of the periods in the lifespan of the organism during
which toxicant exposure is necessary and/or sufficient to produce deleterious
effects. Assessment without such information may be incomplete, and the
availability of such information is of obvious value for prophylaxis.
The present chapter presents a review of the periods of development that have
been examined for lead, mercury, and pesticides (the review covers only pub-
lished articles and omits abstracts). Many of the considerations and directions
*This work supported by NIH-MBS Grant No. 5-S06-RR08066 and the Institute of
Research, New Mexico Highlands University, Las Vegas, New Mexico 87701.
3-1
-------
suggested here also apply to that area of developmental psychopharmacology.
Since this paper is primarily concerned with the consequences of the toxicant
on the developing offspring, only a sampling of studies which administer the
toxicant postweaning are presented. This procedure has been adopted pri-
marily to provide the reader with some insight regarding the ages of the adoles-
cent or adult organism at which each of the toxicants has been studied. In
addition, the paper concentrates on .behavioral studies outside the realm of
electrophysiology/neurophysiology.
Developmental Models of Mercury Exposure
Prior to detailing studies on this toxicant, a brief explanation of the tabular
format is necessary. In Table 3-1, which will serve here as an example, the
fourth column, "Parent," indicates which parent was exposed to the toxicant;
the "Premate Exposure" column reflects whether or not the parent was pre-
exposed to the toxicant prior to mating and the duration of such exposure;
the "Gestation Exposure" and "Nursing Exposure" headings reflect exposure
during those periods with the time of exposure indicated. Horizontal lines
continuing through columns reflect continued exposure over more than one
developmental period; in the "Postweaning Exposure" column, the day indi-
cated reflects when treatment, not behavioral testing, was initiated. In this
column, studies were grouped under "Adult" if only the weight range of sub-
jects was provided with no indication of age.
Most of the behavioral studies on mercury have employed postweaning expo-
sure, including a large number of studies on monkeys (e.g., Evans et a/., 1974;
Helberg and Nystrom, 1972; Vitulli, 1974; Berlin et a/., 1973; and Berliner a/.,
1975) and birds (e.g., Beliles et a/., 1967, 1968; Rosenthal and Sparber, 1972)
not included in this review. Most of the gestation-only studies exposed the
pregnant female during critical days of organogensis. Thus Hughes and Annau
(1976) orally dosed pregnant females on day 8 of gestation, Sobotka et at.
(1974) intubated mothers on days 6-15 of gestation, while Spyker ef a/. (1972)
administered intraperitoneal injections to pregnant mice on days 7 or 9. Both
the Sobotka et al. (1974) and Spyker et a/. (1972) studies point toward the
sensitivity of the neuromotor systems to mercury. Sobotka et al. (1974) re-
vealed that both eye opening and clinging ability developed earlier in neonates
of mercury-exposed dams, while Spyker eta/. (1972) found both differences in
open-field behavior and disrupted swimming ability of mercury offspring.
In terms of assessing contribution of in utero exposure, both of these studies
were confounded in terms of mercury carryover into nursing. Cross-fostering
procedures have been employed to partition out this contribution (Zenick,
1974; Spyker, 1975; Hughes and Annau, 1976). Spyker found that survival
rate was enhanced if offspring exposed during gestation were nursed by con-
trol mothers (74.4 percent) as compared to noncross-fostered counterparts
(57.3 percent). This result suggests that the residual mercury carried over in
the mother's milk was sufficient to have adverse consequences for that off-
spring. This deleterious, residual effect was further confirmed by discerning
a depressed weaning weight in young born of control mothers and cross-fos-
tered to treated mothers. In addition, motor disturbances became increas-
ingly evident with age in both prenatal and postnatal exposed groups. No
3-2
-------
Table 3-1. DEVELOPMENTAL MODELS OF MERCURY EXPOSURE
Author Species
Salvaterra eta/., 1973 Ma
Morganti eta/., 1974
Morganti eta/., 1976
Klein & Atkinson, 1973 R
Klein eta/., 1974 R
Braun & Synder, 1973
Sobotkaef a/., 1974 R
Hughes 81 Annau, 1976 M
Spykereta/., 1972 M
Spyker, 1975
Post eta/., 1973 R
Zenlok, 1974; 1976 R
aM|ce
bRats
Postweaning exposure
(days of age)
1 1 1
Parent Premate Gestation Nursing 34568025
Route 6 9 exposure exposure exposure 00000000 Adult
Ip - - - - 42
- - - - 42
- - - - 42
Diet - - - 25
Water - - - - 25 X
Ip or iv — — — — X
Oral 9 - 6 25
Oral 9 - 8
ip 9 7;9
7;9;12;13
Oral - - - 15;21 60
Water 9 - 1 21
1 21
22-30
-------
behavioral data other than neuromuscular assessment was reported on these
older animals. On the other hand, Zenick (1974) and Hughes and Annau
(1976) found no effects on offspring weight or behavioral tasks in the cross-
fostered young exposed only to this residual; however, these studies did con-
firm the detrimental influence of in utero exposure. In the Hughes and Annau
study (1976), mice exposed to methylmercury in utero exhibited deficits in
a two-way active avoidance shuttle box and in a punishment situation. No dif-
ferences were observed on a conditioned suppression task, in an open field, or
in a water-escape runway. The failure to see effects on the latter two tasks con-
tradicts observations made by Spyker et at. (1972). Hughes and Annau (1976)
also tested the mothers of the progeny as well as offspring of the progeny
themselves. Neither group of subjects exhibited deficits, suggesting that mer-
cury had more obvious deleterious consequences during in utero exposure
leaving the treated mother unaffected; and, furthermore, the effect was res-
tricted to the generation in which exposure occurred.
Zenick (1974) contrasted the effects of mercury exposure administered to
independent groups at different periods of development. Potential mothers
were exposed either during gestation or while nursing (treatment begun at
parturition), or offspring were directly exposed beginning at weaning, day 21.
As noted above, cross-fostered groups were also employed. Results indicated
that offspring of mothers exposed during gestation and offspring exposed
directly to mercury at weaning exhibited learning deficits on a brightness
discrimination task begun at 30 days of age. These deficits also appeared in a
retest session 21 days later. Offspring of mothers exposed during nursing only
and offspring exposed only to the residual mercury in the milk were not dif-
ferent from controls on the learning task. However, in a subsequent study in
which visual evoked potentials were recorded (Zenick, 1976), the nursing-only
offspring were relegated to the "affected" category. The evoked potential
data reinforces the concept of employing a battery of tasks in assessing critical
periods of exposure and suggests that the sensory evoked potential may be as
sensitive, if not more so, than other behavioral assessment techniques.
The effect of exposure during nursing, as contrasted with postweaning, has
also been assessed by Post et al. (1973) who forcefed a single dose of mercury
to either 15-, 21-, or 60-day old rats. No changes in discrimination learning or
open-field behavior were observed in the nursing-exposed pups, while post-
weaning-exposed rats (60 days of age) were significantly different from con-
trols on both tasks. These findings are in agreement with Zenick (1974).
Olson and Boush (1975) examined the effects of prolonged-multiperiod expo-
sure within the same animal by feeding mercury to female rats throughout
gestation and nursing, then maintaining the pups on the same diet throughout
behavioral testing at 68 days of age. These mercury offspring exhibited re-
tarded maturation of swimming behavior and the righting reflex and made
significantly more errors than controls on a symmetrical maze task.
Data on mercury seem to suggest that exposure during gestation or postwean-
ing is sufficient to produce an effect on a number of tasks. The influence of
early postnatal exposure via the mother's milk is still uncertain; however, in
3-4
-------
conjunction with gestation exposure, treatment during nursing may create the
most potentially hazardous setting for the offspring.
Developmental Models of Lead Exposure
A greater variety of developmental models have been employed in examining
the behavioral effects of lead exposure than in studying either mercury or pes-
ticides. There are at least two reasons for this situation: the adult animal has
not served as a good model for examining the effects of low-level lead exposure
because the adult exhibits extreme resistance to the CNS effects of lead poison-
ing, requiring very large dosage of lead to produce even slight clinical manifes-
tations of CNS toxicity. Secondly, because of its early widespread use, lead
toxicity via pica posed a very real environmental threat for the developing
child, especially those living in older, substandard housing.
Frequently employed developmental models are shown in Table 3-2. Although
no lead study has recorded behavioral measures on pups exposed in utero only,
a suggested differential sensitivity of this period as compared to nursing may be
gleaned from a study by Snowden (1973). In that experiment, mothers were
injected ip with lead either thoughout pregnancy or throughout nursing, or
pups were directly exposed to lead by injections for 37 days beginning at 22
or 100 days of age. The postweaning injections had no effects on maze learn-
ing. However, females injected during pregnancy did not give birth to viable
offspring; whereas the same dosage administered to mothers during nursing did
not produce any fatalities, but it did lower the pups' body weights and increase
their errors on a maze task. Although there may be a confounding between
route of exposure, namely ip injection, and critical period, the study does
suggest a need to examine more closely the contribution of gestation exposure
in terms of increased sensitivity to lead.
A developmental model developed by Pentschew and Garro (1966) has become
the prototype for studying developmental lead toxicity. These authors were
able to produce CNS lesions in pups sucking a dam exposed to 4-5 percent lead
carbonate in her daily diet. Unfortunately, direct access to the maternal diet by
the pups was possible at approximately 18 days of age, resulting in the appear-
ance of clinical symptoms (e.g., retarded body weight, paralysis, etc.) in the
offspring at 20-30 days of age. A procedure to correct this methodological flow
and incorporate other needed controls has been presented by Michaelson and
Sauerhoff (1974a, b). However, the Pentschew and Garro model (1966) has
left a definite impression as witnessed by the large number of papers employing
the nursing exposure model to assess the effects of lead (Table 3-2). Many in-
vestigators, therefore, assume that the oral ingestion of lead (via milk) is analo-
gous to the pica route of lead-containing materials by human infants (Michael-
son and Sauerhoff, 1974a).
Sobotka and his colleagues (1974, 1975) have taken a somewhat different
approach to nursing exposure in that they directly intubated the pup beginning
at day 2-3 and continued this exposure until weaning. Direct intubation of
pups during nursing has also been done by Goiter and Michaelson (1975). Pups
exposed to lead exhibited an attenuated response to amphetamine-induced
3-5
-------
Table 3-2. DEVELOPMENTAL MODELS OF LEAD EXPOSURE
Author Species
Avery etal., 1974 Ra
Shapiro et al., 1973 R
Bullock etal., 1966 Rfe
Terzin & Vujkov, 1968 M
Sobotka & Cook, 1974
Sobotka et al., 1975 R
. «n-,E D
C7>
Snowden, 1973 R
Brown etal., 1971 R
Route
Oral
IP
IP
ip
Oral
oral,
or
ip
(pup)
IP
ip
Postweaning exposure
(days of age)
1 1 1
Parent Premate Gestation Nursing 34680258
d ? exposure exposure exposure 00000000 Adult
- - - - X
- . 90
- - - - X
- 120
- 3 21
9 1 01
1 10
11 21
- 1 21
1 21
22
100
8 21 35
(continued)
-------
Table 3-2. DEVELOPMENTAL MODELS OF LEAD EXPOSURE ( continued)
Author Species
Sauerhoff &
Mlchaelson, 1973
Mlchaelson &
Sauerhoff, 1974a,b R
Goiter & Mlchaelson, R
1975
Silbergeld & Goldberg,
1973, 1974a,b, 1975 M
Drlscoll & Stegner, 1976 R
Zenick eta/., 1978a,b R
Padich & Zenick, 1977 R
Reiter, etal., 1975 R
Brady eta/., 1975 R
Route
Diet
Diet
or
oral
(pup)
Water
Water
Water
Water
Water
Oral
Postweaning exposure
(days of age)
1 1 1
Parent Premate Gestation Nursing 34680258
<3 9 exposure exposure exposure 00000000 Adult
9 1 75
- 9
9 1 91
22 42
9 1 91
Rats
Mice
-------
increases in activity and deficits in a two-way shuttle box (Sobotka and Cook,
1974) and in a habit-reversal task (Sobotka et a/., 1975). Sobotka has suggested
that the advantages of this technique were that "this approach avoids the con-
founding variables of maternal effects of lead and also affords a more effec-
tive means of determining the dose of lead that each pup receives {Sobotka
et al., 1974). However, the technique does have its disadvantages in that it
results in the delivery of the lead dosage as a single bolus into the stomach,
altering the parameters of absorption from that which would be observed with
a slower rate of intake spread over a greater period of time each day. Further-
more, we have observed that intubation can be stressful to the animal (Brady
et at. 1975), and this stress may have an indirect, differential effect on lead-
treated pups.
Brown administered lead during nursing either to the mother, by gavage or in
her drinking water, or to the pup directly, ip. Irrespective of route, Brown
(1975) found a significant performance deficit on a T-maze discrimination
task when pups were tested at 8-10 weeks. Furthermore, by contrasting groups
exposed during days 1-10 with pups exposed during days 11-12, Brown noted
that exposure was necessary during the first 10 days in order for lead to exert a
detrimental influence. Since blood lead levels were significantly higher in the
1-10 day old pups, as compared to 11-20 day lead pups or controls. Brown
(1975) suggested that the increased sensitivity of this group may, in part, be a
result of age-related differences in uptake, distribution, and retention of lead.
The influence of nursing exposure has also been investigated by Michaelson and
coworkers (Sauerhoff and Michaelson, 1973; Michaelson and Sauerhoff, 1974a;
Goiter and Michaelson, 1975; Bornschein et al., 1977). These studies, however,
were not nursing-only exposure in the truest sense because the pups were con-
tinued on the treatment following weaning and throughout testing. It is possi-
ble that this direct exposure at what was still a relatively young age (21 days)
may also have exerted detrimental effects. Aside from this consideration,
Michaelson's group found changes in activity levels of the lead-exposed off-
spring as compared to controls. However, they also noted that an interaction of
lead and a state of undernutrition in the pup (induced by the decreased food
intake by the lead mother) must be considered in interpreting the data (Born-
schein et al., 1977). In fact, this retarded body weight of pups appears consis-
tently across laboratories that employ dosages in the range of the Pentschew
and Garro study (1966).
The opportunity to contrast the combined nursing and early postweaning
exposure with nursing-only exposure has been investigated by Silbergeld and
colleagues in a series of papers. Her treatment groups consisted of pups ex-
posed to lead only while nursing via the mother's milk or pups which were
exposed while nursing as well as after weaning, throughout testing. While
Silbergeld reported a decreased weight in the former group until it was
switched to the control fluid, the report on the behavior of these offspring is
still not clear: "Motor activity was also measured in offspring removed from
the lead solutions of 10 mg/ml at weaning. Four out of the six mice measured
3-8
-------
showed increases in motor activity comparable to those seen in mice main-
tained on lead. However, two of these six mice had activity levels not signi-
ficantly higher than those of controls" (Silbergeld and Goldberg, 1973). A
similar statement on this group's behavior was made in a second study
(Silbergeld and Goldberg, 1974) which rather than being a replication of
earlier work, involved an additional manipulation made on the same animals.
It is apparent that additional work with the group exposed only during nursing
(and larger N) is necessary to confirm the deleterious effect of such exposure.
The question of developmental versus direct postweaning exposure has also
been addressed in a study by Padich and Zenick (1977) which employed four
groups: Group Pb/Pb, mothers pretreated with lead and continued on lead
during gestation and nursing with the offspring continued on lead after
weaning and throughout testing (developmental and direct exposure); Group
Pb/C, offspring developmentally exposed only; Group C/Pb, offspring exposed
directly only; and Group C/C, no lead exposure. Significant differences were
found between Groups Pb/Pb and C/C on a fixed ratio (FR20) bar press task;
however, exposure of the offspring during gestation and nursing only (Group
Pb/C) or during postweaning only (Group C/Pb) did not result in any impaired
performance on the FR task, nor was their performance different from con-
trols under fixed-interval schedules (Zenick et a/., 1978a). However, the Pb/C
offspring have been shown to be inferior to controls on discrimination learning
tasks (Brady et a/., 1975; Zenick et a/., 1978b). Thus it appears that for some
tasks a combined developmental/direct exposure is necessary for lead to exert
an effect. As suggested earlier, this may contribute, in part, to the effect seen
by Michaelson in animals continued on treatment following nursing-exposure
(Michaelson and Sauerhoff, 1974a; Sauerhoff and Michaelson, 1973; Goiter
and Michaelson, 1975). In any event, these findings suggest that any conclu-
sions regarding critical exposure periods and lead susceptibility must be quali-
fied with regard to the instrument of measurement.
Encompassed in a large number of developmental lead studies have been some
unique experimental manipulations rarely reflected in the study of other toxi-
cants. These manipulations include assessing the effects of prolonged maternal
exposure (Driscoll and Stegner, 1976; Reiter et a/., 1975a; Brady eta/., 1975;
Padich and Zenick, 1977; Zenick et a/., 1978a, b) and the contribution of
paternal exposure (Brady et a/., 1975). Driscoll and Stegner (1967) have
assessed the influence of chronic exposure by administering lead throughout
gestation and nursing while Reiter et a/. (1975) and Zenick et a/. (1978a, b)
have extended this period backward to include parental exposure prior to
mating. In Reiter's laboratory, this "premating exposure" incorporated pre-
treatment for a period of 40 days to allow lead concentration to approach
steady states. This objective has been included in the rationale employed in
our laboratory; however, it is only a part of a philosophy which assumes that
the potential mother may have existed in the toxic setting for some time prior
to mating, conceivably her entire life. As a consequence, there may be a build-
up and storage of the toxicant in her system, resulting in another potential
source to which the fetus or pup may be exposed in addition to the mother's
daily dosage. Subsequent to mating, the mother continued in this environment
during gestation and nursing. This hypothesis has led to the current model
3-9
-------
employed in our laboratory that is presented in Table 3-3. Such a model incor-
porates premating exposure while contrasting developmental versus direct toxi-
cant exposure on offspring behavior. A cross-fostering manipulation is also in-
cluded. In addition, assessment is made throughout the life of the animal in
order to evaluate permanence as well as detect any late-occurring effects.
Another manipulation in studying developmental lead toxicity has been the
attempt to partial out the parental contribution. A study conducted by Brady
et al., (1975), employing a pretreatment model, contrasted the effects of
exposing both potential parents to lead with exposure of the mother or father
only. Analysis of offspring behavior on a water-escape T-maze task revealed
that the three lead groups made more errors than the controls but did not
differ from one another. However, offspring resulting from dual parental
exposure had longer swimming times than the single-parental exposure groups
who, in turn, were slower than the controls. The finding of a paternal-only
effect has since been replicated in our laboratory. The ramifications of such a
finding may be far-reaching. For instance, limitations on employment of preg-
nant women in industry are common; however, the Brady et al., (1975)
findings suggest that the standards for men may also need further consider-
ation. At the very least, the question of paternal influence merits further in-
vestigation.
Although not reported in Table 3-2, there have been a number of behavioral
studies examining lead toxicity which have employed species other than
rodents as subjects, the most notable effort being a series of investigations by
Van Gelder and colleagues employing sheep as an experimental model. They
examined the effects of lead administration to the adult (Van Gelder et al.,
1973), exposure in utero plus carryover (Carson et al., 1974a), and early
postnatal treatment (Carson et al., 1974b). Employing a series of measures,
the group has found deficits primarily on a visual discrimination task in lambs
that had been exposed in utero. These animals were tested at 10-15 months of
age subsequent to prenatal exposure, reflecting a disruption of behavior in the
absence of elevated blood lead levels, suggesting that lead exposure during
gestation is sufficient to induce permanent debilitation. This aspect of the per-
sistence of an effect in the absence of elevated toxicant levels is a critical issue
that has been neglected by researchers in this area. This point will be addressed
more directly in the final section of this chapter.
, Finally, a study with infant monkeys by Allen et al. (1974) should be men-
tioned, since one phase of the study closely simulated lead poisoning via pica.
In that study, infant rhesus monkeys were housed for three months with
surrogate mothers having a lead base (other infants received lead via their
formula). As Allen et al. (1974) notes, "throughout this period the infants
scratched, licked, and rubbed the lead bases," an interaction with their environ-
ment that closely mimics pica. The research team observed neuromuscular and
visual impairments and altered social behavior with the latter symptom persis-
ting when external lead exposure was eliminated. These effects were not ob-
served in adolescent or adult monkeys receiving similar dosages. These results
again reinforce that view of the increased susceptibility of the developing
organism to low-level exposure as contrasted with its adult counterpart.
3-10
-------
Table 3-3. A MODEL TO CONTRAST DEVELOPMENTAL/DIRECT EXPOSURE
Postweaning exposure
Premate
exposure
Gestation
exposure
Nursing
exposure
Adolescence
Middle age
Old age
Day21-dav90 1 21 1-
-21
C C •
Pe-
eress-fostered to
partial out gesta-
tion from nursing
Influences
T = Toxicant.
In summary, although there have been a number of developmental studies of
lead toxicity, many questions remain unanswered, especially the contribution
of in utero exposure. Future use of cross-fostering techniques may shed some
light on the influence of gestation exposure alone. Furthermore, there is a
need to do more extensive long-term chronological assessment.
Developmental Models of Pesticide Exposure
Review of Table 3-4 will reveal that the number of developmental studies on
the effects of pesticides on behavior is scant compared to the number of
investigations of the heavy metals. This dearth of developmental work is rather
paradoxical, since initial concern over pesticide toxicity has given the overall
area of environmental toxicology possibly its greatest impetus. Many of the
studies of pesticides examined only their anticholinesterase properties for
delineating the role of acetylcholine in the CNS and have ignored their toxic
effects. There are several articles available on these studies, the most recent
being a review by Bignami et a/. (1975). Since these studies have been done
almost exclusively in adult animals, they will not be considered in this chapter.
The influence of DDT and Parathion exposure during gestation has been
systematically studied by AI-Hachim and Fink (1967; 1968a, b, c). In their
investigations, groups of female rats were exposed to DDT during different
3-11
-------
trimesters of pregnancy. They observed a delay in the acquisition of a condi-
tioned avoidance response in offspring of mothers treated during either the
second or third trimester of pregnancy (AI-Hachim and Fink, 1968b). How-
ever, since the animals were tested at 30 days of age, AI-Hachim suggested that
this differential effect may have been a result of greater opportunity for the
accumulation of DDT (via placenta and/or mother's milk) in offspring exposed
at later points in gestation, since they failed to see effects of DDT on open-
field behavior (AI-Hachim and Fink, 1968c) in animals tested after post-
weaning (60-66 days and 70-90 respectively). Unfortunately, rather than
employing different tasks, it would have been necessary to assess conditioned
avoidance responding in older animals before contributing only a transitory,
biochemical effect to gestational DDT exposure. It is possible that the later
stages of pregnancy do reflect an increased vulnerability above and beyond the
simple fact of increased accumulation of DDT.
In the only factorial manipulation of developmental pesticide exposure, Craig
and Ogilvie (1974) contrasted the effects of exposure during gestation and
nursing with exposure to DDT during either gestation or nursing only. This
later manipulation was achieved by employing cross-fostering procedures. One
of the points which was not clear from the study was whether the DDT
mothers remained on treatment during nursing or whether treatment was
discontinued at birth so that nursing-only exposure more accurately reflected
a gestation carryover; (the cross-fostering design is addressed in greater detail
in the next section. Interestingly, Craig and Ogilvie (1974) found that DDT
administered during either gestation or nursing resulted in offspring perfor-
mance that was either equivalent or superior to controls on a series of depen-
dent measures of maze learning. On the other hand, the combined gestation-
nursing exposure resulted in inferior performance as compared to controls.
One conclusion that might be drawn is that DDT is toxic only if exposure
occurs during both gestation and nursing. However, Craig and Ogilvie (1974)
forwarded the hypothesis that the differing performances were a result of a
dose-related, biphasic effect of DDT. Essentially, since the single periods were
of shorter duration, the offspring were exposed to a cumulatively lower dosage
than their dual-period-exposure counterparts. Craig and Ogilvie (1974) sug-
gested a biphasic phenomenon, with lower dosages of DDT (gestation or
nursing only) being stimulatory and higher dosages (gestation and nursing ex-
posure) interfering with performance. This dose-response hypothesis is similar
to that noted above by AI-Hachim. However, one point that should be noted
is that Craig and Ogilvie (1974) began testing their rats at 49 days of age and
retested them one month later. Since treatment was discontinued at weaning,
it was possible that the concentration of DDT was extremely low or absent
during testing. The ability of the compound to exert a stimulatory or inhib-
itory influence in light of this is puzzling. However, the data are suggestive of
a need to more closely examine the relationship between dose and period of
exposure. Furthermore, this study is one of the few where an attempt was
made to assess permanence of an effect via a test-retest design. (See the follow-
ing section).
3-12
-------
Table 3-4. DEVELOPMENTAL MODELS OF PESTICIDE EXPOSURE
Author
Pearson eta/., 1969
Clark & Pearson, 1973
Clark etal., 1971
Russel eta/., 1961
Relter etal,, 1973
Khalry, 1959
Peterle & Peterle, 1971
Sobotka, 1971
V AI-Hachim Si Fink, 1967
CO
AI-Hachim, 1971
Woolley, 1970
Pesticide
Disyston
Dlsulfoton
Systox
Parathlon
DDT
DDT
DDT
DDT
Parathion
Aldrin
DDT
Species
R
M
R
M
R
M
M
M
M
M
R
Parent
Route d 9
Diet -
Diet
Diet
In tub ate —
Diet -
Diet
Intubate —
Intubate 9
Intubate 9
Diet 9
Postweanlng exposure
(days of age)
1 1 1
3468025
Prematlng Gestation Nursing 0000000 Adult
30
_ - - X
42
- - - 100
- - - X
- 120
- 130
_ - - X
- 1-7
7-14
14-21
14-21
1 91
1 21 a
(Several
days?)
aCr oss-fostered.
-------
There has been only one prolonged developmental exposure study using DDT
(Woolley, 1970) in which potential mothers were exposed prior to mating with
exposure continued throughout gestation and nursing. The offspring were
then maintained on the treatment for the duration of the study. Unfor-
tunately, only limited behavioral assessment was done, with the main effect
of DDT being to counteract the ability of strychnine to increase the intensity
of electroshock seizures.
In addition to the rodent, a number of other species have been employed in
assessing the effects of pesticides in postweaning animals. Several studies have
used the monkey (Reiter et at., 1975b; Smith et al., 1976; Lattal and Wilber,
1971; Revzin, 1970), and a series of studies in sheep have been summarized
by Van Gelder (1975).
This review of pesticide literature reinforces the urgency to gain greater insight
into the developmental influences of pesticides. The small number of develop-
mental studies, limited dose-response data, and poverty of behavioral measures
make conclusions regarding the developmental consequences of pesticide ex-
posure premature.
Future Directions in Developmental-Behavioral Toxicology
Although there have been only a few studies which have employed cross-fos-
tering techniques (Zenick, 1974, 1976; Spyker, 1975; Hughes and Annau,
1976; and Craig and Ogilvie, 1974), the necessity of such a manipulation is
essential in evolving a more complete picture of the interaction of a particular
toxicant with the various periods of development. The design for cross-fos-
tering and foster-rearing has been succinctly described by Spyker (1975) and
in essence consists of placing pups at birth with mothers of opposing treat-
ments (cross-fostering) or same treatment (foster-rearing), the latter being a
control manipulation. Other pups may remain with their natural mothers.
Such manipulations may not only partial out the influence of gestation versus
nursing exposure, but also separate the direct effects of the agent on the off-
spring from effects on maternal behavior towards the offspring. Such a mani-
pulation provides a clearer concept of necessary and/or sufficient periods
during which toxicant exposure can have adverse consequences for the off-
spring. This information is vital for prophylaxis. The nature of the prophy-
lactic treatment' will vary depending on whether exposure during gestation
only, nursing only, or both periods is sufficient to cause damage to the off-
spring.
Although the cross-fostering manipulation appears relatively straightforward,
there is some inconsistence in the manner of treating the treatment mother.
Most studies cease treatment for all mothers at parturition. Subsequently,
pups cross-fostered on a treatment mother are exposed only to residual toxi-
cant carried over from gestation. This arrangement is different from ones
involving treated mothers maintained on the toxicant during nursing. Studies
employing the "residual" model cannot as readily conclude in the face of
negative findings that nursing exposure does not exert an effect. The reduced
dosage and length of exposure may mask an effect that would appear with
3-14
-------
actual, full nursing exposure. Future researcners addressing the question of
critical periods need to be alerted to the differences that may be obtained
depending upon which cross-fostering technique is employed.
In terms of assessment there appears to be another gap in our research efforts.
Table 3-5 contains a list of studies that have employed either chronological
or test-retest assessment and/or have run toxicant determinations for the pur-
pose of testing the animal in the absence of elevated levels of the toxicant. The
brevity of this list is quite discouraging although some studies have been
omitted where testing was delayed long enough to assume the absence of the
toxicant (e.g., Sobotka et al., 1975; AI-Hachim, 1968a, c; and Hughes and
Annau, 1976).
Table 3-5. STUDIES USING CHRONOLOGICAL OR TEST-RETEST AND/OR
TESTING IN THE ABSENCE OF HIGH TOXICANT LEVELS
Chronological or Behavioral deficits in
Toxicant test-retest absence of compound
Spyker, 1975 Zenick, 1974, 1976
Mercury
Zenick, 1974, 1976
Sobotka eta/., 1975 Brown, 1975
Lead
Carson eta/,, 1974b
Craig and Ogllvie, 1974
Pesticides
The value of chronological assessment is essential, for it is conceivable that
certain disturbances may be delayed or lie dormant only to appear later in
life. This situation has certainly been documented for mercury (Spyker, 1975).
On the other hand, deficits revealed in young may disappear with age. Reiter
et al. (1975) has suggested that a maturational lag may account for such
behavior. The test-retest design also fits into this scheme in terms of reflecting
permanence of a deficit.
In conjunction with these experimental manipulations, analyses of toxicant
levels need to be conducted at various times during testing. As long as the
compound persists in the system, the possibility of a reversible, biochemical
lesion exists. Deficits persisting in the absence of the compound suggest a
more pathological and likely irreversible effect.
References
AI-Hachim, G. M. and Fink, G. B. (1967). Effects of DDT or Parathion on the
audiogenic seizures of offspring from DDT or Parathion mothers. Psycho/.
Rep. 20, 1183-1187.
AI-Hachim, G. M. and Fink, G. B. (1968a). Effect of DDT or Parathion on the
3-15
-------
open-field behavior of offspring from DDT or Pa rath ion-treated mothers.
Psychol. Rep. 22, 1193-1196.
AI-Hachim, G. M. and Fink, G. B. (1968b). Effect of DDT or Parathion on
condition avoidance response of offspring from DDT or Parathion-treated
mothers. Psychopharmacologia 12, 425-427.
AI-Hachim, G. M. and Fink, G. B. (1968c). Effect of DDT or Parathion on the
minimal electroshock seizure threshold of offspring from DDT or Parathion-
treated mothers. Psychopharmacologia 13, 408-412.
AI-Hachim, G. M. (1971). Effect of aldrin on conditioned avoidance response
and electroshock seizure threshold of offspring from aldrin-treated mothers.
Psychopharmacologia 21, 370-373.
Allen, J. R., McWey, P. J., and Suomi, S. J. (1974). Pathological and behavioral
effects of lead intoxication in infant rhesus monkey. Environ. Health
Perspect. 7, 239-246.
Avery, D. D., Cross, H. A., and Schroeder, T. (1974). The effects of tetraethyl
lead on behavior in the rat. Pharmacol. Biochem. Behav. 2, 473-479.
Bellies, R. P., Clark, R. S., Belluscio, P. R., Yuile, C. L, and Leach, L. J.
(1967). Behavioral effects in pigeons exposed to mercury vapor at a concen-
tration of 0.1 mg/m3. Amer. Ind. Hyg. Assoc. J. 28, 482-484.
Beliles, R. P., Clark, R. S., and Yuile, C. L. (1968). The effects of exposure to
mercury vapor on behavior of rats. Toxicol. Appl. Pharmacol. 12, 15-21.
Berlin, M., Nordberg, G., and Hellberg, J.(1973). The uptake and distribution
of methylmercury in the brain of Saimiri sciureus in relation to behavioral
and morphological changes. In Mercury, Mercurials, and Mercaptans (M. W.
Miller and T. W. Clarkson, eds.), pp. 197-208. Thomas, Springfield, Illinois.
Berlin, M., Grant, C. A., Hellberg, J., Hellstrom, J., and Schutz, A. (1975).
Neurotoxicity of methylmercury in squirrel monkeys. Arch. Environ.
Health 30, 340-348.
Bignami, G., Rosic, N., Michalek, H., Milosevic, M., and Gatti, G. L. (1975).
Behavioral toxicity of anticholinesterase agents: Methodological, neuro-
chemical and neurophysiological aspects. In Behavioral Toxicology. (B.
Weiss and V. G. Laties, eds.), pp. 155-210. Plenum Press, New York.
Bornschein, R. L, Michaelson, I. A., Fox, D. A., and Loch, R. (1977). Evalu-
ation of animal models used to study effects of lead on neurochemistry and
behavior. In Biological Effects of Environmental Pollutants. (S. D. Lee, ed.).
Science Publishers, Inc., Ann Arbor.
Brady, K., Herrera, Y., and Zenick, H. (1975). Influence of parental lead
exposure on subsequent learning ability of offspring. Pharmacol. Biochem.
Behav. 3, 561-565.
3-16
-------
Braun, J. J. and Synder, D. R. (1973). Taste aversion and acute methyl mer-
cury poisoning in rats. Bull. Psychon. Soc. 1, 419-420.
Brimblecombe, R. W. (1968). Behavioural studies. In Modern Trends in Toxi-
cology. (E. Boyland and R. Goulding, eds.), pp. 149-174. Appleton Century
Croft, New York.
Brown, D. R. (1975). Neonatal lead exposure in the rat: Decreased learning as
a function of age and blood lead concentrations. Toxicol. Appl. Pharmacol.
32, 628-637.
Brown, S., Dragann, N., and Vogel, W. H. (1971). Effect of lead acetate on
learning and memory in rats. Arch. Environ. Health 22, 370-372.
Bullock, J. D., Wey, R. J., Zaia, J. A., Zarembok. I., and Schroeder, H. A.
(1966). Effect of tetraethyllead on learning and memory in the rat. Arch.
Environ. Health 13, 21-22.
Carson, T. L, Van Gelder, G. A., Karas, G. G., and Buck, W. B. (1974a).
Slowed learning in lambs prenatally exposed to lead. Arch. Environ. Health
29, 154-156.
Carson, T. L, Van Gelder, G. A., Karas, G. G., and Buck, W. B. (1974b).
Development of behavioral tests for the assessment of neurologic effects
of lead in sheep. Environ. Health Perspect. 7, 233-237.
Clark, G., Koester, A. G., and Pearson, D. W. (1971). Exploratory behavior in
chronic disulfoton poisoning in mice. Psychopharmacologia 20, 169-171.
Clark, G. and Pearson, D. W. (1973). Learning in chronic disulfoton poisoning.
J. Gen. Psycho/. 89, 305-311.
Craig, G. R. and Ogilvie, D. M. (1974). Alteration of T-maze performance in
mice exposed to DDT during pregnancy and lactation. Environ. Physio/.
Biochem. 4, 1 89-199.
Driscoll, J. W. and Stegner, S. E. (1976). Behavioral effects of chronic lead
ingestion on laboratory rats. Pharmacol. Biochem. Behav. 4, 411-417.
Evans, H. L, Latties, V. G., and Weiss, B. (1974). Behavioral effects of methyl-
mercury. Proceedings, First Annual Trace Elements Conference, Oakridge,
Tenn., 534-540.
Goiter, M. and Michaelson, I. A. (1975). Growth, behavior and brain cate-
cholamines in lead-exposed rats: A reappraisal. Science 187, 359-361.
Hellberg, J. and Nystrom, M. (1972). The influence of methyl mercury expo-
sure on learning set behavior of squirrel monkeys (Saimiri sciureus). Psy-
cho/. Res. Bull. 12, 1-9.
Hughes, J. A. and Annau, Z. (1976), Postnatal behavioral effects in mice after
prenatal exposure to methylmercury. Pharmacol. Biochem. Behav. 4,
385-391.
3-17
-------
Khairy, M. (1959). Changes in behaviour associated with a nervous system
poison (D.D.T.) Quat. J. Exp. Psychol. 11, 84-91.
Klein, S. B. and Atkinson, E. J. (1973). Mercuric chloride influence on active
avoidance acquisition in rats. Bull. Psychon. Soc. 1, 437-438.
Klein, S. B., Barter, M. J., Murphy, A. L, and Richardson, J. H. (1974).
Aversion to low doses of mercuric chloride in rats. Physiol. Psychol. 2,
397-400.
Kornetsky, C. (1970). Psychoactive drugs in the immature organism. Psycho-
pharmacologia 17, 105-136.
Latta, K. A. and Wilbur, E. M. (1971). Effects of dieldrin upon the delayed
response performance of rhesus monkeys (Macca mulatta). Edgewood
Arsenal Tech. Rep. EATR 4526.
Michaelson, I. A. and Sauerhoff, M. V. (1974a). An improved model of lead-
induced brain dysfunction in the suckling rat. Toxicol. Appl. Pharmacol.
28, 88-96.
Michaelson, I. A. and Sauerhoff, M. W. (1974b). Animal models of human
disease: Severe and model lead encephalopathy in the neonatal rat. Environ.
Health Perspect. 7, 201-225.
Morganti, J. B., Lown, B. A., Stineman, C., and Massaro, E. J. (1974). Effects
of food deprivation on open field behavior of mice exposed to methyl
mercury. Psychol. Rep. 35,901-902.
Morganti, J. B., Lown, B. A., Salvaterra, P., and Massaro, E. J. (1976). Effects
on open field behavior of mice exposed to multiple doses of methyl mer-
cury. Gen. Pharmacol. 7, 41-44.
Olson, K. and Boush, G. M. (1975). Decreased learning capacity in rats exposed
prenatally and postnatally to low doses of mercury. Bull. Environ. Contam,
Toxicol. 13, 73-79.
Padich, R. and Zenick, H. (1977). The effects of developmental and/or direct
lead exposure on FR behavior in the rat. Pharmacol. Biochem. Behav. 6,
371-375.
Pearson, D. W., Clark, G., and Moore, C. M. (1969). Comparison of behavioral
effects of various levels of chronic di--syston poisoning. Proceedings of 77th
- Annual Conference, APA, 871-872.
Penschew, A., and Garro, F. (1966). Lead-encephalo-myelopathy of the suck-
ling rat and its implications on the porphyrinopathic nervous disease. Acta
Neuropathol. 6, 266-281.
Peterle, A. F. and Peterle, T. J. (1971), The effect of DDT on aggression in
laboratory mice. Bull. Environ. Contam. Toxicol. 6, 401-405.
3-18
-------
Post, E. M., Yang, M. G., King, J. A., and Sanger, V. L (1973). Behavioral
changes of young rats force-fed methyl mercury chloride. Proc. Soc. Exp.
Biol. Med. 143, 1113-1116.
Reiter, L. W., Anderson, G. E., Laskey, J. W., and Cahill, D. F. (1975a). Devel-
opmental and behavioral changes in the rat during chronic exposure to lead.
Environ. Health Perspect. 12, 119-123.
Reiter, L., Talens, G., and Woolley, D., (1973). Acute and subacute Parathion
treatment: Effects on cholinesterase activities and learning in mice. Toxi-
col. Appl. Pharmacol. 25, 582-588.
Reiter, L, Talens, G., and Woolley, D. (1975b). Parathion administration in
the monkey: Time course of inhibition and recovery of blood cholines-
terases and visual discrimination performance. Toxicol. Appl. Pharmacol.
33, 1-3.
Revzin, A. M. (1970). Some acute and chronic effects of endrin on the brains
of pigeons and monkeys. Biological Impact of Pesticides, Environmental
Health Sciences Series No. 1, pp. 134-141. Oregon State University Press,
Corvallis.
Rosenthal, E. and Sparber, S. B. (1972). Methylmercury dicyandiamide: Re-
tardation of detour learning in chicks hatched from injected eggs. Life
Sci. 11,883-893.
Russel, R. W., Watson, R. H., and Frankenhaeuser, M. (1961). Effects of
chronic reduction in brain cholinesterase activity on acquisition and extinc-
tion of a conditioned avoidance response. Scand. J. Psycho/. 2, 21-29.
Salvaterra, P., Lown, B. A., Morganti, J., and Massaro, E. J. (1973). Alterations
in neurochemical and behavior parameters in the mouse induced by low
doses of methyl mercury. Acta pharmacol. toxicol. 33, 177-190.
Sauerhoff, M. W. and Michaelson, I. A. (1973). Hyperactivity and brain cate-
cholamines in lead-exposed developing rats. Science 182, 1022-1024.
Shapiro, M. M., Tritschler, J. M., and Ulm, R. A. (1973). Lead contamination:
Chronic and acute behavioral effects in the albino rat. Bull. Psychon. Soc.
2, 94-96.
Silbergeld, E. K. and Goldberg, A. M. (1973). A lead-induced behavior dis-
order. Life Sci. 13, 1275-1283.
Silbergeld, E. K. and Goldberg, A. M. (1974a). Lead-induced behavior dysfunc-
tion: An animal model of hyperactivity. Exp. Neurol. 42, 146-157.
Silbergeld, E. K. and Goldberg, A. M. (1974b). Hyperactivity: A lead induced
behavior disorder. Environ. Health Perspect. 7, 227-232.
Silbergeld, E. K. and Goldberg, A. M. (1975). Pharmacological and neurochem-
ical investigations of lead-induced hyperactivity. Neuropharmacology 4
431444.
3-19
-------
Smith, R. M. Cunningham, W. L, and Van Gelder, G. A. (1976). Dieldrin toxi-
city and successive discrimination reversal in squirrel monkeys (Saimiri
sciureus). J. Toxicol. Environ. Health 1, 737-747.
Snowden, C. T. (1973). Learning deficits in lead-injected rats. Pharmacol. Bio-
chem. Behav. 1, 599-603.
Sobotka, T. J. (1971). Behavioral effects of low doses of DDT. Proc. Soc. Exp.
Biol. Med. 137, 952-955.
Sobotka, T. J. and Cook, M. P. (1974). Postnatal lead acetate exposure in rats:
Possible relationship to minimal brain dysfunction. Amer. J. /Went. Def. 79,
5-9.
Sobotka, T. J., Cook. M. P., and Brodie, R. E. (1974), Effects of perinatal
exposure to methyl mercury on functional brain development and neuro-
chemistry. Biol. Psychiatry 8, 307-320.
Sobotka, T. J., Brodie, R. E., and Cook, M. P. (1975). Psychophysiologic
effects of early lead expsoure. Toxicology 5, 175-191.
Spyker, J. M., Sparber, S. B., and Goldberg, A. M. (1972). Subtle consequences
of methylmercury exposure: Behavioral deviations in offspring of treated
mothers. Science 177, 621-624.
Spyker, J. M. (1975). Behavioral teratology and toxicology. In Behavioral
Toxicology (B. Weiss and V. G. Laties, eds.), pp 311-350. Plenum Press,
New York.
Terzin, A. L. and Vujkov, V. V. (1968). Alterations of the locomotor activity
of mice poisoned with sublethal doses of lead acetate. Arh. hig. rada. 19,
493-511.
Van Gelder, G. A. (1975). Behavioral toxicologic studies of dieldrin, DDT, and
ruelene in sheep. In Behavioral Toxicology (B. Weiss and V. G. Laties, eds.),
pp. 217-240. Plenum Press, New York.
Van Gelder, G. A., Carson, T., Smith, R. M., and Buck, W. B. (1973). Behav-
ioral toxicologic assessment of the neurologic effect of lead in sheep, din.
Toxicol. 6,405-418.
Vitulli, W. F. (1974). Mercury effects from chronic and acute doses on fixed
interval operant behavior of female squirrel monkeys. Psychol. Rep. 35, 3-9.
Woolley, D. (1970). Effects of DDT on the nervous system of the rat. Biologi-
cal Impact of Pesticides, Environmental Health Sciences Series No. 1, pp.
114-124. Oregon State University Press. Corvallis.
Young, R. D. 11967V Deve\opmenXa\ psvchopharmaco\ogv. A beg'mning. Psy-
cho/. Bull. 67, 73-86.
3-20
-------
Zenick, H., (1978a). Deficits in fixed interval performance following prenatal
and postnatal lead exposure. Develop. Psychobiol. In press.
Zenick, H. Padich, R., Thatcher, T., Santistevan, B., and Aragon, P. (1978b).
Influence of prenatal and postnatal lead exposure on discrimination learning
in rats. Pharmac. Biochem. Beh. 8, 347-350.
Zenick, H. (1976). Evoked potential alterations in methylmercury chloride
toxicity. Pharmacol. Biochem. Behav. 5, 253-255.
Zenick, H. (1974). Behavioral and biochemical consequences in methylmercury
chloride toxicity. Pharmacol. Biochem. Behav. 2,709-713.
3-21
-------
4. METHODOLOGICAL PROBLEMS
ASSOCIATED WITH THE EXPOSURE
OF NEONATAL RODENTS TO LEAD
ROBERT L BORNSCHEIN AND I.ARTHUR MICHAELSON
Laboratories of Behavioral and Neurochemical Toxicology
Division of Toxicology
Department of Environmental Health
University of Cincinnati College of Medicine
Introduction
The more dramatic, harmful effects of pediatric lead encephalopathy are well
recognized. Ample supporting evidence of a dangerous body burden can be
found whenever the blood lead exceeds 80 to 100 fig lead per 100 ml. The
central nervous system is a prime target following acute high level exposures.
Inhibition of heme synthesis and proximal tubular injury can also be detected.
Until recently lead poisoning has been defined as a public health problem only
for those children living in inner city metropolitan slum areas. However, the
definition of the population at risk needs re-evaluation. The concentration of
inorganic lead in our immediate environment is on the increase (Patterson,
1965). Young children are being chronically exposed to low levels of lead in
their diets, in the air they breathe, and in the soil and dust which are found
in their play areas (Cohen et a/., 1973; Vostal et a/., 1974). This chronic low-
level exposure is being reflected in elevated blood lead levels in the range of
30 to 60 jug/100 ml (Caprio et a/., 1974). Furthermore, recent studies indi-
cate that such elevated blood lead levels are not uncommon in children growing
up in rural areas (Fine eta/., 1972). Therefore the population at risk as a result
of undue lead exposure is not restricted to just those children living in the
inner city.
*Supported by NIEHS Grants ES-02614 (RLB) and ES-01077 (IAM).
4-1
-------
A major question from the public health point of view is: What effect does
excessive lead absorption have on mental development among those who
never develop encephalopathy or any other overt symptoms of poisoning? At
present there is good reason to believe that adverse effects may be particularly
marked when exposure occurs during the period of central nervous system
development in early childhood. For example, Byers and Lord (1943) reported
retarded mental development in lead-poisoned children without encephalo-
pathy, and Albert et al. (1974) found deficits in intellectual maturation in
asymptomatic lead-exposed children. However, we have very little reliable
information concerning the potential neurotoxic effects of chronic low-level
lead exposures, even though there have been indications from clinical studies
that such exposures during early childhood may produce behavioral and intel-
lectual impairment (de la Burde and Choate, 1972).
Clinical studies which attempt to derive data from asymptomatic pediatric
populations are fraught with numerous methodological problems. For example,
although a child may exhibit an elevated blood lead level at the time of exami-
nation, we usually have no indication of that past history of lead exposure,
either with respect to the maximum level obtained, the duration of the lead
exposure, or the timing of the exposure relative to various critical aspects of
central nervous system development. Furthermore, we usually have no measure
of the child's intellectual development prior to the onset of lead exposure.
Because of these problems, there has been considerable research effort devoted
to the development of an animal model for asymptomatic neonatal lead expo-
sure. Most experimental animals are resistant to the effects of lead on the
brain. A review of the literature on this subject shows an amazing resistance of
the adult animals to lead poisoning (Pentschew, 1958). However, our work
and that of others indicate that if lead is given to newborn rodents, they suffer
impairment in central nervous system function.
Table 4-1 provides a brief overview of some of the animal models used to in-
vestigate the effects of lead on the developing nervous system. It is apparent
that there is a great diversity in exposure parameters such as route, level, and
duration as well as timing of exposure relative to age. Unfortunately, the
blood, brain, and whole body burdens which result from these various expo-
sure techniques have not been well documented. In addition, the range of
behaviors examined in lead-exposed animals has been rather limited. This limi-
tation is unfortunate since, as mentioned previously, animal models may pro-
vide insight into the relationships between level and timing of lead exposure
and resultant body burden, behavioral effects, and duration of effects. The
purpose of this paper is to identify some of the inherent methodological
problems associated with experimental neonatal lead exposure. We also intend
to suggest strategies which should be of use in overcoming some of these
problems.
Methods of Lead Exposure
A common method of exposure is to treat the neonate by daily oral intuba-
tions, permitting a relatively accurate estimate of daily exposure. This tech-
nique has several inherent disadvantages, including: (a) delivery of a single
4-2
-------
Table 4-1. PROTOCOLS USED TO STUDY EFFECTS OF LEAD ON ANIMAL BEHAVIOR
CO
Reference
Brown eta/., 1971
Brown, 1975
Brady eta/., 1975
Goiter 81
Mlchaelson, 1975
Mlchaelson &
Sauerhoff, 1974
Reiter eta/., 1975
Shapiro eta/., 1973
Snowden, 1973
Sobotka ef al., 1975
Morrison et al., 1975
Silbergeld Si
Goldberg, 1973
Terzin Si Vujkov, 1968
Carson eta/., 1974
Van Gelder eta/., 1973
Allen eta/., 1974
Goods &
Calandra, 1973
amg/kg of Pb ip
bmg/kg of Pb po
cppm of Pb in diet
dppm of Pb in water
No.
1
2
3
4
5
6
7
a
9
10
11
12
13
14
15
16
R =
M =
S =
P =
Lactatlng Pb concentration JLtg%
dam Neonate Weanling Adult Blood Brain M||k
R 100a
R 4.1a 198-257
R 500b
R 1b 40C 225 48
R 27,300° 25C 50 2500
R 5 & 50d 5 Si 50d 5 & 50
R 0.16-0.96a
R 5, 8, 12a 5, 8, 12a
R 5, 15, 44b 18-71
M 1090-5500d 1090-5500d 1090-5500d
M 1090-5500d 1090-5500d 1090-5500d
M 7-36a
S 5,19,35
S 55b
P 2.1-5.1° 100-500
P 0.05, 0.5 0.05, 0.5 45-82
Si5.0b & 5.0b
Rat
Mouse
Sheep
Monkey
Behavioral
measu res
Water maze and
shuttle box learning
T-maze learning,
locomotor activity
Water maze learning
Locomotor activity
Locomotor activity
Reflex development.
locomotor activity
V.I. 30 sec operant
discrimination
Hebb-Wllllams maze
Locomotor activity,
shuttle box learning
Consumatory behavior
Locomotor activity
Locomotor activity
Visual discrimination
learning
Auditory signal
detection task
Social behavior
Delayed and condi-
tioned response
-------
bolus of the agent with a short transit time through the gastrointestinal tract,
(b) large periodicity in resultant blood lead concentrations, (c) trauma asso-
ciated with daily forced feeding, and (d) considerable expenditure of time and
effort by the experimenter. Our interest in chronic exposure and the detection
of any resultant behavioral alterations precludes such a method of exposure.
For these reasons we have chosen to expose neonatal rodents to toxic agents
via the dam's milk. This approach is currently utilized by several investigators
exposing rats and mice to inorganic lead (Brown, 1975; Maker et al., 1975;
Silbergeld and Goldberg, 1973; Goiter and Michaelson, 197.5), DDT (Fahim
et al., 1970), and methylmercury (Brown et al., 1972). Although this indirect
method of exposure mimics the usual oral route of entry into the body and
avoids the disadvantages of oral intubation, it is difficult to estimate the daily
lead exposure of the rapidly developing neonate. This difficulty in turn has
hampered development of animal models with realistic levels of exposure
relative to the human pediatric population. Furthermore, it has been difficult
to compare this model to others using direct injection or intubation proce-
dures.
Investigators who use the lactating dam as a vehicle for delivery of lead usually
report only the concentration of the agent in the dam's drinking water or food
(Brown, 1975; Maker et al., 1975; Silbergeld and Goldberg, 1973; Goiter and
Michaelson, 1975). Few report milk lead concentrations (Michaelson and
Sauerhoff, 1974; Goiter and Michaelson, 1975). None have attempted to esti-
mate the daily amount of lead ingestion by the suckling. While it is relatively
easy to determine milk lead concentrations, it is much more difficult to esti-
mate the amount of milk consumed by the suckling. There are several tech-
niques available for estimating milk production by the lactating dam. These
include: (a) determination of differences in weight between freshly dissected,
milk-filled glands of dams separated from their litter for a fixed interval
(usually 8 hours) and the mean weight of mammary glands from unseparated
control animals at the same stage of lactation (Hanwell and Linzell, 1972),
(b) biochemical determination of lactose content of mammary tissues in
suckled versus nonsuckled lactating dams (Mutch and Hurley, 1974), and (c)
maximum milk yield during artificial milking after a fixed interval of sepa-
ration from pups (Hanrahan and Eisen, 1970). These techniques provide only
an indirect estimate of how much the neonates may be ingesting. The most
commonly employed technique for directly estimating milk consumption is
to allow the pups to remove milk from the dam and then determine the
amount suckled during a specified interval either by estimating the changes in
body weight before and after suckling, or by removing and weighing the
stomach and its contents at completion of the suckling period (Grosvenor and
Turner, 1956, 1957, 1959; Grosvenor and Mena, 1967; Moon and Turner,
1959; Mena et al., 1974). Another approach is to hand-feed (Miller and
Dymsza, 1963) or infuse (Messer et al., 1969) the neonates with normal rat
milk and determine the quantity per 24 hours necessary for maintenance of a
normal growth pattern. Intubation or infusion techniques probably result in an
overestimate: of actual milk consumption by normal suckling neonates. Over-
estimates can result from: (a) increased caloric requirements arising from ex-
cessive handling during artificial feeding, (b) ineffective utilization of caloric
intake due to differences between artificial patterns of feeding and natural
4-4
-------
daily rhythms in suckling behavior (Grota and Ader, 1974; Redman and
Sweeney, 1976), and (c) differences between the caloric values of infused or
intubated milk, usually obtained during peak lactation periods (10 to 12 days,
1.3 Kcal/ml) by artificial milking and that milk obtained by neonates during
natural suckling (approximately 1.9, 1.3, 1.5 Kcal/ml during weeks 1, 2, and 3
respectively) (Bornschein et a/., 1977a). It is important to emphasize that
thermoregulatory mechanisms are not well developed in newborn rats and do
not become operational until after the first week of life (Fowler and Kellogg,
1975). Close contact between litter and dam is necessary for maintenance of
normal body temperature. The procedure of intubation or infusion upsets the
normal litter-dam relationship, and this insult can be amplified by artificial
feeding of late lactation (low calorie) rat milk, when milk normally produced
in the early stages of lactation is of relatively high caloric content.
Because of these various methodological drawbacks, we sought a different
approach to the estimation of daily milk consumption. As a result of this
search, we have developed a way by which one can make a fairly good approxi-
mation of the volume of milk and, consequently, the amount of lead consumed
each day (Fox et a/., 1975; Michaelson et al., 1976; Bornschein etal., 1977b).
The volume of milk consumed each day can be estimated with knowledge of:
(a) the caloric requirements necessary to maintain normal daily growth during
the suckling period and (b) the caloric value of milk during lactation. Esti-
mated caloric requirements per gram body weight in neonatal rats range from
0.29 Kcal on day 1 to 0.53 teal on day 18. Caloric values of milk range from
2.62 Kcal/ml to 1.48 Kcal/ml on the 1st through 18th day of lactation. Neo-
nates are highly efficient in extracting energy from the milk, the efficiency of
utilization being over 90 percent. Estimated milk consumption per suckling
pup increases from 0.7 ml on day 1 to 17.1 ml on day 18. Details pertaining to
the above discussion have been published (Bornschein etal. 1977b).
A sample calculation appears below:
Grams of milk = Body wt ^ x caloric requirement (Kcal/g)
Caloric value of milk (Kcal/g) x eff. utilization
For example, on day 5: Neonate's body wt = 12 g
Caloric requirement = 0.29 Kcal/g
Caloric value of milk = 1.25 Kcal/ml
Efficiency of utilization = 0.9
Grams of milk = 02 9> (0.29 Kcal/g) = 3 Q
1.30 Kcal/ml x 0.9
Our theoretical estimates of daily milk consumption are in good agreement
with our expectation based on limited reports of milk production in rodents.
We estimate that on day 10 of lactation a 300 gram rat nursing six pups yields
45 ml of milk in 24 hours. A recent study by Hanwell and Linzell (1972) on
the rate of milk secretion in rats on the tenth day of lactation indicated that
240-280 gram rats produce 42 ml of milk per day. Thus, our estimates during
4-5
-------
late lactation are in general agreement with the limited data available. We
recognize the need for experimental validation of our theoretical estimates
and have conducted a 24-hour fostering experiment in which normal non-
exposed sucklings received lead from an exposed lactating rat. Analytical
estimation of the lead-load in these sucklings yielded 90 percent to 126 per-
cent recovery relative to the expected value (Table 4-2).
Table 4-2. ACCUMULATION OF LEAD IN 12-13-DAY-OLD RATS FOLLOWING
24 HOURS OF EXPOSURE VIA CONTAMINATED MILK
Pup
No.
1
2
3
4
5
6
A
Body
weight
(9)
28.7
30.6
28.7
27.1
30.9
29.8
B3
Milk
consumed
(g)
10.3
11.2
10.3
9.7
11.0
10.9
C°
MaPb
consumed
B x D
14.9
16,2
14.9
23.8
27.0
26.7
D°.d
Mg Pb/ml
milk
1.45+ 0.056
1.45 + 0.05e
1.45+ 0.05e
2.45 + 0.45f
2.45+ 0,45f
2.45+ 0.45f
E<=
Pb body
burden JUg
(corrected)
15.8
15.1
14.0
24.5
27.8
26.6
F
% Recovery
E/C
106
93
94
103
103
100
100+5
aTheoretical estimate.
bPredicted Intake.
cExperimentally determined.
dTwo milk samples per dam, each sample analyzed in triplicate.
eMjlk lead content of lead-exposed dam No. 1.
fMllk lead content of lead-exposed dam No. 2.
We concluded that this indirect (noninvasive) method is applicable for esti-
mating the actual exposure of the neonate during lactation.
Although estimates of daily milk ingestion appear to be valid for control rats,
it is necessary to consider what effects the presence of lead might have on
milk quality, quantity, or utilization by the neonate. As an indirect indica-
tion of the status of these parameters, food and fluid intake by the dam and
growth rate of the neonates have been closely monitored. The effects of lead
acetate in lactating dams' drinking water were examined over a range of con-
centrations from 0.02 percent to 2.0 percent. We found that alteration in
food and fluid intake occurred at all concentrations above 0.2 percent. At
these same concentrations depression in neonatal growth rate was evident.
However, at a concentration of 0.2 percent lead acetate, all of the above para-
meters were unaltered. Therefore, it would appear that the presence of lead at
the concentration used in this study does not significantly alter the validity of
the assumptions underlying the theoretical estimates of daily lead intake.
4-6
-------
Employing this methodology we have estimated the daily exposure of the
Pentschew and Garro (1966) model wherein neonatal rats suckle dams pro-
ducing 40 ppm lead-containing milk from birth to 21 days of age. The pups
then wean to the dam's diet which contains 31,000 ppm lead. This protocol
is commonly used as an animal model of lead encephalopathy. Prior to wean-
ing, the exposure of the neonate is about 1.5 /ig Pb/g/day (Figure 4-1, upper
solid curve). Following weaning, daily lead exposure jumps to 3,000 jug
Pb/g/day, a 2,000-fold increase.
4% Pb CO3 DIET (PENTSCHEW & GARRO)
1000
100
I 10 h
s
I 1-°
Q.
3 o.i
0.01
0.001
).5% Pb (Ac)2 H20 (SILBERGELD & GOLDBERG)
RAT
MOUSE
J-L—SOBOTKA AND COOK (INTUBATION)
-+• BROWN (ip)
1 OX)2% Pb (Ac)2H20 (MICHAELSON AND BORNSCHEIN)
-------
percent lead acetate results in an estimated exposure of the neonate to about
0.5 //g Pb/g body wt/day, resulting in a blood lead concentration of about
65 M9 percent. The difference in resultant blood lead concentrations is most
likely a reflection of higher absorption levels obtained with the milk vehicle
{Kostial et al., 1971) and of a higher body burden due to a more nearly contin-
uous daily exposure to lead.
Using these estimation techniques, it is also possible to compare exposure levels
in rats and mice with those seen in the human population. Recommendations
pertaining to maximum daily permissible intake D.P.I, for various segments of
the human population range from 4 to 10 Mg Pb/kg body weight (Barltrop,
1973). Most investigators using the lactating dam as a vehicle for lead adminis-
tration are exposing the neonates to lead levels 10 to 100 times higher than the
recommended daily permissible intake for the human population. The ability
to make such comparisons, i.e., between various exposure techniques and
between animal models and the human population, should enhance our ability
to develop relevant animal models of lead exposure for the human pediatric
population.
Our investigations of biochemical and behavioral effects of lead exposure are
being carried out in animals receiving lead at a level approximately one order
of magnitude greater than the level recommended for children (see Figure 4-1).
The exposure level or external dose is of course only one of several important
descriptive indices. Perhaps of greater importance is the internal dose as re-
flected by blood lead concentrations and brain lead concentrations. We have
undertaken an analysis of the transfer of lead from the dam's water to the
blood and brain of the neonate. A summary of the results appears in Table 4-3,
which reveals that the blood lead levels span those concentrations of greatest
interest, i.e., 20 to 80 M9 percent.
Table 4-3. LEAD CONCENTRATION IN RAT BLOOD AND BRAIN AFTER
20 DAYS OF LACTATIONS
Concentration of
Lead acetate In
dam's water
Tap water
0.02%
0.10%
0.20%
0.50%
Dam's blood
lead (#g%)
9+3
24+5
41+7
69+21
104+ 2
Dam's milk
lead (Atg%)
<1
21+2
87+7
139 + 11
264 + 40
Pup's blood
lead (jUg%)
11 + 4
29+5
42+4
65 + 25
137 +, 7
Pup's brain
lead (jUg%)
12+2
29+7
47 + 10
65 + 6
1 36 +_ 3 1
aValues represent X + SEM.
We have also assessed blood and brain lead levels at various points following
termination of exposure at day 20. Tables 4-4 and 4-5 summarize the effects
seen in rats and mice, respectively. There are several points of interest in
Table 4-5. First, irrespective of the treatment group, i.e., acute or chronic
exposure, the blood lead concentrations were highest at the end of the lacta-
tion period which is probably a reflection of the high absorptive capacity of
4-8
-------
the neonate. In the acutely exposed group, blood lead concentrations declined
but did not reach control values. The chronically exposed group tended to
reach steady state blood lead concentrations of about 120-140 jug percent. A
second interesting finding is that brain lead levels in the acutely exposed group
slowly decreased following termination of exposure. In the chronically exposed
group, lead continued to accumulate in brain tissue despite a relatively steady-
state blood lead. The potential pitfalls in any attempts to relate a specific
biochemical or behavioral effect to a single determination of blood lead con-
centration at a single point in time are readily apparent.
Table 4-4. LEAD CONCENTRATION IN RAT BLOOD AND BRAIN
AGE: 20, 60, and 270 DAYS
Concentration of
lead acetate
Tap water
0.02%
0.1%
Tap water
0.02%
0.1%
Blood lead concentration (jllg%)
20
11 +_ 4a
29 +_ 5
42 ±_ 4
60
4 + 1
5+0
9+3
270
9 +.0.3
9 ±0.7
9 +_ 1.0
Brain lead concentration (;Ug%)
12 +_ 2
29 +_ 7
47 ± 10
14+0
8+0
14+0
12 +_ 1
14 +_ 0.3
20 +_ 4
aN = 3 to 5 subject for each treatment x age group.
Values represent X+_ SEM.
Table 4-5. LEAD CONCENTRATION" IN MICE AS A FUNCTION OF
AGE AND TREATMENT
Concentration of lead in dam's water
Age, days
15
20
40
100
(0.05
Blood
15
17
4
ppm Pb
— Brain
5
5
4
109 pprr
Blood —
42
79
29
20
i Pbb
Brain
16
24
24
10
2780 ppm Pbc
Blood — Brain
190 200
120 306
143 584
aPooled samples—4-5 mice per sample—reported asjUg%,
Exposure via maternal milk only — exposure terminated at day 21.
°Continuous exposure — during first 3 weeks via maternal milk and postweanlng via
dams water supply.
4-9
-------
Lead Exposures and Nutritional Status
The effect of lead ingestion on the normal pattern of food and water intake by
rodents has been well documented (Michaelson and Sauerhoff, 19/4; Goiter
and Michaelson, 1975; Morrison eta/., 1975; Maker eta/., 1975). Furthermore,
we have recently reported on the effects of developmental undernutrition on
motor activity, a behavioral parameter also reported to be affected by lead
exposure (Loch et a/., 1976). The results of these studies indicate the need to
consider the dietary status of animals as a major variable in experimental de-
sign. Dietary factors can be regarded as both a previously uncontrolled variable
which can confound the interpretation of experimental findings and as a po-
tential independent variable which adds a critically important dimension to the
development of an animal model for asymptomatic exposure.
In demonstrating the role of dietary status as an uncontrolled variable in most
lead toxicity studies, it should first be pointed out that numerous dietary con-
stituents have been shown to influence lead metabolism. The list is extensive
and includes: protein and fat content; trace metals, e.g., iron, copper, and zinc;
vitamins, e.g., vitamin D; and essential minerals, e.g., calcium and phosphate
(Barltrop and Khoo, 1975).
Absolute nutrient concentrations as well as concentrations relative to other
dietary components can drastically alter the toxic effects which may ensue
following exposure to lead. Commercially available animal chows are charac-
teristically oversupplemented with essential nutrients. Often they exceed by
5- to 10-fold the levels recommended by the National Research Council (1972)
for normal growth and development. The concentration of any particular
dietary constituent may vary considerably between different manufacturers as
well as between different lots of chow obtained at different points in time
from the same manufacturer. However, concentrations seldom fall below the
NRC recommendations. The extent of variability between different suppliers
and the extent of oversupplementation can be seen in Table 4-6. In this table
are listed several dietary constituents known to influence lead metabolism.
Recommendations of the NRC can be compared with dietary concentrations
reported by several manufacturers. Table 4-7 illustrates the extent of the varia-
bility between advertised metal content and that actually found in various
lots obtained over the course of 3 years. Note also the high concentrations of
lead and cadmium in successive shipments from the same supplier. The obvious
conclusion to be drawn from these results is that it is essential that investi-
gators control, document, and report those aspects of the animal's diet which
are known to influence lead metabolism. Until this is done, we will continue
to encounter difficulties replicating work within our laboratories as well as that
work carried out in other laboratories.
4-10
-------
Table 4-6. NUTRIENT COMPOSITION OF COMMERCIAL RAT DIETS
Dietary
Constituent
Fat (%)
Ca (%)
Cu (ppm)
Fe (ppm)
Zn (ppm)
Vit D (I.U./g)
NRC
Requirements
7.6
0.6
5.6
39.0
13.0
1.11
Supplier number
1
9.0
1.2
18.0
198.0
58.0
3.3
2
6.0
2.0
13.0
370.0
38.0
5.1
3
7.0
1.1
15.5
370,0
2.4
4
10.0
0.9
6.5
37.0
20.0
4.0
Table 4-7. TRACE METAL AND HEAVY METAL CONTENT OF
COMMERCIAL DIET NO. 1a
Metal content (ppm)
NRC
Advertised
Sample 1 (1974)
2 (1975)
3 (1975)
4 (1975)
5 (1975)
6 (1976)
Zn
13
58
53
55
60
68
58
54
Cu
6
18
11
17
17
13
13
9
Cr
1.4
1.8
2.1
1.3
2.4
Mn
56
51
60
62
60
57
53
Fe
39
198
278
209
465
475
415
363
Cd
0.1
1.3
0.8
0.9
1.1
Pb
3.0
2.8
3.1
3.0
2.5
1,0
aUnpublished data from Dr. L. Murthy and
Environmental Health, University of Cincinnati.
H. Petering of the Department of
In addition to concerns pertaining to dietary status as a control variable, we
also argue for its consideration as a significant independent variable in any
attempt to model asymptomatic pediatric lead toxicity. As noted previously,
rats consuming commercial lab chows are being subjected to a diet which is
heavily fortified with essential nutrients. Such diets are seldom encountered in
the human population. In fact, it has been well documented that a substantial
portion of our population is deficient in one or more of the nutrients con-
sidered essential for normal health. The parameter of greatest concern with
respect to the lead problem may be iron deficiency. The recent Ten State Nu-
tritional Survey (1968-1970) indicates that in children under three years of
age, 50 percent were deficient in iron intake irrespective of ethnic group or
socio-economic status. Approximately 40 percent of this population had iron
intakes less than one-half the Recommended Dietary Allowance. The findings
are of considerable importance because of the demonstrated biochemical inter-
actions of lead and iron in important biochemical pathways involving electron
transport, catecholamine metabolism, and porphyrin synthesis. Similar argu-
ments could be made for several other dietary constituents.
Iron deficiency alone can produce significant biochemical and behavioral
changes, many of which bear a striking resemblance to the clinical picture
4-11
-------
obtained with elevated lead exposure (Pollit and Leibel, 1976). Dietary intake
of other trace metals, e.g., zinc and copper, is also considered to be marginal in
some segments of our population. Animal models of asymptomatic lead expo-
sure should provide the capability for controlling and manipulating dietary
status. Dietary manipulations in conjunction with various levels and durations
of lead exposure can reveal much about the mechanisms of lead neurotoxicity
as well as suggest possible methods for preventing or correcting lead's effects
on critically sensitive segments of the population. For these reasons our
animals are maintained on a semipurified diet which closely approximates the
NRC recommendations.
During the 21-day period prior to weaning, the neonate's nutritional status is
directly related to the amount of milk consumed per unit time per gram body
weight and the concentration of essential nutrients in the dam's milk at that
time. Utilizing the same methodology as that used to estimate daily lead in-
take, it is possible to estimate the daily intake of iron, copper, zinc, or other
essential nutrients by the neonate at any point during the lactation period. For
example, we have determined the zinc content of control dam's milk at 5-day
intervals during lactation. These data, together with estimated daily milk con-
sumption, permit us to estimate daily zinc intake during the lactation period.
Table 4-8 illustrates the result of such a calculation. By employing such a pro-
cedure, it is possible to define the daily intake of critical nutrients such as
iron, copper, zinc, and other metals known to interact with lead. It may also
be possible to manipulate the dam's diet in order to produce specific trace
metal deficiencies in the neonates. Such manipulations may permit an evalu-
ation of lead exposure in conjunction with specific dietary deficiencies in the
neonate.
Table 4-8. ESTIMATED DAILY ZINC INTAKE
Day of lactation
Pup body wt (g)
Zinc content of milk (ppm)
Milk ingestion (g)
Zinc Intake (jUg absolute)
Zinc Intake (jUg/g body wt)
5
12
17
2.8
47.6
4.0
10
24
13
7.4
96.2
4.0
15
37
11
12.8
140.1
3.8
20
50
10
18.2
182.0
3.6
Unintentional Sources of Exposure
As indicated in Table 4-6, commercial laboratory chows contain considerable
concentrations of lead (Fox et a/,, 1976). This concentration poses several
problems in regard to the preparation of adequate control animals. Long-term
consumption of these chows can result in an elevated lead body burden. During
pregnancy, the fetus is subjected to unintentional and, at this point, unpre-
ventable exposure to lead. Furthermore, at parturition the neonates are sub-
jected to a substantial increase in lead body burden as a result of elevated lead
concentration in dam's milk. Finally, beginning on about day 17 of life, the
neonate begins to consume solid food, producing about a 15-fold elevation in
4-12
-------
daily lead exposure after weaning if the chow contains 0.5 ppm lead (see
Figure 4-1). Since gastrointestinal absorption is still very high in the 17-to 30-
day-old rat (Forbes and Reina, 1972), a marked elevation in body burden can
occur. Of far greater consequence is the lead exposure which can result if the
15- to 20-day-old rat pup should gain access to the dam's lead-containing water
supply. For example, a dam drinking 0.2 percent lead acetate in water pro-
duces milk containing about 1.4 ppm lead. An 18-day-old rat pup consuming
this milk is exposed to about 0.5 jug lead/g body wt/day. If this pup should
ingest one ml of the dam's water supply, its exposure is increased to 22 jug
lead/g body wt/day. Fortunately, this type of unintended lead exposure can
be prevented with the appropriate choice of cages. The greatest potential
source of unintended lead exposure is that lead which is introduced into the
neonate's environment by the dam. The lactating dam drinking 0.2 percent
lead acetate-water consumes about 80 ml of water or about 87,200 jug of lead
per day. The dam retains less than 5 percent of this total (Kostial eta/., 1971).
The remainder, about 82,000 ;ug of lead, is excreted through the urine and
feces into the neonate's immediate environment. If cage bedding is changed
only every 5 days, lead contamination can increase to 410,000 nq of lead. How
much of this lead is available for intake by the neonate is unknown. Most of
this lead is in fecal material, and some is adsorbed onto the surface of the
bedding material. Clean ground corn cob bedding contains 0.06 ppm lead.
After 5 days, the bedding, free of fecal material, contains 36.0 ppm lead. The
high level of licking, preening, and oral exploratory behavior in which neonatal
rats normally engage can result in the unwanted transfer of some of this lead
into the neonate. In experiments where bedding material was changed only
once a week, we found that within 3 weeks the neonates had a lead body
burden 4 times greater than would be predicted on the basis of lead intake
via milk only.
In this paper we have attempted to highlight some of the major problems
associated with the experimental exposure of neonatal animals to inorganic
lead. It is our hope that appropriate attention to these problems will lead to
a further refinement of current models of low-level chronic neonatal lead
exposure. We further hope that our methodology for estimating daily expo-
sure levels will generate new dialogue concerning the adequacy and relevance of
animal models currently used to assess the effects of low-level lead exposure
on the developing nervous system.
References
Albert, R. E., Shore, R. E., Sayer, A. J., Strethlow, C., Kneip, T. J., Pasternack,
B. S., Friedhoff, A. J., Covan, F., and Cimino, J. A. (1974). Followup of
children overexposed to lead. Environ. Health Perspect. 7, 33-40.
Allen, J. R., McWay, P. J., and Suomi, S. J. (1974). Pathobiological and behav-
ioral effects of lead intoxication in the infant rhesus monkey. Environ.
Health Perspect. 1,239-246.
Barltrop, D. (1973). Sources and significance of environmental lead for chil-
dren. In Proceedings of International Symposium: Environmental Health
4-13
-------
Aspects of Lead. Amsterdam, Netherlands, Oct. 2-6, 1972, pp. 675-681.
Commission European Committees, Luxembourg.
Barltrop, D. and Khoo, H. E. (1975). The influence of nutritional factors on
lead absorption. Postgrad. Med. J. 51, 795-800.
Bornschein, R. L, Fox, D. A., and Michaelson, I. A. (1977a). Estimation of
daily exposure in neonatal rats receiving lead via dam's milk. Toxicol. Appl.
Pharmacol. 40, 577-587.
Bornschein, R. L, Michaelson, I. A., Fox, D. A., and Loch, R. (1977b). Evalu-
ation of animal models used to study lead effects on neurochemistry and
behavior. In Biochemical Effects of Environmental Pollutants: EPA Sympo-
sium, Cincinnati, Ohio, June 1-3, 1976. Ann Arbor Science Publishers, Inc.,
Ann Arbor, Michigan.
Brady, K., Herrera, Y., and Zenick, H. (1975). Influence of parental lead expo-
sure on subsequent learning ability of offspring. Pharmacol. Biochem.
Behav. 3, 561-565.
Brown, S., Dragann, N., and Vogel, W. H. (1971). Effects of lead acetate on
learning and memory in rats. Arch. Environ. Health. 22, 370-372.
Brown, R. V., Zenick, H., Cox, V., and Fahim, M. S. (1972). The effects of
methylmercury chloride on maze learning in mice. Fed. Proc. 31, 552.
Brown, D. R. (1975). Neonatal lead exposure in the rat: Decreased learning as
a function of age and blood lead concentrations. Toxicol. Appl. Pharmacol.
32, 628-637.
Byers, R. K. and Lord, E. E. (1943). Late effects of lead poisoning on mental
development. Amer. J. Dis. Child. 66, 471-494.
de la Burde, B. and Choate, M. S. (1972). Does asymptomatic lead exposure in
children have latent sequelae? J. Pediatr. 81, 1088-1091.
Caprio, R. J., Margulis, H. L., and Joselow, M. M. (1974). Lead absorption in
children and its relationship to urban traffic density. Arch. Environ. Health.
28, 195-197.
Carson, T., Van Gelder, G., Buck, W., and Hoffman, L. (1973). Effects of low
level lead ingestion in sheep. Clin. Toxicol. 6, 389-403.
Carson, T. L., Van Gelder, G. A., Karas, G. C. and Buck, W. B. (1974). Slowed
learning in lambs prenatally exposed to lead. Arch. Environ. Health 29,
154-156.
Cohen, C. J., Bowers, G.-N., and Lepow, M. L. (1973). Epidemiology of lead
poisoning: a comparison between urban and rural children. J. Amer. Med.
Assoc. 226, 1430-1433.
4-14
-------
Fahim, M. S., Bennet, R., and Hall, D. G. (1970). Effects of DDT on the
nursing neonate. Nature 228, 1222-1223.
Fine, P. R., Thomas, C. W., Suhs, R. J., Cohnberg, R. E., and Flashner, B. A.
(1972). Pediatric blood lead levels: A study in 14 Illinois cities of interme-
diate population. J. Amer. Med. Assoc. 221, 1475-1479.
Forbes, G. B. and Reina, J. C. (1972). Effect of age on gastrointestinal absorp-
tion (Fe, Sr, Pb) in the rat. J. Nutr. 102, 647-652.
Fowler, S. J. and Kellogg, C. (1975). Ontogeny of thermoregulatory mecha-
nism in the rat. J. Comp. Physiological Psychology 89, 738-746.
Fox, D. A., Michaelson, I. A., and Bornschein, R. L (1975). A method for
estimating degree of exposure to toxic agents transmitted to neonates via
maternal milk: lead. Pharmacologist 17, 212.
Fox, J. G., Aldrich, F. D., and Boylen, G. W., Jr. (1976). Lead in animal foods.
J. Toxicol. Environ. Health 1, 461-467.
Goode, J. W., Johnson, S., and Calandra, J. C. (1973). Evaluation of chronic
oral administration of lead acetate to rhesus monkeys. Toxicol. Appl.
Pharmacol. 24, 53.
Goiter, M. and Michaelson, I. A. (1975). Growth, behavior, and brain catechol-
amines in lead-exposed neonatal rats: a reappraisal. Science 187, 359-361.
Grosvenor, C. E. and Turner, C. W. (1956). Ergotamine, oxytocin and milk
let-down in lactating rat. Proc. Soc. Exp. Biol. Med. 93, 466-468.
Grosvenor, C. E. and Turner, C. W. (1957). A method for evaluation of milk
"let-down" in lactating rat. Proc. Soc. Exp. Biol. Med. 94, 816-817.
Grosvenor, C. E. and Turner, C. W. (1959). Effect of growth hormone and
oxytocin upon milk yield in the lactating rat. Proc. Soc. Exp. Biol. Med.
100, 158-161.
Grosvenor, C. E. and Mena, F. (1967). Effect of auditory, olfactory and optic
stimuli upon milk ejection and suckling-induced release of prolactin in
lactating rats. Endocrinology 80, 840-846.
Grota, L. J. and Ader, R. (1974). Behavior of lactating rats in a dual-cham-
bered maternity cage. Hormones and Behavior 5, 275-282.
Hanrahan, J. P. and Eisen, E. J. (1970). A lactation curve for mice. Lab. Anim.
Care 20, 101-104.
Hanwell, A. and Linzell, J. L. (1972). A simple technique for measuring the
rate of milk secretion in the rat. Comp. Biochem. Physio/. 43A, 259-270.
Kostial, K., Simonovic, I., and Pisonic, M. (1971). Lead absorption from the
intestine in newborn rats. Nature 233, 564-565.
4-15
-------
Loch, R., Bornschein, R. L, and Michaelson, I. A. (1976). Role of undernutri-
tion in the paradoxical response of lead exposed "hyperactive" mice to
amphetamine and phenobarbital. Pharmacologist 18, 124.
Maker, H. S., Lehrer, G. M., and Silides, D. J. (1975). The effect of lead on
mouse brain development. Environ. Res. 10, 76-91.
Mena, R., Beyer, C., and Grosvenor, C. E. (1974). On the mechanism by which
oxytocin depresses milk ejection and milk secretion in rats. Amer. J. Phy-
sio/. 227, 1249-1254.
Messer, M., Thoman, E. B., Terrasa, A. G., and Dallman, P. R. (1969). Artifi-
cial feeding of infant rats by continuous gastric infusion. J. Nutr. 98,
404-410.
Michaelson, I. A. and Sauerhoff, M. W. (1974a). An improved model of lead-
induced brain dysfunction in the suckling rat. Toxicol. Appl. Pharmacol.
28, 88-96.
Michaelson, I. A. and Sauerhoff, M. W. (1974b). Animal models of human
disease: Severe and mild lead encephalopathy in the neonatal rat. Environ.
Health Perspect. 7, 201-225.
Michaelson, I. A., Bornschein, R. L, and Fox, D. A. (1976). Verification of
theoretical estimation of daily lead (Pb) exposure in suckling neonatal
rodents. Pharmacologist 18, 125.
Miller, S. A. and Dymsza, H. A. (1963). Artificial feeding of neonatal rats.
Science 141, 517-518.
Moon, R. C. and Turner, C. W. (1959). Effects of reserpine on oxytocin and
lactogen discharge in lactating rats. Proc. Soc. Exp. Bio/. Meet. 101, 332-
335.
Morrison, J. H., Olton, D. S., Goldberg, A. M., and Silbergeld, E. K., (1975).
Alterations in consumatory behavior of mice produced by dietary exposure
to inorganic lead. Devel. Psychobiol. 8, 389-396.
Mutch, P. B. and Hurley, L. S. (1974). Effect of zinc deficiency during lacta-
tion on postnatal growth and development of rats. J. Nutr. 104, 828-842.
National Research Council (1972). Number 10: Nutrient requirements of
laboratory animals, 2nd rev. ed. National Academy of Sciences, Washington,
D. C.
Patterson, C. C. (1965). Contaminated and natural lead environments of man.
Arch. Environ. Health 11, 344-363.
Pentschew, A. (1958). Intoxikationen: In Handbuch der speziellen patholo-
geschen Anatomie und Histologie (F. Henke, O. Lubarsch und R Rossle
eds.). Vol. 13, Part B, pp. 1910-1914, 1929-1971. Springer, Berlin! Gottin-
gen, Heidelberg.
4-16
-------
Pentschew, A. and Garro, F. (1966). Lead encephalo-myelopathy of the suck-
ling rat and its implications on the porphyrinopathic nervous diseases with
special reference to the permeability disorders of the nervous system's
capillaries. Acta Neuropathol. 6, 266-278.
Pollit, E. and Leibel, R. L (1976). Iron deficiency and behavior. J. Pediat 88,
372-381.
Redman, R. S. and Sweney, L. R. (1976). Changes in diet and patterns of
feeding activity of developing rats. J. Nutr. 106(5), 615-626.
Reiter, L. W., Anderson, G. E., Ladkey, J. W., and Cahill, D. F (1975). Devel-
opmental and behavioral changes in the rat during chronic exposure to lead.
Environ Health Perspect. 12, 119-123.
Shapiro, M. A., Tritschler, J. M., and Ulm, R. A. (1973). Lead contamination:
Chronic and acute behavioral effects in the albino rat. Bull. Psychon. Soc.
2(2), 94-96.
Silbergeld, E. K. and Goldberg, A. M. (1973). A lead-induced behavioral dis-
order. Life Sci. 13, 1275-1283.
Silbergeld, E. K. and Goldberg, A. M. (1974). Lead-induced behavioral dys-
function: An animal model of hyperactivity. Exp. Neurol. 42, 146-157.
Snowden, C. T. (1973). Learning deficits in lead-injected rats. Pharmacol.
Biochem. Behav. 1, 599-603.
Sobotka, T. J., Brodie, R. E., and Cook, M. P. (1975). Psychophysiologic
effects of early lead exposure. Toxicology 5, 175-191.
Ten State Nutritional Survey (1968-1970). U.S. DHEW Publication No. (HSM)
72-8133.
Terzin, A. L. and Vujkov, V. V. (1968). Alteration of the locomotive activity
of mice poisoned with sublethal doses of lead acetate. Arh. hig. rada. 19,
493-511.
Van Gelder, G. A., Carson, T., Smith, R. M., and Buck, W. B. (1973). Behav-
ioral toxicologic assessment of the neurologic effect of lead in sheep, din.
Toxicol. 6(3), 405-418.
Vostal, J. J., Taves, E., Sayre, J. W., and Charney, E. (1974). Lead analysis of
the house dust: A method for the detection of another source of lead expo-
sure in inner city children. Environ. Health Perspect. 7, 91-98.
Acknowledgments
The authors would like to thank Donald A. Fox for his assistance in data
collection and discussion of material presented in this paper. We also thank
Drs. H. Petering and L. Murthy for their discussion related to the use of semi-
purified diets and for the data presented in Table 4-7.
4-17
-------
5. OBSERVATIONAL TECHNIQUES
IN BEHAVIORAL TOXICOLOGY
STATA NORTON
Ralph L Smith Mental Retardation Research Center
and
Department of Pharmacology
University of Kansas Medical Center
Introduction
Observational techniques, at their simplest, require no more equipment than
pencil and paper and a willingness on the part of an observer to describe as
much of the behavior as interests him or as his sensitive, but sometimes unre-
liable, eyes and ears can capture. In this guise, observational techniques have
had a long history of use by lovers of the natural history of animals, by field
ecologists, and, more recently, by the school of ethologists, led into the field
by Konrad Lorenz to capture the natural history of animals in the snares of
theory, quantification, and laboratory experimentation. After generations of
scientific indifference or, even more, disdain for the presumed anthropomor-
phic and subjective elements in all observation of animal behavior, the value
of controlled observation of animals—freely expressing the subtleties of social,
maternal, or even simple exploratory behavior—is now accepted and exploited
as a way of investigating the function of the central nervous system. Table 5-1
covers the areas which have been most extensively investigated in recent years.
Table 5-1. BEHAVIORS STUDIES BY OBSERVATIONAL TECHNIQUES
Locomotor
acts
Exploration
Hypoactlvity
Hyperactlvlty
Social
behavior
Dominance roles
Territorial ity
Communication
Reproductive
behavior
Mating
Maternal
Early
development
Sensory reflexes
Motor function
Rhythmic
behavior
Diurnal rhythms
Circadian rhythms
5-1
-------
Given that observational techniques have achieved acceptance as powerful
quantitative methods of studying behavior, do these techniques have a useful
role in the study of behavioral toxicology? The answer is an unequivocal "yes"
for reasons to be explained.
It is a truism, or should be, that behavior is the output of the central nervous
system. Behavior is produced by the CNS in response to various external and
internal stimuli received from the external environment and the body. In addi-
tion, behavior is presumably produced by spontaneous activity of the CNS
itself without direct stimuli, at least in higher animals. The sum of these acti-
vities of the CNS, when expressed as changes in motor function, becomes the
output known as behavior. If behavior is defined as spatially directed, complex
motor acts, the following are excluded: the amorphous phenomena known as
emotions, volition, and consciousness; and specific nondirected muscle move-
ments such as tremors, twitches, and seizures. In the study of toxicants which
affect the CNS, the potential alterations in behavior are many. The effects of
toxicants can range in duration from rapidly reversible, as in the case of thera-
peutic drugs which last for a few hours, to permanent changes, as in the effects
of x-irradiation on the developing brain. Toxicants can directly affect sensory,
motor, or integrating systems in the CNS and hence alter behavior. Behavior
can also be altered when the CNS is not the target organ for the toxicant but
is affected indirectly as a result of a toxic effect elsewhere in the body.
Examples of indirect effects on behavior are numerous. Kidney and liver dam-
age obviously can affect behavior as can cardiovascular effects. Altered levels
of circulating hormones and damage to peripheral sensory and motor systems
modify behavior. In fact, it is difficult to conceive of damage to the body any-
where which cannot affect behavior. Thus behavioral toxicology cannot be
limited to specific kinds of effects, sites of action in the CNS, or to reversible
phenomena, and in these three ways behavioral toxicology differs from behavi-
oral pharmacology. These considerations bring us to the use of observational
techniques in behavioral toxicology. If one is interested in whether or not the
behavior of an animal has been modified in any way by exposure to a toxic
substance, such a global question could require examination of a wide range of
behaviors in order to get a reliable answer. Alternatively, one could look for a
global answer by examining a few behaviors which reflect a generalized func-
tion of the CNS. The difficulty here is knowing which behaviors to study. This
problem leads to the argument which favors observational techniques as
methods for obtaining global answers. If behavior is the consequence of the
processing and integration by the CNS of the three sources of information
listed before (external and internal stimuli and spontaneous CNS activity),
the easiest of these to control experimentally are the external stimuli. If the
external stimuli are kept constant and if the external environment is as neutral
as possible (i.e., if the animal is not pressured experimentally to perform any
particular motor acts), then the remaining two factors predominate. Since
these are the two factors which will be altered by toxic substances, it follows
that observation of spontaneous behavior in a neutral environment should
reflect changes in various CNS functions brought on by the toxicant. This
logic suggests that observational techniques should be sensitive for detection of
5-2
-------
many types of toxicants. The next question concerns the reliability of obser-
vational techniques and the methods by which quantification of spontaneous
behavior can be achieved.
The most common method for quantification of spontaneous behavior is to
record the frequency with which an act occurs. Thus many studies have re-
ported the number of incidents of given acts per unit time. An example of
this method, drawn from pharmacology, is found in the study by Silverman
(1965) of the effects of drugs on the frequency of various acts performed by
rats before and after drug administration. In addition to frequency analysis,
a more elaborate analysis will include the duration of behavior acts as well
as frequency, introducing another concept. It is inherent in the use of fre-
quency and duration in reference to behavior acts that such acts must be dis-
crete quantities which can be defined in some way as starting and stopping
and hence having a unique appearance in the stream of movement of an ani-
mal from birth to death. It has been argued that behavior acts are not "real,"
i.e., are not discrete, and that, therefore, quantitative analysis of behavior acts
is not possible or, at least, is not reliable. This argument is equivalent to the
criticism that observational tests produce highly "subjective" data. This point
deserves careful consideration since the discrete nature of behavior is intui-
tively favored by most observers of animal behavior. The evidence for the
concept has been discussed in detail previously (Norton, 1968) and will
not be repeated here, but the ability to define specific behavior acts in terms
accurately recognized by a computer from video film (Norton and Servoss,
1972) is a strong argument in favor of the nonsubjective nature of spontaneous
behavior.
The consequence of the discrete nature of behavior acts is that observers of
spontaneous behavior have at their disposal rich and varied sets of data which
can be analyzed in many ways that have hardly been explored at the present
time. In general terms, spontaneous behavior can be viewed as having both a
macrostructure and a microstructure, depending on the time scale in which the
analysis is carried out. Moderately long-term trends are apparent in diurnal and
circadian rhythms associated with the light-dark cycle, and short patterns of
behavior appear in the rapid cycling of acts in many situations, such as explor-
ation of a novel environment. Figure 5-1 presents a simple illustration of these
relationships. Is there any real difference in the macro-and microstructure of
behavior other than the time frame? The answer to this question is not clear at
present. The distinction is made here to point out the relationship between
these areas of research, but there may also be a valid physiological distinction.
For example the long-term behavioral sequences in the macrostructure, such
as sleep-wakefulness, may be dependent on internal changes such as diurnal
shifts in adrenal cortical hormones; whereas the microstructural changes are
more dependent on rapid local modification of CNS activity which is res-
ponding to rapidly changing sensory stimuli in the environment. Generally,
different techniques are used to record these two major divisions of behavioral
structure. Because the macrostructure is concerned with slow behavioral
processes, the usual forms of analysis are concerned with recording only on-off
timing of the behavior, as in sleep-wakefulness cycles. Another common tech-
nique is to measure circadian alterations in activity, a technique exemplified
5-3
-------
by the use of photocells to monitor locomotion in a residential cage through-
out the day (Barnett and McEwan, 1973; Norton eta/., 1975; Reiter and Ash,
1976).
OBSERVATIONAL TECHNIQUES
DESCRIPTIVE
QUANTITATIVE
Figure 5-1. Schematic relationship of macro- and microstructure of behavior.
The most precise technique presently available for studying the microstructure
of behavior is unquestionably to record the behavior on photographic film or
video tape, a procedure which gives the observer a permanent record from
which to study the behaviors of interest in any desired detail. The drawback to
working with a permanent visual record is that it can contain a tremendous
amount of information.
The first step in data reduction should take place before the permanent record
is even made. That is, before the observer records the animal's behavior, he
should select an environment appropriate to produce the desired behavior. As
Suomi and Harlow (1969) pointed out several years ago, all cages in which
animals are placed are instruments which affect the behavior. If it were pos-
sible, it would be ideal to have time stand still in the analysis of short bouts
of behavior, but time is also a variable to the CNS, and the brain clearly res-
ponds to what has just transpired as well as to immediate events. In this way,
the output of the CNS resembles a partial Markov chain in which each event
has a relationship to the preceding event in the chain. An example of such a
chain of behavioral events is shown in Figure 5-2. This figure depicts the type
of behavior obtained when a rat is placed in an exploratory situation, for
example, in a simple plastic box of the type shown in Figure 5-3, and is photo-
graphed with a time-lapse cinecamera at one frame per second. Placing the rat
in the box triggers off a sequence of behavioral acts. These discrete acts average
about 3 seconds in duration, and transition times between discrete acts are very
5-4
-------
short compared to the duration of the acts. This difference in duration and
transition times allows the use of time lapse to reduce data acquisition with-
out much loss of significant information.
MICROSTRUCTURE OF BEHAVIOR
PHOTOGRAPHIC ANALYSIS
RE
SM
RE
PW
RE
PW
ST
LO
ST
LO
ST
LO
ST
HT
ST
HT
SI
WF
SI
WF
4567
TIME, seconds
10
Figure 5-2. Example of microstructure of behavior of rats in an
exploratory situation as recorded by time-lapse photography at
one frame per second. RE, rearing; SM, smelling; PW, pawing; ST,
standing; LO, looking; HT, head turning; SI, sitting; WF, washing
face.
Figure 5-3. Photograph of chamber used for time-lapse analysis of exploratory behavior of rats.
Having decided upon the behaviors of interest (in the experiment described
here it is the exploratory behavior of the rat) and the method of producing a
permanent record of that behavior, three quantitative aspects of the behavior
are obtained from the record: the durations, the frequencies, and the associ-
ations of behavioral acts. When pairs of rats are photographed at one frame
5-5
-------
per second for 15 minutes, the photographic record for each rat is 900 frames
of film. In order to handle even this amount of data, computer programs are
needed to analyze the results. The behavioral act of each rat in each frame of
film is identified by the observer. The body position (standing, sitting, rearing,
walking, turning, or lying down) and head or paw position (looking, smelling,
head turning, sniffing, bobbing, washing face, grooming, scratching, or pawing)
is identified for each second. These behavioral acts of rats have been defined
previously (Norton, 1973) although in general the words follow conventional
English usage.
The following computations are made from the data obtained on groups of
rats (i.e., experimental and control groups):
1. The duration (D) in seconds of each act listed above is determined. Aver-
age durations and standard deviations are calculated for each act. In the
data examined so far, the durations are normally distributed or have a
gamma distribution.
2. The frequency (F) of each act is determined as the number of times each
act is initiated, regardless of duration. Average frequency per rat and stan-
dard deviation are determined conventionally.
3. The total number of seconds that each act is performed is calculated per
rat. This number is obviously the product, D x F.
4. Behavior patterns (behavior sequences) are estimated. The computations
involved here will be described in some detail since this represents the least
familiar (and most revealing) parameter measured.
Determination of Behavior Patterns
In the long history of observation of animal life, behavior patterns of animals
have been described often. As an example, drawn from classical biology, many
observers have described the patterns in mating behavior of various species.
Mating patterns have been seen in some insects where the behavior is a determi-
nistic chain in which each step follows a rigid form dependent on the previous
act and from which there is no deviation. Even in vertebrates-for example, in
some fish (Nelson, 1964; Barlow, 1968)-the pattern of mating behavior has
a highly rigid structure. In most of these studies, the behavior pattern is so pre-
cisely executed that the pattern has been described without reference to alter-
nate patterns or variability in the patterns. In other studies, such as those on
verbal communication among rhesus monkeys by Altmann (1965), the in-
herent flexibility in sequences of acts is so great that probability measurements
are required to identify patterns. It is not surprising that both situations have
been described: that some behavior patterns are "hard-wired" so that some
behaviors are performed in an unvarying sequence following the initiation of
the sequence by one of a pair of animals, and, alternatively, that some
behaviors are performed in sequences with a probability that is dependent on
many variables including the preceding behavior. Then are there any random
5-6
-------
acts performed without dependence on preceding acts? There is no unequivocal
answer to this question because the available data are not adequate. It can be
predicted only on the basis of present data that there may not be any truly
random behavior acts, at least not among the higher vertebrate groups.
In the case of behaviors which follow each other with sequential probabilities,
i.e., those in which the behavior pattern is not completely determined, the
strategy for detecting the pattern needs to be considered. This is particularly
true of complex patterns which, to the unaided eye, may superficially appear
to be random sequences. The most straightforward approach is to calculate
probabilities directly for various sequences of acts in which the observer is
interested. While this is an accurate and efficient method, it has two draw-
backs. First, it requires that the sequence of interest be predicted by the ob-
server. Second, it is limited to short sequences, i.e., three or four acts, because
the calculations required to obtain absolute probabilities of all possible com-
binations and permutations of longer sequences are prohibitive.
An alternate strategy to calculate sequences for large numbers of acts is to
compute chi square values for observed and expected frequencies of each pair
of behavior acts as they are obtained in successive time bins in the photo-
graphic record. For example, if there are 14 behavior acts in 900 seconds of
exploratory behavior of each rat, observed and expected occurrence of pairs of
the 14 acts in successive 2- to 30-second intervals or longer can be readily
determined. This method has been used to detect three patterns of exploratory
behavior in rats under these circumstances (Norton, 1968; Norton, 1973).
As an example of the data obtained with these calculations. Figure 5-4 con-
tains the details for a pair of acts, walking and rearing. This figure shows that
the number of simultaneous occurrences of walking and rearing is less than
expected in intervals shorter than 5 seconds, while in intervals 5 to 30 seconds
long, walking and rearing are paired more frequently than expected. The dif-
ference in observed and expected frequencies is significant in intervals from 9
to 20 seconds using a chi square test for goodness of fit of the observed to the
expected occurrences. Sincewalking and rearing are acts which each last several
seconds, once initiated, it is not surprising that they are found together with
less than random expectation in intervals less than 5 seconds. The duration
histograms in Figure 5-4 show the actual durations of each of these acts. The
significant association of walking and rearing as a pair in the longer intervals
up to 30 seconds shows that the initiation of one of the pair increases the
probability of occurrence of the other act. If these calculations are continued
for all possible pairs of the 14 behavior acts identified from film in these
experiments, the most probable sequence of behavior can be determined.
Table 5-2 shows this sequence for a group of 28 control rats. The interval
used to determine the sequence can be any time bin from 5 to 20 seconds in
these experiments with rat behavior, and the same sequence will be obtained.
5-7
-------
o
cc"
22
z
180 •
140
100
MICROSTRUCTURE OF BEHAVIOR
LINKAGE OF TWO BEHAVIOR ACTS
120.
O 40
HI
DC
U.
DURATION
HISTOGRAMS
WALKING
40 ••
REARING
SECONDS
6 10
SECONDS
10
15
20
30
Figure 5-4. Example of nonrandom association of 2 behavior acts, walking and rearing, in
exploratory behavior of controls rats. In line graph, X = expected occurrences of walking
and rearing together, in time intervals from 2 to 30 sec, calculated from observed fre-
quency of occurrence of each act independently; 0 = observed occurrences of walking and
rearing in the same intervals. Arrow points to time intervals where difference between
observed and expected was significant (p = (0.05, chi square test for goodness of fit).
Table 5-2. MICROSTRUCTURE OF BEHAVIOR SEQUENCE OF
BEHAVIOR ACTS IN CONTROL RATS
(Abbreviations are defined in the text)
WF SI GR SC
Groom ing
BO WA TU RE PW SN
Exploratory
HT SM ST LO
Attention
Table 5-3 shows the data from which a complete sequence of 14 acts is ob-
tained, using the calculations of chi square values for observed and expected
pairs. This table shows the chi squares for pairs of behavior acts in a group of
control rats (upper right half of matrix) and a group of rats with lesions in the
globus pallidus (lower left half of matrix). The linear arrangement of the acts is
the same as that in the sequence of acts in Table 5-2 and is obtained by sorting
the acts for control rats so that the diagonal adjacent to the main diagonal of
the matrix reaches the maximum numerical value of the chi squares in which
observed frequency is greater than expected. The effect of this arrangement is
to put acts together which have the highest chi squares for the pairs occurring
together and also to separate pairs with large chi squares which were less
5-8
-------
frequent than expected. The negative sign is used in Table 5-3 to indicate the
latter chi squares. Only one such arrangement with the maximum numerical
value for the summed diagonal is obtained for each matrix, and this arrange-
ment determines the most probable behavior sequence. The labels given to
three clusters of acts—grooming, exploratory, and attention—are for conven-
ience in identifying the types of acts involved. The abbreviations for the acts
are: WF, washing face; SI, sitting; GR, grooming; SC, scratching; BO, bobbing;
WA, walking; TU, turning; RE, rearing; PW, pawing; SN, sniffing; HT, head
turning; SM, smelling; ST, standing; and LO, looking.
Table 5-3. MATRIX OF CHI SQUARES FOR PAIRS OF BEHAVIOR ITEMS IN
SUCCESSIVE 5 SEC INTERVALS IN 4 CONTROL AND 4 LESIONED RATS3
ARRANGED TO MAXIMIZE NUMERICAL VALUE OF CONTROL DIAGONAL
ADJACENT TO MAIN DIAGONAL (MINUS SIGN USED TO INDICATE OBSERVED
BEHAVIOR WAS LESS FREQUENT THAN EXPECTED)
WF
SI
GR
SC
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Pairs of acts in control rats
WF SI GR SC
Grooming
45 15 2
31 31 0
5 7 0
000
Groom Ing-Exploratory
-3 2-10
-5 -3 -1 0
-7 0-10
-8 0-20
-1500
0000
Groom Ing- Attention
-10 -5 -7 0
-8 -12-5 0
-6 -4 -3 0
—7 0-10
BO WA TU RE PW SN
Exploratory-Grooming
-6 -16 -7 -5 -3 -8
0 -9-2-24 0
-2 -20-5 0 -2
-2 -20-5 0 -2
Exploratory
0252 3
0 1 0-1-3
1 0 24 1 0
7 1 14 5 2
1 006 2
0 0000
Exploratory- Attention
0 0000 0
-5001-1 0
0 00-10 0
0 0000 0
HT SM ST LO
Attention-Grooming
-12 -39 -28 -5
-13 -39 -25 -3
-1 -7 0 -1
-1 -7 0 -1
Attention-Exploratory
0021
1220
021-1
0400
0000
0002
Attention
750
0 12 0
20 4
1-8 2
aControl rat data in upper right half of matrix; lesloned rat data In lower left half of
matrix.
In the experiment in Table 5-3, each lesioned rat was photographed with a con-
trol rat as shown in Figure 5-3, Although there is some effect of one rat on the
other in these paired circumstances, the effect is to make the animals' behavior
more alike, thus slightly reducing the difference between control and experi-
mental animals (Norton, unpublished observations). The data in Table 5-3
show that the effect of the lesion was to reduce the chi square values in each
cluster of acts. The interpretation of this phenomenon is that the lesion
affected the microstructure of behavior, causing the rats to perform acts in a
more random fashion. This experiment has been reported previously in greater
detail (Norton, 1976).
5-9
-------
The significance of the microstructure of behavior and the effect of toxic
conditions on it need consideration. First, it appears that there is a structure
to behavior even under conditions where an observer may not readily detect it.
The example given here is the analysis of a simple exploratory situation in
which a rat is placed in a plastic cage for 15 minutes. Under these conditions
control rats show a distinct patterning of behavior. Furthermore, this pat-
terning is disrupted by certain types of brain damage. In the case shown here,
damage was caused by stereotaxic placement of bilateral lesions in the globus
pallidus.
In an analysis of observed behavior, the frequency and duration of behavior
acts should be examined as well as the sequences. Such an analysis of behavior
should be able to detect toxic effects on behavior in experiments in which a
less complete analysis might miss one or more of these alterations. Although it
has not been directly demonstrated, it is possible that certain types of brain
damage could affect the structure of spontaneous behavior and yet cause
little or no effect on other behavioral parameters. If spontaneous behavior of
man and other animals is observed in various naturally occurring and environ-
mentally limited situations, and if the structure of the behavior is examined,
it should be possible to widen the range of toxic effects which can be experi-
mentally quantified. This ability may be valuable both in the detection of
toxic behavioral effects and in understanding the ways in which the output of
the central nervous system is organized.
References
Altmann, S. A. (1965). Sociobiology of rhesus monkeys-ll. Stochastics of
social communication. J. Theor. Biol. 8, 490-522.
Barlow, G. W. (1968). Ethological units of behavior. In The Central Nervous
System and Fish Behavior (D. Ingle, ed), Univ. of Chicago Press, Chicago.
Barnett, S. A. and McEwan, I. M. (1973). Movements of virgin, pregnant, and
lactating mice in a residential maze. Physio/. Behav. 10, 741-746.
Nelson, K. (1964). The temporal patterning of courtship behaviour in the
glandulocaudine fishes (Ostariophysi, Characidae). Behaviour 24, 90-144.
Norton, S. (1968). On the discontinuous nature of behavior.,/. Theor. Biol.
21, 229-243.
Norton, S. (1973). Amphetamine as a model for hyperactivity in the rat.
Physiol. Behav. 11, 181-186.
Norton, S. (1976). Hyperactive behavior of rats after lesions of the globus
pallidus. Brain Res. Bull. 1, 193-202.
Norton, S., Culver, B., and Mullenix, P. (1975). Development of nocturnal
behavior in albino rats. Behav. Biol. 15, 317-331.
5-10
-------
Norton, S. and Servoss, W. (1972). Pattern recognition in rat behavior using a
laboratory computer. Fifth International Congress of Pharmacology, San
Francisco, July, 1972.
Reiter, L. W. and Ash, M. E. (1976). Neurotoxicity during lead exposure in the
rat. Toxicol. Appl. Pharmacol. 37, 160.
Silverman, A. P. (1965). Ethological and statistical analysis of drug effects on
the social behaviour of laboratory rats. Brit. J. Pharmacol. 24, 579-490.
Suomi, S. J. and Harlow, H. F. (1969). Apparatus conceptualization for psy-
chopathological research in monkeys. Behav. Res. Methods Instrum. 1, 247-
250.
5-11
-------
6. LOCOMOTOR ACTIVITY
MEASUREMENTS IN BEHAVIORAL
TOXICOLOGY: EFFECTS OF LEAD
ADMINISTRATION ON RESIDENTIAL
MAZE BEHAVIOR
LAWRENCE W. REITER, GEORGE E. ANDERSON,
MIRIAM E. ASH, AND L EARL GRAY, JR.
United States Environmental Protection Agency
Health Effects Research Laboratory
Introduction
Various measures of locomotor activity have been widely used in behavioral
toxicology research. The popularity of this endpoint is probably due to the
ease with which "some" measure of activity can be made. Several of the
methods currently employed have been described by Finger (1972). Typically,
these methods measure locomotor activity over short periods of time. Open
field activity measurements, for example, range from 2-5 minutes in duration
(Archer, 1973) and therefore, record what may be a unique activity, generally
referred to as "exploratory activity." Although this is a useful measure, it
does nothing to describe the animal's "Spontaneous" activity levels which re-
quire measurement in a more familiar environment. Further, chemically-in-
duced changes in exploratory activity may differ from those recorded in a
more familiar environment.
Biological rhythms are another factor which influences an animal's responsive-
ness to chemical agents. Reinberg and Halberg (1971) have introduced the
field of chronopharmacology which is concerned with the effects of drugs
(and toxins) as a function of biologic timing. It is not surprising that many
classes of drugs, such as stimulants and depressants, show circadian rhythms of
potency in various species including man. By limiting behavioral measure-
ments, especially measurements of locomotor activity, to a single time of day
(typically the diurnal period), one overlooks the possible interaction of bio-
rhythms with the toxic effects of an agent.
6-1
-------
Norton et al. (1975a) have recently described a residential maze system de-
signed to measure locomotor activity in either grouped or isolated animals
over extended periods of time. This system utilizes a number of photocells to
record passage of an animal through a series of interconnected alleys (see
Figure 6-1) and has been shown to be capable of detecting behavioral disrup-
tion by a variety of factors (Norton et al., 1976; Reiter et al., 1975). Our
laboratory has implemented a modification of this system which allows for
computer controlled data acquisition and analysis, thus making the maze par-
ticularly useful for long-term activity measurements.
WATERING TUBE
Figure 6-1. Schematic representation of the residential maze used to measure locomotor
activity. The maze consists of a series of interconnected alleys (10 cm wide, 10cm high)
converging on a central open field and covered with transparent acrylic plastic equipped
with a central raised area over the open field. The overall dimensions are 76 cm wide x 60
cm deep. Locomotion is detected by eight phototransistors (+)/infrared light-emitting
diode (o) pairs (Reiter eta/., 1975).
The present paper describes a series of experiments which measured the effects
of continuous pre- and/or postnatal exposure to lead on locomotor activity.
Some short-term measures of activity were made in preweanling animals, but
in general, activity was measured in adults (120 days old) over extended
periods ranging from 5-14 days. In this way, we could study various compo-
nents of an animal's activity, including exploratory, diurnal, and nocturnal
activity, as well as ultradian and circadian rhythms. The results indicate that
lead administration affects these components of activity in a dissimilar manner
and serve to illustrate the utility of this approach.
6-2
-------
Methods
Locomotor activity of adult males was recorded at approximately 120 days of
age in the residential maze (Figure 6-1). Groups of three littermates were
tested in the maze for a period of either 5 or 14 days. Four mazes were housed
in a sound-attenuated room maintained on a 12-hour light/12-hour dark cycle
beginning at 6:00 a.m. Food and water were available ad libitum. The daily
experimental procedure was as follows: Groups were introduced into the mazes
at 10:00 a.m. and removed at 9:00 a.m. the following day. The mazes were
cleaned, food and water were replenished, and the animals were then reintro-
duced into the mazes at 10:00 a.m. Activity counts, indicating the number of
times a rat passed a photodetector in an arm of the maze, were recorded hourly
with the use of a laboratory minicomputer (Computer Automation, Inc.).
Measurements of residential maze activity were also obtained following subcu-
taneous administration of d-amphetamine, which the animals received 20 min-
utes prior to being reintroduced into the maze. In Experiment 1, we gave only
one dose of amphetamine (4.0 mg/kg), which we administered on the fifth
day in the maze. In Experiment 2, utilizing doses of 1.0, 2.0 and 4.0 mg/kg,
a dose response to amphetamine was obtained. Animals receive each dose of
the drug, with two days allowed for recovery between drug administrations.
The sequence of drug administration was balanced both within and across
treatment groups.
Social interaction was also measured in the residential mazes. This behavior was
recorded on video tape, under dim red light, approximately 2 hours into the
nocturnal period on the third day in the maze. Interaction was measured as
the number of contacts and the total time in contact between any two of the
three animals during a 10-minute observation period. A contact was defined
simply as two animals physically touching and required a break of at least one
second between contacts. Due to availability of equipment, only two groups
per treatment group were recorded.
Male offspring in Experiment 2 were also tested for acute exploratory activity
at various ages of development, both prior to and following weaning. This
activity was measured in individual animals placed in a jiggle-cage (Lafayette
Instrument Company, Model A 501) for a 4-minute test period at 13, 16, 29,
and 44, days of age. This method measures movement of the animals on a
springloaded platform and is, therefore, affected by the animal's body weight.
Since there were marked differences in body weight between treatment groups,
comparisons of jiggle-cage activity were possible only between lead-treated and
pair-fed controls.
Experiment 1
Male and female Sprague-Dawley rats, obtained from Blue Spruce Farms
(Altamont, N. Y.), were randomly assigned to one of three treatment groups:
0, 5, or 50 ppm lead administered in the drinking water in the acetate form.
Purina lab chow was available ad libitum. Starting at 50-60 days of age, both
males and females were pretreated for a period of 40 days, which allowed lead
6-3
-------
concentrations to approach steady-state levels in soft tissues (Castellmo and
Aloj 1964). Following pretreatment, animals were mated, and pregnant re-
male's were continued on lead throughout gestation and lactation. At weaning,
the offspring were similarly exposed through 180 days. These F-] generation
rats were mated at 90 days of age, and the resulting F2 generation animals were
exposed to the same lead regimen. No differences between F-| and F2 gener-
ations were observed in any parameter, and results for both generations have
been combined.
Experiment 2
Timed-pregnant Sprague-Dawley rats, obtained from Charles River Breeding
Laboratory (Willmington, Massachusetts), were used. The lead exposure regi-
men was similar to that of Sauerhoff and Michaelson (1973). Briefly, ground
Purina lab chow containing 5 percent lead carbonate was provided to nursing
mothers at parturition. Two control groups were employed: one group re-
ceived food ad libitum, and a second group was pair-fed to lead-treated
mothers. Pair feeding was accomplished by measuring daily food consumption
in lead-treated mothers and providing a matched pair-fed mother with that
amount of food on the following day. This pair-fed group served as a control
for the decreased food consumption observed in the lead-treated animals. Lead
was removed from the chow on the sixteenth day postpartum and added to the
drinking water, as lead acetate, in a concentration of 50 ppm. At 21 days,
offspring were weaned onto lead-containing water and continued on treat-
ment throughout the experiment.
Results
The 23-hour activity of control male Blue Spruce Sprague-Dawley rats, taken
on the fourth day in the maze, is shown in the upper portion of Figure 6-2.
Following an initial period of exploratory activity, during the first hour after
reintroducing the rats into the maze, the control animals showed a normal
circadian rhythm with low diurnal activity and high nocturnal activity. The
mean activity levels (counts/hour) obtained during these various periods,
i.e., exploratory, diurnal, and nocturnal, as well as total activity, were used to
test for treatment effects.
Figure 6-3 represents control data for the adult male Sprague-Dawley rats
obtained from either Blue Spruce (Experiment 1) or Charles River (Experi-
ment 2). Data are presented for the first four days in the maze during the time
periods just described. Locomotor activity was higher on the first day than on
subsequent days. This elevated activity represents the animal's increased
reactivity due to the novel environment. In Blue Spruce animals, activity was
elevated through the nocturnal period of the first day. Any activity measure-
ments of 23 hours or less will therefore include some component of explora-
tory activity. Animals become "established" in the maze by the second day,
showing consistent activity levels on subsequent days.
Although both control groups showed the same general pattern of activity in
the maze, we observed some differences. First, Blue Spruce animals were more
active than Charles River animals. On day 1, the Blue Spruce animals were
6-4
-------
40 percent and 160 percent more active than Charles River animals during the
nocturnal and diurnal periods, respectively. By day 4, this difference had
diminished to 3 percent and 50 percent for the same periods. Second, activity
of both groups fell on day 2, but this fall in activity was greater in Blue Spruce
animals, such that the ratio of day 1/day 2 activities was greater in the Blue
Spruce. The ratio of day 1/day 2 activity for the diurnal period was 3.5 for
Blue Spruce and 2.0 for Charles River rats. This elevated ratio indicates that
Blue Spruce animals were not only more active, but were also different in their
reactivity, showing a greater percentage drop in activity as the novelty of the
environment wore off.
800
600
400
200
0
600
400
C 200
0
600
400
200
H EXPLORATORY
• DIURNAL
D NOCTURNAL
CONTROL
5 ppm LEAD
50 ppm LEAD
10a.m. 2p.m. 6p.m. 10p.m. 2a.m. 6a.m.
LIGHTS OFF LIGHTSON
TIME OF DAY
Figure 6-2. Representative time-interval histograms from control and lead-treated animals
(Experiment 1). Data for each histogram represent means for 5 groups of adult males (3
animals/group) obtained on day 4 in the maze. For purposes of analysis, the data has been
divided into various activity periods represented by the shaded areas.
6-5
-------
400
200
100
§ o
TOTAL
•>4-i
H 1 r-
EXPLORATORY
DIURNAL
BLUE SPRUCG
SPRAQUE-DAWLEY
CMARLGSRIV6R —
SPRAGUG-DAWLEY
NOCTURNAL
34 12
DAY IN MAZE
Figure 6-3. Control locomotor activity for adult molu Spruguo-Dawloy rots obtolnod from
two suppliers. Animals worn IBS tod in groups of 3, und oach point represents tho moon
counts/hr (+SE) for oither 4 (Charles Rivor) or 5 (Bluo Spruco) groups of unimals. Dtitu
are prosoniocl (or Iho various activity periods for tho first A days In tho maze.
Experiment 1
The residential maze activity levels (expressed as moan counts/hour) of the
various treatment groups for the first four successive days of testing appear
in Figure 6-4. Load administration resulted in a significant depression of loco-
motor activity which ranged From 27-42 percent on day 1. This loud-induced
6-6
-------
hypoactivity was not dose-related although tissue lead levels demonstrated a
dose-related internal exposure (Cahill et a/., 1976). Both the total and the
nocturnal activity remained depressed throughout the 4-day period. By day 4,
neither lead-treated group differed from control during the exploratory period,
and following the first day, no treatment differences were found in the diurnal
period. Therefore, the treatment differences seen in total activity are primarily
a reflection of the depressed activity during the nocturnal period.
500
400
300
200
| 100
§
°. o
>
>2500
u
<2000
1500
1000
500
TOTAL - -
EXPLORATORY
NOCTURNAL
500
400
300
200
100
0
500
400
300
200
100
234 12
DAYS IN MAZE
Figure 6-4. Effects of lead administration on locomotor activity in adult male rats
(Experiment 1): (•) control; (A) 5 ppm Pb; (•) 50 ppm Pb. Animals were tested in groups
of 3 and each point represents the mean (counts/hr ±SE) for 5 groups of animals. Data are
presented for various activity periods for the first 4 days in the maze. Points marked with
"a" are significantly different from control (t-test), p < 0.05) (Reiter et a/., 1975).
The social interaction measured on the third night in the maze, approximately
two hours into the nocturnal period, is represented in Figure 6-5 as the total
number of contacts between any two of the three animals as well as total time
in contact during a 10-minute observation period. Although these data repre-
sent only two groups of animals per treatment and were not subjected to sta-
tistical analysis, they suggest that social interaction is increased with lead treat-
ment. These differences are especially striking for the total number of contacts,
where the range of values for lead-treated groups does not overlap the con-
trol range.
6-7
-------
Finally, locomotor activity was measured following 4.0 mg/kg d-amphetamine
(sc) given 20 minutes prior to maze testing on day 5. Figure 6-6 presents the
results as change in activity (A) following amphetamine administration. The A
activity was determined by subtracting the activity obtained on day 4 (predrug,
control levels) from activity following amphetamine administration and there-
fore represents the drug-induced change in activity.
90 -
70 -
8
tc.
UJ
CO
3
50 -
30 -
10
400
I
z
uj
5 200
100
5 50 0
LEAD IN DRINKING WATER, ppm
50
Figure 6-5. Effects of lead administration on social behavior in the adult male rat. (Exper-
iment 1). Animals were scored for number of contacts and total time in contact in the resi-
dential maze during a 10-minute period of nocturnal activity on day 3 in the maze. Verti-
cal bars represent the range of values.
In control animals, amphetamine produced the expected hyperactivity. Since
the drug was administered during the diurnal period, when activity is normally
low, the major effect appeared during this time. Activity in control animals
was elevated 444 percent and 682 percent of predrug levels during the explor-
atory and diurnal periods, respectively. These increases were reflected in an
6-8
-------
elevation of total activity to 224 percent of predrug control level. During the
nocturnal period, the control activity fell to 80 percent of the predrug level,
demonstrating a partial compensation for the drug-induced hyperactivity.
A ACTIVITY, counts/hr
-
T
_i_
EXPLORATORY
a
T
_L
a
T
-±-
a
T
1
c
IURNAL
a
T
_L
-
750
600 t>
>
0
H
<
450 .<
1
300 -
150
0
0 5 50 0 5 50
LEAD IN DRINKING WATER,ppm
Figure 6-6. Effects of d-amphetamine administration (4.0 mg/kg sc) on locomotor activity
of control and lead-treated rats (Experiment 1). Values are presented as ^-activity (counts/
hr±SE) following amphetamine; A = (activity on drug day) - (activity on predrug day).
Each bar, therefore, represents the drug-induced locomotor activity for 5 groups of ani-
mals. Bars marked with "a" are significantly different from control (t-test, p < 0.05).
Lead-treated animals showed a diminution in the amphetamine response
(Figure 6-6). When compared to the A activity of the control group, both
treatment groups showed a significant (p < 0.05) decrease in amphetamine-
induced activity. This decreased responsiveness was also apparent when activity
was expressed as percentage of predrug levels. Whereas control activity was
elevated to 444 and 682 percent during the exploratory and diurnal periods,
respectively, the 50 ppm group was elevated 323 and 298 percent during these
same periods. As with the controls, treatment groups showed less activity
during the nocturnal period, but no significant differences between treatment
groups were present during this period.
Experiment 2
The acute exploratory activity of male offspring from mothers receiving 5 per-
cent lead carbonate at parturition was measured in a jiggle cage at various
ages. Figure 6-7 shows the results of these measurements, expressed as percent
of pair-fed controls at these ages. Lead exposure produced a transient hyper-
activity (208 percent of control at 13 days of age) which continued through
29 days of age and returned to normal by 44 days of age.
6-9
-------
250
100
12
16
20
24 28
AGE, days
32
36
40
44
Figure 6-7. Effects of lead administration on acute exploratory activity in rats (Experi-
ment 2), measured at various ages. Values are expressed as percentage of pair-fed controls.
Each point represents the mean activity (± SE) for 5 litters (2 animals/litter). Points
marked with "a" are significantly different from control (t-test, p { 0.05) Reiter, 1977).
Adult animals tested on the first 4 successive days in the maze showed no
treatment effects relative to the total counts per 23 hours (Figures 6-8 and
6-9). However, there were notable treatment differences in activity following
various doses of d-amphetamine, findings consistent with those of Experi-
ment 1. Figure 6-10 shows the dose-response curves of d-amphetamine for
the treatment groups. Again, the data are presented as ^activity and there-
fore represent drug-induced changes in locomotor activity. The three treat-
ment groups responded differently to d-amphetamine administration. In
general, the amphetamine response of treatment groups fell in the following
order: pair-fed > ad libitum controls) lead-treated. Analysis of variance indi-
cated a significant treatment-drug interaction (p { 0.002) during the diurnal
period. Comparison of the lead-treated groups with pair-fed controls showed
significant differences (t-test; p < 0.01) at the 2 and 4 mg/kg doses, during
the exploratory and diurnal periods. Similar comparisons with ad libitum
6-10
-------
controls were significantly different only at the 4 mg/kg dose. It appears then
that early undernutrition (pair-fed group) facilitates the response to ampheta-
mine, shifting the dose-response curve to the left; whereas lead treatment
attenuates the response, shifting the dose-response curve to the right. Since
undernutrition is present in the lead group, a comparison between pair-fed
and lead-treated groups is probably the more meaningful one.
800
600
400
200
600
' 400
200
600
400
200
|EXPLORATORY
DIURNAL
I I NOCTURNAL
AD LIB CONTROL
PAIR-FED CONTROL
LEAD-TREATED
10a.m. 2p.m. 6p.m. 10p.m.
LIGHTS OFF
TIME OF DAY
2 a.m. 6 a.m.
LIGHTS ON
Figure 6-8. Representative time-interval histograms for control, pair-fed control and
lead-treated animals (Experiment 2). Data for each histogram represent means for 4 groups
of adult males (3 animals/group) obtained on day 4 in the maze.
6-11
-------
»
>-
u
<
350
300
250
200
150
100
.CONTROL
• PAIR-FED
* LEAD-TREATED
DIURNAL
~i 1 r
NOCTURNAL
-PT
4 1
DAY IN MAZE
350
300
250
200
150
100 f
c
50 8
(-
450 >
Q
400 <
350
300
250
200
150
100
50
Figure 6-9. Residential maze activity of control and lead-treated adult male rats, tested
in groups of 3 (Experiment 2). Each point represents the mean activity (count/hr ± SE)
for four groups of animals taken on the first 4 days in the maze. No treatment differences
were observed (Reiter, 1977).
6-12
-------
.CONTROL
.PAIR-FED
_ * LEAD-TREATED
= 3000
8
H2000
1000
EXPLORATORY
DIURNAL
400
1.0 2.0 4.0 1.0 2.0 4.0
DOSE OF d-AMPHETAMINE, mg/kgs.c.
Figure 6-10. Dose-response curves for d-amphetamine in control and lead-treated animals
(Experiment 2). Values are presented as A activity (counts/hr ±SE) for 4 groups of ani-
mals (A-activity calculated as in Figure 6-6). An analysis of variance indicated a significant
drug-treatment interaction during the diurnal period (p < 0.002) with pair-fed animals
showing a facilitated drug response and lead-treated animals showing an attenuated
response.
During the nocturnal period, activity levels showed an apparent ultradian
rhythm having an approximate 3-4 hour cycle (Figure 6-8). This rhythm was
present in the control groups, partially disrupted in pair-fed controls, and com-
pletely disrupted in the lead-treated animals. Figure 6-11 shows predicted
curves obtained for the various treatment groups pooled over three predrug
days (days 7, 10, and 13 in the maze). Logio transformations of the data have
been performed to correct for heterogeneity of variance. Control animals
showed a clear triphasic distribution of activity during the nocturnal period,
with peaks occurring at 4-hour intervals. An analysis of variance indicated a
highly significant effect of time (p < 0.0001) in these control animals, which
accounted for 71 percent of the variability in the activity levels. Pair-fed con-
trols showed some disruption in the later peaks. Nevertheless, an analysis of
variance showed a significant effect of time (p < 0.03) in this group, with time
accounting for 27 percent of the variability. Finally, activity in lead-treated
animals showed no effect of time (p = 0.88), accounting for only 9 percent of
the variability in this group. As previously indicated, we observed no treatment
effects on activity levels during the nocturnal period, but there was a signifi-
cant treatment x time interaction (p = 0.05), demonstrating that the nocturnal
activity curves were not parallel. A further comparison of control with lead-
treated animals, using an analysis of variance, indicated a significant treatment
x time interaction (p { 0.02), thus demonstrating a difference between the
predicted curves of these two groups.
6-13
-------
2.63
2.49
^2.35
u
<
o
o
2.21
2.07
.CONTROL
•PAIR-FED
•LEAD
L /
I
6 p.m.
10 p.m.
2a.m.
6a.m.
TIME
Figure 6-11. Predicted nocturnal activity curves for control and lead-treated adult male
rats, tested in groups of 3 (Experiment 2). Log-jQ transformations have been performed
to correct for heterogeneity of variance. An analysis of variance indicates an overall treat-
ment x time interaction (p ( 0.05) demonstrating that the curves are nonparallel.
Discussion
In the previous chapter of this book. Dr. Norton separated spontaneous behav-
ior into two categories, microstructure and macrostructure, based on the time
scale in which the behavioral analysis is carried out. In the present study, the
residential mazes serve as a useful tool in measuring the macrostructure of
activity since a continuous recording of the animal's activity is obtained over
extended periods of time. The present study observed several interesting facets
of this macrostructure, and although the measurements were limited primarily
to activity in the residential mazes, they are likely to occur to some extent in
environments of different complexity.
6-14
-------
There has been a tendency in the area of behavioral toxicology to use data
taken from short-term measurements of ambulation to make general state-
ments about locomotor activity with the assumption that a short-term mea-
surement of activity will be representative of the macrostructure of activity
in general. The data presented here do not support that assumption. The differ-
ent facets of activity—i.e., exploratory, diurnal, and nocturnal—as well as the
rhythms, appear to be different in both normal animals and animals exposed
to lead.
Exploratory activity is operationally defined here as the ambulation occurring
in a novel environment which exceeds the animal's baseline level of activity
once established in that environment. As defined, exploratory activity in the
residential maze extended well into the first day and, in the case of Blue
Spruce animals, into the nocturnal period. Charles River animals, on the other
hand, did not show elevated activity during the day-1 nocturnal period, indi-
cating a difference in the duration of exploratory behavior in these animals.
The two groups also differed in the relative degrees to which they responded
with elevated activity on day 1. Whereas Blue Spruce animals were 350 percent
more active on day 1 than on day 2, Charles River animals showed only a 200
percent elevation during the same period. Therefore, both the amplitude and
duration of exploratory behavior were greater in the Blue Spruce animals.
Removing animals from the maze for one hour per day repeatedly produced an
"exploratory burst" of activity when the animals were reintroduced into the
maze. In both experiments, activity was elevated to approximately 500 per-
cent of the normal diurnal activity level (see Figure 6-9). Therefore, daily one-
hour activity measurements taken on consecutive days, although stable, will
still represent exploratory behavior.
In addition to the familiar circadian rhythm in activity levels, the Sprague-
Dawley rat also shows an ultradian rhythm. This rhythm, which has a 4-hour
cycle, was pronounced during the nocturnal period. A similar pattern of
resential maze activity was reported by Norton et a/. (1975b) in adult female
Sprague-Dawley rats obtained from Charles River. Although the data on ultra-
dian rhythm were not presented for the Blue Spruce animals (Experiment 1),
a similar pattern of activity was observed (see Figure 6-2). The apparent ab-
sence of ultradian rhythm during the diurnal period may be related to the
experimental design. Since diurnal activity level is very low, it may be difficult
to detect any small fluctuations, but more importantly, removal of the animals
from the maze for one hour per day during the diurnal period may disrupt the
rhythm. However, by the nocturnal period the animals have readjusted and the
rhythm returns.
In Experiment 1, lead exposure produced changes in locomotor activity which
differed during the various time periods. Exploratory activity was initially
depressed in lead-treated animals, but this difference disappeared by the fourth
day in the maze. Also, the treatment effect observed on day-1 diurnal activity
probably reflects a change in the exploratory component of this activity which
is present on the first day. On subsequent days, when exploratory behavior is
no longer a part of the diurnal activity, the treatment effects disappear. The
6-15
-------
final treatment difference in activity levels, observed throughout the 4-day
period, was the depressed nocturnal activity. It seems likely, therefore, that
the depressed total activity is a reflection of this difference. A selective dis-
ruption of nocturnal activity in the rat has also been reported by Norton
(1976) following lesions of the globus pallidus. In her experiments, residential
maze activity was increased approximately 200 percent during the nocturnal
period.
The lack of a lead effect on adult locomotor activity levels in Experiment 2
may have been due to the choice of the animal supplier (Charles River). Since
Charles River animals are normally less active in the maze, it may be difficult
to lower their activity further with treatment. This reasoning, however, is
speculative, and certainly other differences in experimental protocol (i.e., dose
and period of exposure) may account for these differences in observed activity
between the two experiments.
Although there was no treatment effect on activity levels perse, lead exposure
in Experiment 2 did result in a disruption of the nocturnal ultradian rhythm.
However, this disruption cannot be attributed solely to lead treatment, since
some change in the rhythm was observed in the pair-fed (undernourished)
animals. Nevertheless, these results demonstrate that an alteration in the bio-
logical activity rhythm can occur with no accompanying alteration in the
overall activity level. This finding presents an interesting situation. If one were
to sample activity from rats in an established environment at 10 p.m. (4 hours
into the nocturnal period), he would observe hypoactivity. Sampling at mid-
night would yield no difference in the activity levels. Finally, if measurements
were made at 5 a.m., it would appear that treatment had produced a hyper-
activity (see Figure 6-11).
The hyperactivity observed in young animals exposed to lead (Experiment 2)
agrees well with the findings of Sauerhoff and Michaelson (1973) who reported
increased activity in 28-day-old animals. Sauerhoff (1974) also found that
lead-induced hyperactivity disappeared as the animals matured. This transient
increase in activity may represent a delay in the animal's normal development.
Some reports indicate that during the first 3 weeks of life, rats undergo normal
developmental changes in activity with the peak level of activity occurring
between 15 and 20 days of age (Campbell et a/., 1969). Since lead exposure
has been shown to delay normal brain development in the rat, as indicated by
delayed reflex development (Reiter et al.f 1975; Reiter and Ash, 1976), the
possibility exists that lead is simply delaying this normal developmental pattern
of activity.
In both experiments, the locomotor response to d-amphetamine was atten-
uated in lead-treated animals. This attenuated drug response has been pre-
viously reported in mice (Silbergeld and Goldberg, 1974) and rats (Sobatka
and Cook, 1974). Furthermore, this diminished responsiveness to amphetamine
has been observed when baseline activity levels of lead-treated animals were
elevated (Silbergeld and Goldberg, 1974), normal (Sobatka and Cook, 1974;
Experiment 2 of the present study), or depressed (Experiment 1 of the present
study). This dissociation between the attenuated amphetamine response and
6-16
-------
the baseline level of activity in lead-treated animals would suggest that differ-
ent neural systems are involved in these two responses. The work of Creese
and Inversen (1975) has demonstrated that the integrity of the nigro-striatal
pathway is essential for the locomotor response to amphetamine. It has also
been suggested that a cholinergic link is involved in the regulation of the
nigro-striatal dopaminergic pathway (Bartholini et a/., 1975). Since lead expo-
sure has been reported to affect both the aminergic (Sauerhoff and Michaelson,
1973; Goiter and Michaelson, 1975; Silbergeld and Goldberg, 1975) and cho-
linergic (Modak et a/., 1975) systems in the brain, it is possible that lead is
affecting the nigro-striatal pathway resulting in an altered amphetamine res-
ponse. However, until such an effect can be demonstrated directly, we must
also entertain the possibility that lead treatment is altering the pharmacody-
namics of amphetamine. By modifying the distribution and/or metabolism
of amphetamine, lead treatment could be producing an attenuated response
irrespective of a CNS effect.
References
Archer, J. (1973). Tests for emotionality in rats and mice: A review. Anim.
Behav. 21, 205-235.
Bartholini, G., Stadler, H., and Loyd, K. G. (1975). Cholinergic-dopaminergic
interregulations within the extrapyramidal system. In Cholinergic Mecha-
nisms (P. G. Waser, ed.), pp. 411-418. Raven Press, New York.
Cahill, D. F., Reiter, L. W., Santolucito, J. A., Rehnberg, G. I., Ash, M. E.,
Favor, M. J., Bursian, S. J., Wright, J. F., and Laskey, J. W. (1976). Bio-
logical assessment of continuous exposure to tritium and lead in the rat.
In Proceedings of International Atomic Energy Symposium on Biological
Effects of Low-Level Radiation Pertinent to Protection of Man and His
Environment 2, 65-78.
Cambell, B. A., Lytle, L. D., and Fibiger, H. C. (1969). Ontogeny of adrener-
gic arousal and cholinergic inhibitory mechanisms in the rat. Science 166,
635-637.
Castellino, N. and Aloj, S. (1964). Kinetics of distribution and execretion of
lead in the rat. Brit. J. Ind. Med. 21, 308-314.
Creese, I. and Iverson, S, D. (1975). The pharmacological and anatomical sub-
strates of the amphetamine response in the rat. Brain Res. 83, 419-436.
Finger, F. W. (1972). Measuring behavioral activity. In Methods in Psycho-
biology (R. D. Myers, ed.). Vol. 2, pp. 1-9.
Goiter, M. and Michaelson, I.A. (1975). Growth, behavior, and brain cate-
cholamines in lead-exposed neonatal rats: A reappraisal. Science 178,
359-361.
Modak, A. T. Weintraub, S. T., and Stavinaha, W. B. (1975). Effects of chronic
ingestion of lead on the central cholinergic system in rat brain legions.
Toxicol. Appl. Pharmacol. 34, 340-347.
6-17
-------
Norton, S. (1976). Hyperactive behavior of rats after lesions of the globus palli-
dus.'Brain Res. Bull. 1: 193-202.
Norton, S., Culver, B., and Mullenix, R. (1975a). Measurements of the effects
of drugs on activity of permanent groups of rats. Psychopharmacologia 1,
131-138.
Norton, S., Culver, B., and Mullenix, P. (1975b). The development of noctur-
nal behavior in albino rats. Behav. Bio/. 15, 317-331.
Norton, S., Mullenix, P., and Culver, B. (1976). Comparison of the structure
of hyperactive behavior in rats after brain damage from x-irradiation, car-
bon monoxide, and pallidal lesions. Brain Res. 116: 49-67.
Reinberg, A. and Halberg I. (1971). Orcadian chromopharmacology. Annu.
Rev. Pharmacol. 11, 455-492.
Reiter, L. W. (1977). Effects of early postnatal exposure to neurotoxins.
J. Occup. Med. 19, 201-204.
Reiter, L W., Anderson, G. E., Laskey, J. W., and Cahill, D. F. (1975). Devel-
opment and behavioral changes in the rat during chronic exposure to lead.
Environ. Health Perspect. 12, 119-123.
Reiter, L. W. and Ash, M. E. (1976). Neurotoxicity during lead exposure in
the rat. Toxicol. Appl. Pharmacol. 37, 160.
Sauerhoff, M. W. (1974). Neurochemical correlates of lead encephalopathy in
the developing rat. Ph. D. Dissertation. Univ. of Cincinnati, Ann Arbor
Univ. Microfilm.
Sauerhoff, M. W. and Michaelson, I. A. (1973). Hyperactivity and brain cate-
cholamines in lead-exposed developing rats. Science 182, 1022-1024.
Silbergeld, E. K. and Goldberg, A. M. (1974). Lead-induced behavioral dys-
function: An aminal model of hyperactivity. Exp. Neurol. 42, 146-157.
Silbergeld, E. K. and Goldberg, A. M. (1975). Pharmacological and neuro-
chemical investigations of lead-induced hyperactivity. Neuropharmacology
14,431-444.
Sobatka, T. J. and Cook, M. P. (1974). Postnatal lead acetate exposure in rats:
Possible relationship to minimal brain dysfunction. Amer. J. Mem Def. 79
5-9.
Acknowledgment
We wish to thank Mr. Ronnie McLamb and Ms, Karen Kidd for their technical
assistance in various phases of this work.
6-18
-------
7. ALTERED BEHAVIORAL
PATTERNING IN RATS POSTNATALLY
EXPOSED TO LEAD: THE USE OF
TIME-LAPSE PHOTOGRAPHIC
ANALYSIS
PHYLLIS MULLENIX
Department of Environmental Health Sciences
The Johns Hopkins School of Hygiene and Public Health
Introduction
Researchers commonly measure changes in levels of spontaneous locomotor
activity to assess effects of CNS toxicants, drugs, and specific brain lesions on
behavior. Most methods measuring activity, however, are restricted to quanti-
fying changes in frequency, or possibly even duration, of a few select behav-
iors. This restriction allows the all-too-common possibility of finding an ani-
mal hyperactive with one method but not with another. For example, rats
chronically exposed to low levels of lead were found hyperactive in a Y-maze
and in a tilt-box, but not in a running wheel (Kostas et a/,, 1976). Here, a
toxicant's qualitative effect on locomotor activity depends upon the method
of measurement; the method's capabilities, rather than the animals' behavior
are dictating the results. This is a recurrent problem in behavioral toxicology,
even when parameters such as habituation, presence of other animals, and
timing during the light-dark cycle are kept constant.
Perhaps in a situation where a toxicant's effect on locomotor activity is incon-
sistent between methods, all possible parameters comprising changes in activity
should be measured. Norton (1973) has postulated that changes in activity may
consist of changes in sequential patterning as well as changes in frequency and
duration of behavioral acts. All three—patterning, frequency, and dura-
tion—were altered in hyperactivities induced by amphetamine (Norton, 1973),
•Supported in part by the EHS Grants 00034 and 00454.
7-1
-------
globus pallidus lesions, carbon monoxide, and x-irradiation (Norton et al.,
1976). However, hyperactivity induced by morphine consisted only of changes
in frequency and duration of behavioral acts (Norton, personal communi-
cation). Theoretically, it is possible that any one, either patterning, frequency,
or duration, could be altered independently of the other two. Distinguishing
changes in the patterning, frequency, and duration of many behavioral acts
would certainly define alterations in locomotor activity more accurately and
perhaps result in more consistent findings.
The purpose of this study is to measure the frequencies, durations, and pat-
terning of behaviors in rats exposed postnatally to lead and their responses to
the CNS stimulant, amphetamine. Exposure of rats postnatally to lead appar-
ently produces an inconsistent effect on spontaneous locomotor activity; both
hyperactivity (Sauerhoff and Michaelson, 1973) and no change in activity
(Sobotka and Cook, 1974; Grant et al., 1976) have been reported. As indi-
cated in the study by Kostas and coworkers (1976), the method selected to
monitor general activity in lead-treated rats may itself contribute to incon-
sistent findings. Perhaps the measurements proposed in this study can further
understanding of the possible role a method of measurement may play in the
inconsistent effect of lead upon spontaneous locomotor activity.
Methods
Charles River female rats (Sprague-Dawley derived proven breeders, 300-350 g)
were mated with proven breeder males of the same strain. At parturition all
litters were reduced to eight pups, and 5 mg/ml lead acetate were dissolved
in boiled distilled water which was supplied ad libitum as drinking water to
one group of mothers. This method of lead administration is routinely used at
this concentration to produce hyperactivity in mice (Silbergeld and Gold-
berg, 1973, 1974, 1975). Drinking water containing lead was replaced with
regular tap water after the first 10 days of the pup's suckling period. Re-
stricting administration of lead-treated drinking water to the first 10 days
after parturition was an attempt to minimize the effect of lead on body
weights of suckling pups and yet still expose the pups during their critical
period of sensitivity to lead (Brown, 1975). The other group of mothers re-
ceived tap water during this period. At 28 days of age, all pups were weaned,
separated according to sex, and housed in small groups in standard animal
facilities. All rats received Purina Rat Chow and water ad libitum.
The animals were tested when they were 4 or 6 weeks, or 4 or 5 months old.
At each age their body weights were recorded. A total of 22 groups of lead-
treated and 22 groups of controls were studied, including 6 groups of lead-
exposed and 6 groups of controls at each age, except at 4 months when only
4 lead-exposed and 4 control groups were tested. The groups, each consisting
of 4 rats of one sex from a single litter, were independent and randomly
selected at each age. An equal number of male and female groups were studied.
Frequency, duration, and patterning of animals' behaviors were measured using
time-lapse photography. Experimental procedure was similar to that described
by Norton (1973). One lead-exposed and one control rat were placed in oppo-
site sides of a plexiglas cage, which was divided by a clear center panel. The
7-2
-------
rats were then filmed simultaneously as they explored the novel environment.
They were filmed at a rate of one frame per second for the first 15 minutes
they were in the cage. All rats filmed were identified by number only, so that
each animal's behavior was analyzed without reference to its treatment. Re-
searchers viewed films of the animals and recorded one of the following body
positions of the rat for each frame: standing (ST), walking (WA), sitting (SI),
rearing (RE), or lying down (LD). Furthermore, if present, one of the fol-
lowing behaviors was also recorded for each frame: scratching (SC), grooming
(GR), washing face (WF), head bobbing (BO), turning (TU), pawing (PW), snif-
fing (SN)> head turning (HT), smelling (SM), or looking (LO). Norton (1973)
has previously described these behavioral acts in detail, and good inter- and
intra-observer reliability for recognizing each individual act has been reported
(Norton eta/., 1976).
Computer programs, similar to those used in the study by Norton and co-
workers (1976), facilitated determinations of frequency, duration, and pat-
terning of the various behaviors. The frequency of each behavior was deter-
mined from the total number of frames (or seconds)-the behavior was observed
during the 15-minute film of each rat. Average duration of each behavior was
determined from the average number of consecutive frames (or seconds) that
a behavior continued, once it was initiated. Frequency and average duration of
each behavioral act were compared in lead-exposed and control rats using the
t-test.
Patterning of behaviors in control and lead-exposed rats was compared by first
determining the number of intervals where any two different behaviors oc-
curred within 5 seconds of each other. This determination was accomplished
by dividing the 15-minute film of each rat into 5 consecutive second intervals,
then recording the number of 5-second intervals where two different, specified
behaviors occurred. This determination was combined for all rats in a group.
The occurrence of two different behaviors within 5 seconds of each other was
defined as a "linkage" or "pairing" of those behaviors. For example, in the
first 5 seconds of a film, if the behavior, rearing, occurred and if the behavior,
pawing, also occurred at any time within the first five seconds, then one inci-
dent of rearing linked with pawing was recorded. This scoring was continued
for remaining 5-second intervals and for all possible pairs of behaviors. In this
study, since 15 different behaviors were scored, n(n-1)/2 or 105 pairs of two
different behaviors were possible.
Observed occurrences of paired acts within 5-second intervals were compared
with expected occurrences, and the chi-square value (goodness-of-fit) was ob-
tained for each pair for all groups of rats. A matrix of so determined chi-square
values for control groups was set up with the rows and columns arranged to
obtain the maximum numerical value in the diagonal adjacent to the main
diagonal of the matrix. This optimal arrangement was obtained in order to
determine the most probable or preferred sequence of behaviors in control
rats. The optimal sequence for control rats in this study was the same as that
previously reported for controls (Norton et al., 1976): LD-WF-SI-GR-SC-BO-
WA-TU-RE-PW-SN-HT-SM-ST-LO. To determine if lead-exposed rats linked
7-3
-------
behaviors in this sequence as frequently as did controls, the observed fre-
quencies of each pair of acts in this sequence in lead-exposed groups were
averaged and compared with those for corresponding control groups using the
Wilcoxon signed rank test. This comparison is referred to in the following re-
sults as the "main diagonal" comparison.
From the matrix of chi-square values, ,it was observed (Norton et a/., 1976)
that within the optimal sequence of behaviors, there were three clusters of
closely associated acts, based upon the high chi-square values for association
of certain pairs. These clusters were labeled grooming, exploratory, and atten-
tion clusters. The grooming cluster included the behaviors lying down, washing
face, sitting, grooming, and scratching. The exploratory cluster included head
bobbing, walking, turning, rearing, pawing, and sniffing; the attention cluster
included head turning, smelling, standing, and looking. In this study, the
average frequency or number of 5-second intervals where each act was linked
with other acts in the same cluster was compared in lead-exposed and control
groups using the Wilcoxon signed rank test. These results are presented as
comparisons of paired behaviors within clusters (grooming, exploratory, and
attention). Furthermore, using the same statistical test, the average frequency
or number of 5-second intervals where each act of one cluster was linked to
each act in another cluster was also compared in lead-exposed and control
groups. These results are presented as comparisons of paired behaviors between
clusters (grooming linked with exploratory, grooming linked with attention,
and exploratory linked with attention).
At six weeks of age, control and lead-exposed rats were filmed before receiving
amphetamine and again the next day after receiving a subcutaneous injection
of 0.5 mg/kg d-amphetamine hydrochloride in normal saline. The injections
were given 30 minutes prior to filming so that during the time the animals were
filmed, between 30 and 45 minutes after injection, the maximum effect of
amphetamine at this dose would be observed (Norton, 1973). The frequency,
duration, and patterning of behaviors after amphetamine were compared in
control and lead-exposed rats as previously described.
Results
Average body weight of lead-exposed rats was not significantly lower than con-
trols at any age their behavior was tested (Table 7-1). In general, weight differ-
ences between control and lead-exposed rats were insignificant, except for the
slightly heavier weight of 6-week-old lead-exposed females.
Significant differences in frequency and duration (using the f-test) and in pat-
terning (using the Wilcoxon signed rank test) of behaviors were observed
between male and female rats at all ages, whether they were control or lead-
exposed. Such differences precluded combining male and female data. How-
ever, since this study was not specifically designed to delineate sex differences,
further studies must be conducted to determine if these differences were real.
Frequency of Behavior Acts
Frequency (total seconds of occurrence) of most behavior acts in lead-exposed
animals, both males (Table 7-2) and females (Table 7-3), did not differ signifi-
cantly from that in controls, whether they were 4 or 6 weeks or 4 or 5 months
of age.
7.4
-------
Table 7-1. AVERAGE BODY WEIGHTS (g±_SE)
Sex Age
Males 4 wks
6 wks
4 mo
5 mo
Females 4 wks
6 wks
4 mo
5 mo
Controls
No. of
animals
35
22
8
17
20
16
8
16
Average
weight
76 t 2
195 ±_ 2
498 ±_ 8
577 t 9
73 t 3
147 t 4
282 t 10
301 ±.8
Lead-treated
No. of
animals
33
19
8
19
30
21
8
18
Average
weight
76 t 3
193 t 4
508 t 9
575 t 9
71 t 2
157 t 3a
295 t 10
316 *• 6
Significant difference between lead-treated and controls (t-test, p ( 0.05).
Table 7-2. TOTAL OCCURRENCES OF BEHAVIORAL ACTS IN MALE RATS
(Av+_SE)a
Age
Behaviors
Controls
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Lead-treated
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
4 Weeks
0
6 t 2
69 t 15
165 t 30
86 t 14
15 t 2
33 t 3
44 ±.3
72 t 9
16 t 3
19 t 3
88 t 7
224 ±_ 19
638 +_ 35
96 t 11
0
1 ±_ 1
32 ±_ 4b
112 t 10
75 ±_ 9
11 t 2
35 t 4
43 t 3
87 t 15
15 t 3
30 ± 6
89 ±9
276 ±_ 16°
667 t 14
100 t 11
6 Weeks
13 ±_ 9
4 t 2
1 02 t 1 2
168 t 13
62 t 5
8 t 1
45 t 5
38 t 4
117 ± 14
19 t 3
46 t 7
85 t 5
205 t 14
557 ± 21
91 t 8
39 ± 18
1 t 0
57 t 9b
1 33 t 1 5
72 t 7
8±_ 1
32 ±_ 5
32 t 4
96 t 13
16 ±_ 6
40 ± 9
95 ±6
193 ± 16
600 ±_ 22
80 t 7
4 Months
68 t 22
2 t 2
95 t 15
200 ±_ 29
93 ± 24
8±_ 1
12 t 3
21 t 4
1 1 1 t 20
26 ±. 9
23 t 6
65 t 8
183 ±. 22
509 t 27
86 t 1 1
52 t 28
1 t 1
1 12 t 23
200 ±_ 31
82 t 12
6 t 1
13 ±_ 2
21 t 4
109 t 15
22 ± 5
51 t 17
63 t 5
212 ±_ 12
527 ±_ 38
69 t 12
5 Months
77 t 44
4+3
112 ± 23
195 ±_ 37
69 ±_ 10
8 t 2
8 +_ 1
15 ±_ 2
89 ± 13
25 t 5
21+7
49 ±_ 4
130 t 12
531 t 39
99 +_ 1 1
81 ±_ 25
1 + 1
105 ±23
196 t 34
74 t 9
10 ± 1
9 t 3
18 ±_ 3
81 ±_ 20
17 ±- 6
31 t 8
51 ±_ 3
165 ± 15
534 ±_ 41
57 +_ 5b
aControl N = 12 and lead-treated N = 12 per age group, except control N = 8 and lead-
treated N = 8 at 4 months.
DSignificant difference between lead-treated and control groups (t-test, p {o,05).
7-5
-------
Table 7-3. TOTAL OCCURRENCES OF BEHAVIORAL ACTS IN FEMALE RATS
(Av ±_ SE)a
Behaviors
Controls
UD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Lead-treated
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Age
4 Weeks
0
5 t 1
42 i 7
97 t 11
56 t 6
11 t 2
46 t 4
51 i 4
84 t 10
15 t 3
38 t 5
1 04 t 7
254 t 16
673 t 16
96 t 9
0
5 t 1
44 t 7
131 i 13b
77 t 10
8 t 2
43 t 5
46 ±_3
88 t 14
16 t 3
40 t 5
89 t 8
258 t 22
638 t 18
102 ±_ 12
6 Weeks
2 t 1
5t 3
54 t 11
107 t 16
52 t 6
9 t 2
55 t 7
45 t 6
122 t 14
23 t 3
52 t 8
87 t 7
256 t 21
614 t 29
81 t 11
5 t 5
7 ±.3
73 t 13
140 t 20
66 ±8
9 t 2
42 t 4
44 t 5
102 t 8
13 ±_2D
47 t 9
83 t 6
240 t 21
612 t 14
95 t 8
4 Months
37 t 29
5 ±_3
83 t 21
150 t 32
62 t 12
6 ±_2
27 t 2
24 t 3
186 t 41
16 t 3
24 t 3
63 t 6
268 t 16
504 ±_ 45
100 t 16
19 t 15
1 t 1
69 t 14
141 t 25
62 t 11
4 t 1
26 t 2
32 t 20
185 t 19
30 t 5b
50 t 13
66 t 6
318 dL 20
528 ±_ 24
93 t 9
5 Months
10 t 5
5 t 2
64 t 17
1 24 t 23
49 ±- 6
5 t 1
28 t 2
31 t 2
244 t 22
30 t 4
29 t 6
60 t 4
282 t 12
495 t 24
118 t 12
31 ±.18
6 t 3
84 t 14
155 ±.21
56 t 8
5 t 1
26 t 3
31 t 3
223 +.27
35 t 6
34 t 9
55 ±- 4
282 ±- 19
465 ±- 25
87 ±-9b
aControl N = 12 and lead-treated N = 12 per age group, except control N = 8 and lead-
treated N = 8 at 4 months.
"Significant difference between lead-treated and control groups (t-test, p (o.05).
Duration of Behavior Acts
For all ages at which the animals were tested, average durations of most behav-
iors in lead-exposed males (Table 7-4) and females Table 7-5) were not signifi-
cantly different from those in male and female controls respectively.
Patterning of Behavior
Patterning of behavior in lead-exposed groups was significantly different from
that in controls. At all ages, significantly fewer occurrences of the pairing of
different behaviors within 5-second intervals were observed in lead-exposed
males (Table 7-6). This difference was maximum in 6-week-old animals. At
that age, there were significantly fewer occurrences (1) of behaviors being
linked in the optimal sequence characteristic of controls, (2) of each explor-
atory behavior being linked with other exploratory behaviors, (3) of attention
7-6
-------
behaviors being linked with other attention behaviors, and (4) of exploratory
behaviors being linked with attention behaviors within 5-second intervals.
Fewer occurrences of paired exploratory behaviors were still observed at 4
months, and decreased pairing of behaviors within the grooming cluster was
also observed in 4-week- and 5-month-old lead-exposed males.
Table 7-4. DURATIONS OF BEHAVIORAL ACTS IN MALE RATS
(Av Sec/Rat ± SE)a
Behaviors
Controls
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Lead-treated
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
4 Weeks
0
2.3 i 0.7
3.9 t 0.4
12.4 t 1.9
5.5 i 0.5
1.0 t 0.01
1.2 t 0.02
1.1 ±. 0.02
2.1 t 0.1
1.3 ± 0.1
1.5 t 0.2
1.2 t 0.01
2.1 i 0.1
10.2 t 1.6
1 .4 t 0.04
0
1.0 ±_ 0.3
3.1 ±.0.5
10.2 t 1.3
5.3 t 0.4
.1 t-0. 03
.2 t 0.04
.0 t 0.01b
2.4 t 0.1
.3 t 0. 1
.6 t 0.1
.1 ± 0.02
2.4 t 0.1
10.3 ±_ 1.0
1.5 t 0.05
A
6 Weeks
8.9 t 6.9
1.6 i 0.5
6.3 t 0.6
18. 3 ± 1.7
4.7 t 0.4
1.0 tO
1.2 t 0.03
1.1 t 0.02
3.0 t 0.2
1.6 ± 0.1
2.0 ±0.2
1.1 ±.0.01
2.0 ±.0.04
7.9 t 0.8
1.4 t 0.04
17.0 t 8.7
1.0 t 0.3
4.5 t 0.5b
14.7 t 1.4
5,5 t 0,4
1,0 t 0.02
1.2 t 0.02
1.1 i 0.02
2.7 t 0.1
1.5 t 0.2
2.0 ±.0.2
1.1 t 0.02
2.0 t 0,1
10.4 t 1.1
1.4 t 0.05
ge
4 Months
27.7 i 8.0
1.0 ±.0.5
7.3 t 0.9
19.8 t 2.1
6.7 t 0.9
1.0 ±.0.01
1.2 ±.0.1
1.2 t 0.03
5.6 t 0.5
2.2t 0.2
1.8 ±. 0.3
1.1 ±. 0.02
2.4 t 0.2
14.4 t 2.4
1.7t 0.1
14.8 t 8,1
1.0 ±.0.4
10.1 t 0.8b
25.2 t 4.5
8.3 ±. 1.5
1.0 to
1.2 t 0.1
1.1 ± 0.03
5.3 t 0.5
1.9 t 0.2
3,0 t 0.4b
1.1 t 0.02
2.6 t 0.1
14.9 t 2.2
1.5 t 0.1
5 Months
43.3 t 35.3
1.0 t 0.3
7.5t 0.9
23. 7 t 3.7
6.0t 1.0
1.1 ± 0.04
1.2t 0.1
1.1 t 0.01
5.0 t 0.5
2.3t 0.2
2.2 t 0.3
1.1 t 0.02
2.5 t 0.2
22. 7 t 5.8
2.5 t 0.4
19.0 t 5.0
1.0 t 0.2
9.7 t 1.2
28.2 ±. 4.6
8.8 t 0.8b
1.1 t 0.03
1.0 t 0.1
1.1 t 0.02
4.8 t 0.4
1.7 t 0.2
2.3 t 0.3
1.1 t 0.02
2.5 t 0.1
21.7 t 5.0
1.6 t 0.1b
Control N = 12 and lead-treated N 12 per age group, except control N = 8 and lead-
treated N = 8 at 4 months.
'•'Significant difference between lead-treated and control groups (t-test, p (0.05).
7-7
-------
Table 7-5. DURATIONS OF BEHAVIORAL ACTS IN FEMALE RATS
(Av Sec/Rat ±SE) a
Behaviors
Controls
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Lead-treated
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
'Age1
4 Weeks
0
1.8 t 0.4
3.0 t 0.4
8.1 ± 1.1
4,3 ±0.3
1.0 t 0.01
1.2 tO. 02
1.1 ± 0.01
2.4 ± 0.1
1.3 ± 0.1
1.7 t 0.1
1.2 ± 0.03
2.1 t 0.1
9.2 ± 1.0
1.4 ± 0.04
0
2.1 ± 0.5
3.4 ± 0.4
10.3 t 1.4
4.9 t 0.7
1.0 ± 0.1
1.2 ± 0.03
Lit 0.02
2.3 t 0.1
1.3 t 0.1
1.6 t 0.1
1.2 t 0.03
2.2 t 0.1
8.8 t 1.1
1.5 t 0.05
6 Weeks
1.8 t 1.5
1.1 t 0.4
3.6 t 0.5
15.4 t 2.6
6.6 t 1.8
1.0 t 0.1
1.2 t 0.02
1.1 t 0.02
3.0 t 0.1
1.5 t 0.1
1.9 t 0.2
1.1 t 0.02
2.2 t 0.1
11.4 t 4.3
1.4 t 0.05
5.2 t '4.6
1.3 t 0.3
4.9 t 0.5
13.0 t 1.7
4.8 t 0.5
1.0 t 0.02
1.2 t 0.02
1.1 ± 0.02
2.7 t 0.1
1.3 t 0.1
2.0 t 0.1
1.1 t 0.02
2.2± 0.1
8.6 t 0.5
1.5 t 0.1
4 Months
11.1 t 8.6
1.1 t 0.4
6.4 t 0.9
19.4 t 3.5
5.9 t 1.0
1.0 t 0.02
1.2 t 0.03
1.1 t 0.03
5.0 t 0.5
1.6 t 0.1
2.0 t 0,1
1.1 t 0.03
2.8 t 0.1
10.4 t 2.5
1.6 t 0.1
10.2 t 7.6
1.0 t 0.3
7.4 t 1.3
19,0 t 4.9
4.6 t 0.4
1.0 t 0.04
1.1 t 0.03
1.1 t 0.02
5,2 t 0.3
2.0 t 0.3
2.5 t 0.3
1.1 t 0.02
2.9 t 0.1
8.9 t 0.6
1.5 t 0.1
5 Months
8.3 t 4.8
2.5 t 0.9
5.7 ± 0.8
13.9 t 2.1
4.7 t 0.3
1.0 t 0.02
1.1 t 0.02
1.1 ± 0.02
5.7 t 0.4
2.0 t 0.2
2.3 t 0.3
1.1 t 0.01
2,7 t 0.1
7.7 t 0.5
1.7 t 0.1
9.0 t 4.3
1.0 t 0.4
6.3 t 0.6
18.3 t 3,2
4.9 t 0.4
1.0 t 0.02
1.2 t 0.03
1.1 t 0.02
5.9 t 0.5
2.1 t 0.1
2.5 t 0.3
1.1 t 0.2
2. 7 ±0.1
8.5± 1,4
1.4 ±0.1
aControl N 12 and lead-treated N = 12 per age group, except control N = 8 and lead-
treated N = 8 at 4 months.
bSignificant difference between lead-treated and control groups (t-test, p { 0.05).
Table 7-6. SIGNIFICANT EFFECTS OF POSTNATAL LEAD ON PATTERNING
OF BEHAVIOR IN MALESa
Age
Main diagonal
Grooming
Exploratory
Attention
Grooming linked
with exploratory
Grooming linked
with attention
Exploratory linked
with attention
4 Weeks
Decreased
P (0.05
6 Weeks
Decreased
P <0.01
Decreased
p (o.01
Decreased
p <0.05
Decreased
p<0.01
4 Months
Decreased
p<0.02
5 Months
Decreased
p<0.05
Significant differences using the Wilcoxon signed rank test, Occurrences of linked behav-
iors within 5-second Intervals In lead-treated groups are indicated as "decreased" or "in-
creased" In comparison to controls.
7-8
-------
Patterning of behavior in lead-exposed females at all ages tested was signifi-
cantly different from that in corresponding controls (Table 7-7). As in lead-
exposed males, more patterning differences in pairing of behaviors within 5-
second intervals were observed at 6 weeks of age than at any other age tested.
Also as in lead-exposed males, there were significantly fewer occurrences of
paired exploratory behaviors and paired exploratory and attention behaviors
in 6-week-old lead-exposed females than in controls. However, unlike pattern-
ing differences between lead-exposed and control males, pairing of different
behaviors in lead-exposed females did not always decrease with respect to con-
trols. More frequent occurrences of different behaviors linking within 5-second
intervals were apparent in lead-exposed females at 6 weeks and at 4 and 5
months of age.
Table 7-7. SIGNIFICANT EFFECTS OF POSTNATAL LEAD ON PATTERNING
OF BEHAVIOR IN FEMALES3
Age
Main diagonal
Grooming
Exploratory
Attention
Grooming linked
with exploratory
Grooming linked
with attention
Exploratory linked
with attention
4 Weeks
Decreased
p <0.05
Decreased
p <0.02
6 Weeks
Decreased
p <0.01
Increased
p <0.01
Decreased
p <0.01
4 Months
Increased
p (o.05
I ncreased
p <0.05
5 Months
Increased
p (0.05
Decreased
P (0.05
aSlgnificant differences using the Wilcoxon signed rank test. Occurrences of linked behav-
iors within 5-second intervals in lead-treated groups are indicated as "decreased" or "in-
creased" in comparison to controls.
Behavior After Amphetamine
There was very little difference in the behavior of controls and the behavior of
lead-exposed rats after both had received amphetamine at 6 weeks of age. With
respect to frequency (Table 7-8) and duration (Table 7-9) of most behaviors,
there were no significant differences after amphetamine between control and
lead-exposed rats. This result was similar to that observed prior to administra-
tion of drug. Whereas there were many significant differences in patterning of
behavior between control and lead-exposed male groups prior to drug (Table
7-6), after amphetamine there appeared to be none. The frequencies with
which behaviors were paired in the control optimal sequence and the frequen-
cies with which different behaviors within and between clusters were paired
7-9
-------
were not significantly different between control and lead-exposed males after
both had received amphetamine. However, unlike the males, there were signi-
ficant differences between control and lead-exposed females in their behavioral
patterning both before (Table 7-7) and after amphetamine. The differences
after amphetamine consisted of significantly more occurrences, using the
Wilcoxon signed rank test, of linked grooming and exploratory behaviors
(p < 0.05) and of linked exploratory and attention behaviors (p ( 0.05).
Table 7-8. TOTAL OCCURRENCES OF BEHAVIOR ACTS IN 6-WEEK-OLD RATS
AFTER AMPHETAMINE (0.5 mg/kg) AVERAGE t STANDARD ERROR
Behaviors
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Males
Control
0
3 t 2
73 t 12
140 £ 16
60 ±. 6
15 t 3
56 t 9
38 t 6
93 t 11
11 ± 1
28 t 6
108 t 8
212 t 26
610 t 14
146 t 20
Lead-treated
0
2 t 1
41 t 1 1a
108 £ 26
63 t 15
10 t 2
42 t 8
40 t 8
87 ±_ 22
13 t 3
70 i 24
97 t 8
220 t 22
663 t 27
127 t 28
Females
Control
0
3 t 1
43 ± 12
97 t 17
50 t 7
12 t 3
66 t 5
44 ±_ 5
1 08 t 1 3
13 t 2
37 t 10
110 t 9
270 t 16
629 t 23
1 32 t 1 1
Lead-treated
0
2 t 1
31 t 10
74 t 16
41 ± 6
9 t 1
63 t 5
56 ± 4
119 t 20
9 t 2
52 t 11
98 t 4
256 i 14
645 ±_ 21
168 t 148
"Significant difference between control and lead-treated groups after amphetamine (t-test,
p (0.05).
Table 7-9. DURATIONS OF BEHAVIOR ACTS IN 6-WEEK-OLD RATS AFTER
AMPHETAMINE (0.5 mg/kg) AVERAGE DURATION IN
SECONDS JL STANDARD ERROR
Behaviors
LD
SC
GR
SI
WF
BO
WA
TU
RE
PW
SN
HT
SM
ST
LO
Males
Control
0
1.0 t 0.3
4.4 ±_ 0.5
13.8 t 1.5
3.8 t 0.3
1.0 ±_ 0.03
1.3 t 0.05
1.1 t 0.02
2.3 ±_ 0.2
1.1 ±. 0.05
1.3 ± 0.1
1.1 t 0,02
1.8 t 0.1
7.8 ±_ 1.1
1.6 t 0.1
Lead-treated
0
1.5 t 0.8
3.9 ±_ 0.5
12.7 ±. 1.3
4.5 ± 0.4
1.0 t 0
1.2 t 0.04
1.1 t 0.03
2.3 t 0.1
1.2 t 0.1
2.2 t 0.2a
1.1 ±_ 0.01
2.0 ±_ 0.1
1 1.5 t 2.3
1.7 t 0.3
Females
Control
0
1.2 t 0.4
3.5± 0.5
10.5 t 1.5
4.2 t 0.4
1.0 t 0
1.2 i 0.04
1.1 t 0.01
2.5 t 0.2
1.2 t 0.05
1.4 t 0.1
1.2 ±_ 0.03
2.0 t 0.1
7.2 i 0.8
1.5 t 0.1
Lead-treated
0
1.5 t 0.7
3.2 t 0.4
6.6 t 0,6B
3.6 t 0.2
1.0 t 0.01
1.2 t 0.02
1.1 ± 0.02
2.2 t 0.1
1.2 t 0.1
1.7 t 0.1
1.1 t 0.02
1.9 t 0.1
7.2 t 0.7
1.6 t 0.1
Significant difference between control and lead-treated groups after amphetarnln (t t
p \ 0.05). '
7-10
-------
Discussion
This study was carried out to measure the frequency, duration, and patterning
of behaviors in lead-exposed and control rats at various ages and after adminis-
tration of amphetamine. Since the only major difference between control and
lead-exposed rats was in the patterning of their behaviors, routine methods for
monitoring locomotor activity would not have been useful in defining this
difference. The insignificant differences in frequency and duration of most
behaviors in lead-exposed rats in this study were definitely not consistent with
frequency and duration changes observed in rats found hyperactive in a resi-
dential maze. Rats that were hyperactive in a residential maze (Norton et a/.,
1976) demonstrated significant increases in frequency of locomotor behaviors,
such as bobbing, walking, turning, rearing, pawing, and sniffing. Concomi-
tantly, frequencies of grooming behaviors—lying down, scratching, grooming,
sitting, and washing face—decreased, as did the frequencies of attention behav-
iors—standing, head turning, smelling, and looking. Duration of most behaviors,
except that of walking, was shortened in hyperactivity. Perhaps when the fre-
quency and duration changes characteristic of hyperactivity occur, methods
measuring activity would be sufficient to consistently detect differences. But
in the absence of such changes and when differences exist in behavioral patter-
ning only, results from various activity methods may vary because the methods
monitor different behaviors with varying sensitivities for a behavior's frequency
and duration.
There is the possibility that if lead-exposed rats in this study had been filmed
during the nocturnal period of their light-dark cycle, frequency and duration
of their behaviors may have changed in accordance with hyperactivity. This
possibility is lessened, however, by the observation that rats with globus palli-
dus lesions are hyperactive only at night in a residential maze (Norton, 1976);
yet when they are filmed during the day, the frequency and duration of their
behavior acts are still indicative of hyperactivity. This observation suggests
that regardless of when rats are found hyperactive with a maze activity
method, basic components of hyperactivity, which can be quantified by time-
lapse photographic analysis of diurnal exploratory behavior, may exist through-
out the animals' light-dark cycles.
Behavior pattern analysis indicated that the number of intervals when certain
different behaviors occurred within 5 seconds of each other in lead-exposed
animals was often significantly different from that in corresponding controls.
This difference was maximum in the 6-week-old animals. In lead-exposed males
at that age, there were fewer 5-second intervals when exploratory behaviors
were linked with other exploratory behaviors, when attention behaviors
were linked with other attention behaviors, and when attention behaviors
were linked with exploratory behaviors. Furthermore, fewer occurrences of
behavior pairs comprising the control optimal sequence were observed, perhaps
indicating that lead-exposed males had established a different optimal sequence
of behaviors. However, further pattern analyses are necessary to test this possi-
bility. There were fewer patterning differences between control and lead-ex-
posed males at the other ages. This peaking of altered behavior pattern at 6
weeks of age may perhaps be comparable to hyperactivity that peaked in
7-11
-------
young rats postnatally exposed to carbon monoxide (Culver and Norton, 1976)
and in children (Weiss et a/., 1971); then, in both instances, the hyperactivity
diminished with age.
Certain patterning differences between lead-exposed females and their corre-
sponding controls were dissimilar to those between control and lead-exposed
males. More frequent occurrences of certain behavior pairs were observed in
6 week and older lead-exposed females, but at no age was this increased fre-
quency of behavior pairing observed in lead-exposed males. These results indi-
cate the necessity of observing lead's effect on behavior in both sexes.
Behavior pattern differences between 6-week-old control and lead-exposed
male rats disappeared after administration of amphetamine. After 0.5 mg/kg
amphetamine, there were practically no differences in frequency, duration, or
patterning of behaviors between control and lead-exposed males. Conse-
quently, the "paradoxical" effect of amphetamine, consisting of decreasing
rather than increasing activity in lead-exposed mice (Silbergeld and Goldberg,
1974), was not apparent in lead-exposed rats. There were also no significant
differences in frequency and duration of most behaviors and fewer differences
in patterning between control and lead-exposed females after amphetamine.
However, some patterning differences between lead-exposed and control fe-
males were still observed after amphetamine. The higher frequencies of linked
grooming and exploratory behaviors and of linked exploratory and attention
behaviors suggested that the qualitative effect of amphetamine in lead-exposed
females, with respect to their controls, was not the same as that in lead-ex-
posed males.
This study demonstrates a situation in which quantifying changes in behav-
ioral patterning is paramount to simply quantifying changes in activity levels.
It is possible that other CNS toxicants besides lead predominantly affect behav-
ioral patterning, and, if so, the analysis of patterning with time-lapse photo-
graphy becomes more than just a sensitive technique; it becomes a necessity.
References
Brown, D. R. (1975). Neonatal lead exposure in the rat: Decreased learning
as a function of age and blood lead concentrations. Toxicol. Appl. Phar-
macol. 32, 628-637.
Culver, B. and Norton, S. (1976). Juvenile hyperactivity in rats after acute
exposure to carbon monoxide. Exp. Neurol. 50, 80-98.
Grant, L. D., Breese, G., Howard, J. L, Krigman, M. R., and Mushak, P.
(1976). Neurobiology of lead-intoxication in the developing rat. Fed Proc
35, 1640.
Kostas, J., McFarland, D. J., and Drew, W. G. (1976). Lead-induced hyper-
activity. Chronic exposure during the neonatal period in the rat. Pharma-
cology 14, 435-442.
7-12
-------
Norton, S. (1973). Amphetamine as a model for hyperactivity in the rat.
Physiol. Behav. 11, 181-186.
Norton, S., (1976). Hyperactive behavior of rats after lesions of the globus
pallidus. Brain Res. Bull. 1,193-202.
Norton, S., Mullenix, P., and Culver, B. (1976). Comparison of the structure
of hyperactive behavior in rats after brain damage from x—irradiation,
carbon monoxide and pallidal lesions. Brain Res. 116, 49-67.
Sauerhoff, M. W. and Michaelson, I. A. (1973). Hyperactivity and brain cate-
cholamines in lead-exposed developing rats. Science 182, 1022-1024.
Silbergeld, E. K. and Goldberg, A. M. (1973). A lead-induced behavioral dis-
order. Life Sci. 13, 1275-1283.
Silbergeld, E. K. and Goldberg, A. M. (1974). Lead-induced behavioral dys-
function: An animal model of hyperactivity. Exp. Neurol. 42, 146-157.
Silbergeld, E. K. and Goldberg, A. M. (1975). Pharmacological and neuro-
chemical investigations of lead-induced hyperactivity. Neuropharmacology
14,431-444.
Sobotka, T. J. and Cook. M. P. (1974). Postnatal lead acetate exposure in rats:
Possible relationship to minimal brain dysfunction. Amer J. Ment. Def. 79,
5-9.
Weiss, G., Minde, K., Werry, J. S., Douglas, V., and Nemeth, E. (1971). Studies
on the hyperactive child. VIII. Five-year follow-up. Arch. Gen. Psychiatry
24,409-419.
7-13
-------
8. BEHAVIORAL ASSESSMENT
IN SENSORY TOXICOLOGY
HUGH L EVANS**
Environmental Health Sciences Center and
Department of Radiation Biology and Biophysics
School of Medicine and Dentistry
University of Rochester
Sensory impairment, such as numbness or blurred vision, is a consequence of
exposure to any of a large variety of substances (see Grant, 1974, or Foulds,
1974). In the toxicology of sensory impairment, the behavioral scientist may
be able to offer more precise definition of impairment, including improved
sensitivity and reliability of tests, with the ultimate hope of providing better
differential diagnosis, early detection of intoxication, and more intelligent
environmental health standards. This chapter reviews early indications of this
potential and examines behavioral principles likely to aid the precise defini-
tion of sensory impairment caused by toxins.
Sensory function is a particularly apt behavioral process for toxicology because
of well developed psychometric procedures relating the behavioral response to
physical qualities of the stimulus, while segregating nonsensory variables. Some
of the earliest yet most complete behavioral studies involved auditory and
visual impairment (see reviews by Stebbins and Coombs, 1975; and by Hanson,
1975). Orderly psychometric curves demonstrated a selective sensory impair-
ment related to a lesion caused by the drugs; they illustrate the utility of
Supported in part by grants ES-01247 and ES-01248 from the National Institute of
Environmental Health Sciences, MH-11752 from the National Institute of Mental
Health, and by a contract with the U S. Energy Research and Development Adminis-
tration at the University of Rochester Biomedical and Environmental Research Pro-
ject (Report No. UR-3490-991).
* *Present address: Institute of Environmental Medicine, New York University Medical
Center, 550 First Ave., New York, N.Y. 10016.
8-1
-------
probing the controlling parameters and the use of psychophysics to specify
sensory impairment after initial stages of drug screening. It is not necessary to
review this body of work again here, but a reading of these studies and of the
book Animal Psychophysics (Stebbins, 1970), should precede behavioral
studies involving sensory toxicology.
Reviewing behavioral studies of sensory impairment reveals some principles of
general utility in the diverse endeavors known as Behavioral Toxicology. The
paramount principle is the importance of determining specific mechanisms of
action, both at the behavioral and biological levels of analysis. Unfortunately,
too many studies in behavioral toxicology are purely descriptive, with neither
an attack on underlying mechanisms nor a clear extrapolation to human health
questions. Our sibling discipline of sensory psychopharmacology evokes criti-
cism on similar grounds (Robinson and Sabat, 1975). Some behavioral toxi-
cologists appear to be neglecting mechanisms while seeking a threshold dose in
hope of observing an effect at a dose lower than previously reported. The study
of basic mechanisms is not incompatible with the search for threshold doses.
The pursuit of the mechanisms (i.e., the controlling variables) will ultimately
improve a behavioral test's sensitivity to low doses. This is my message.
Why examine specific behavioral mechanisms of action? One reason is to im-
prove the basis for pursuing underlying neural mechanisms and for extrapo-
lating to humans. Since humans cannot be subjects for toxicity experiments,
the applicability of animal data to humans must be assessed upon some basis
other than direct species comparison. A similar limitation has not prevented
great progress in the neurophysiology of vision, particularly since the pioneer-
ing work of Hubel and Wiesel (1962). Behavioral tests of sensory function
identified behavioral mechanisms common to humans and laboratory animals.
Then brain recordings and manipulations, with experimental animals, suggested
human neural mechanisms as well as new paths of human behavioral research.
Behavioral toxicology should provide a similar liaison between basic laboratory
experiments and clinical studies with humans, with cross-reference between
behavioral and biological analyses.
Using this approach in our studies of methylmercury, we identified similar
visual functions in humans and normal monkeys, then determined the specific
visual impairment in monkeys that suggest early warning signs of intoxication
in humans (Evans, 1975b). The pattern of mercury distribution in the brain
and the rate of elimination also are being explored in search of biologic mecha-
nisms that would account for the primate's greater sensitivity to methylmer-
cury as well as its greater susceptibility to sensory impairment compared to
subprimate species (Evans et a/., 1977; Evans and Laties, in preparation).
Knowledge of basic behavioral mechanisms also is useful in dealing with special
kinds of problems encountered in assessing toxins, problems seldom encoun-
tered in assessing psychopharmacologic agents. A problem in examining both
biological and behavioral measures is to define a "toxic effect," i.e., to discrim-
inate a true impairment from a "change." Several recent papers, that I found
to be interesting and better than average in other respects, illustrate this point
8-2
-------
The Iowa group used a variation of Stebbins' auditory procedure and found
that the pesticide, parathion, increased the variability of the monkey's auditory
threshold (Reischl et a/., 1975). Should this be considered a toxic effect in the
sense of greatest concern in making health decisions? A finding of this sort
should be examined with one of the procedures discussed below that can help
separate specific sensory effects from nonsensory effects. Reischl's monkey
was to press a lever when he heard a tone and to do nothing in the absence of
a tone. There is a hazard associated with using two markedly different kinds
of response to indicate the two choices in a discrimination. Behavioral pharma-
cologists have found that many drugs can change the rate of probability of
some behaviors' occurrence, depending upon the details of the test situation
(see reviews by Sanger and Blackman, 1976; Kelleher and Morse, 1968). Re-
sponse-rate changes induced by toxins need not be related to sensory impair-
ment. Suppose the toxin changed the probability of occurrence of all behav-
iors. This would change the proportion of trials in which the monkey does
nothing, i.e., indicates "I don't hear anything/' without necessarily changing
auditory sensitivity. Forcing the monkey actively to select one of two levers
to indicate the presence or absence of the stimulus would have permitted the
calculation of the percentage of correct discriminations independently of
changes in response probability. Then, no default choice need be inferred from
a withheld response. This approach has been illustrated with anticholinergic
drugs (Evans, 1975a).
The findings of Reischl et al. (1975) suggest a characteristic of threshold or
near-threshold doses: the problem of weak and variable results. An apparently
toxic effect may wax and wane before becoming clear (e.g., Berlin et al.,
1975). If the basic mechanism of action were pursued first, even if higher doses
were required to do so, we would have a basis for deciding whether the weak
effect, seen after the administration of a low dose, represents a logical extra-
polation of the clear effects determined previously.
As the discussion above suggests, attempts to determine whether an effect is a
true "toxic" effect lead to a related problem: determining specificity of the
result. Behavioral measures such as "spontaneous" locomotor activity or multi-
ple (FI/FR) response rates make excellent screening devices because they may
be influenced by any toxin that produces "sickness." Used in this way, changes
in response rates reveal little about controlling mechanisms. The skills and
interests of the behavioral scientist are more fully engaged by questions such
as "Can a toxin's effect be said to involve sensory, motor, or some intellectual
mechanism?"
Categories of behavioral disruption (e.g., impairment of discriminative
learning), like disease symptoms, are not specific to a single causal factor but
may occur as the result of a variety of causes. Young rats treated with lead
showed an intriguing impairment in the learning of a brightness discrimination
although they ultimately performed at a level characteristic of the control
animals (Brown, 1975). As with earlier studies in this area (Carson et al.,
1974), perhaps the next question should be "Could this represent a sensory
deficit, with the brightness discrimination learned more slowly because the
animals didn't see very well?" There seems to be some evidence that lead may,
8-3
-------
some visual functions (Bushnell et al., 1977), as does another heavy metal,
methylmercury (Evans et al., 1975). Since Brown's rats were required to run
through a T-maze, the learning impairment may also reflect changes in moti-
vation or motor coordination. In a similar experimental situation, impaired
learning of a brightness discrimination by rats exposed to tellurium was attri-
buted to motivational rather than sensory impairment, since impairment was
also present in a one-way shock avoidance that involved no discriminative
stimuli (Drueta/., 1972).
Acquisition of maze-running is influenced by many variables, such as baseline
level of motor activity, which may be changed by lead exposure (Silbergeld
and Goldberg, 1974) without reflecting either learning or sensory capabilities.
The influence of activity level upon the speed of rats' maze-learning has been
more fully examined in the area of genetic differences (Ray and Barrett, 1975).
Thus, Brown's report (1975) leaves us far from a conclusion that early expo-
sure to lead retards discriminative ability, although many people would agree
that whatever the ultimate nature of the effect, it is likely to be classified as
toxic.
We must pursue these basic mechanisms to insure the greatest success beyond
the initial screening stage. Psychologists long have been struggling with ques-
tions such as learning versus memory versus performance. Behavioral toxi-
cology needs input from those who are conversant with this literature, Several
contributors to this volume have indicated the wisdom of pausing to check
the literature instead of "reinventing the wheel." The following pages review
some principles of behavior that can improve our behavioral assay by increasing
sensitivity to low doses and by permitting greater specification of response
classes. Only then can we say with confidence exactly what aspect of behavior
is changing and whether the change is good, bad, or indifferent.
The Signal Detection Approach
Signal detection theory arose from engineering and illustrates many ways in
which we can be confused when estimating a change in sensory capability (e.g.,
Green and Swets, 1966; Swets, 1973; Wright and Nevin, 1974). Possible sen-
sory changes can be confounded with changes in nonsensory factors, such as
response position bias, motivation, guessing, or in the motor coordination in
the execution of the required response. These changes can alter the reporting
of a sensory stimulus and thus can be confounded with the true sensory impair-
ment.
One of the main advances of the signal detection approach was to yield a pure
measure of sensory capability (d'), with the nonsensory factors lumped to-
gether in a second measure, termed "bias." However, even when the stimulus
is held constant, this supposedly pure, quantal measure of sensitivity varies
with variations in the reinforcement contingencies employed in the test (Nevin,
1975), as does practically every other behavioral measure. Even if the signal
detection paradigm is not perfect, its emphasis upon segregating the influence
of bias provides a basis for improving some of the previous approaches to sen-
sory toxicology.
8-4
-------
The only application of signal detection methods thus far in behavioral toxi-
cology is the suggested use of olfactory detection to improve the monitoring
of air pollution (Berglund and Berglund, 1974). More experiments employing
traditional signal detection are found in pharmacology. In the most thorough
animal experiment to date, LSD was found to alter response bias of rats
without changing the auditory sensitivity (Dykstra and Appel, 1974). These
drug-induced changes resembled those produced by manipulations of rein-
forcement probabilities. This study illustrates the need for caution in applying
signal detection procedures, developed by using trained human observers, to
studies with animals exposed to drugs and toxins. The common tendency of
animals to withhold responding on a proportion of the trials and for drugs and
toxins further to reduce the response probability can result in a paucity of data
with which the originators of human signal detection methods were not forced
to contend.
Signal detection methods may help to separate changes in cutaneous sensitivity
from changes in verbal reporting following analgesia (Chapman et a/., 1973;
Clark and Yang, 1974). However, the meaning and validity of these examples
have been questioned in letters from numerous authors (see Science 189, 65-
68, 1975). Pastore and Scheirer (1974) have reviewed other limitations in
signal detection application.
The signal detection analysis, as most commonly employed, has additional
limitations. Although the paradigm yields quantal results, one must be cautious
not to violate the assumptions on which the statistical calculations are based.
One hopes that the opportunity to obtain a quantal index of sensitivity, by
plugging numbers into one of the formulae available, doesn't promote un-
warranted complacency.
The signal detection approach most commonly employs an extremely simple
behavioral task, with a stimulus either "on" or "off," and two responses equiv-
alent to "yes I see it," "no I don't see it." Many questions in behavioral toxi-
cology require manipulation of other stimulus parameters and of reinforcement
contingencies in order to probe fully the variables of interest. The following
sections describe these alternatives. In summary, signal detection theory is
stimulating a helpful reappraisal of experimental design. The classic procedure
offers an alternative approach in assessing sensory impairment; its utility in
behavioral toxicology is still uncertain.
Complex Discriminative Behavior
A discriminative technique of long standing in primate research, the WGTA
(Wisconsin General Test Apparatus) was employed in assessing chronic expo-
sure of squirrel monkeys to methymercury (Berlin et a/., 1973; 1975). The
monkey was required to discriminate between two stimuli, which were three-
dimensional objects, by pulling a chain attached to the stimulus. This proce-
dure, like the maze discussed above, may reflect changes in both sensory and
motor systems, requiring an astute experimenter to distinguish between
changes in response speed, visual-motor coordination, accuracy of discrimi-
nation, etc.
8-5
-------
It will be useful in further discussions to consider the WGTA task in two com-
ponents: (1) the visual stimuli and (2) the required response. Berlin's discrimi-
nation could be considered relatively easy, since the two stimuli differed in
each of three visual dimensions: color, form, and size. The choice response
could be considered relatively complex and difficult, since the monkey first
had to select the end of one of two chains attached to the stimuli, then use the
chain to pull the stimulus object toward him and thus retrieve a piece of food.
The point I wish to make here is that the accuracy of the visual discrimination
was no more revealing of the early effects of methylmercury than were the
performance measures embodied in the chain-pulling sequence.
More subtle types of visual impairment were identified by subsequent experi-
ments as the earliest signs of methylmercury intoxication in primates (Evans
et al. 1975; Berlin et a/., 1975). The critical results were obtained by manipu-
lating the strength of the discriminative stimuli. Since motor coordination had
been shown not to be among the early toxic signs in primates (Evans ef al.,
1975), a simple operant response (button-pressing) was employed in order to
minimize the motor complications in evaluating visual impairment. Exposure
to methylmercury, in doses much like those used previously, increased the
critical fusion intensity; i.e., the luminance required for a flickering light to
be discriminated from a steady light increased as exposure progressed (Berlin
ef al., 1975). The frequency of flicker was held constant while luminance was
varied. This reaction was interpreted as a loss of scotopic (night vision) sensi-
tivity to light.
The critical flicker frequency test (CFF) is complementary to the critical
fusion intensity test described above. In determining the CFF, luminance is
held constant while frequency is manipulated until the flicker disappears.
Although toxins have not been widely studied with CFF, the majority of
psychoactive drugs tested in humans cause a decrease in the critical frequency
(Smith and Misiak, 1976).
Another aspect of visual impairment occurring during the early phase of
methylmercury intoxication was assessed with a visual discrimination test de-
vised in our Rochester laboratories. The procedure was planned to facilitate
the distinction between sensory and nonsensory effects illustrated in dis-
cussing the signal detection approach. We tried to make the discrimination
procedure more informative than the "yes-no" detection paradigm by re-
quiring the animals to choose from among three discriminative stimuli, instead
of between the customary two choices. This change expanded the range of
scores between perfect (100 percent) and chance (33 percent correct with
three choices but 50 percent correct with two choices). Similarly, the three
response choices provide an expanded baseline for identifying nonsensory
factors such as response position bias (Evans, 1975a). We achieved a quanta!
index of strength of the controlling stimuli by varying the luminance of the
stimuli in steps instead of in an all or none fashion, as was customary in pre-
vious drug studies. The resultant psychometric functions can be used to esti-
mate the strength of the stimulus control of behavior and to relate this to a
quantal property of the stimulus, (Evans, 1975a,b).
8-6
-------
Strength of Stimulus Control as a Predisposing Factor
Experiments with operant behavior have contributed a substantial understand-
ing of how environmental stimuli can influence (i.e., "control") behavior. Ex-
periments in this tradition have indicated that effects of drugs and toxins vary
with the extent to which behavior is controlled by stimuli and reinforcing con-
tingencies (Laties, 1975; Evans and Weiss, 1978). The following examples
illustrate how strength of control by visual stimuli can be a predisposing factor
in behavioral toxicology. I believe the principle applies to other sensory modali-
ties, but more evidence is required.
Visual discrimination under the control of weak stimuli (low luminance) was
more vulnerable to disruption by drugs and toxins than was similar discrimi-
native behavior controlled by strong stimuli. Impairment of discrimination
with weak stimuli (low luminance) was the first sign to emerge during chronic
methylmercury exposure (Evans et at., 1975; Evans, 1975b), which was in
agreement with findings reported at the same time by Berlin et al. (1975). Gen-
erality for the conclusion was obtained in acute studies with anticholinergic
drugs as reference componds (Evans, 1975a). Discriminative accuracy with
weak controlling stimuli was impaired by a dose of scopolamine one-fourth
that required to impair accuracy slightly with a strong stimuli. These findings
may be related to the fact that visual deficits caused by brain lesions vary,
depending upon the eye's adaptation to ambient illumination (Glickstein et
al., 1970; Weiskrantz and Cowey, 1967).
Another type of two-choice visual discrimination procedure was used to assess
the effects of parathion (Reiter et al., 1975). All effective doses abolished
responding and markedly reduced blood cholinesterases within 5 hours; com-
plete recovery usually occurred within several days. Several factors suggest
that visual processes probably were not involved in these results. First, para-
thion reduced response probability without markedly changing accuracy of
discrimination. Several variants of the discrimination problem, each with dif-
ferent stimulus patterns, revealed no differential effect relating to stimulus
complexity or task difficulty. These predisposing factors will be discussed
below. Second, the abolition of responding was accompanied by marked, non-
visual toxic signs. Third, both reference drugs, scopolamine and methylscopo-
lamine, produced an abrupt cessation of responding similar to the effect of
parathion. Involvement of the eye or the brain would be indicated by a marked
difference in the effectiveness of the two drugs (see Evans, 1975a).
Strengthening the stimulus control in an operant task with pigeons could im-
prove the performance that had been degraded by methylmercury (Laties,
1975). Manipulating stimulus control showed that the toxic effect involved
stimulus, rather than motor, mechanisms, since the pigeons' response patterns
returned to baseline following an increase in the strength of the controlling
stimulus.
The utility of manipulating an aspect of stimulus strength is also illustrated in
a preliminary report of a selective decrement of visual acuity. Lead-exposed
monkeys, that had been considered "normal" after showing no deficits under
8-7
-------
high illumination, showed a deficit when later tested under low illumination
(Bushnell et a/., 1977). This effect of lead resembles that reported for methyl-
mercury (Evans et at., 1975; Evans, 1975b), a surprising finding since neither
behavioral nor biological mechanisms of inorganic lead intoxication has been
reported to resemble those of methylmercury.
The above evidence indicates two benefits in manipulating strength of stimuli.
First, it increases the sensitivity of the test to low doses and to early, preclini-
cal effects in chronic exposure, such as those with methylmercury (Evans,
1975b). Second, manipulations of stimulus strength facilitate conclusions
about specific sensory mechanisms, with nonsensory factors considered sepa-
rately. For example, demonstrating a relationship between a toxic effect and a
specific stimulus parameter, e.g., luminance, argues against an influence of
memory, motivation, or motor coordination, since these factors can be con-
trolled during manipulation of the stimulus parameter (e.g., Evans, 1975a;
Evans eta/.. 1975).
Behavioral toxicology in the Soviet Union may contain evidence supportive of
the above conclusions. Medved et al. (1964) review evidence that conditioned
responses elicited by weak stimuli are more easily disrupted by a pesticide than
are responses to stronger stimuli. Trakhtenberg (1974) summarizes studies
characterizing early effects of mercury vapor, including a greater increase in
response time to "a weak stimulus" (a white light) than to a "strong stimulus''
(buzzer). Later stages of intoxication were characterized by a "release of differ-
entiation" (differential responses were no longer controlled by the discrimi-
native stimuli), followed by anorexia, lethargy, etc. Soviet scientists also have
employed olfactory and visual sensitivity in assessing sulfur dioxide and related
gases (see Xintaras and Johnson, 1976, for a summary). Appraisal of the Soviet
literature is difficult for the average Western scientist because of the unfamiliar
vocubulary, which represents an insurmountable barrier if sufficient details of
methods and results are not provided. One is uncertain as to whether "weak-
ening of internal active inhibition" or "cortical inhibition spreading to the
sub-cortical region" (Trakhtenberg, 1974) refers to any of the mechanisms
discussed here. Trakhtenberg cites some startling positive behavioral findings
with extremely low concentrations of mercury vapor, but these reports appar-
ently are not published in the archival scientific literature.
The Complexity of the Stimuli
Thus far we have focused on the strength or intensity of stimuli as a predis-
posing factor in determining sensitivity to toxins. It may be profitable to con-
sider complexity of the stimuli as an additional predisposing factor. Together,
strength and complexity of the stimuli seems to account for much of what
might be referred to as the "difficulty" of sensory tasks. Difficulty could be
related to the rate of reinforcement under control conditions, i.e., the percent
of correct choices. Our work with methylmercury and visual discrimination is
accumulating evidence of this relationship. When stimulus strength (luminance)
is held constant, discrimination of relatively complex stimuli (3 stimuli of
equal area but different shape) seems more vulnerable to .disruption than is the
discrimination of less complex stimuli (selection of the brightest of 3 stimuli
without regard to form). Our monkeys are tested equally often on both dis-
crimination tasks. Psychophysical data from normal monkeys and humans, as
8-8
-------
well as from monkeys exposed to methylmercury, suggest an interesting inter-
action between strength and complexity of the stimuli.
Rats showed impaired visual pattern discrimination after exposure to tellurium,
although a simple brightness discrimination was not impaired (Dru et a/.,
1972). As before, more evidence exists in drug studies of sensory impairment
reviewed by Laties (1975). Additional studies have shown that alcohol selec-
tively impaired detection of both auditory (Moskowitz and DePry, 1968) and
visual (Moskowitz and Sharma, 1973) stimuli only when the task complexity
was increased. Replication of this work is described by Xintaras and Johnson
(1976). Additional manipulations of task difficulty in discrimination research
have been described by Ettlinger and Rashbass (1976).
Behavioral toxicologists should consider these and other variables influencing
task difficulty to be of basic importance as well as of practical utility. Task
difficulty may be a determinant of the behavioral effects of toxins as it has
long been recognized as a determinant of impairment following brain lesions:
performance of a very easy task is refractory to lesions that are sufficient to
eliminate performance of very difficult tasks (Lashley, 1929; Chow, 1967).
Most psychologists recognize Lashley's efforts to identify behavioral impair-
ment associated with removal of large areas of the rat's brain; he found that as
much as one-half of the brain could be removed without altering the rat's
ability to solve some mazes, illustrating that performance decrement is a joint
function of task difficulty and the amount of brain damage (Lashley, 1929),
a 50-year-old fundamental principle that should be of utility to the fledging
discipline of Behavioral Toxicology.
Conclusions
Sensory function has a good potential for behavioral studies in toxicology,
whether the prime interest is in developing a very sensitive assay or in defining
specific sensory mechanisms. A review of recent reports with methylmercury,
lead, and parathion reveals that, empirically, sensory function seems to be
sensitive to a wide variety of toxic agents. Compared with other psychological
phenomena, changes in sensory performance can be more clearly interpreted
as "toxic" and seem to be relevant to questions of human health and safety.
Manipulation of the behavioral parameters should be considered of importance
equal to the manipulation of the dose parameters. The parametric curve, such
as the psychophysical function, should have a standard place along with the
dose-effect curve in behavioral toxicology.
The concept exemplified by the signal detection approach merit inclusion
in all assessments of sensory function instead of the classic detection proce-
dure, even if the experiment requires one of the discriminative tasks des-
cribed herein. All of the procedures employ a discrete-trial, rather than free-
operant schedule, to obviate the need to infer sensory function from changes
in response rates.
8-9
-------
Our legacies in neuropsychology and in behavioral pharmacology offer useful
discriminative procedures and suggest that the task difficulty, resulting from
strength and complexity of the stimuli, will be an important factor in assessing
sensory impairment.
References
Berglund, B. and Berglund U. (1974). A psychological detection method in
environmental research. Environ. Res. 7, 342-352.
Berlin, M., Nordberg, G., and Hellberg, J. (1973). The uptake and distribution
of methylmercury in the brain of Saimiri sciureus in relation to behavioral
and morphological changes. In Mercury, Mercurials and Mercapatans (M. W.
Miller and T. W. Clarkson, eds.), pp. 187-208. Charles C. Thomas, Spring-
field, Illinois.
Berlin, M., Grant, C. A., Hellberg, J., Hellstrom, J., and Schutz, A. (1975).
Neurotoxicity of methylmercury in squirrel monkeys. Arch. Environ.
Health 30, 340-348.
Brown, D. R. (1975). Neonatal lead exposure in the rat: Decreased learning as
a function of age and blood lead concentrations. Toxicol. Appl. Pharmacol.
32, 628-637.
Bushnell, P. J., Bowman, R. E., Allen, J. R., and Mariar, R. J. (1977). Scotopic
vision deficits in young monkeys exposed to lead. Science 196, 333-335.
Carson, T. L, Van Gelder, G. A., Karas, G. G., and Buck, W. B. (1974). Slowed
learning in lambs prenatally exposed to lead. Arch. Environ. Health 29, 154-
156.
Chapman, C. R., Murphy, T. M., and Butler, S. H. (1973). Analgesic strength
of 33 percent nitrous oxide: A signal detection theory evaluation. Science
179, 1246-1248.
Chow, K. L. (1967). Effects of ablation. In The Neurosciences (G. C.Quarton,
T. Melnechuk, and F D. Schmitt, eds.), pp. 705-713. Rockefeller Univ.
Press, New York.
Clark, W. C. and Yang, J. C. (1974). Acupunctural analgesia? Evaluation by
signal detection theory. Science 184, 1096-1097.
Dru, D., Agnew, D. F., and Greene, E. (1972). Effects of tellurium ingestion
on learning capacity of the rat. Psychopharmacologia 24, 508-515.
Dykstra, L. A. and Appel, J. B. (1974). Effects of LSD on auditory perception:
A signal detection analysis. Psychopharmacologia 34, 289-307
Ettlinger, G. and Rashbass, C. (1976). Adjusting the amount of overtraining to
the difficulty of discrimination learning. Neuropsychologia 14, 257-260.
8-10
-------
Evans, H. L. (1975a). Scopolamine effects on visual discrimination: Modifica-
tions related to stimulus control J. Pharmacol. Exp. Therap. 195, 105-113.
Evans, H. L. (1975b). Early methylmercury signs revealed in invisual tests. In
International Conference on Heavy Metals in the Environment, 1975, Pro-
ceedings Vol. 3 (T. C. Hutchinson, ed.). Institute of Environmental Studies,
Toronto, pp. 241-256.
Evans, H. L. and Laties, V. G, (in preparation). Neurotoxicity of methylmer-
cury in pigeon.
Evans, H. L., Laties, V. G., and Weiss, B. (1975). Behavioral effects of mercury
and methylmercury. Fed Proc. 34, 1858-1867.
Evans, H. L., Garman, R. H., and Weiss, B. (1977). Methylmercury: Exposure
duration and regional distribution as determinants of neurotoxicity in non-
human primates. Toxicol. Appl, Pharmacol. 41, 15-33.
Evans, H. L. and Weiss, B. (1978). Behavioral toxicology. In Contemporary
Research in Behavioral Pharmacology (D. E. Blackman and D. J. Sanger,
eds.), pp. 449487. Plenum Press, New York.
Foulds, W. S. (1974). The toxic amblyopias. In Aspects of Neuro-ophthal-
mology (S. I. Davidson, ed.), pp. 86-163. Butterworths, London.
Glickstein, M., Barrow, S., and Luschei, E. (1970). Vision in monkeys with
lesions of the striate cortex. In Animal Psychophysics: The Design and Con-
duct of Sensory Experiments (W. C. Stebbins, ed.), pp. 249-258. Appleton
Century Crofts, New York.
Grant, W. M. (1974). Toxicology of the Eye, 2nd ed., Charles C. Thomas Co.,
Springfield, Illinois.
Green, D. M. and Swets, J. A. (1966). Signal Detection Theory and Psycho-
physics. John Wiley and Sons, New York.
Hanson, H. M. (1975). Psychophysical evaluation of toxic effects on sensory
systems. Fed, Proc. 34, 1852-1857.
Hubel, D. H. and Wiesel, T. N. (1962). Receptive fields, binocular interaction,
and functional architecture in the cat's visual cortex. J. Physio/, 160, 106-
154.
Kelleher, R. T. and Morse, W. H. (1968). Determinants of the specificity of
behavioral effects of drugs. Ergeb. Physiol. 60, 1-56.
Lashley, K. S. (1929). Brain mechanisms and Intelligence. Univ. of Chicago
Press, Chicago.
Laties, V. G. (1975). The role of discriminative stimuli in modulating drug
action. Fed. Proc. 34, 1880-1888.
8-11
-------
Medved, L I., Spynu, E. I., and Kagan, I. S. (1964). The method of condi-
tioned reflexes in toxicology and its application for determining the toxicity
of small quantities of pesticides. Residue Rev. 6, 42-74.
Moskowitz, H. and DePry, D. (1968). The effect of alcohol upon auditory
vigilance and divided attention tasks. Q. J. Stud. Alcohol 29, 54-63.
Moskowitz, H. and Sharma, S. (1973). Effect of alcohol on the visual auto-
kinetic phenomenon. Percept. Mot. Ski/Is 36, 801-802.
Nevin, J. A. (1975). Differential reincorcement and signal detection. J. Exp.
Anal. Behav. 24, 355-367.
Pastore, R. E. and Scheirer, C. J. (1974). Signal detection theory: Consider-
ations for general application. Psycho/. Bull. 81, 945-958.
Ray, 0. S. and Barrett, R. J. (1975). Behavioral, pharmacological and biochem-
ical analysis of genetic differences in rats. Behav. Biol. 15, 391-417.
Reischl, P., Van Gelder, G. A., and Karas, G. G. (1975). Auditory detection
behavior in parathion-treated squirrel monkeys. Toxicol. Appl. Pharmacol.
34,88-101.
Reiter, L. W., Talens, G. M., and Woolley, D. C. (1975). Parathion adminstra-
tion in the monkey: Time course of inhibition and recovery of blood cho-
linesterases and visual discrimination performance. Toxicol. Appl. Pharma-
col. 33, 1 -13.
Robinson, D. N. and Sabat, S. R. (1975). Sensory psychopharmacology. In
Current Developments in Psychopharmacology, Vol. 2 (W. B. Essman and
L. Valzelli, eds.), pp. 185-203. Spectrum Publications, New York.
Sanger, D. J. and Blackman, D. E. (1976). Rate-dependent effects of drugs: A
review of the literature. Pharmacol. Biochem. Behav. 4, 73-83.
Silbergeld, E. K. and Goldberg, A. M. (1974). Lead-induced behavioral dys-
function: An animal model of hyperactivity. Exp. Neurol. 142, 46-157.
Smith, J. P. and Misiak, H. (1976). Critical flicker frequency (CFF) and psy-
chotropic drugs in normal human subjects—A review, Psychopharmaco/ogia
47, 175-182.
Stebbins, W. C., (ed.). (1970). Animal Psychophysics: The design and conduct
of sensory experiments. Appleton Century Crofts, New York.
Stebbins, W. C. and Coombs, S. (1975). Behavioral assessment of ototoxicity in
nonhuman primates. In Behavioral Toxicology (B. Weiss and V. G. Laties,
eds.), pp. 401-428. Plenum, New York.
Swets, J. A. (1973). The relative operating characteristic in psychology.
Science 182,990-1000.
8-12
-------
Trakhtenberg, I. M. (1974). Chronic Effects of Mercury on Organisms. English
translation, U.S. DREW Publication No. (NIH) 74-473.
Weiskrantz, L. and Cowey, A. A. (1967). Comparison of the effects of striate
cortex and retinal lesions on visual acuity in the monkey. Science 155, 104-
105.
Wright, A. A. and Nevin, J. A. (1974). Signal detection methods for measure-
ment of utility in animals. J. Anal. Behav. 21, 373-380.
Xintaras, C. and Johnson, B. L. (1976). Behavioral toxicology: Early warning
and worker safety and health. Essays in Toxicol. 7, 155-201.
8-13
-------
9. ELECTROPHYSIOLOGICAL
TECHNIQUES IN TOXICOLOGY
DOROTHY E. WOOLLEY
Departments of Animal Physiology and Environmental Toxicology
University of California, Davis
An Overview of "Electrotoxicology"
Most poisons are neuropoisons. Poisons which act upon the nervous system are
particularly effective in producing toxicity because the nervous system is more
sensitive to disruption of function than are other physiological systems. At the
same time, the organism is dependent for survival on an intact and functioning
nervous system,^Thus, hypoxia will result in death of neurons before death of
other cells in the body, which, in turn, may lead to death of the organism if
the destroyed neurons normally regulate vital functions, such as respiration.
This combined sensitivity and vital importance of the nervous system has been
known for a long time. It has become evident, particularly in the last decade,
that low-level exposure to a toxicant may produce deleterious neural effects
which are discovered only upon the utilization of sensitive, sophisticated
testing techniques. The concern today is not so much with neurotoxicants
which produce death immediately, but with the fear that environmental toxi-
cants may reduce functioning of the nervous system in a less obvious, but still
important, way, so that intelligence, memory, or other higher central neural
functions are affected. With this realization has come the birth of behavioral
toxicology.
Behavioral toxicologists are frequencly under pressure from other toxicolo-
gists, who often either don't understand or don't trust behavioral tests, to
relate changes in behavior to changes in other physiological or biochemical
endpoints. Therefore, it is to the advantage of the behavioral toxicologist to
measure as many parameters of nervous system activity as is feasible and to
relate these to changes in other physiological systems. Among these other
*0riginal research described in this report was supported by NIH grant ES-00163.
9-1
-------
parameters of nervous system functioning which may be utilized to detect
and understand neurotoxic effects are electrophysiological measurements.
This chapter is a brief attempt to point out some of the advantages and limi-
tations of electrophysiological recording techniques.
Electrophysiological techniques may be relatively more sensitive indicators
of toxic action than are other parameters. Of course, to be sure that this is
true, other parameters must be measured along with the electrophysiological
recordings. Electrophysiology may be a useful tool to determine some aspects
of the mechanisms of action of toxicants. It is possible to implant electrodes in
various cortical and subcortical brain structures and then to utilize evoked
potential techniques to test the integrity of specific brain systems or parts of
brain systems. The use of microelectrode techniques to determine the effects
of toxicants at the cellular level has proven to be an immensely successful pro-
cedure for elucidating specific types of membrane or synaptic effects of toxi-
cants. For example, research has demonstrated that DDT is an axon labilizer,
i.e., it prolongs the action potential negative afterpotential or causes repetitive
firing of action potentials, probably by prolonging the increased sodium
conductance which accompanies the spike potential and which normally de-
clines during the falling phase of the action potential (Narahashi, 1971). Also,
lead ions have been shown to block transmission in the superior cervical or
other sympathetic ganglion and neuromuscular junction, probably by reducing
release of acetylcholine from presynaptic terminals (Kostial and Vouk, 1957;
Cooper and Manalis, 1974; Silbergeld eta/., 1974). However, the electrophysio-
logical technique most useful to a behavioral toxicologist is probably that of
using animals with chronically implanted brain electrodes,so that the effects
of an agent on spontaneous and evoked brain electrical activity may be re-
corded in awake unanesthetized animals, preferably in association with a be-
havioral task. Recently, event-related slow potentials recorded from the cor-
tical surface with nonpolarizable electrodes have been used to investigate the
effects of drugs on behavior and electrophysiology (Pirch and Osterholm,
1975). These potentials should prove useful for the study of toxic agents.
Because chronically implanted rats are so useful for neurotoxicological studies,
a description of some of the procedures for implantation of brain electrodes
which have been developed in the author's laboratory is presented below.
Descriptions of additional procedures for electrophysiological studies are
also available, e.g., Skinner, 1971.
Preparation of Rats with Chronically Implanted Brain Electrodes for
Studies in Toxicology
The use of animals with chronically implanted brain electrodes makes it possi-
ble to study the effects of an agent in an unanesthetized animal. This technique
has the obvious advantage of eliminating the complicating factor of inter-
action between toxicant and anesthetic, and it permits comparison of changes
in brain electrical activity with changes in spontaneous or experimentally
modified behavior. Rats are particularly convenient to study because a chron-
ically implanted rat may be "plugged in" for recording (Figure 9-1) in its home
9-2
-------
cage and still have the same degree of freedom to move about that it normally
does (Figure 9-2). In this way the novelty of the recording situation is mini-
mized.
Figure 9-1. Rat with chronically implanted brain electrodes plugged in for
recording. The electrodes in the brain are connected to the terminals of a
subminiature connector and are held in place by dental acrylic. The record-
ing cable connects the recording equipment to the brain electrodes via a
subminiature connector which is the mate to the one on the rat's head. The
initial portion of the recording cable is covered with a light weight rubber-
ized material to give some rigidity to the cable.
9-3
-------
Figure 9-2. Tall cages are used as home cages so that implanted rats cannot bump their con-
nectors against the top of the cage. Rats may be plugged in and recorded in these cages to
minimize the novelty of the recording situation.
-------
For chronic implantation, side-by-side electrodes with 0.5-1.5 mm tip sepa-
ration (Figure 9-3) are usually used. Thirty-two gauge stainless steel wire
coated with insulating varnish (Diamel) is stretched slightly to straighten it, and
lengths are cut with fingernail clippers, which readily cut wire of this gauge
without bending the tips. Two wires are held together with a drop of epoxy
adhesive (fast curing, 2-part epoxy) so that one of the tips is longer than the
other (Figure 9-3). The wires are dipped in an instantly drying adhesive con-
taining cyanoacrylate (Krazy Glue) to hold the wires below the bead together.
Next, they are dipped in Epoxylite (No. 6001-M, electrode insulator, Epoxy-
lite Corporation) and baked (3 hours, 150°C). After repeating the dipping and
baking, the long electrode tip is cut off with the clippers to give the desired tip
separation, usually 1.0 mm, and the tips of the electrodes are bared with a
grinding wheel, using a hand-held hobby drill. The other ends of the electrodes
are scraped clean, and each is inserted into a steel tube epoxyed onto a thin
piece of bakelite (or other plastic) held by epoxy to a steel bar (Figure 9-3),
which fits into the standard electrode carriers associated with any stereotaxic
apparatus. Wires from the steel tubing connect to stimulating and recording
equipment so that any bipolar electrode may be used for either electrical
stimulation or recording during positioning of the electrodes. Electrodes are
positioned in various brain structures with the aid of stereotaxic equipment.
STEEL ROD TO
FIT INTO.
ELECTRODE CARRIER
30 GAUGE
STAINLESS
STEEL WIRE
EPOXY BEAD
INSULATED WIRES
STEEL TUBING.
. 0.5-1.5 mm
TIP SEPARATION
SIDE-BY-SIDE
BIPOLAR ELECTRODE
BAKELITE-
ELECTRODE HOLDER
Figure 9-3. A bipolar, side-by-side electrode and electrode holder constructed as de-
scribed in the text. The upper portion of each of the two wires which from the electrode
is scraped clean and inserted into the tubing on the bakelite section of the electrode
holder. The tubing is connected via insulated wires to the stimulating and recording
equipment. The section of bakelite is narrow so that two or more electrodes may be
positioned simultaneously over the rat's head.
9-5
-------
Evoked potential techniques are utilized to assist in accurate placement of both
stimulating and recording electrodes. First, the stimulating electrode is posi-
tioned approximately using stereotaxic coordinates, and the bipolar recording
electrode is lowered until the evoked potential recorded from each of the re-
cording tips is nearly optimal. Then, the stimulating electrode is adjusted
slightly to improve further the elicited response. In other words, the stimu-
lating and recording electrodes are first positioned approximately with stereo-
taxic coordinates, and the positions of each are refined using evoked potential
techniques. By stimulating the olfactory bulb or lateral olfactory tract, it is
possible to place bipolar recording electrodes so that the tips straddle the
prepyriform pyramidal cell bodies and the evoked potentials recorded from
the electrode tips are mirror images of each other, i.e., are 180 degrees out of
phase (Woolley and Timiras, 1965). This requires placing the long electrode
tip in the prepyriform superficial dendritic layer with an accuracy of a few
tenths of a millimeter, a feat which would be impossible using stereotaxic
coordinates only. Similarly, accurate placement in the hippocampus may be
achieved by stimulating the prepyriform cortex and lowering a bipolar record-
ing electrode into the hippocampus until a phase reversal of the evoked poten-
tial between the two recording tips is observed (Woolley and Barren, 1967).
For positioning electrodes in the superior colliculus, light flashes may be used
to elicit visual evoked potentials during implantation. An electrode is lowered
into the superior colliculus until a phase reversal of the elicited potential indi-
cates that the long electrode has penetrated the stratum opticum of the
superior colliculus. The reference electrode used for comparing monopolar
recordings from each tip of a bipolar electrode is usually formed by a wire
wrapped around each of 4 screws positioned at the outer edges of the top of
the skull.
After the electrodes are positioned, small screws are threaded into holes drilled
into the skull. The electrodes are held in place by flowing dental acrylic around
the electrodes and screws. Ends of the electrodes are connected to the pins
of a subminiature socket or plug connector (for example, ITT Cannon sub-
miniature connectors MD1-9SL1 or MD1-9PL1), either by soldering or by
tightly wrapping the electrodes around the connector pins. During recordings,
after recovery from the implantation procedure, another subminiature connec-
tor, which is the mate to the connector on the rat's head, is used to plug in the
rat and connect the implanted electrodes to amplifying, stimulating, and re-
cording equipment via a recording cable (Figure 9-1). Implanted rats are
housed individually in specially tall plastic cages (14" high x 12" x 12") so
that the rat does not bump the connector on top if its head against the top of
the cage. Side panels of the cages are painted black so that the animals are not
distracted by seeing rats in adjacent cages. The front panel is clear so that the
experimenter may view the animal from the front as well as the top (Figure
9-2).
In examining the effects of toxicants on brain electrical activity, both spon-
taneous electrical activity (SEA), i.e., EEC type of activity, and evoked poten-
tials are recorded. For both, recordings are usually bipolar. Monopolar re-
cordings are sometimes carried out in addition to determine if recordings from
one of the bipolar electrode tips is altered more than the other. For monopolar
9-6
-------
recordings, the reference electrode is formed by a wire wrapped around several
screws in the skull, as described above. For the findings described below, most
recordings were bipolar.
It is well known that changes in behavioral state are correlated with changes
in spontaneous and evoked brain electrical activity. Therefore, in evaluating
the effects of an agent on brain electrical activity, the behavioral state during
the recordings must be noted. It is not unusual for a toxicant to alter the level
of arousal. In order to be able to evaluate whether or not the changes in brain
electrical activity produced by an agent are related to changes in arousal, it is
necessary to carry out control recordings during inattentiveness and arousal
so that the patterns characteristic of these states can be recognized and com-
pared with the patterns produced by the toxicant.
Examples of Electrophysiological Studies in Toxicology
Effects of DDT on Brain Spontaneous Electrical Activity (SEA)
The effects of a single dose of DDT, dissolved in corn oil administered per os
by intubation, on brain SEA and on spontaneous behavior were studied in rats
with chronically implanted brain electrodes (Woolley and Barren, 1968;
Woolley, 1970). The single dose LD$Q for DDT was 250 mg/kg, and a dose of
50 mg/kg was the threshold dose for producing grossly observable symptoms.
Hyperexcitability, as evidenced by a startle response to sudden sound, such as
a hand clap, was usually present at the time of maximal effect of DDT, which
was about 8 to 12 hours after administration of 50 mg/kg. In about half of the
animals, mild, whole-body tremors were also evident at this dose.
When 50 mg/kg DDT were administered, the frequency of the slow waves in
the SEA of the olfactory bulb increased steadily from 4 through 12 hours
(Figure 9-4) and then decreased again to reach control levels at 24 hours.
These slow waves track respiration (Woolley and Timiras, 1965,), and an in-
crease in respiration is a sensitive early effect of low doses of DDT in the adult
rat (Henderson and Woolley, 1969). Frequency of the respiratory waves was
330/min 8 hours after DDT in the example shown in Figure 9-4, compared
with 120/min during the control period.
Another significant and sensitive effect of DDT was an increased occurrence
and amplitude of the high frequency "arousal" waves in olfactory electrical
activity, as shown in Figure 9-5. These fast (55 to 75 cycles/sec) waves are
called arousal waves because they occur when the animal is excited, decrease
when he is inattentive, and disappear when he is asleep. During the stimu-
latory stages of DDT poisoning, these waves were consistently present and
increased dramatically for hours at a time. This was, in fact, one of the most
unique effects of DDT. DDT administration also increased the occurrence and
regularity of 7/sec frequency theta waves and increased the amplitude of a
fast frequency of about 40/sec superimposed over the theta waves in the hip-
pocampal SEA (Figure 9-4).
9-7
-------
HOURS AFTER DDT, 50 mg/kg
0 (CONTROLS) 4
BULB0™"™ ^•"^"^VMX^^ ^^-^^^^/^
v > i I r i
CORTEX1 F°R^Xfc|Vlt^Y%^^ |^M^^
tO
00 0.2 sec
H IPPOCAMPUS v^^^Y\AlA\A^/VYv^^ '^\^^A^
500 ,jV
8 (TREMORS) 12
^
^
Figure 9-4. Spontaneous electrical activity recorded from the olfactory bulb, prepyriform
cortex, and hippocampus in awake, unrestrained rats before and 4, 8, and 12 hours after
per os administration by intubation of p,p'-DDT dissolved in corn oil.
-------
OLFACTORY BULB
T500 jiV
CONTROL °-'sec
6 HOURS AFTER 100 mg/kg ODT
Figure 9-5. High frequency "arousal" waves recorded in the spontaneous electrical activ-
ity of the olfactory bulb of the rat before and 6 hours after DDT.
Still other effects of 50 mg/kg DDT on the SEA are shown in Figure 9-6.
During the control period, a sudden noise caused the high amplitude slow
waves characteristic of an innattentive animal to change suddenly to low
amplitude, high frequency waves in the SEA of the frontal cortex, occipital
cortex, reticular formation, and vermis of the cerebellum. These are the typical
arousal patterns for the SEA of these brain areas. The effects of the relatively
low dose of DDT were to cause high frequency "arousal" waves to be continu-
ously present, even in the absence of an arousing stimulus, but these waves
had a significantly higher amplitude than in the usual arousal state. In addition,
DDT also caused the appearance of "spiking" in the SEA of the cerebellum.
High amplitude spikes with about the same frequency as that of the whole-
body tremor were pronounced in cerebellar recordings even in the mild tremors
produced by 50 mg/kg DDT (Figure 9-6), but the spikes were not usually pres-
ent in recordings from other brain areas. The changes in cerebellar electrical
activity appeared to support earlier suggestions that the cerebellum is a primary
target for the action of DDT and that cerebellar dysfunction is basically in-
volved in the symptomatology of DDT poisoning (see review by Hayes, 1959).
Comparison of the effects of a higher dose of DDT (100 mg/kg per os) on the
SEA of the neocerebellum (ansiform lobule), the cerebellar vermis (lobulus
simplex), and the medial and lateral portions of the somesthetic cortex (Figure
9-7), revealed that amplitude of the electrical activity of the neocerebellum
was increased 1 hour after DDT administration, which was before amplitude
of the electrical activity of the vermis was increased and before tremoring was
grossly evident 4-13 hours after DDT. Amplitudes of electrical activities were
maximal in the vermis and neocerebellum at 4 to 7 hours. High amplitude
spiking was still evident in the vermis at 9 and 13 hours, whereas activity in
the neocerebellum at these times was largely restored toward normal. During
the periods of maximal increase in electrical activity, spikes appeared syn-
chronously in the vermis and neocerebellum. This study showed that cerebellar
SEA was increased before gross behavioral changes, such as tremoring, were
9-9
-------
evident, so that increased amplitude of cerebellar SEA was another sensitive
index of DDT action.
CONTROLS
NOISE
FRONTAL CORTEX , | ,' ( (l
>^v%VW-Vi^^
OCCIPITAL CORTEX ' i ' 1
;*v^
RETICULAR FORMATION / ',';
V#';V^^
• ' ' ' ' '' ' \' '
CEREBELLUM, VERMIS ' |
8 HOURS AFTER 50 mg/kg DDT. TREMORS
I ^W^H^
I '^^
T ;0^:ft'^
',,' ,, i,' 'I'' )/ ' '''.'i'l' M'
i^^^',l::fe!'i'^r^;:^
2001 /j V
1 sec
Figure 9-6. The spontaneous electrical activity of 4 brain areas before and 8 hours after
DDT per os in the rat. A sudden noise during the control recording session changed the
patterns of the SEA from those typical of inattention to those typical of arousal. After
DDT administration arousal patterns prevailed and "spiking" occurred in cerebellar
electrical activity even in the absence of external "arousing" stimuli.
9-10
-------
HOURS AFTER DDT, 100 mg/kg per os
0 (CONTROLS) 1
NEOCORTEX,
MEDIAL
2%
2 sec
Figure 9-7. Comparison of the changes in electrical activity of 2 cerebral cortical (somes-
thetic) areas and 2 cerebellar areas produced by DDT administration. Amplitude of
electrical activity increased in the neocerebellum sooner than in the vemnis. Amplitudes
of both cerebellar electrical activities increased more than cortical activities at the height
of the effect of DDT. Spiking continued in the vermis for a longer period of time than in
the neocerebellum during recovery.
The principal conclusions which may be drawn from studies presented thus far
illustrate some of the advantages and limitations of electrophysiological studies
in toxicology. The studies demonstrated that after DDT administration in the
awake, unrestrained animal: (1) changes in spontaneous brain electrical activity
may be detected before changes in spontaneous behavior; (2) the electrical
activity of some brain areas changes before and to a greater extent than that of
other areas; (3) changes in the respiratory waves in the electrical activity of the
olfactory system show that low levels of DDT stimulate respiration and that
this is a sensitive, early index of DDT action; (4) marked increases in the
arousal waves of the olfactory system and in the arousal patterns of other
brain areas show that low levels of DDT are strongly excitatory to the CNS,
in agreement with the behavioral hyperexcitability caused by DDT; and (5) the
greatest changes in brain electrical activity occur in the cerebellum.
9-11
-------
Effects of DDT on Evoked Potentials
The striking effects of DDT in increasing the electrical activity of the cere-
bellum suggest that DDT has a direct stimulatory effect on the cerebellum.
Another possibility is that the increase in cerebellar electrical activity merely
reflects the greater proprioceptive input to the cerebellum because of in-
creased muscle activity during tremors. Therefore, the effect of DDT on visual
and auditory potentials in the cerebellum and other brain areas were investi-
gated. A greater effect of DDT on potentials evoked by light and sound in the
cerebellum than in other brain areas would be compatible with the hypothesis
that DDT preferentially affects the cerebellum.
Auditory potentials were evoked by a 1,000 Hz tone of 30 msec duration and
60 db above threshold as judged by oscilloscope tracings. Light flashes (10
jizsec duration, maximal intensity, 18 inches above the animals) for visual stimu-
lation were produced with a Grass photostimulator (model PS-2). Responses
of various brain areas were amplified and averaged with a Northern Scientific
online computer (model No. 544). Stimulation rate was 1/sec and 50 responses
were summed to produce each averaged evoked potential.
Figure 9-8 shows the typical effects of 100 mg/kg of DDT on auditory evoked
potentials in cerebral cortex, midbrain reticular formation, and vermis of the
cerebellum. In frontal and occipital cortex the responses were depressed at
the height of the effects of DDT which occurred 6 to 12 hours after adminis-
tration. On the other hand, some components of both visual and auditory re-
sponses in the cerebellum and reticular formation were significantly increased
in amplitude. Early components of the cerebellar response, which were often
not present during the control period, appeared after DDT administration,
increased in amplitude during the maximal effects of DDT, and then disap-
peared again during the recovery period (Figure 9-8).
The depression by DDT of the evoked responses in the cerebral cortex may be
explained by a "line-busy" effect, i.e., that the increased spontaneous or back-
ground electrical activity caused by DDT interfered with the neuronal response
to the light or sound stimulus. On the other hand, the effect of DDT on the
cerebellum was unique because both spontaneous and evoked electrical activ-
ities were increased.
The effects of DDT were compared with those of tremorine, which is also a
tremorigenic agent. Responses evoked by sound recorded at several positions
in the cerebellum before and after administration of tremorine clearly showed
that sound-evoked cerebellar potentials were completely eliminated by tre-
morine, in marked contrast to the effect of DDT in increasing amplitudes of
these potentials in the cerebellum. Behaviorally, tremorine produced an ab-
sence of the startle response, the rats appearing to be almost oblivious to visual
and auditory stimuli, in marked contrast to the effects of DDT in increasing
the startle response.
The relation between the potentiation of the startle response and visual and
auditory responses in the cerebellum, as produced by DDT, and the elimination
of the startle response and visual and auditory cerebellar responses, as pro-
duced by tremorine, may be more than coincidental. It has been reported that
9-12
-------
SOUND-EVOKED, AVERAGED POTENTIALS
HOURS AFTER DDT, 100 mg/kg per os
0
1
3
6
10
12
24
48
CEREBELLUM.VERMIS !~ ' OCCIPITAL CORTEX
20 msec
0
1
3
6
10
12
24
48
RETICULAR
FORMATION, MIDBRAIN
FRONTAL CORTEX
20 msec
Figure 9-8. Effects of a single dose of DDT on averaged auditory evoked potentials re-
corded simultaneously from four different brain areas in the awake, unrestrained rat. Note
the changing scale for amplitude of cerebellar evoked potentials. Amplitudes of the re-
sponses recorded from the frontal and occipital cortex were depressed by DDT, especially
6 hours post administration, whereas some components of the responses evoked in the
cerebellum and reticular formation increased in amplitude. The major effect of DDT was
to cause the appearance and increase the amplitude of an early component of the response
recorded from the cerebellum. This early response was first evident 6 hours after DDT,
reached maximal amplitude at 10 hours, and had disappeared by 48 hours post admini-
stration.
9-13
-------
lesions in portions of the vermis, which also receive visual and auditory inputs,
result in absence of the startle response (Chambers and Sprague, 1955).
DDT and tremorine produced similar increases in cerebellar SEA. These in-
creases are attributed to increased proprioceptive input to the cerebellum
resulting from tremoring and increased muscle activity.
Another drug which has been shown to increase visual and auditory responses
in the cerebellum, at least in the cat, is a-chloralose (Fadiga and Pupilla, 1964).
Amplitudes of visual responses recorded from several electrode positions in
the cerebellum were clearly increased by the drug. This is particularly interest-
ing because a-chloralose also caused an exaggerated startle response to sudden
loud noises. This lends further support to the hypothesis that DDT increases
the startle response by an action on the cerebellum, and that the cerebellum
is basically related to the startle response.
The pathways by which visual and auditory potentials reach the cerebellum are
not completely understood. However, there is evidence that these teleceptive
impulses relay in the tectum (superior and inferior colliculi) and in the brain-
stem enroute to the cerebellum (Fadiga and Pupilla, 1964). The next study
then undertook to investigate the effect of DDT and a-chloralose on responses
evoked in the cerebellum by electrical stimulation of the superior colliculus
and reticular formation (Figure 9-9). During the control period it was almost
impossible to elicit anything except a low-amplitude response in the cerebellum
by stimulating this way. However, after administration of either chemical
agent, the cerebellar response was increased many-fold. Additional studies
showed that the responses to light and sound in the superior colliculus and re-
ticular formation were not themselves potentiated by DDT. Thus, the poten-
tiation by DDT of the cerebellar response to electrical stimulation of the su-
perior colliculus and reticular formation appears to be due to a direct effect on
the cerebellum or on a brain area relaying the impulses from the lectum to the
cerebellum.
The studies presented so far represent an effort to understand the types of
effects produced on the brain by acute exposure to DDT. An important ques-
tion was whether or not DDT preferentially affected some brain areas more
than others, and, if so, whether this could be related to the symptomatology
of DDT poisoning. The evoked potential studies provide evidence that DDT
in sublethal doses affects the cerebellum particularly, that this effect is one
of increased excitability, and that this effect on the cerebellum may be a factor
contributing to the exaggerated startle response characteristic of DDT poison-
ing. Comparison of the effects of DDT with those of drugs which produce
some of the clinical symptoms also produced by DDT helped delineate some of
the rather unique features of the effects of DDT on the brain.
Distribution of DDT in Brain and Spinal Cord
Next, the concentrations of DDT in several brain areas and spinal cord of the
rat were determined at various times after a single dose of 100 mg/kgperos of
DDT dissolved in oil (Table 9-1; Woolley and Runnells, 1967). The hope was
that a correlation between symptomatology and CIMS DDT concentrations
9-14
-------
EFFECTS OF a-CHLORALOSE
EFFECTS OF DDT
20 msec
PRETREATMENT
POSTTREATMENT
Figure 9-9. Effects of alpha-chloralose and DDT on responses
evoked in the cerebellar vermis by electrical stimulation of the
superior colliculus (SC) andmidbrain reticular formation (RF).
Both chemical agents potentiated the responses recorded in the
cerebellum.
9-15
-------
could be established. Because of the well-known lipid solubility of DDT, it
was of interest to determine the pattern of uptake and release of DDT in a
sample of neocortex, which is gray matter only, and the pattern in bramstem
(medulla and pons) and spinal cord, which have a high percent of white matter
and thus of lipid.
When DDT concentrations were expressed in terms of wet tissue weight (Table
9-1), concentrations were 40 percent higher in the neocortex than in the
spinal cord at 6 hours. However, at 12 and 24 hours DDT concentrations were
significantly higher in brainstem and spinal cord than in neocortex and cere-
bellum, suggesting that CNS areas with high lipid (i.e., high myelin) concen-
trations took up DDT more slowly but retained it longer than did CNS areas
which were primarily gray matter only. When DDT concentrations were cal-
culated for the whole brain and compared with spinal cord concentrations,
it was evident that DDT concentrations per unit fresh weight were about the
same at 6 and 12 hours but were about twice as high in the spinal cord as in
the brain at 24 hours.
Table 9-1. DDT CONCENTRATIONS IN TISSUES OF THE RAT WITH TIME AFTER
A SINGLE DOSE OF DDT (100 mg/kgperos)
Tissue
Neocortex
Cerebellum
Brainstem
Remaining brain
Whole brain
Spinal cord
Liver
Plasma
DDT concentration (l^g/g fresh tissue)
Hours after DDT administration
6
35.2a
±4.0
29.0
±3.1
27.2
±2.2
20.0
±1.9
22.0
±2.4
24.5
±2,0
216
±17
±5.9
±0.9
12
20.2
± 2.5
20.9
±.2.3
34,9
±4.2
23.2
±2.7
22.2
±2.5
27.2
±2.8
165
±9
3.2
±0.3
24
13.1
±1.7
13.6
±1.5
17.5
±1.9
11.2
±0.5
11.8
±0.5
25,1
±3.1
43
±5
3.4
±0.4
% Lipid In
fresh tissue
6.0
±0.5
6.8
±0.5
16.1
±0.7
7.2
±0.4
7.2
±0.3
17.9
±0.6
4.2
±0.6
—
3Mean ±SE for 8 rats per group. From Woolley and Runnells, 1967.
In these animals, hyperirritability and tremor, the outstanding symptoms pro-
duced by the dose of DDT used, were generally more intense at 6 than at 12
hours and so appeared to correlate with the peak concentrations in the neo-
cortex and cerebellum. However, because the time periods selected were so
9-16
-------
widely spaced, this correlation cannot be made with certainty. According to
the literature, reviewed by Hayes (1959) and Shankland (1964), during DDT
poisoning in the intact animal, a fine tremor is present initially and is even-
tually replaced by a coarse, low frequency, tremor. In spinal animals the
tremor is usually coarser in the body parts innervated by the spinal cord caudal
to the lesion than it is in the intact animal. The fine tremor may occur when
DDT effects on gray matter in brain predominate, whereas the coarse tremor
may be indicative of predominant effects on the spinal cord.
Effects of Parathion
Parathion is a widely used, highly toxic organophosphate insecticide, which is
oxidatively desulfurated in vivo to its active metabolite, paraoxon. Paraoxon
is a powerful acetylcholinesterase (AChE) inhibitor and so is believed to affect
cholinergic synapses or junctions by reducing the rate of breakdown of acetyl-
choline (ACh), which in turn, permits buildup of excess ACh.
The effects of parathion administration (3 mg/kg sc) on brain electrical activity
were determined in six female Sprague-Dawley rats with chronically implanted
brain electrodes (Woolley, 1976). Visual and auditory evoked potentials were
recorded in the visual cortex, superior colliculus, midbrain reticular formation,
and cerebellum. Potentials were elicited and averaged as described above.
The typical effects of parathion on visual evoked potentials recorded in the
superior colliculus (Figure 9-10) and the visual cortex were to increase latency,
i.e., the time from the flash until the beginning of the response, and to de-
crease amplitude of the waves of the response. Latency was increased 40 per-
cent or more in responses recorded from the visual cortex and superior colli-
culus, whereas latency of responses recorded simultaneously in the cerebellum
and reticular formation were affected less. Amplitudes of the main components
of the responses of the visual cortex and superior colliculus were reduced to
one-third or less of control amplitudes. Amplitudes of the visual evoked
responses recorded in the cerebellum and reticular formation also tended to
be reduced, but variability was greater for these responses than for those re-
corded in the cortex and colliculus. Effects on amplitude and latency were
usually marked 2 and 4 hours after parathion administration, although they
were sometimes pronounced as early as 45 minutes after administration. The
responses had returned to pretreatment values by 8 hours after parathion ad-
ministration.
The effects of parathion administration on sound-evoked potentials in the
reticular formation (Figure 9-11) were to decrease markedly the amplitude
2 and 4 hours after administration with some recovery at 8 and 20 hours. In
contrast to effects on visual evoked potentials (Figure 9-10), marked changes
in latency did not occur in auditory evoked potentials. Thus, the effects of
parathion administration on auditory evoked potentials were less marked than
on visual evoked potentials.
These observations permit two conclusions: (1) there must be a cholinergic
link somewhere in the generation of these visual and auditory evoked poten-
tials, and (2) recovery of visual evoked potentials from the effects of parathion
9-17
-------
administration is surprisingly fast when one considers that paraoxon is believed
to produce irreversible, or at least only slowly reversible, inhibition of AChE
activity (O'Brien, 1967).
TIME AFTER PARATHION, 3 mg/kg sc
45 min
2 his
4hrs
8hrs
20hrs
RAT Bl 10/8/68
Figure 9-10. Effects of parathion on averaged flash-evoked potentials in the
superior colliculus of the adult female rat. Latency was increased and ampli-
tude of the major component was decreased 2 and 4 hours after administra-
tion of parathion. Recovery was evident at 8 hours.
9-18
-------
TIME AFTER PARATHION, 3 mg/kg sc
45 min
2hrs
4hrs
8hrs
20hrs
Bl RFL 10-68
20 msec
Figure 9-11. Effects of parathion administration on averaged sound-evoked po-
tentials in the midbrain reticular formation of the adult female rat. Amplitude
was decreased at 2 and 4 hours post parathion but latency was not altered
markedly.
Regarding the first point, histological evidence for involvement of cholinergic
synapses in generation of visual evoked potentials has been provided by Shute
and Lewis (1967), who have described AChE-containing fibers rising from the
midbrain and passing via the dorsal and ventral tegmental pathways directly
and indirectly to the anterior and posterior colliculi, lateral and medial genic-
ulate bodies, pretectal area, and cerebral cortex.
Some controversy exists concerning the rate of recovery of brain AChE activ-
ity after parathion adminstration in the rat. Although an early report described
recovery of brain AChE activity 4 hours after administration of parathion in
the rat (Du Bois et al., 1949), another stated that several days are required
(Davison, 1953), and still another that a month is required for recovery
9-19
-------
(Giachetti et al., 1966). Therefore, we determined the time-course of inhibi-
tion and recovery of AChE activity after intraperitoneal (ip) and subcutaneous
(sc) administration of 2 mg/kg parathion in the adult female rat. The colori-
metric method of Ellman et al. (1961) was used to measure AChE and ChE
(pseudocholinesterase) activities. The results (Figure 9-12) showed that blood
AChE and plasma ChE activities recovered rather rapidly, i.e., within the 1
week of this study. However, brain and spinal cord AChE activities reached
50
45
40
I 35
si
1 30
c
£ 25
CAUDATE AChE
CORTEX AChE
i i i
234567 01
7.5
BRAIN STEM AChE
234567
SPINAL CORD AChE
234567
PLASMA ChE
01234567 0123456
DAYS AFTER PARATHION (2.0 mg/kg) ADMINISTRATION
IP
SC
Figure 9-12. Effects of ip and sc administration of parathion in the rat on AChE activities
in brain, spinal cord and blood and on ChE activity in plasma. Each point represents the
mean, and the vertical bracketed lines represent the standard errors about the mean for
6 animals in this and the following figure.
9-20
-------
peak inhibition about 9 hours after parathion administration and showed
essentially no recovery during the study. When the experiment was repeated
and recovery followed over a longer time period (Figure 9-13), it was again
observed that blood AChE and plasma ChE activities had recovered by 1 week
after ip administration of parathion (2.25 mg/kg). AChE activities did not
become inhibited as rapidly in brain as in blood and required a longer time to
recover in brain than in blood. Inhibition of AChE activities in caudate, cere-
bral cortex, and brainstem was greater at 1 week than at 6 hours post para-
thion administration; recovery was complete between 2 and 4 weeks after
administration of parathion.
From these results it is evident that the effect of parathion administration on
visual evoked potentials disappeared despite continuing inhibition of AChE
CORTEX AChE
01 234 01 234
WEEKS AFTER PARATHION (2.25 mg/kg IP) ADMINISTRATION
Figure 9-13. Effects of ip administration of parathion on inhibition and recovery of
AChE and ChE activities in various brain areas, blood or plasma of the rat.
9-21
-------
activity in the rat. Visual evoked potentials recovered within 8 hours after
parathion administration, but blood AChE activity required about a week to
recover, and brain AChE activities required 2-4 weeks for recovery. Effects
on evoked potentials were maximal when CNS AChE activities were falling,
and these effects disappeared after AChE inhibition had reached its maximum
and stabilized.
One interpretation of these findings is that development of adaption or toler-
ance to depression of AChE activity in the CNS occurred after parathion ad-
ministration in rats and monkeys. The problem of tolerance in relation to a
number of anticholinesterase agents has been reviewed recently (Bignami
et a/., 1975). Usually, an AChE inhibitor is administered and cholinergic symp-
toms are immediately evident, but despite continued AChE inhibition, the
symptoms soon disappear. The major present hypotheses regarding the mech-
anism of onset of tolerance to AChE inhibition are as follows: (1) ACh levels,
which become elevated immediately after AChE inhibition, become lower
with time despite the persisting depression of AChE activity; (2) prolonged
exposure to elevated ACh levels brings about a reduction in receptor sensi-
tivity, thus counteracting the effects of AChE inhibition and allowing a recov-
ery of function; and (3) the activity of other neurophysiological systems,
either agonistic or antagonistic, alters to counteract the overstimulation of
cholinergic systems.
Still other interpretations of the dissociation between the neurological effects
and AChE inhibition after parathion administration remain to be explored.
AChE and the ACh receptor must share some structural similarities because
both combine with ACh. Paraoxon also combines with AChE at the active
esteratic site and therefore probably also has the structural requirements to
react with the ACh receptor. It is possible that paraoxon first reacts with AChE
to inhibit it and thus to permit a buildup of ACh at cholinergic synapses. After
some delay paraoxon may next combine with the ACh receptor and in this way
reduce receptor sensitivity to the elevated ACh levels and permit a return to
more normal functioning, even though AChE activity is strongly inhibited.
Still another explanation has been provided recently by observations that
synaptosomes prepared from rats shortly after injection of a single dose of
paraoxon showed significantly increased release and synthesis of ACh. How-
ever, when a slightly smaller dose of paraoxon was injected once daily for five
days for an even greater total injection, the rate of synaptosomal synthesis and
release of ACh did not differ from controls even though AChE activity was
strongly inhibited.
This may mean that paraoxon initially acts on synaptosomes to increase ACh
release and synthesis, thus causing cholinergic symptoms, but that the synapto-
somes adapt to this effect of paraoxon, and so the cholinergic symptoms disap-
pear (Speth eta/., 1975). If further work substantiates this new finding, it
would suggest that the development of "tolerance" to the AChE inhibiting
effects of parathion administration is not an actual tolerance but merely re-
flects the fact that the mechanism of action is primarily on neuretransmitter
release and that AChE inhibition is of lesser importance in the symptoma-
tology of parathion poisoning, at least at low doses of parathion.
9-22
-------
The Value of a Multifaceted Approach in Neurotoxicology
A strong argument for using a multifaceted approach for investigating the ef-
fects of a toxicant is that an investigator may simply be misled regarding the in-
tensity, time, and duration of action of a toxicant if he considers only one
endpoint. Each of the various methods for assessing neural function has
its advantages and limitations. Behavioral tests are usually very sensitive in
their ability to detect the effects of low levels of toxins. A major reason for
this sensitivity also accounts for one of the more frustrating aspects of behav-
ioral testing. In behavioral tests the animal is frequently treated like a "black
box." An input of some sort is applied to this "box," and the output, behavior,
is measured. Modifications in behavior may be produced by toxins or other
agents because of changes in any one or more of the many links in the neural
chain of events between input and output or by changes in non-neural systems
which secondarily alter the nervous system. The result istfhatthe behavioralist
may be left with the satisfying feeling of having recorded a change in behavior
but with the frustrating feeling of not knowing why the change has occurred.
To determine "why," it is usually necessary to perform additional tests. These
other tests may involve electrophysiological techniques or biochemical proce-
dures.
In a clinical setting it is frequently easier to measure some biochemical end-
point, e.g., blood AChE activity, than to conduct extensive behavioral or
electrophysiological testing. Hence, experimental analyses for relevant bio-
chemical changes or for levels of a toxin or its metabolites should be carried
out in readily obtainable tissues such as blood, urine, or hair, along with be-
havioral or electrophysiological testing. From the data thus obtained, it may
be possible to predict biochemical measurements which may be carried out
clinically to provide a warning that nervous sytem functioning may be impaired
at a certain level of change in tissue components. Naturally, there may be pit-
falls to such an approach. In the study reported here, the early effects of
parathion poisoning on brain evoked potentials were correlated with falling
levels of blood AChE activity but not with the absolute level of AChE activity.
The principal advantage of electrophysiological testing, using animals with
chronically implanted brain electrodes and recording while the animals are
awake, unanesthetized, and relatively unrestrained, is that the differential ef-
fects of a toxicant on various brain areas or brain systems may be investigated.
This, in turn, should lead to a better understanding of the mechanism of action
of the toxicant. Using these procedures in previous work in this laboratory, it
was possible to demonstrate the selective activation of the cerebellum and ol-
factory brain areas by DDT (Woolley and Barren, 1968). It was also demon-
strated that parathion administration had greater effects on visual evoked than
on auditory evoked potentials in rats with chronically implanted brain elec-
trodes and that there was a dissociation between the time course of effects of
parathion administration on evoked potentials and on brain AChE activities.
Electrophysiological techniques may be more sensitive than some other end-
points to the effects of a toxin. Changes in spontaneous brain electrical activ-
ity after DDT were evident before changes in spontaneous behavior, for
9-23
-------
example. Brain electrical activity may be recorded continuously before, during,
and after administration of a toxin, so that each animal may serve as its own
control.
Finally, however, an obvious limitation of electrophysiological techniques is
that the investigator must have specialized equipment and training to be able
to record and interpret electrophysiological data. Perhaps, this is the reason
for the relative rarity of "electrotoxicologists."
References
Bignami, G., Rosic, N., Michalek, H., Milosevic, M., and Gatti, G. L. (1975).
Behavioral toxicity of anticholinesterase agents: Methodological, neuro-
chemical, and neuropsychological aspects. In Behavioral Toxicology (B.
Weiss and V. G. Laties, eds.), pp. 155-215. Plenum Press, New York.
Chambers, W. W. and Sprague, J. M. (1955). Functional localization in the
cerebellum. II. Somatotopic organization in cortex and nuclei. Arch.
Neurol. Psychiatry 74, 653-680.
Cooper, G. and Manalis, R. (1974). Effects of polyvalent cations on synaptic
transmission in frog neuromuscular junction and frog sympathetic ganglion.
In Behavioral Toxicology (C. Xintaras, B. Johnson, and I. deGroot, eds.),
pp. 267-276. U.S. DHEW Publication No. (N10SH) 74-126.
Davison, A. N. (1953). Return of cholinesterase activity in the rat after inhi-
bition by organophosphorus compounds. 1. Diethyl p-nitrophenyl phos-
phate (E600, paraoxon). Biochem. J. 54, 583-590.
Dubois, K. P., Doull, J., Salerno, P. R., and Coon, J. M. (1949). Studies on the
toxicity and mechanism of action of p-nitrophenyl diethyl thionophosphate
(parathion), J. Pharmacol. Exp. Therap, 95, 79-91.
Ellman, G. L., Courtney, K. D., Andres, Jr., V., and Featherstone, R. M.
(1961). A new and rapid colorimetric determination of acetylcholinesterase
activity. Biochem. Pharmacol. 7, 88-95.
Fadiga, E. and Pupilla, G. C. (1964). Teleceptive components of cerebellar
function. Physiol. Rev. 44, 432-486.
Giachetti, A., Ciappi, O., and Baccini, C. (1966). Persistence of cholinesterase
inhibition in rat brain after treatment with parathion: Pesticide residues in
relation to public health. Ric. Sci. (Italian) 36, 1081-1082.
Hayes, W. J., Jr. (1959). Pharmacology and toxicology of DDT. In DDT.
The Insecticide Dichlorodiphenyltrichloroethane and Its Significance
(P. Muller, ed.), Voll. II. Birkhauser, Basel.
Henderson, G. L and Woolley, D. E. (1969). Mechanisms of neurotoxic action
of 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (DDT) in immature and
adult rats. J. Pharmacol. Exp. Therap. 175, 60-68.
9-24
-------
Kostial, K. and Vouk, V. (1957). Lead ions and synaptic transmission in the
superior cervical ganglion of the cat. Brit J. Pharmacol. 12, 219-222.
Narahashi, T. (1971). Effects of insecticides on excitable tissues. Adv. Insect.
Physiol. 8, 1 -93.
Pirch, J. H. and Osterholm, K. C. (1975). Drug-induced alterations of slow
potential responses in the rat. Pharmacologist 17, 189.
O'Brien, R. D. (1967). Insecticides. Action and Metabolism. Academic Press,
New York.
Shankland, D. L. (1964). Involvement of spinal cord and peripheral nerves in
DDT-poisoning syndrome. Toxicol. Appl. Pharmacol. 6, 197-213.
Shute, C. and Lewis, P. (1967). The ascending cholinergic reticular system:
Neocortical, olfactory and subcortical projections. Brain 90, 497-520.
Silbergeld, E., Fales, J., and Goldberg, A. (1974). The effects of inorganic
lead on the neuromuscular junction. Neuropharmacology 13, 795-801.
Skinner, J. E. (1971). Neuroscience: A Laboratory Manual. Saunders,
Philadelphia.
Speth, R., Dettbarn, W., and Schmidt, D. (1975). Cholinesterase (ChE) inhi-
bition and synaptosomal acetylcholine (ACh) synthesis and release follow-
ing in vivo paraoxon (Px) adminstration. Pharmacologist 17, 388.
Woolley, D. E. (1970). Effects of DDT on the nervous system of the rat. In
The Biological Impact of Pesticides in the Environment (J. W. Gillett, ed.),
Enviornmental Sciences Series No. 1, Oregon State University Press,
Corvallis.
Woolley, D. E. (1976). Some aspects of the neurophysiological basis of in-
secticide action. Fed. Proc. 35, 2610-2617.
Woolley, D. E. and Barren, B. (1967). Hippocampal responses evoked by
stimulation of the prepyriform cortex in the rat. Electroencephalog. din.
Neurophysiol. 24, 63-74.
Woolley, D. E. and Barron, B. (1968). Effects of DDT on brain electrical
activity in awake, unrestrained rats. Toxicol. Appl. Pharmacol. 12, 440-
454.
Woolley, D. E. and Runnells, A. (1967). Distribution of DDT in brain and
spinal cord of the rat. Toxicol. Appl. Pharmacol. 11, 389-395.
Woolley, D. E. and Timiras, P. (1965). Prepyriform electrical activity in the
rat during high altitude exposure. Electroencephalog. din. Neurophysiol.
18, 680-690.
9-25
-------
10. NEUROBEHAVIORAL
TOXICOLOGY: PROBLEMS AND
METHODS IN HUMAN RESEARCH
DAVID A. OTTO
U. S. Environmental Protection Agency
University of North Carolina
Other papers in this volume describe methods for studying the effects of toxi-
cants on the behavior of species ranging from mice to monkeys and sheep.
Research on the human species, however, imposes a complex set of contraints,
both ethical and empirical, which essentially limit the investigator to noninva-
sive procedures. This chapter will review these constraints and describe a useful
method for studying the behavioral and electrophysiological effects of toxicant
exposure in man.
The effect of chemical insult on central nervous system (CNS) function has
often been inferred from behavioral measures such as visual threshold
(McFarland et a/., 1944) or temporal discrimination (Beard and Wertheim,
1967). Direct measures of CNS function are ultimately required to validate
behavioral inferences and to elaborate the neurophysiological effects of toxi-
cant exposure on brain tissue. Electrocortical events recordable from the scalp
of humans provide a noninvasive measure of electrical activity of the brain. If
these events are recorded during the performance of sensorimotor or cognitive
tasks, changes in behavior and CNS function can be measured concurrently.
Moreover, neuroelectric events can be recorded from the scalp or from elec-
trodes implanted in specific brain structures (cf. Woolley, this volume). Event-
related potentials of the brain thus offer a promising bridge for the compara-
tive study of neurobehavioral toxicant effects in man and animal.
Methodological Problems in Human Research
Individual Differences
Philosophers have devoted volumes to the concept of free will, or the tendency
of man to behave in a nondeterministic, unpredictable manner. In the labora-
tory this translates to individual differences and intersubject variability. If free
10-1
-------
will is the unique characteristic which differentiate man from beast, it poses a
great nuisance in behavioral toxicology. Application of stringent criteria based
on standard personality and medical inventories, physicals, and interviews to
secure "normal, healthy" populations has been remarkably unsuccessful in
reducing intersubject variability. No matter how rigorously subjects are
screened in preparation for human experiments, there is no hope of ever
attaining a pure laboratory strain of.Homo sapiens.
This problem is particularly severe when the objective is to define the threshold
level at which a given toxicant produces a significant impairment in function.
Threshold effects, by definition, are "just noticeable differences."
How can individual variability be minimized in the study of toxicant thresh-
olds? The standard approach is to use a repeated-measures design in which each
subject is exposed double-blind to multiple levels of a toxicant, including a "no
drug" control or baseline condition. The variability attributable to individual
differences can then be partialled out by subtracting control values from expo-
sure values. The residual, theoretically, represents the effect of toxicant expo-
sure. One disadvantage of the repeated-measures design, compared to an inde-
pendent-measures design, is a loss of degrees of freedom and, therefore, a loss
in statistical power. Repeated-measures designs also place restrictions on expo-
sure regimens due to "carry-over" effects, a problem which will be discussed
later.
Phylogenetic Differences in Brain Function
Language Capacity
One of the complaints voiced frequently by behavioral toxicologists is the
length of time required to train a rat or monkey to perform a simple dis-
crimination. Homo sapiens, with his advanced language capacity, can be
handed a written set of instructions instead of patiently shaping his behav-
ior with operant conditioning techniques. The training period is certainly
shorter in man, but the advantage gained is often illusory.
In communicating verbally or in writing, the investigator is faced in human
research with a dismaying problem called "instructional set." Instuctions must
be carefully standardized across subjects. The unique blend of expectancies,
motivation, and experience which each subject brings to the laboratory has
considerable bearing on how the individual responds to instuctions. More
than likely, the subject will be a college student, peering skeptically at the
instructions and trying to second-guess the investigator.
Cortical Development
Language capacity is an example of cortical development which differentiates
Homo sapiens on the phylogenetic scale. Man has a proportionately larger
cortex, particularly in the frontal region, than most animals. This capacity
permits the study of more complex "cognitive" tasks than might be con-
sidered with other species. Again, the advantages are probably illusory, since
10-2
-------
complex tasks yield complex results which may obscure the simple effect of
toxicant exposure.
The extensive CO studies of Beard and Grandstaff (1975) provide a good illus-
tration. Impairments in performance due to CO exposure were observed in
simple visual and auditory signal detection tasks, but when subjects were re-
quired to make a complex spatial rotation, no CO effects were observed. Beard
and Grandstaff concluded that the physiological arousal produced by the com-
plex, challenging task completely compensated for the subtle impairment ob-
served during CO exposure in other simple, monotonous tasks.
Sensory Dominance
Another dimension of brain function which varies widely among species is
sensory specialization. Olfaction is the dominant modality in rats, while vision
and hearing are dominant in man. Differences in sensory dominance compli-
cate interspecies comparisons and must be taken into account in experimental
design. Sensory discrimination tasks appropriate for one species may not be
appropriate for another. Cats, for instance, have very poor color vision, while
primates have excellent color vision. Unless the toxicant under study is known
to have a modality-specific effect, it would seem prudent to challenge a domi-
nant, rather than a minor sensory function.
In practice, the investigator frequently does not know which sensory modality,
if any, will be sensitive to toxicant exposure. A good strategy is to challenge
multiple modalities to determine comparative sensitivities. Misconceptions can
easily arise if observation is limited to a single modality, particularly when the
results are negative. Dyer and Annau (in press), for instance, failed to find any
low level CO effect on visual evoked potentials in the rat. Groll-Knapp et a/.
(in press), however, compared the effects of low level CO on somatosensory
(SEP), auditory (AEP), and visual evoked potentials (VEP) in humans. The
SEP was maximally sensitive, the AEP was less sensitive, and the VEP was
insensitive to CO effects. Although it is risky to extrapolate across species, the
significance of Dyer and Annau's negative findings is difficult to evaluate with-
out comparative evidence from other sensory modalities in the rat.
Ethical Considerations
Federal Requirements
Some scientists would prefer to leave ethical considerations to philosophers or
theologians, but the federal government has legislated otherwise in the case of
human research. The National Research Act of 1975 (PL 93-348) established
guidelines to safeguard the rights and welfare of human subjects. The Depart-
ment of Health, Education, and Welfare has assumed primary responsibility
for implementation and enforcement of PL 93-348 in accordance with the
Code of Federal Regulations (45CFR46). DHEW-supported activities conduct-
ing human research are required to establish an Institutional Review Board to
assure the protection of human subjects. Investigators must obtain approval
from the Board before initiating any study of human subjects and must submit
periodic reports to the Board to assure compliance with federal guidelines.
10-3
-------
Informed consent must be obtained from subjects (or legally authorized repre-
sentatives in the case of minors, prisoners, or individuals who are mentally or
physically incompetent). To illustrate the stringency of requirements, the fol-
lowing information must be provided to satisfy the definition of "informed
consent" (45CFR§46.103):
o A fair explanation of the procedures to be followed and their pur-
poses, including identification of any procedures which are experi-
mental;
o A description of any attendant discomforts and risks reasonably
to be expected;
o A description of any benefits reasonably to be expected;
o A disclosure of any appropriate alternative procedures that might
be advantageous for the subject;
o An offer to answer any inquiries concerning the procedures; and
o An instruction that the person is free to withdraw his consent and
to discontinue participation in the project or activity at any time
without prejudice to the subject.
Federal agencies engaged in human research are subject to further constraints
on the type of information that may be collected (Privacy Act of 1974) and
the manner in which the information may be disseminated (Freedom of Infor-
mation Act of 1966). The objectives of these regulations are laudable, although
the practical results are increased paperwork and slower implementation of
research.
Exposure Limits
Ethical considerations limit the study of toxicant effects to relatively low
exposure levels in adult populations. Therefore, only the lower segment of a
dose-response curve can be experimentally defined in humans, although acci-
dental or occupational exposures may supplement data obtained in the labora-
tory.
The usual objective in studying the effects of low concentrations is to deter-
mine the threshold toxicologic dose (TTD) or "no effect" level. The TTD may
be defined as the concentration below which no functional impairment can be
demonstrated. In principle, experimental procedures to determine TTD and
LDso are similar, i.e., methods to estimate the dose at which 50 percent of
the true population will exhibit a specified effect within prescribed statistical
confidence limits. Death is an absolute, unambiguous criterion of effect in
determining the LD50. Functional impairment, on the other hand, is the
relatively ambiguous criterion of effect in determining TTD. This criterion is
operationally defined relative to performance in a nonexposed control copu-
lation.
10-4
-------
If an overt behavioral parameter—such as the error rate in a vigilance task- is
chosen as the criterion function, then the physiological significance of any
observed functional impairment must be inferred from behavior. For example,
what is the physiological significance of an increase in the number of visual
targets missed during exposure to 75 ppm carbon monoxide compared to clean
air? Although the observed effect may be significant at the 0.05 level of confi-
dence, behavioral, statistical, and physiological significance are not synony-
mous.
Furthermore, the toxicologic threshold is a function of methodological sensi-
tivity as well as absolute chemical potency. That is, the absence of an observed
effect may be due to the use of an inappropriate index or a crude measurement
device which is incapable of detecting an actual toxicologic effect. The choice
of an appropriate test assumes that the investigator knows in advance which
function is most likely to be affected. With new drugs or untested toxicologic
substances, this assumption is frequently untenable. In this case, preliminary
study of multiple functions must be undertaken to identify which is most
sensitive to the substance under investigation. When the appropriate function
has been identified, the investigator must then select the most sensitive mea-
surement available. In the behavioral sciences, this choice is problematic since
there are as many different tests as there are pineftrees in Georgia. The need for
standardized, properly-validated test batteries is critical in behavioral toxi-
cology and in the behavioral sciences in general.
The investigator is limited in the duration as well as the level of toxicologic
exposure in human research. Economic, social, and ethical considerations limit
the length of time that humans can be confined in an experimental chamber.
Human studies generally consist of short-term exposures lasting no more than
2-4 hours. Repeated acute exposures may be made at periodic intervals of days,
weeks, or months. Between exposures, however, the subject generally departs
from the laboratory and the sphere of experimental control.
Acute high-level exposures occur occasionally in humans as a result of acci-
dents or natural disasters. Chronic low-level exposures are also encountered in
occupational environments. The investigator has no control over such expo-
sures, and the conclusions which can be drawn from these isolated instances are
limited. Therefore, chronic, in utero, and longitudinal toxicologic studies must
be undertaken in species other than Homo sapiens.
Noninvasive Measures
Human research in normal adult populations is limited to measurements which
do not require any major surgical intervention. If there is no added risk to
patient health, basic researchers can often collaborate with physicians treating
patients with acute disorders which entail surgical intervention. Penfield and
Jasper (1954), for instance, contributed greatly to the knowledge of neuro-
physiology by electrically stimulating the cortex of patients undergoing stereo-
taxic craniotomy for treatment of intractible epilepsy. This kind of fruitful
collaboration, however, is rarely available to the behavioral toxicologist except
in the case of accidental poisoning. Thus, systematic studies of physiological
10-5
-------
parameters which require implantation of cannulae of electrodes must be
undertaken with nonhumans.
Sex Differences
The variable of sex is often ignored in human research. The potential danger to
an unborn fetus, however, precludes the use of pregnant women in toxicolog-
ical research. Nor can the investigator rely on verbal report, since a prospective
subject may not know if she is pregnant. Women should be given pregnancy
tests before exposure.
Menstrual cycles also pose a problem which must be taken into account, par-
ticularly if the experiment requires repeated exposures at fixed intervals (e.g.
weekly). Since the behavioral effects of toxicants, particularly at low concen-
trations, tend to be extremely subtle, the physiological and psychological
states of subjects must be maintained as constant as possible.
In view of these problems, the simplest recourse is to limit human toxicological
research to male subjects. This alternative can pose problems in the present
climate of sexual egalitarianism. For instance, three young ladies walked into
our laboratories recently, demanding equal time and equal pay. When they
were shown what the experiment would entail—pedaling a bicycle ergometer in
a transparent plastic chamber while stripped to the waist for ECG recording-
two blushed and left the laboratory. The third, however, smiled and asked
when she could begin. A practical difficulty in accommodating such egalitari-
anism is that sample size must be doubled to control for sex differences.
Experimental Control in Human Research
Many variables such as diet, length and pattern of sleep, and learning history
can be easily controlled in animal populations. These variables are extremely
difficult to control in human subjects who come into the laboratory for short-
term acute exposures. Even with elaborate instructions and questionnaires, it
is impossible to control or determine unequivocally what subjects do outside
the laboratory.
One approach to this problem is to continously monitor the behavioral and
physiological state of subjects for an extended period of time before, during,
and after exposure. Telemetry systems offer a flexible tool for remote moni-
toring of free-roving subjects in work or home environments. Live-in facilities
which permit continuous, intensive study for periods of days or even weeks,
however, provide the most effective control of variables such as diet, sleep,
exercise, and toxicant exposure.
The Clinical Laboratory Evaluation and Assessment of Noxious Substances
(CLEANS) facility in Chapel Hill, North Carolina provides an example of live-
in chambers designed specifically for studying the health effects of environ-
mental toxicants in humans. The CLEANS facility consists of two large expo-
sure chambers with a computerized physiological data acquisition system as
diagrammed in Figure 10-1. Up to four subjects can be studied simultaneously
in each chamber, permitting rapid collection of data for characterizing the
10-6
-------
DIGITAL CONTROL LINES
O
vj
It
LUNG MECHANICS
GAS EXCHANGE
NEUROBEHAVIOR
EXERCISE
OXYGEN CONSUMPTION (
RESPIRATORY QUOTIENT
ELECTROCARDIOGRAM (ECG)
ELECTROENCEPHALOGRAM (EEC) ^OOD PRESSURE (BP)
EVENT-RELATED POTENTIALS (ERP)CARDIAC OUTPUT (QC'
.INFORMATION PROCESSING
SIGNAL DETECTION (SD)
REACTION-TIME (RT)
SUBJECTIVE RATINGS
COMMUNICATIONS LINES
CARDIOVASCULAR
ELECTROCARDIOGRAM (ECG)
2 (R. BLOOD PRESSURE (BP)
SYSTOLIC TIME INTERVAL (STI)
ECHOCARDIOGRAPHY (ECHO)
CARDIAC OUTPUT (QC>
OPERATOR
Figure 10-1. Schematic representation of the testing facility and psychophysiological functions to be studied in humans during
prolonged exposure to low-level environmental insult. The computerized exposure, monitoring, and physiological data acqui-
sition system will permit precise control of experimental variables. Four subjects can be tested simultaneously in each live-in
chamber. (Drawing courtesy of R. W. Stacy, Ph.D.
-------
toxicologic effects of oxidants, acid aerosols, odors, noise, and microwave
radiation.
Event-Related Potentials in Human Toxicological Research
Event-related potentials (ERPs) of the brain may be recorded noninvasively
from the scalp of man or from electrodes implanted in specific neural struc-
tures of animals. Neuroelectric "events" include sensory evoked potentials,
association potentials, and motor-related potentials. Sensory evoked potentials
are elicited by external stimulation of auditory, visual, or somatosensory recep-
tors and reflect the arrival of afferent signals from peripheral receptors. Associ-
ation potentials reflect a later stage of information processing related to the
evaluation of incoming signals which have already been registered. Association
potentials presumably index the activity of corticocortical or thalamocortical
circuits involved in central decision-making processes. Motor-related potentials
reflect cortical processes associated with the preparation for and execution of
motor responses, processes which include signals emitted from the central
nervous system to peripheral effectors.
William Grey Walter (1964), a British neurophysiologist, first described a slow
negative shift in electrical activity of the brain related to sensorimotor condi-
tioning in man. This discovery spawned the rapidly growing study of event-
related potentials. ERP investigators encompass many disciplines including
psychology, psychiatry, neurology, physiology, and pharmacology. To the
psychologist, ERPs offer a long-sought quantitative handle on mental proces-
sing, an index of activity in the "black box'' intervening between stimulus and
response. To the behavioral toxicologist, ERPs offer a noninvasive index of
chemical insult at the various stages of information processing in the central
nervous system. The advantages of ERP techniques in human toxicological
research will be reviewed below.
Noninvasive Measure of CNS Function
Ethical considerations preclude the use in human subjects of techniques which
require surgical intervention or which pose a threat to safety or health. ERPs
provide a noninvasive measure of CNS function which can be used without risk
in human research. Some neurotoxins such as industrial solvents produce
peripheral rather than central neuropathy (Seppalainen, 1975; Zappoli et al.,
in press). Traditional electroneuromyographic measures, such as maximal
nerve conduction velocity, may be more sensitive than ERPs with substances
which produce peripheral neuropathy. On the other hand, peripheral deficits
should be reflected in decreased amplitude and/or increased latency of early
sensory evoked potential components in the impaired modality. Lead poison-
ing, for instance, produces marked peripheral neuropathy. Somatosensory
evoked potentials (SEP) elicited by stimulation of the damaged afferent path-
way should reflect this deficit centrally. Supportive evidence was recently
reported by Seppalainen (in press) who found that the SEP was more sensi-
tive than peripheral measures in diagnosing subclinical neuropathy in workers
chronically exposed to lead in a storage battery factory.
10-8
-------
Sensitivity to Drug Effects
Many substances consumed for pleasure or medicinal purposes may be toxic
if used in excess or may produce altered states of consciousness in lesser con-
centrations. Alcohol is a CNS depressant whose intoxicating and anesthetic
qualities have been used and abused for millennia. The tragic consequences of
drug abuse and addiction are a direct concern of the clinician. On the other
hand, the systematic administration of drugs whose behavioral and neuro-
physiological mechanisms are known can provide the experimental toxicologist
with a useful tool for comparative evaluation of unknown chemical substances.
The prospective sensitivity of event-related potentials to neurotoxins can be
similarly gauged by examining the effects on ERPs of commonly used psycho-
tropic drugs including stimulants, depressants, and tranquilizers.
Considerable evidence is available on the effects of psychotropic drugs on the
contingent negative variation (CNV), a slow negative ERP observed during a
signaled interval preparatory to motor response or cognitive decision. In
general, CNV amplitude is enhanced by stimulants—amphetamine (Kopell
et al., 1974) and caffeine (Ashton et a/., 1974)—and decreased by depres-
sants and tranquilizers—barbiturates (Kopell et a/., 1974), chlorpromazine
(Tecce et al., 1975), carbon monoxide (Groll-Knapp et al., 1972), ethanol
(Kopell et al., 1972) and nitrazepam (Ashton et al., 1974). In cases where the
behavioral effect of drugs has been paradoxical, CNV changes have consistently
mirrored behavioral observations. That is, when a given drug enhanced per-
formance in some subjects but depressed performance in others, CNV ampli-
tude varied in the same direction as performance. Paradoxical effects have been
reported with d-amphetamine (Tecce and Cole, 1974) and nicotine (Ashton
et al., 1973). This evidence suggests that the CNV should provide a useful
index of central-acting neurotoxic drug effects.
Convergent and Discriminant Validity
ERP techniques offer the advantage of simultaneous behavioral and electro-
physiological measurement of toxicant effect. That is, ERPs are usually re-
corded during performance of sensorimotor conditioning, signal detection, or
vigilance tasks which yield concomitant behavioral measures. Multiple measures
lend convergent validity to results when two or more measures vary consis-
tently or correlate highly. Groll-Knapp et al. (1972) demonstrated convergent
validation of a vigilance decrement during low-level carbon monoxide expo-
sure. Signal detection and the amplitude of a slow negative brain potential
both varied inversely with CO concentration.
On the other hand, it would be redundant to collect both types of information
if behavioral and electrophysiological measures always yielded convergent
results. The data would be more informative if the measures were orthogonal,
i.e., indexed independent dimensions. The discriminant validity of these mea-
sures is the degree to which ERP and behavioral parameters reflect orthogonal
dimensions. ERP and behavioral measures frequently do not covary. Reaction-
time, a traditional behavioral index of central processing time, can be easily
dissociated from the CNV observed during the reaction-time fore-period
10-9
-------
Rebert and Tecce, 1973). Similarly, the observed electrophysiological effects
of toxicant exposure might be orthogonal to behavior, i.e., ERP changes might
be observed in the absence of a behavioral deficit. Winneke and Kaska (1974),
for instance, observed a decrease in amplitude of the AEP without impairment
of vigilance performance in subjects exposed to trichloroethylene. The irony
of the Winneke and Kaska finding is that the functional significance of the
ERP deficit is difficult to evaluate in the absence of convergent behavioral
validation.
Localization of Neurotoxic Effects
Although the precise neuroanatomical substrate of most event-related poten-
tials is presently unknown, it is generally assumed that the temporal sequence
of early components in sensory evoked potentials reflects the serial transmis-
sion of signals along afferent pathways from distal (brainstem) to proximal
(cortical) structures. Evidence is rapidly accumulating that later components
reflect successive stages of information processing at the cortical level. Picton
et al. (1974), for instance, have identified fifteen distinct components in the
auditory evoked potential as shown in Figure 10-2. They attribute early com-
ponents (within 8 msec of stimulus onset) to activation of cochlea and brain-
stem auditory nuclei, middle components (8-50 msec) to activation of the
auditory thalamus and cortex, and late components (50-300 msec) to diffuse
activation of frontal cortex.
HUMAN AUDITORY EVOKED POTENTIALS
60 dB CLICK STIMULUS, VERTEX TO MASTOID RECORDING
VERTEX 0.5
POSITIVE
10 20 50 10 20 50 100 200
LATENCY, ms
Figure 10-2. Diagram of auditory evoked potential components plotted on logarithmic
scales (from Picton et al., 1974 courtesy of Elsevier Scientific Publishing Company).
10-10
-------
Table 10-1 summarizes contemporary evidence on the neuroanatomical origin
of AEP components. Note that components with latencies greater than 20
msec are presumed to be of cortical origin. The topographic recording and
mapping of middle and late components permit at least a crude localization of
the source of these components within the cortical mantle. As the neuroana-
tomic origins of ERP components are elaborated, ERP techniques will become
an increasingly important tool for localizing the site of neurotoxic action.
Table 10-1. PROBABLE NEUROANATOMICAL SOURCE OF AUDITORY
EVOKED POTENTIAL COMPONENTS3
Latency,
Component
I
II - IV
V, VI
N , P N
pa. Nb
msec
<2
2-5
6-7
8-20
21-49
Probable source
Auditory nerve fiber
Cochlea Si superior olivary
nuclei
Lateral lemniscusor
inferior colliculus
Medial genicuiate & non-
specific thalamlc nuclei
Anterior Si posterior
Reference
Jewett Si Willlston, 1971
Galamboset a/., 1959
Webster, 1971
Kitahata eta/., 1969
Picton eta/., 1974
Ibid.
association cortex
P2, N2
50-300
Widespread activation of
frontal cortex
Goff eta/., 1969
Kooi eta/., 1971
Hardin & Castellucci, 1970
aBased on T. W. Picton eta/., 1974.
Neurotoxic Effect on Information Processing
Neurotoxic effects on complex integrative (or cognitive) functions of the brain
have been inferred from behavioral measures in the past (e.g., Beard and
Grandstaff, 1975). Event-related potentials offer a more direct, remarkably
sensitive index of cognitive processing in the central nervous system. Theories
of signal detection (Green and Swets, 1966) and information processing
(Broadbent 1958, 1971; Treisman, 1969) have provided the conceptual frame-
work for investigating the relationship of ERP components and cognitive
processes in man. Tueting (in press) has critically reviewed this rapidly-expand-
ing field of research. Hillyard et al. (1973), Picton et a/. (1974), Picton and
Hillyard (1974), and Schwent and Hillyard (1975) have intensively studied the
functional significance of auditory ERPs. Evidence from these studies suggests
that components with latencies in the 60-120 msec range reflect attentional
processes related to stimulus selection or "stimulus set," while later compo-
nents (160-500 msec) reflect decisional processes of response selection or
"response set."
10-11
-------
The amplitude of a large positive polysensory* wave, peaking 200-400 msec
after stimulus onset (designated P300), varies inversely with the probability
of stimulus occurrence (Tueting et a/., 1971). Probability of occurrence is
also a theoretical definition of information content. The P300 has been asso-
ciated in other studies with decision-making (Davis, 1964), information
delivery (Button et a/., 1967), and response inhibition (Papakostopoulos and
Crow, 1976). In brief, considerable evidence has accumulated that ERPs
reflect multiple parameters of information processing in man. ERPs should,
therefore, provide an excellent, noninvasive electrophysiological index of
neurotoxic insult on cognitive function.
The Neurobehavioral Effects of Low Frequency Noise in Humans:
An Application of ERP Methods
The following study was designed to assess the effects of low frequency ran-
dom noise (11.5-325 Hz) at moderate intensity (80 db) on ERPs and behavior.
The results have been reported in detail elsewhere (Benignus et a/., 1975 and
Otto et a/., in press). Selective aspects will be elaborated here to illustrate
basic principles of ERP methodology.
Rationale
Although discussion in this chapter has been oriented toward chemical sub-
stances, the behavioral toxicologist is also concerned with physical insults to
the brain including noise and electromagnetic radiations. Noise is a staple in
urban man's diet of carbon monoxide, oxidants, acid aerosols, and heavy
metal particulates. Noise is an omnipresent stressor capable of severely dis-
rupting behavior or physically damaging the auditory system at high intensities.
Many researchers routinely use white noise to mask extraneous laboratory
sounds during testing. What is the effect of masking noise on human perfor-
mance? That question underlies the following experiment in which the level
of noise was not perceived as particularly irritating or distracting. The syner-
gistic effects of noise and noxious chemical substances which frequently occur
together in experimental as well as real-life situations are also or empirical
interest.
Vigilance tests have been widely used in studies of stress (cf. Broadbent, 1971).
The vigilance decrement-a measure of target signals missed over time-appears
to be extremely sensitive to both physical and chemical stressors. A continuous
performance vigilance test designed by Rosvold et al. (1956) to minimize
fluctuations in alertness was used as the behavioral context. Mirsky and
Rosvold (1960) found this test to be sensitive to psychoactive drug effects.
McCallum (1975) demonstrated its utility in recording event-related potentials.
"That is, the observed effects are independent of sensory modality.
10-12
-------
Methods
Subjects
Twenty-seven male university students, aged 18-31, were paid to participate.
All subjects were carefully screened to determine that hearing (Bekesy audio-
meter), personality (Minnesota Multiphasic Personality Inventory), and health
(physical examination) were within normal limits.
Continuous Performance Test (CPT)
Single-digit numerals were displayed at one-second intervals in pseudorandom
series. Subjects were instructed to press a button whenever a sequence of three
consecutive even digits or three consecutive odd digits (a three-string) was
observed. Digits were presented in nine 12.5-minute series, each containing
about 111 target three-strings. Three-minute rest breaks were interspersed
between work periods.
The CPT imposes greater information processing demands than classical vigi-
lance or signal detection tasks. Stimuli in a standard vigilance task require a
simple binary decision—target or nontarget. The numeric monitoring task
described here includes a short-term memory component. In order to identify
a target correctly, subjects must recall the previous two stimuli. Target identi-
fication also requires more than simple physical matching of stimuli. Each
numeral must be coded as "even" or "odd" and this "polarity" dimension
must 'be held in short-term memory. No single stimulus or simple template
match provides the solution.
More complex questions can be addressed with a three-string sequential depen-
dency than with a binary-choice problem. For instance, the proportion of
target and neutral stimuli cannot be varied independently in a binary vigilance
problem. One factor is always the converse of the other. In the ternary prob-
lem, however, the value of one factor does not uniquely define the values of
the other two factors. Benignus et al. (1976) have exploited this feature to
reassess the effect of signal rate on vigilance performance.
If the reader is momentarily experiencing a decline in vigilance, the discussion
bears out Occam's razor: parsimony is the essence of good experimental
design! The high level information-processing capabilities of Homo sapiens may
tempt the behavioral toxicologist to needless levels of complexity. The theo-
retical challenge of the CPT notwithstanding, a simple binary-choice vigi-
lance task would be better suited for interspecies comparison of toxicological
effects!
Experimental Design
Subjects completed orientation, stabilization, and testing sessions on three
consecutive days. The first day consisted of an abbreviated training session
to familiarize subjects with the task and laboratory setting. On the second
10-13
-------
day electrophysiological recording procedures were introduced and the subjects
were trained to fixate on a visual target to minimize eye movements.*
Subjects also completed nine numeric-monitoring series during which perfor-
mance reached asymptote. The same procedure was repeated on the third day.
Analyses are based on stabilized performance recorded on the fin-al day.
The output of a random-noise generator was bandpass-filtered to obtain a
91-325 Hz noise spectrum and was then tape-recorded at 15 inches per second
(IPS). A second noise spectrum (11.4-41 Hz) was obtained by playing the tape
back at 1 7/8 IPS. These bandwiths will be designated hereafter as upper (U)
and lower (L). The noise was presented free-field through a loudspeaker placed
about 1.9 m in front of subjects. **
Noise and no-noise control (C) conditions were alternated during successive
work periods on days 2 and 3 as follows: CUCLCLCUC or CLCUCUCLC.
Noise exposure schedules were counterbalanced across subjects with each one
exposed to a different schedule on succeeding days. Subjects were instructed to
ignore the noise.
An alternating repeated-measures design is useful only when the effects of a
stressor are presumed to be rapidly reversible. It \snot practical with chemical
compounds which accumulate in body tissue during exposure. In any repeated-
measures study, sufficient time for a complete return to physiological baseline
must be allowed between successive exposures. This may be a matter of hours,
days, or months, depending on the biological half-life of the test compound.
Liposoluble substances such as DDT and dieldrin, for instance, accumulate in
the fatty tissues of mammals (cf. Walker, 1974) where residues may persist for
years. Independent groups are more practical (or mandatory) with persistent
substances such as heavy metals or organochlorine insecticides.
Electrophysiological Procedures
The recording of electrophysiological parameters is a relatively simple proce-
dure but requires sophisticated electronic equipment. Figure 10-3 is a sche-
matic diagram of the basic components required to record psychophysiological
parameters such as the electroencephalogram (EEC), electro-oculogram (EOG),
electromyogram (EMG), and electrodermal response (EDR). Excellent des-
criptions of recording techniques are available elsewhere (e.g., Venables and
Martin, 1967; Greenfield and Sternbach, 1972; Thompson and Patterson,
1974). Only a brief overview is provided here.
*Eye movement artifact can severely contaminate ERP recordings if subjects do not
fixate on a visual target. Normal adults can be easily trained to fixate. Eye movement
control poses a serious problem in young children and abnormal adult populations
(i.e., psychotics or sociopaths) where subjects are unable or refuse to fixate. Hillyard
(1974) describes techniques that can be used to compensate for eye movement atri-
fact under these circumstances.
**See Benignusef a/. (1975) for technical details of noise generation.
10-14
-------
o
Ul
AMPS AND
PREAMPLIFIERS
BAND-PASS FILTERS
ANALOG CHART
RECORDER
ANALOG TAPE
RECORDER
DATA ACQUISITION
AND STORAGE |
MINI COMPUTER
ANALOG-DIGITAL
CONVERSION,
SIGNAL AVERAGING
I AND PLOTTING |
MASS STORAGE
COMPUTER
SPECTRUM AND
STATISTICAL ANALYSIS
[SIGNAL AVERAGING)
/CUMULATIVE^
( RESPONSE
VRECORDER
RESPONSE^
RECORDER/
REACTION TIME
^RECORDER
> ?
*S r*r
PROGRAMMABLE
LOGIC
:ATA ACQUISITION,
EXPERIMENTAL f
CONTROL, STIMULUS--
JRESPONSE CODING|_
U
Figure 10-3. Diagram of basic equipment needed to record event-related potentials of the brain and other psychophysiological
parameters of human performance.
-------
Nonpolarizing electrodes are attached to the scalp and appropriate skin areas
with commercially-available pastes or adhesive collars for short-term exposures.
Collodion provides a more stable bond for longer exposures. The International
10-20 Electrode System (Jasper, 1958) is generally used as a reference for EEC
electrode placement to standardize procedures among laboratories. It is manda-
tory in ERP research to record vertical EOG in order to identify records con-
taining eye movement artifact. If movement-related potentials are to be mea-
sured it is also important to record the EMG of responding muscle groups for
comp'arision of central waveforms and the temporal pattern of distal motor
unit discharge.
In the present study, nonpolarizable silver-silver chloride Beckman biopotential
pellet electrodes were placed at the vertex (Cz) and over the right (P3> and left
(P4) mid-parietal association areas. Electrodes were also placed above the inner
canthus and below the outer canthus of the left eye to monitor and record
eye movements. Scalp electrodes were referred to a linked pair of electrodes
clipped to the earlobes. The earlobes are considered to be electrically neutral
so that electrical potentials recorded differentially between scalp electrodes
and linked ears can be assumed to reflect activity at the scalp recording site.
Parietal recordings were made to investigate the effects of low-frequency noise
on the coherence of spontaneous EEG rhythms in the left and right hemi-
spheres. This report will be limited to ERP components observed at the vertex.
The amplitude range of spontaneous EEG rhythms in normal adults is 2-
200 /-iV (see Table 10-2). It is necessary, therefore, to amplify the spontaneous
EEG by as much as 1Q9 to reach the minimum 1 V peak-to-peak deflection
required by IRIG, *Telemetry Standards (1969). Amplified signals are then
recorded on FM tape or channelled directly to a computer for analysis. If
laboratory facilities permit, electrophysiological data should be processed on-
line for immediate availability of results. Frequently, on-line computer access
is not practical, and the data is recorded on analog tape for off-line processing.
Signal Averaging
How are EEG signals processed? In signal averaging the first step is to convert
the continuous EEG record to a series of discrete voltage samples—a process
called analog-to-digital conversion. The digitizing rate determines the minimum
frequency of ERP components which can be resolved. That is, the digitizing
rate should be at least twice the frequency of the fastest component to be
measured. Since most ERP components are slower than 50 Hz, a digitizing rate
of 100 Hz is adequate in most cases. Analysis of brainstem evoked potentials
which occur within the first 10 msec following stimulus onset, however, re-
quires a digitizing rate of at least 2000 Hz.
The second step is signal averaging in which digitized arrays of data from suc-
cessive trials time-locked to a repetitive event are summed and divided by
the number of trials. The triggering event is usually an external stimulus but
"Inter-Range Instrumentation Group.
10-16
-------
may also be a response mechanogram, EMG burst, or any discriminable physio-
logical signal.
Table 10-2. CHARACTERISTICS OF SPONTANEOUS EEC RHYTHMS IN NORMAL
HUMAN ADULTS"
Type
Delta
Theta
Alpha
Beta
Frequency
(c/sec)
0.5-4
5-7
8-12
18-30
Amplitude
(JUV)
20-200
5-100
5-100
2-20
Level of consciousness
: Slow wave sleep
Slow wave sleep
Awake, relaxed
Awake, concentrating
or rapid eye
movement (REM)
sleep
aAdapted from Lindsley and Wicke (1974, p 26).
Why is signal averaging necessary? Scalp-recorded ERP components are fre-
quently submerged in a turbulent sea of spontaneous EEC rhythms. For
instance, brainstem evoked potentials are 1 /iV or less, sensory evoked poten-
tials are 5-15 /xV, and association potentials are 5-25 [iV in amplitude. Spon-
taneous EEG rhythms range from 2-200 /iV in normal adults as shown in
Table 10-2. Signal averaging is an effective method of increasing the ERP
signal-to-EEG noise ratio.
Signal averaging assumes that the temporal characteristics of ERPs are invari-
ant with respect to the triggering event. That is, the signal is assumed to be
stationary in time. In statistical terms, the mean and variance of the signal
should be constant within the averaging epoch and across successive trials. The
validity of this assumption is dependent on the length of time required to col-
lect trials for the average. Unfortunately, Homo sapiens has little respect for
statistical theory and tends to be notoriously variable. ERP changes due to
alterations in arousal, attention, fatigue, habituation, learning, or other dy-
namic neural processes are obscured by signal averaging. This problem is
minimized by using as few trials as possible in the construction of averages.
How many trials should be included in a given signal average? As a rule of
thumb, the amplitude ratio of spontaneous EEG to ERP is proportional to
the square root of the number (N) of trials (EEG/ERP = \/N). Table 10-3 pro-
vides a simple guide to determine N. ERP and EEG values represent peak-to-
peak measurements.
Rigorous treatment of the theory, methods, and problems of signal averaging
are beyond the scope of this review. Procedures and problems specific to ERP
research are discussed in depth in Donchin and Lindsley (1969), McCallum
and Knott (1973, 1976) and Thompson and Patterson (1974). Vaughan (1974)
provides a particularly lucid account.
10-17
-------
Table 10-3. MINIMUM NUMBER OF TRIALS REQUIRED FOR SIGNAL AVERAGING
AS A FUNCTION OF ESTIMATED ERP AMPLITUDE AND AVERAGE AMPLITUDE
OF SPONTANEOUS EEG
Estimated
ERP
amplitude
(MV>
0.5
1.0
2.0
5.0
10.0
25.0
Number of trials/average
EEG Amplitude (jUV)
25 50 100
2,500 10,000 40,000 "^
625 2,500 10,000
156 625 2,500
25 100 400 "1
6 25 100 >
1 4 16 J
Type of ERP
Brain stem
evoked potentials
V Sensory evoked potentials
Association potentials
Results
Changes in numeric-monitoring efficiency and ERP amplitude were observed
during exposure to low-frequency noise. Figure 10-4 illustrates the behavioral
effects.* In brief, a vigilance decrement across time was observed in all condi-
tions. Subjects missed more target strings during both upper (p < 0.005) afid
lower band (p < 0.039) series than during control series. The difference in
performance between upper and lower noise series was not significant.
Complex relationships between noise conditions, performance, and ERP com-
ponents were also observed. ERP averages were constructed separately for hit,
miss, and correct rejection trials. Due to excessive EOG artifact in some sub-
jects, only 21 subjects were used in ERP analyses. Large intersubject variance
m performance further reduced the number of usable subjects in hit and miss
analysis. To illustrate the range of performance, subjects were rank-ordered
according to the total number of misses and then divided into three equal
groups. The average number of misses per series for subjects in high medium
and low performance groups is shown in Table 10-4. Subjects in the high per-
formance group missed so few trials that meaningful averages of "miss" data
could not be constructed. The converse problem was encountered to a lesser
extent with 'hit" data in the low performance group.
Hit Analysis
SuffirTm h?^ ^ three:String sec'uences which subjects correctly detected.
no e^L If r6 ^I3'"6" t0 C°nStmCt ™ra^ for ™* control and
is Town n F ^n f ^ ^^ ^^ ^°m r6C°rded at the vertex
is shown ,n Figure 10-5 to illustrate ERP measurement procedures
detailS °f ^ ««I«i«' Analysis of behavioral
10-18
-------
20 -
16 -
12-
O
tc
uj 8
to
D
Z
4-
O CONTROL
O U NOISE
A L NOISE
345
RUN NUMBER
Figure 10-4. Number of misses across time for control, upper and lower noise bands.
(Reprinted with permission of publisher from : Benignus, V. A., Otto, D. A., and
Knelson, J. H. (1975). Effect of low-frequency random noises on performance of a
numeric monitoring task. Perceptual and Motor Ski/Is 40, 231-239.
Table 10-4. MISS RATES DURING CONTROL VS. NOISE CONDITIONS FOR HIGH,
MEDIU M & LOW PERFORMANCE GROUP
Performance
level
High (N = 7)
Medium (N = 7)
Low (N = 7)
t 1.96 x SEb
Control
2.2
7.0
26.9
±.2;69
Noise3
3.0
8.4
38.8
t4.68
Average misses/series
a Upper and lower bands combined.
b95% ponfidence Interval = mean i 1.96 x standard error of mean.
10-19
-------
BASELINE
Figure 10-5. Measurement procedures for components of a typical event-related
potential waveform. S2 and S3 designated the intervals during which the second
and third digits of 3-string sequences were displayed. Contingent negative varia-
tion (CNV) was measured as the average voltage during the specified interval
relative to the pre-S2 baseline. Baseline-to-peak measurements were made of
negative (N110) and positive (P456) components following S3. Timebase: one
sec interstimulus interval (S2 onset - S3 onset).
Three ERP components designated CNV, N110, and P456 were measured. The
CNV was computed as the mean amplitude of a 256 msec epoch preceding the
third digit (S3) relative to a 256 msec baseline epoch preceding the second digit
(S2). N110 and P456 were measured from the peaks in each individual average
to the same baseline. The convention used to label N110 and P456 components
specifies the polarity—negative (N) and positive (P)—and the average peak
latency from stimulus (S3) onset expressed in msec.
Table 10-5 shows ERP amplitudes during hit trials in control, upper, and
lower band noise series. Small decrements in the amplitude of each measure
were observed in noise compared to control conditions. In order to reduce the
dimensionality of the data (and thereby increase the power of statistical tests),
control means were subtracted from upper and lower band means for each
subject. The effect of noise on ERPs was tested against the null hypothesis
that none of the differential ERP scores differed from zero. A single-factor
multivariate test with Wilks-Lambda criterion converted to an F ratio by Rao's
procedure (Cooley and Lohnes, 1971) was used to reject the overall null
hypothesis (Fgji = 6.40, p (0.004).
10-20
-------
Table 10-5. HIT TRIALS: ERP AMPLITUDE (/UV) AS A FUNCTION OF
Control
Upper
Band
Lower
Band
i 1.96 x SEa
CNV
-3.9
-3.8
-3.1
0.43
N110
-5.7
-4.7
-4.4
0.77
P456
13.1
12.0
12.7
0.98
a95% confidence interval = mean ±_ 1.96 x standard error of mean.
Table 10-6 provides the results of post hoc univariate tests for each compo-
nent. N110 amplitude decreased significantly during upper and lower band
noise exposure; CNV decreased during lower band noise only; and P456 was
not significantly affected by noise. In order to preclude the possiblity that
the N110 measure was confounded by the slow negative shift preceding S3,
CNV values were subtracted from N110 values of each subject prior to statis-
tical testing.
Table 10-6. SUMMARY OF UNIVARIATE F TESTS TO EVALUATE EFFECTS
OF NOISE ON ERP COMPONENTS IN HIT AVERAGES
Null hypothesis
(Lower-Band CNV) = 0
(Upper-Band CNV) = 0
(Lower-Band N1 10) = 0
(Upper-Band N110) = 0
(Lower-Band P456) = 0
(Upper-Band P456) = 0
Mean Sq
11.35
0.41
2.97
10.20
3,17
13.34
F(1,16)
8.53
0.31
4.35
10.93
0.67
1.54
P
(0.010
(0. 585
<0.053
<0.004
<0.426
<0.233
Correct Rejection (CR) Analysis
CR trails were defined as two-strings to which subjects did not respond. The
''no-response" contingency, it should be noted, provided no indication of
whether or not the subjects processed the digits. That is, the absence of re-
sponse may have reflected either correct identification of a two-string sequence
or a lapse in attention. Twenty-one subjects were used in this analysis. CNV,
IM110, and P456 components were measured in CR averages in the same man-
ner as in hit averages. None of these measures, however, significantly discrimi-
nated between noise and control conditions. The effects on ERPs of perfor-
mance was also examined. Table 10-7 shows the amplitude of CNV, N110 and
P456 averaged across control and noise conditions for high, medium, and low
10-21
-------
performance groups. Although decrements in all ERP components were ob-
served in the low performance group compared to medium and high perfor-
mance subjects, only P456 differences approach significance.
Table 10-7. CORRECT REJECTION TRIALS: ERP AMPLITUDE (/JV) AS
A FUNCTION OF MONITORING PERFORMANCE
Performance Level
High (N = 7)
Medium (N = 7)
Low (N = 7)
+.1.96 x SE
CNV
-4.2
-4.1
-32
±_0.63
N1 10
-6.7
-5.8
-5,2
±0.74
P4S6a
12.0
10.5
6.7
±1.22
aF(2,1S) = 3.09S, p (0.07.
Miss Analysis
In view of the ambiguous nature of two-string events, a direct comparison of
hit and miss trials was undertaken to further assess the effects of performance
on ERPs. Miss trials were precisely defined as three-string sequences which
subjects failed to detect. In previous analyses, averages for each successive
control and noise series were constructed without saving single trials. This
procedure could not be followed in this analysis since miss trials had to be
pooled across multiple series in order to accumulate sufficient trials for aver-
aging. Even pooling across series did not yield meaningful miss averages in
many high performance subjects. Therefore, the miss analysis was restricted to
the low performance group. Hit and miss trials of these subjects were redigi-
tized and saved individually. Successive averages of 24 trials were then con-
structed for hits and misses, irrespective of control and noise conditions.
Hit and miss trials averaged across all conditions and subjects in the low per-
formance group are superimposed in Figure 10-6. Note the diminution of CNV,
N110, and the absence of P456 in the average of miss trials. Mean ampli-
tudes of ERP components in hit and miss averages are shown in Table 10-8. A
multivariate test was used to evaluate the null hypothesis of no difference be-
tween hit and miss ERPs (F-|2,25 = 13.665, p < 0.001).* Univariate F-tests
were significant for all hit-miss ERP comparisons as indicated in Table 10-8.
*Data obtained from left and right parietal electrodes, as well as the vertex, were in-
cluded in this analysis. Consistent results were observed at all recording sites.
10-22
-------
HIT
MISS
IM110
P456
Figure 10-6. Summary waveforms of hit and miss trials averaged across all noise and
control series of 7 low performance subjects. Note the attenuation of CIMV, N110 and
absence of P456 during miss-trials. Abbreviations as in Figure 10-5.
Table 10-8. LOW PERFORMANCE SUBJECTS (N = 7): ERP AMPLITUDE (|UV)
AS A FUNCTION OF HITS VS. MISSES
Hits
M isses
F(1,36)
P<
CNV
-4.3
0.5
7.197
0.011
N1 10
-10.5
- 0.9
26.956
0.001
P456
12.2
1.8
15.243
0.001
Discussion
Results of this study indicate that low-frequency continuous noise at moderate
intensity affects vigilance performance and at least one ERP component. What
is the functional significance of these findings? Interpretation of the behavioral
results is unequivocal: vigilance performance is slightly impaired. Extrapola-
tion from a monotonous numeric-monitoring task in a small isolated chamber
to real-life requires only a minimal leap in imagination. An individual driving
cross-country on a superhighway may experience analogous conditions of
monotony and social isolation. Engine hum, tires, and air-conditioning may
10-23
-------
generate low-frequency noise of comparable intensity. High-speed driving cer-
tainly demands continuous vigilance to visual signals. The consequence of even
a slight impairment in driving vigilance needs no elaboration.
The effect of noise on the N110 component was observed in an analysis of
"hit" trials. The interpretation of the N110-noise relationship is weakened by
the lack of effect in a subsequent analysis of "correct rejection" trials. The
ambiguous behavioral significance of CR events, however, renders the inter-
pretation of associated ERPs difficult.
The final comparison of "hit" versus "miss" averages in the low performance
group illustrates the sensitivity of ERPs to behavioral contingencies. CNV and
N110 components were sharply reduced, and the P456 component was com-
pletely absent in miss trials. Are these differences merely a reflection of the
presence or absence of an overt behavioral response (i.e., motor artifacts)?
CR averages rule out this possibility. CNV, N110, and P456 components were
clearly present in CR averages.
What is the significance of the change in 1X1110 amplitude observed during
exposure to low-frequency noise? The neurophysiological mechanism under-
lying the N110 is unknown. Cautious inferences may be drawn from the
associated behavioral decrement. Convergent decrements in observed behav-
ioral and electrophysiological measures could reflect an impairment in either
selective or general attention.
Available evidence does not support a general attentional or arousal decrement
hypothesis. First, a decrease in P456 amplitude would be expected if the effect
were nonspecific. Second, there is little reason to suspect that N110 amplitude
would decrease as a function of arousal level. Picton et al. (1974) did not ob-
serve any change in auditory ERP components earlier than 250 msec during
the transition from wakefulness to sleep. In fact, the predominant change in
auditory ERPs as subjects fell asleep was a dramatic increase in the amplitude
of a negative component peaking at 290 msec. Results of a spectrum analysis
of the present data (Benignus ef al., in preparation) also failed to support the
arousal hypothesis. If low-frequency noise induced drowsiness, a decrease in
beta and alpha frequencies accompanied by an increase in theta and delta
frequencies would be expected. No difference in the frequency distribution
of spontaneous EEG during control versus noise series was observed.
What evidence can be marshalled in support of a selective attentional hypo-
thesis? Schechter and Buchsbaum (1973) observed consistent changes in P110-
N140 (peak-to-peak) amplitudes of auditory and visual evoked potentials in
a series of tasks with graded attentional demands. The P100-N140 component
was smallest in a "distraction" condition. Schwent and Hillyard (1975) provide
convincing evidence of intramodal selective attentional effects in an auditory
N1 component analagous to N110. They observed a marked enhancement of
N1 in "attend" compared to "ignore" conditions.
These studies indicate that a polysensory negative ERP component with a
latency of about 100 msec may reflect selective attentional processes. The
convergent behavioral and ERP results of the present study suggest, therefore,
10-24
-------
that low-frequency noise selectively interferes with the ability of subjects to
"tune in" or concentrate on incoming sensory signals. That is, low-frequency
noise at moderate intensity may impair selective attention but does not seem
to affect general arousal in a vigilance-like task.
Concluding Comment
Event-related potentials provide an intriguing, noninvasive window on sensori-
motor and cognitive functions in man. There are optimistic signs that research
during the next decade will resolve many of the current mysteries concerning
the neuroanatomical mechanisms and functional significance of ERPs. Consid-
erable progress has been made in the study of the brain stem evoked potentials,
the P300, and the CNV through the close alliance of behavioral and electro-
physiological testing procedures.
Do ERP techniques currently offer a useful tool to the behavioral toxicologist?
This article has presented arguments that ERPs provide a noninvasive index of
neurotoxic effect on CNS function and information processing. The conviction
of this author is not shared universally, however, and a dash of skepticism is
appropriate to complete the story.
S. S. Stevens, the late Harvard psychophysicist eloquently summarized the
negative view (1971, p. 502):
An electrode on the skull may or may not be able to reflect
the operation of the sensory transducer in a meaningful way.
The full answer to that question remains to be discovered... .
Granting all the difficulties, I would like to believe that
the knowledge that can be gleaned about sensory systems by
studying evoked potentials is greater than zero.
H. E. Stokinger, Chairman of the Threshold Limits Committee at the National
Institute of Occupational Safety and Health, bluntly challenges the utility of
ERP data in environmental toxicology: "Changes in the visually evoked re-
sponse (VER) are presently not sufficiently well understood in over-all physio-
logical terms to serve as basic criteria for air standards" (1974, p. 20).
Stokinger's position is questionable on theoretical grounds because it assumes
that evidence of physiological effects requires knowledge of causal mech-
anisms. In medicine, drugs known to relieve specific symptoms are frequently
administered without precise knowledge of physiological mechanisms. Am-
phetamines, for instance, may be used to reduce hyperactivity, although the
mechanism underlying this paradoxical effect is not understood. Behavioral
observations are not admissible by Stokinger's logic either, since the physio-
logical mechanisms underlying behavioral decrements are seldom known. The
functional significance of behavioral observations, however, is generally self-
evident, whereas the functional significance of ERP changes per se may be
ambiguous. In cases where neither the physiological nor the behavioral signifi-
cance of an ERP change can be established, Stokinger's position is difficult to
refute.
10-25
-------
ERP investigators are vigorously amassing evidence to dispel this skepticism.
ERP techniques have been applied with modest success to a variety of clinical
problems in audiometry (Davis, 1976) neurology (Starr and Achor, 1975; Low
et a/., 1976), psychiatry (Shagass 1972; Timsett et a/., 1973), and develop-
mental disorders (Karrer, 1976). The evidence reviewed in this chapter suggests
that ERP techniques can provide a useful research tool in behavioral toxicology
as well.
References
Ashton, H., Millman, J. E., Telford, R., and Thompson, J. W. (1973). Stimu-
lant and depressant effects of cigarette smoking on brain activity in man.
Brit. J. Pharmacol. 48, 715-717.
Ashton, H., Millman, J. E., Telford, R., and Thompson, J. W. (1974). The
effect of caffeine, nitrazepam, and cigarette smoking on the CNV in man.
Electroenceph. din. Neurophysiol. 37, 59-71.
Beard, R. R. and Grandstaff, N. (1975). Carbon monoxide and human func-
tions. In Behavioral Toxicology (B. Weiss and V. Laties, eds.), pp. 1-26.
Plenum Press, New York.
Beard, R. R. and Wertheim, G. A. (1967). Behavioral impairment associated
with small doses of carbon monoxide. Amer. J. Public Health 57, 2012-
2022.
Benignus, V. A., Otto, D. A., and Knelson, J. H. (1975). Effects of low-fre-
quency random noises on performance of a numeric monitoring task.
Percept Mot. Skills 40, 231-239.
Benignus, V. A., Benignus, G., Otto, D. A., and Prah, J. D. (in preparation).
EEG spectra and vigilance performance in the presence of low frequency
noise.
Benignus, V. A., Otto, D. A., Prah, J. D., and Ryan, L. J. (1976). Monitoring
performance as a function of the rate of ready signals Percept. Mot Skills
43,815-821.
Broadbent, D. E. (1958). Perception and Communication. Pergamon Press,
London.
Broadbent, D. E. (1971). Decision and Stress. Academic Press, New York.
Cooley, W. W. and Lohnes, P. R. (1971). Multivariate Data Analysis. Wiley,
New York.
Davis, H. (1964). Enhancement of evoked cortical potentials in humans related
to a task requiring a decision. Science 145, 182-183.
Davis, H. (1976). Principles of electric response audiometry. Ann. Otol.
Rhino/. Laryngol. Suppl. 28, 1-97.
10-26
-------
Donchin, E. And Lindsley, D. B. (1969). Average Evoked Potentials: Methods,
Results, and Evaluations. NASA SP-191.
Dyer, R. S. and Annau, Z, (in press). Carbon monoxide and evoked potentials.
In Multi'disciplinary Perspectives in Event-related Brain Potential Research
(D. A. Otto, ed.), U.S. Government Printing Office, Washington, D.C.
Galambos, R., Schwartzkopff, J., and Rupert, A. (1959). Microelectrode study
of superior olivary nuclei. Amer. J. Physiol. 197, 527-536.
Goff, W. R., Matsumiya, Y., Allison, T., and Goff, G. D. (1969). Cross-modal-
ity comparisons of average evoked potentials. In Average Evoked Potentials:
Methods, Results, and Evaluations. (E. Donchin and D. B. Lindsley, eds.),
95, 141. NASA, Washington, D.C.
Green, D. M. and Swets, I. A. (1966). Signal Detection Theory and Psycho-
physics. Wiley, New York.
Greenfield, N. S. and Sternbach, R. A. (eds.) (1972). Handbook of Psycho-
physiology. Holt, Rhinehart, and Winston, New York.
Groll-Knapp, E., Haider, M., Holler, H., Jenkner, G., and Stidl, H. (in press).
Neuro- and psychophysiological effects of moderate CO-doses. In Multi-
disciplinary Perspectives in Event-related Brain Potential Research (D. A.
Otto, ed.), U.S. Government Printing Office, Washington, D.C.
Groll-Knapp, E., Wagner, H., Hauck, H., and Haider, M. (1972). Effects of low
carbon monoxide concentrations on vigilance and computer-analyzed brain
potentials. Staub-Reinhalt. Luft 32, 64-68.
Hardin, W. B. and Castellucci, V. F. (1970). Analysis of somatosensory, audi-
tory and visual averaged transcortical and scalp responses in the monkey.
Electroenceph. din. Neurophysiol. 28, 488-498.
Hillyard, S. A. (1974). Methodological issues in CNV research. In Bioelectric
Recording Techniques (R. F. Thompson and M. M. Patterson, eds.), Part B,
281-304.
Hillyard, S. A., Hink, R. F., Schwent, V. L, and Picton, T. W. (1973). Elec-
trical signs of selective attention in the human brain. Science 182, 177-
180.
Inter-Range Instrumentation Group (1969). Document 106-69, Telemetry
Standards, White Sands Missile Range, New Mexico.
Jasper, H. H. (1958). The ten twenty electrode system of the International
Federation. Electroenceph. din. Neurophysiol. 10,371-375.
Jewett, D. L. and Williston, J. S. (1971). Auditory-evoked far fields averaged
from the scalp of humans. Brain 94, 681-696.
10-27
-------
Karrer, R. (ed.) (1976). Developmental Psychophysiology of Mental Retar-
dation. Thomas, Springfield.
Kitahata, L. M., Amakata, Y., and Galambos, R. (1969). Effects of halothane
upon auditory recovery function in cats. J. Pharmacol. Exp. Therap. 167,
14-25.
Kooi, K. A., Tipton, A. C., and Marshall, R. E. (1971). Polarities and field con-
figurations of the vertex components of the human auditory evoked re-
sponse: A reinterpretation. Electroenceph. din. Neurophysiol. 31, 166-169.
Kopell, B. S., Tinklenberg, J. R., and Hollister, L. E. (1972). Contingent nega-
tive variation amplitudes: Marihuana and alcohol. Arch. Gen. Psychiatry
27,809-811.
Kopell, B. S., Wittner, W. K., Lunde, D. T., Wolcott, L J., and Tinklenberg,
J. R. (1974). The effects of metamphetamine and secobarbital on CNV
amplitude. Psychopharmacologia (Berl.) 34, 55-62.
Lindsley, D. B. and Wicke, J. D. (1974). The electroencephalogram: Autono-
mous electrical activity in man and animals. In Bioelectric Recording Tech-
niques (R. F. Thompson and M. M. Patterson, eds.). Part B, 3-83. Academic
Press, New York.
Low, M. D., Wada, J. A., and Fox, M. (1976). Electroencephalographic locali-
zation of conative aspects of language production in the human brain. In
The Responsive Brain (W. C. McCallum and J. R. Knott, eds.), pp. 165-171.
Wright, Bristol.
McCallum, W. C. (1976). Brain slow potential changes and motor response in
a vigilance situation. In The Responsive Brain (W. C. McCallum and J. R.
Knott, eds.), pp. 46-50. Wright, Bristol.
McCallum, W. C. and Knott, J. R. (eds.) (1973). Event-related Slow Poten-
tials of the Brain: Their Relations to Behavior. Elsevier, New York.
McCallum, W. C. and Knott, J. R. (eds.) (1976). The Responsive Brain. Wright,
Bristol.
McFarland, R. A., Roughton, F. J., Halperin, M. H., and Iven, J. J. (1944).
The effects of carbon monoxide and altitude on visual thresholds J Aviat.
Med. 15,381-394.
Mirsky, A. F and Rosvold, H. E. (1960). The use of psychoactive drugs as a
neuropsychological tool in studies of attention in man. In Drugs and
Behavior (L. Uhr and J. G. Miller, eds.). Wiley and Sons, New York.
Otto, D. A., Benignus, V. A., Prah, J. D., and Ryan, L. J., (in press). Low fre-
quency noise, vigilance performance and the vertex evoked potential. In
Multidisciplinary Perspectives in Event-related Brain Potential Research
(D. A. Otto, ed.). U.S. Government Printing Office, Washington, D.C.
10-28
-------
Papakostopoulos, D. And Crow, H. J. (1976). Electrocorticographic studies of
the contingent Negative Variation and 'P300' in man. In The Responsive
Brain (W. C. McCallum and J. R. Knott (eds.), pp. 201-204. Wright, Bristol.
Penfield, W. and Jasper, H. (1954). Epilepsy and the Functional Anatomy of
the Human Brain. Little, Brown, Boston.
Picton, T. W., Hillyard, S. A., Krausz, H. I., and Galambos, R. (1974). Human
auditory evoked potentials. I: Evaluation of components. Electroenceph.
din. Neurophysiol. 36, 179-190.
Picton, T. W. and Hillyard, S. A. (1974). Human auditory evoked potentials.
II. Effects of attention. Electroenceph. din. Neurophysiol. 36, 191-199.
Rebert, C. S. and Tecce, J. J. (1973). A summary of CNV and reaction time.
Electroenceph. din. Neurophysiol. 33, Suppl. 173-178.
Rosvold, H. E., Mirsky, A. F., Sarason, I., Bransome, E. D., and Beck, L H.
(1956). A continuous performance test of brain damage. J. Consult. Psy-
cho/. 20, 343-350.
Schechter, G. and Buchsbaum, M. (1973). The effects of attention, stimulus
intensity and individual differences on the average evoked response. Psy-
ch ophysiology. 10, 392-400.
Schwent, V. L. and Hillyard, S. A. (1975). Evoked potential correlates of
selective attention with multi-channel auditory inputs. Electroenceph.
din. Neurophysiol. 38, 131-138.
Seppalainen, A. M. (1975). Applications of neurophysiological methods in
occupational medicine: A review. Scand. J. Work Environ, and Health 1,
1-14.
Seppalainen, A. M. (in press). Diagnostic utility of ERPs and other neuro-
electric measures in environmental and occupational medicine. In Multi-
disciplinary Perspectives in Event-related Brain Potential Research. (D. A.
Otto, ed.), U.S. Government Printing Office, Washington, D.C.
Shagass, C. (1972). Evoked Brain Potentials in Psychiatry. Plenum Press, New
York.
Starr, A. and Achor, J. (1975). Auditory brain stem responses in neurological
disease. Arch. Neurol. 32, 761-768.
Stevens, S. S. (1971). Reply to Monte Buchsbaum. Science 172, 502.
Stewart, R. D., Peterson, J. E., Baretta, E. D., Bachand, R. T., Hosko, M. J.,
and Herrmann, A. A. (1970). Experimental human exposure to carbon
monoxide. Arch. Environ. Health 21, 154-164.
10-29
-------
Stokinger, H. E. (1974). Behavioral toxicology in the development of threshold
limit values. In Behavioral Toxicology: Early Detection of Occupational
Hazards (E. Xintaras, B. L. Johnson, and I. de Grott, eds.). U.S. DHEW Pub.
No. (NIOSH) 74-126.
Sutton, S., Tueting, P., Zubin, J., and John, E. R. (1967). Information delivery
and the sensory evoked potential. Science 155, 1436-1439.
Tecce, J. J. and Cole, J. O. (1974). Amphetamine effects in man: Paradoxical
drowsiness and lowered electrical brain activity (CNV) Science 185, 451-
453.
Tecce, J. J., Cole, J. O., and Savignano-Bowman, J. (1975). Chlorpromazine
effects on brain activity (CNV) and reaction time in normal women.
Psychopharmacologia (Berl.) 43, 293-295.
Thompson, R. F. and Patterson, M. M. (eds.) (1974). Bioelectric Recording
Techniques. Part B: Electroencephalography and human brain potentials.
Academic Press, New York.
Timsit-Berthier, M., Delaunoy, J., Koninckx, N., and Rousseau, J. C. (1973).
Slow potential changes in psychiatry. I. Contingent negative variation.
Electroenceph. din, Neurophysiol. 35, 355-361.
Treisman, A. M. (1969). Strategies and models of selective attention. Psycho/.
Rev. 76, 282-299.
Tueting, P. (in press). Event-related potentials, cognitive events, and informa-
tion processing. In Multidisciplinary Perspectives in Event-related Brain
-Potential Research (D. A. Otto, ed.), U.S. Government Printing Office,
Washington, D.C.
Tueting, P., Sutton, S., and Zubin, S. (1971). Quantitative evoked potential
correlates of the probability of events. Psychophysiology 7, 385-394.
Vaughan, H. J., Jr., (1974). The analysis of scalp-recorded brain potentials. In
Bioelectrical Recording Techniques (R. F Thompson and M. M. Patterson,
eds.), Part B, pp. 157-207. Academic Press, New York.
Venables, P. H. and Martin, I. (eds.) (1967). A Manual of Psychophysiological
Methods. John Wiley, New York.
Walker, C. H. (1974). Comparative aspects of the metabolism of pesticides.
In Environmental Quality and Safety (F. Coulston and F Korte, eds.),
pp. 113-153. Academic Press, New York.
Walter, W. G. (1967). Slow potential changes in human brain associated with
expectancy, decision and intention. Electroenceph. din Neurophvsiol
Suppl. 26, 123-130.
10-30
-------
Walter, W. G., Cooper, R., Aldridge, V. J., McCallum, W. C., and Winter, A. L.
(1964). Contingent negative variation: An electric sign of sensorimotor
association and expectancy in the human brain. Nature (London) 203
380-384.
Webster, W. R. (1971). The effects of repetitive stimulation on auditory
evoked potentials. Electroenceph. din. Neurophysiol. 30, 318-330.
Winneke, G. and Kaska, J. (1974). Effects of trichlorethylene on signal detec-
tion and auditory evoked potentials in man. Paper presented at First World
Congress of Environmental Medicine and Biology, Paris.
Zappoli, R., Giulianb, G., Rossi, L. Papini, M., and Ronchi, 0. (in press). CNV
and SEP in shoe-industry workers affected with neuropathy due to the
toxic effect of adhesive solvents. In Multidisciplinary Perspectives in Event-
related Brain Potential Research (D. A. Otto, ed.), U.S. Government Print-
ing Office, Washington, D.C.
Acknowledgement
The low-frequency noise study was carried out in collaboration with V. A.
Benignus, J. H. Knelson, and L. J. Ryan. The technical assistance of J. D.
Prah in data analysis and in preparation of figures and the critical comments
of V. A. Benignus are gratefully acknowledged.
U.S. Government Printing Office: 1978—640-903/F7490 Region No. 4
10-31
------- |