NEUROBEHAVIORAL TOXICOLOGY
VOLUME 1 SUPPLEMENT 1
Test Methods for Definition
of Effects of Toxic Substances on
Behavior and Neuromotor Function
Edited by
I. Geller, W. C. Stebbins and M. J. Wayner
1979
Published by
ANKHO International Inc
Fayetteville, New York
ISSN 0191-3581
REPORT NO. EPA 560/11-79-010
-------
Test Methods for Definition
of Effects of Toxic Substances on
Behavior and Neuromotor Function
Report No. EPA 560/11-79-010
-------
Test Methods for Definition
of Effects of Toxic Substances on
Behavior and Neuromotor Function
NEUROBEHAVIORAL TOXICOLOGY
Volume I, Supplement 1, 1979
This Supplement to Volume 1 ofNeurobehavioral Toxicology con-
tains the Proceedings of the Workshop on "Test Methods for
Definition of Effects of Toxic Substances on Behavior and
Neuromotor Function," which was held April 1-4, 1979, at the
Hilton Palacio del Rio Hotel in San Antonio, Texas. This meeting
was organized by the Southwest Foundation for Research and
Education with the direction of Dr. Irving Geller; and was spon-
sored by the United States Environmental Protection Agency,
Office of Toxic Substances, under Contract No. EPA 68-01-4870.
These Proceedings serve as the Final Report No. EPA 560/11-79-
010. The opinions expressed in these articles, however, are those
of the authors and are not necessarily endorsed nor shared by the
Environmental Protection Agency.
Library of Congress Catalog Card Number 79-53843
ISBN 0-916086-02-2
ISSN 0191-3581
-------
Proceedings of the Workshop on
Test Methods for Definition
of Effects of Toxic Substances on
Behavior and Neuromotor Function
Sponsored by the
United States Environmental Protecton Agency
Office of Toxic Substances
Under Contract No. EPA 68-01-4870
(Final Report No. EPA 560/11-79-010)
San Antonio, Texas, April 1-4, 1979
Edited by
I. GELLER
Southwest Foundation for Research and Eductation, San Antonio, Texas
W. C. STEBBINS
University of Michigan
and
M. J. WAYNER
Syracuse University
ANKHO INTERNATIONAL INC.
-------
Test Methods for Definition of Effects of Toxic Substances
on Behavior and Neuromotor Function
Supplement to NEUROBEHAV1ORAL TOXICOLOGY
CONTENTS
Preface. GOULD, D. H 1
Opening Remarks. PAGE, N. P 3
Introduction and Overview. GELLER, I. and W. C. STEBBINS 7
Use of discrimination behavior for the evaluation of toxicants. GELLER, I., E. GAUSE, R. J.
HARTMANN and J. SEIFTER 9
Effects of toxicants on visual systems. MERIGAN, W. H 15
Effects of toxicants on the somatosensory system. MAURISSEN, J. P. J 23
Comparative behavioral toxicology. STEBBINS, W. C. and D. B. MOODY 33
Trialwise tracking method for measuring drug-affected sensory threshold changes in animals. ANDO, K.
and K. TAKADA 45
Motor activity: A survey of methods with potential use in toxicity testing. REITER, L. W. and R. C.
MACPHAIL 53
Reinforcing properties of inhaled substances. WOOD, R. W 67
Behavioral assessment of risk-taking and psychophysical functions in the baboon. BRADY, J. V.,
L. D. BRADFORD and R. D. HIENZ 73
Operant conditioning of infant monkeys (Macaca fascicularis) for toxicity testing. RICE, D. C. . . 85
Effects of pre- and post-natal lead on affective behavior and learning in the rat. FLYNN, J. C., E. R.
FLYNN and J. H. PATTON 93
Performance and acquisition of serial position sequences by pigeons as measures of behavioral toxicity.
MCMILLAN, D. E 105
Effects of solvents on schedule-controlled behavior. COLOTLA, V. A., S. BAUTISTA, M.
LORENZANA-JIMENEZ and R. RODRIGUEZ 113
Testing for behavioral effects of agents. DEWS, P. B. and G. R. WENGER 119
Some problems in interpreting the behavioral effects of lead and methylmercury. LATIES, V. G. and
D. A. CORY-SLECHTA 129
Screening for neurobehavioral toxicity: The need for and examples of validation of testing procedures.
TILSON, H. A., C. L. MITCHELL and P. A. CABE 137
Behavioral epidemiology of food additives. WEISS, B., C. COX, M. YOUNG, S. MARGEN and
J. H. WILLIAMS 149
Psychological test methods: Sensitivity to long term chemical exposure at work. HANNINEN, H. . 157
Quantitative analysis of rat behavior patterns in a residential maze. ELSNER, J., R. LOOSER and
G. ZBINDEN 163
-------
Comparison of neurobehavioral effects induced by various experimental models of ataxia in the rat.
JOLICOEUR, F. B., D. B. RONDEAU, A. BARBEAU and M. J. WAYNER 175
Methodological problems in the analysis of behavioral tolerance in toxicology. BIGNAMI, G 179
Morphological studies of toxic distal axonopathy. SCHAUMBURG, H. H 187
Cellular responses to neurotoxic compounds of environmental significance. SPENCER, P. S 189
Physiological and neurobehavioral alterations during development in lead exposed rats. FOX, D. A. 193
A preliminary test battery for the investigation of the behavioral teratology of selected psychotropic drugs.
BUTCHER, R. E. and C. V. VORHEES 207
Assays for behavioral toxicity: A strategy for the environmental protection agency. WEISS, B. and
V. G. LATIES 213
Final Comments. SEIFTER, J 217
Author Index 219
Subject Index 221
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, p. 1, ANKHO International Inc., 1979.
Test Methods for Definition of
Effects of Toxic Substances on
Behavior and Neuromotor Function
Preface
The subject of this Workshop and the desired output is
clearly stated in the title. It was held "to obtain a scientific
assessment for regulatory decision-making of the currently
available methodologies necessary to determine the toxic
threat to human health and the environment posed by all
chemicals in commerce." The Office of Toxic Substances of
the Environmental Protection Agency is required in its en-
abling act, the Toxic Substances Control Act (TSCA), to
carry out such determinations in the area of behavioral dis-
orders, among others. This meeting was therefore initiated
with the inspiration of Dr. Joseph Seifter.
When I became Project Officer for this Workshop, I soon
realized how fortunate the Office of Toxic Substances was to
have Irving Geller as the guiding hand to organize the meet-
ing. It is indeed true to say that the meeting could not have
succeeded without his outstanding knowledge of the behav-
ioral field, and his untiring efforts to pull all the loose ends
together. With the invaluable assistance of Dr. William
Stebbins, he achieved a program which included the best
international thinking and research in behavioral toxicology,
and gave the EPA its best chance of succeeding in its
prescribed responsibilities.
The various opinions I have heard indicate that this
Workshop did succeed in delineating areas of agreement. To
this extent, EPA can be confident that meaningful tests are
available for measuring sensory and functional impairment,
and beyond this, changes in the more complex areas of dis-
criminant and learning processes and reproductive behavior.
It has, however, clearly been confirmed that the full breadth
of behavioral toxicology with its integration of complex in-
teractions and compensatory mechanisms may never be fully
circumscribable. In any case, since EPA must take into ac-
count the economic realities, it could not prescribe every one
of a completely predictive set of tests since there would be so
many, even if such a set could be developed.
The answer to achieving adequate guidelines for regula-
tory consistency may lie in some approaches which merit
further exploration. The quantitative activities of various
reference compounds in a selection of well-characterized
tests treated with multivariate statistical methods may give
profiles which delineate different types of neurobehavioral
toxicity, and which may be useful in determining the qualita-
tive toxicities of unknowns. Each such profile might then
suggest the need for more extensive testing in certain critical
areas of behavior. Then, as suggested by Weiss and Laties in
the last paper, EPA by consultation with a scientific advisory
board could decide whether the sponsor's proposed set of
tests properly covers the likely areas of concern in light of
the latest understanding of the discipline.
Finally, I must express my thanks for the valuable contri-
butions of the session chairmen and all of the speakers who,
of course, were the raison d'etre of the meeting. And for this
report, thanks again to Irv Geller, Bill Stebbins and Matt
Wayner for the efficient editing process which has led so
expeditiously to this publication. Only one contribution
could not be included in this publication and the final
material was received by the Publisher on August 13, 1979.
DAVID H. GOULD
Office of Toxic Substances
U. S. Environmental Protection Agency
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neummotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 3-5. ANKHO International Inc., 1979.
Opening Remarks:
TSCA Requirements for Testing Chemicals for
Behavioral Effects and Neurotoxicity
NORBERT P. PAGE
Environmental Protection Agency
ON behalf of the Environmental Protection Agency (EPA) I
join with Mr. Goland in welcoming you to this important
workshop on Behavioral and Neurotoxicologic Test Meth-
ods. I particularly want to thank Dr. Irving Geller, Dr.
William Stebbins, and Dr. David Gould for developing such
a promising program. The actual need for this workshop was
identified over two years ago, Dr. Joseph Seifter of the EPA
and Dr. Geller of the Southwest Foundation for Research
and Education providing the initiative to organize and enter
the workshop into the EPA's program. The need for a
successful workshop is even greater today than was realized
then.
We have with us many national and international experts
in the behavioral sciences and other areas of toxicology, and
I look forward to an enlightening discussion of testing meth-
ods in behavioral and neurotoxicology. To provide some
groundwork for this discussion and at the same time keep
introductory remarks brief, my comments will be restricted
primarily to three subjects: (1) the regulatory framework
within which the workshop results might be utilized; (2) the
EPA's responsibilities for testing under the Toxic Sub-
stances Control Act (TSCA) and its approach to implement-
ing standards; and (3) the specific objectives or issues for
consideration by this workshop.
The overall objective for the human behavioral workshop
described in the announcement follows: "The workshop
is to obtain a scientific assessment for regulatory decision-
making of the currently available methodologies necessary
to determine the toxic threat to human health and the en-
vironment posed by all chemicals in commerce." I would
like to stress one part of that objective—the need for scien-
tific assessment of currently available methodologies. Under
TSCA, EPA has the authority and the responsibility to en-
sure manufacturers provide data on which the assessment of
unreasonable risk can be made. Congress singled out several
key health effects of concern, one of which was behavioral
disorders.
CHEMICALS AS BEHAVIORAL AND NEUROTOXICITY DETER-
MINANTS
How big a role do chemicals play in the human behavioral
disorder problem? Some scientists say their role is minimal.
Others claim chemicals play a very major role which gener-
ally has gone unrecognized. To me it seems likely that behav-
ioral disorders are the result of a complex array of many
factors including genetics, nutritional aspects, our socio-
economic factors, diseases, and of course, chemicals. It is
certain that some chemicals play a major role in the etiology
of neurological or behavioral disorders. Several come to
mind immediately: the role of heavy metals, such as lead and
mercury, a number of the pesticides, including many organo-
phosphates, and some of the chlorinated chemicals such as
kepone. An important group that has been incriminated is
the chlorinated solvents. I am sure as we move through this
workshop we will hear of many other chemicals which have
been shown to produce behavioral or neurotoxic effects.
METHODS OF ASSESSMENT
How can we assess for the behavioral or neurotoxicology
potential of chemicals? One very obvious method is the as-
sociation of conditions observed in humans with exposure to
specific chemicals. Like all epidemiological studies, this ap-
proach is expensive to conduct. Moreover, it is difficult to
sort through subjective complaints where exposures may be
to several chemicals, and where many modifying factors
interact. Indeed, this is not an easy task. Some papers to be
presented in this workshop will deal with this aspect of
human observation and surveillance. The other main cate-
gory of assessment methods is the whole animal tests of
various types. These range from general toxicity studies
where behavior or neurological effects may be observed as a
part of a routine health examination to the more sophisti-
cated and specific or tailored tests which are designed to
assess behavioral and neuromotor function in a more consis-
tent manner.
REGULATORY FRAMEWORK FOR TSCA, SECTIONS 4 AND 5
The regulatory framework of TSCA is unique for Fed-
eral legislation in that it provides specific authority in Sec-
tions 4 and 5 to require manufacturers or processors to pro-
vide data to EPA for assessment purposes. Section 4 per-
tains to existing chemicals or categories of chemicals;
whereas Section 5 pertains to new chemicals which are to be
manufactured. Under Section 4, EPA issues chemical test
rules that define a chemical or categories of chemicals which
must be tested, the effect to be tested for, and the standards
-------
PAGE
by which the testing will be performed. Under Section 5, the
Agency has no specific testing authority; however, a pre-
manufacturing notification must be submitted to EPA 90
days prior to manufacture. In that premanufacturing notifi-
cation, the manufacturer must provide data on which the
Agency can make an assessment of risk to human health and
the environment.
The approach we are using with Section 4 is to develop
and place into the. Code of Federal Regulations generic stan-
dards for various health effects. These standards will then be
referenced at a later time when chemical test rules are pro-
posed. At the time of proposing a chemical test rule, specific
modifications to the generic standards can be made so as to
customize the standards to the chemicals or category of
chemicals to be tested. Moreover, the standards must be
reviewed annually and revised as appropriate to assure their
currency with scientific development.
A number of health effects testing standards have been
developed and will shortly be proposed in the Federal Regis-
ter. Four of these should be published by the end of April.
These are testing for oncogenicity, nononcogenic chronic
effects, all chronic effects and good laboratory practices. A
number of other health standards should be proposed in the
early summer. These are for acute toxicity including lethal-
ity, eye irritation, dermal sensitization and dermal irritation,
subchronic toxicity testing, teratogenicity, reproductive ef-
fects, and mutagenicity. This latter group of standards will be
proposed basically as they appear in the guidelines proposed
last August for use in registration of pesticides under the
Federal Insecticide, Fungicide and Rodenticide Act (FIFRA).
We have not made a decision on behavioral or
neurotoxicity standards. The reason for this is simple. We
are uncertain as to which test systems are well enough
validated and acceptable to the scientific community. There-
fore, results of this workshop will be carefully reviewed by
EPA in making its decision on how to proceed with these
particular standards.
COORDINATION WITH U.S. AND INTERNATIONAL ORGAN-
IZATIONS IN DEVELOPMENT OF TEST STANDARDS
The Office of Toxic Substances is aware that EPA is not
the only organization which has the responsibility for devel-
oping test methods. Within EPA we have joined with the
Office of Pesticide Programs to form a joint work group to
develop test methods which are basically consistent for the
TSCA and for FIFRA. We have also joined with the three
other U.S. Federal Regulatory Agencies in forming the In-
teragency Regulatory Liaison Group (IRLG). The IRLG
committee on guidelines and standards is attempting to de-
velop consistent standards and guidelines for EPA, FDA,
CPSC, and OSHA. In addition, we are a partner or member
of the Organization for Economic Cooperation and Devel-
opment (OECD). The OECD will propose test standards for
use by the international community in the testing of chemi-
cals for toxic chemical control. The EPA is attempting to
harmonize our test standards with the OECD.
SPECIFIC BEHAVIORAL OR NEUROTOXICOLOGICAL STANDARD
I would like now to turn my attention to specific behav-
ioral or neurotoxicological standards that have been pro-
posed or are under development. Under the proposed
FIFRA guidlines a very minimal set of specific tests are pro-
posed in this area. They consist of basic observations in the
general toxicity tests and two specific tests for delayed
neurotoxicity using hens. These are also observational tests,
primarily directed to testing organophosphates and esterase
inhibitors. Nothing is proposed or planned at this time by the
IRLG. The OECD so far has completely avoided discussion
of the needs or methods of testing for behavioral effects.
What chemicals should be tested? Of the thousands in the
environment, there are probably some "sleepers" that are
responsible for some of the bazaar, behavioral or neurolog-
ical conditions which exist in the human population, but for
which a specific cause and effect relationship has not been
established. These chemicals must be identified for testing.
Since Congress recognized that EPA was not the only
Agency that had a concern for proper and useful data gener-
ation, it provided a provision for an interagency committee
to select chemicals for EPA to test, the Interagency Testing
Committee (ITC). This committee is composed for represen-
tatives of eight different Federal agencies as well as partici-
pating observers from a number of other agencies. The
committee is authorized to designate up to fifty chemicals or
categories of chemicals for testing at any one time. As of this
time, the ITC has designated approximately 25 chemicals
and categories of chemicals to EPA.
Few of the designated chemicals are proposed for testing
for specific behavioral or neurotoxic effects. Several, how-
ever, are proposed for basically a general toxicity profile in
which neurologic effects would be one effect to be tested for.
A few of those which are proposed, for example acrylamide
and arylphosphates, are already known to have neurotoxi-
cology effects. We would welcome the review and comments
by the work group participants on the 25 different chemicals
or categories. If you do not have this list of chemicals, we
would be happy to provide this to you.
As I indicated earlier, under Section 5, there is no provi-
sion for specific testing requirements for new chemicals, un-
less they fall within a category which has been proposed for
testing under Section 4. Such a chemical, cannot be man-
ufactured until it has been appropriately tested. The pre-
manufacturing notification under Section 5 will commence
30 days after the publication of the inventory of chemicals
currently in commerce. We expect that the inventory will be
published in late May or early June, and therefore pre-
manufacturing notification will begin in June or July.
We have little concept at this time as to the amount of
testing that will be performed on new chemicals by the
manufacturer or the type of test data that we may be provid-
ed, especially in the health area. It could be rather minimal
and consist primarily of acute toxicity tests, mutagenicity
tests and perhaps some subchronic toxicity testing. This will,
of course, depend upon the volume of the chemical the
manufacturer expects to produce and market, its release into
the environment and the anticipated human exposure. In the
event that the Agency does not have sufficient data provided
on potential health or environmental effects and thus is un-
able to conduct a meaningful risk assessment, the EPA can
undertake legal proceedings to prevent manufacture of the
chemical.
In our attempts to impliment TSCA, we recognize that we
are pushing the state of the science in developing test stan-
dards. No other legislation requires actual standards to be
placed in the Code of Federal Regulations. The FIFRA re-
quires that the Agency develop testing guidelines for the
registration of pesticides. The FDA reviews Pharmaceuticals
and food additives but does not have an assigned responsi-
-------
OPENING REMARKS
bility to provide standards for test methods to be used for the
development of the toxic effects data.
I would like to mention one aspect in relation to the test
methods program of the Office of Toxic Substances. We are
attempting to provide for the validation of many tests that are
in current use or proposed for use. A number of interagency
agreements or contracts have been awarded to validate a
number of test methods in certain areas, in particular that of
mutagenicity or short-term tests, in acute effects and sub-
chronic tests. We have issued a contract "Request for Pro-
posal" to validate several of the promising tests on behav-
ioral or neurotoxic effects. This "Request for Proposal"
closes on April 30.1 would encourage those of you that have
the interest and scientific resources to undertake such a
validation program on behavioral or neurotoxicology test
methods to obtain a copy of this RFP. It is possible that some
of the findings of this workshop will be useful in deciding the
nature of the validation program.
ISSUES FOR CONSIDERATION BY THE WORKSHOP
I would like to conclude by proposing a list of issues that I
think the workshop should consider during these next two
days. These issues are directed toward the three different
forms of tests which may be required under TSCA. One of
these would be routine observational assessments that can
be made during routine or general toxicity tests. Such obser-
vations can enhance the quality of information derived for
assessment purposes. The second type of testing is the
neuropathology, neurophysiology or neurochemistry exam-
inations, which may also be conducted as a part of routine
general toxicity tests or perhaps the specifics for those ef-
fects. The third type of test include those which are very
specific and are designed for behavioral or neurological
functions including that on the developing fetus.
The questions to be considered by the workshop are as
follows: (1) How can we strengthen the acute, subchronic or
chronic toxicity tests to provide the best possible indication
of potential behavior of neurotoxicology effects? (2) Can and
should we require a greater level of pathology examinations?
How sensitive is pathology in detecting behavioral or
neurotoxic effects? (3) Can we propose meaningful and
validated neurochemical, physiological or neuropathology
parameters which can be used in testing for neurologic or
behavioral effects? (4) What existing tests for sensory, motor
or cognitive effects are well enough developed and validated
to be used as standards at this time? (5) Are there tests now
in the research stage that need further development and
validation? (6) What areas need further research on test
methods? (7) How can we best group the various tests to
provide for a safety assessment scheme? Should we go with
a battery or a sequential scheme, and if so what would be the
criteria for choosing the various tests? (For example: use
pattern, structure relationships, results of prior tests, pro-
duction level, etc.)? (8) Is there a logical and scientific
scheme which can be used to test for certain classes of chem-
icals.
These are the kinds of questions OTS must address as it
proceeds with its program for developing test standards
under TSCA. We are hopeful that this workshop will con-
sider these aspects as we discuss papers to be presented
during the next couple of days. In looking over the agenda,
we have some excellent papers on existing test methods and
new test methods under development.
In closing, I would like to direct our attention to another
requirement of the EPA under TSCA, we must not only
consider the scientific value of the tests, but we must also
consider the economic and resource limitations in applying
proposed test methods to the testing of chemical substances.
I would encourage discussion of these aspects along with the
scientific utility of the test methods.
-------
Test Met hods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, p. 7. ANKHO International Inc., 1979.
Introduction and Overview
This Workshop was convened by the Environmental Pro-
tection Agency for the purpose of assessing the current
capabilities of the discipline of Behavioral Toxicology for
predicting neural and behavioral toxicity of environmental
chemicals. Behavioral techniques can be employed to detect
and establish dose-response relationships for toxicants for
which the critical target is the nervous system. It has been
pointed out in this meeting that behavioral tests may also
detect effects upon systems other than the nervous system;
i.e., establish indirect effects of toxicants, as well as make
possible the identification of populations at greatest risk
from a given toxicant.
Doctors Page, Tilson, Reiter, Gage and Seifter have em-
phasized the needs and priorities of regulatory agencies with
regard to tests and methodologies for the detection and
assessment of health effects. These requirements provide a
challenge to behavioral scientists concerned with adapting
the science of behavior to rigorous screening for behavioral
toxicity. This challenge perhaps begins with the problem of
definition of behavioral toxicity, which is obviously not an
easy task. For convenience, a working definition probably
should distinguish between acute, functional, reversible ef-
fects upon behavior, and chronic, structural damage to the
nervous system which may or may not be associated with
some degree of functional recovery from the primary deficit.
Other Aspects of the Challenge to Behaviorists Include
1. The need to maximize the amount of information ob-
tainable from a behavioral test approach in order to evaluate
effects upon overlapping functions, and to detect any degree
of possible functional recovery. A corollary of this is the
need for investigators to look for delayed effects upon behav-
ior and for recovery from observed toxicities.
2. The need to refine estimates of the dose of toxicant to
the nervous system. This may require an understanding of
the metabolism of the initial exposure agent, identification of
the specific neurotoxic chemical species, and establishment
of the pharmacodynamic relationships involved.
3. The need to determine whether or not a dose-response
relationship actually exists for all classes of neurotoxicants
and what the limitations of specific behaviors are for the
measurement of dose-response. There is some indication
that acute or sub-chronic functional effects may plateau at
low doses—perhaps due to saturation of receptors, or with
the establishment of steady-state circulating and tissue res-
ervoir levels. These considerations may be particularly im-
portant for substances administered by inhalation.
4. The need to evaluate and allow for differences be-
tween individual test animals in susceptibility to effects of a
toxicant.
5. The need to consider whether enzyme induction or
other effects of prior exposure to the test agent or to other
compounds, treatments, etc., are affecting the outcome of
the behavioral test.
6. The importance of testing for possible potentiation (or
diminishment) of effects due to a given agent by the presence
of other substances likely to be encountered in mixtures or
adventitiously in the environment.
7. The sensitivity of the behavioral test and the relevance
to the human situation may both be increased by the incor-
poration of a pharmacological challenge into the test.
8. The need to examine other possible treatment-
behavior interactions including behavioral tolerance,
sensitization-desensitization, hypersensitization, and com-
pensation.
It is anticipated that the need of regulatory agencies for
validated and comprehensive behavioral tests of nervous
system function will provide clear-cut goals and objectives
for behavioral scientists, and that the insights gained will
result in greater coordination of research efforts—both with
other behaviorists and with scientists in auxiliary disciplines
such as biochemistry and pharmacology.
IRVING GELLER
Southwest Foundation for Research
and Education
San Antonio, TX
and
WILLIAM C. STEBBINS
University of Michigan
Ann Arbor, MI
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 9-13. ANKHO International Inc., 1979.
Use of Discrimination Behavior for the
Evaluation of Toxicants1'2 3
I. GELLER, E. CAUSE, R. J. HARTMANN AND J. SEIFTER*
Department of Behavioral and Environmental Sciences, Southwest Foundation for Research and Education
8848 West Commerce, P.O. Box 28147, San Antonio, TX 78284 and
*Office of Toxic Substances, TS 792, U.S. Environmental Protection Agency
Washington, DC 20460
GELLER, I., E. CAUSE, R. J. HARTMANN AND J. SEIFTER. Use of discrimination behavior for the evaluation of
toxicants. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 9-13, 1979.—This study involved the application of discrimination
behavior for the study of effects of environmental contaminants on the behavior of laboratory animals. Polybrominated
biphenyl (PBB) was evaluated for effects on the acquisition and performance of a simple auditory discrimination by rats.
Methyl ethyl ketone (MEK), methyl isobutyl ketone (MIBK) and carbon monoxide (CO) were evaluated for effects on a
delayed match-to-sample discrimination task in the juvenile baboon. All of the contaminants slowed response times and
increased extra responses. These findings suggest that discrimination behavior may be of value for the evaluation of
environmental contaminants for effects on the central nervous system.
Polybrominated biphenyl
Auditory discrimination
Ketones Carbon monoxide Environmental contaminants
Delayed match-to-sample discrimination
THIS workshop focuses upon behavioral observations
which may be useful for early detection of neurotoxicity at-
tributable to environmental agents. Assessment of potential
neurotoxicity becomes a formidable task because such a
large number of functions are under nervous system control
and these various functions may be inhibited differentially by
any given neurotoxicant. In the detection of neuroactivity,
discrimination tasks are useful because they lend themselves
to the simultaneous measurement of a number of CNS
mediated functions. For example, the delayed match-to-
sample discrimination task can be said to include associative
learning, visual reproduction (short-term memory),
similarities or dissimilarities in stimuli, psychomotor func-
tion and response or reaction time. The potential value of
such discrimination behavior in screening for neurotoxicity
is indicated by the work of Hanninen who reported in this
workshop [8] that humans exposed to toluene showed
marked impairment of associative learning, visual reproduc-
tion, similarities and psychomotor function. This
neurotoxicity of toluene appears to be quite specific since
exposure to styrene, a structurally similar compound, did
not affect results of the cognitive tests but did alter
psychomotor function [8].
In the study of any type of toxicity, the relevance of ro-
dent level observations for extrapolation to humans must be
continually considered. Discrimination behavioral assays
may be employed in both sub-human primates and rodents
offering a direct inter-species comparison of the effects of a
given agent as well as optimal relevance.
We have employed discrimination behavior to evaluate
the toxicity of substances that are abused through inhalation
and are also encountered as environmental or potential
spacecraft contaminants. The absence of any clear cut in-
formation relative to the central nervous system effects of
Polybrominated biphenyl provided the impetus for a study
on the effects of this toxicant on the acquisition and per-
formance of a simple discrimination task by rats.
EXPERIMENT 1
METHOD
The animals were male Holtzman, Sprague-Dawley rats,
approximately three months old at the start of the experi-
ment.
A purified and analyzed sample of hexabrominated
biphenyl, obtained from NIEHS, was prepared in lecithin-
liposomes suspended in saline. It was administered orally at
1 mg/kg to 12 rats Monday through Friday of each week
during a one-month period for a total of 20 doses of PBB.
Twelve additional rats received 20 administrations of the
control vehicle during the same one-month period.
For the discrimination task, hungry rats in Skinner boxes
had to select the right or left lever as correct as a function of
'Request for reprints should be addressed to Dr. I. Geller.
"Supported in part by grants and contracts from APHA no. 68-01-3859, NIDA no. DA01339, NASA no. NAS 9-14743, and NIEHS no.
ES01246.
•The authors are indebted to Murray Hamilton for tissue analysis of PBB's.
-------
10
GELLER ETAL.
the presence of a tone or clicker stimulus, respectively. The
auditory stimuli occurred at random intervals on the average
of once every two minutes (2-min VI). By making the correct
choice, animals obtained milk rewards. Perfect discrimina-
tion was reflected in 100% correct responding to stimulus
presentations with minimal or no responding in the absence
of stimuli. Responses which occurred in the absence of
stimuli reflected the general activity of the animal as well as a
lack of efficiency.
Training on the discrimination task was as follows: all
animals were gradually reduced to 80% of their original start-
ing weights. They were then placed in the chambers for 1/2
hr, during which time the feeders were activated every 90
sec. On the following three days rats were placed in the
chamber and given access to only the left lever. Pressing the
lever would activate the clicker stimulus and produce a food
reward. Animals remained in the chamber for 1/2 hr or until
they made 100 responses. The right lever was then substi-
tuted for the left lever and animals received similar training
in which a tone stimulus was paired with lever presses. After
three days on this procedure, acquisition training for the
discrimination task began. Tones or clicker stimuli occurred
in a mixed order on the average of once every two minutes
(2-min VI). Pressing the correct lever turned off the stimulus
and activated the milk feeder. Pressing the incorrect lever
simply turned off the stimulus. Response latencies for the
entire session were cumulated on a running time meter.
RESULTS
PBB rats did not differ from controls with respect to accu-
racy on the discrimination task during the first four weeks of
training. During Weeks 5-8 acquisition was more rapid for
the control animals; however, the difference between PBB
and control data was significant only on the eighth week of
acquisition (p<0.05). From Week 9 onward, all rats per-
formed at the 90% criterion level.
Throughout 24 weeks of discrimination training PBB rats
were less efficient than controls in that they made many
more extra responses. PBB rats also showed a trend toward
longer response times throughout the experiment. Figure 1
shows these data for 18 weeks of training. Averaged weekly
response time for PBB (solid lines) or control animals
(broken lines) indicate PBB animals generally were slower to
respond to either tone or clicker stimuli. The effect is most
striking for response time measured on the right lever.
Ten months after the last PBB administration rats were
sacrificed and analyzed for PBB levels by electron capture
gas liquid chromatography. PBB was found in whole brain in
concentrations ranging from 0.038 to 0.40 /u.g/g wet weight.
PBB was also found in plasma in concentrations ranging
from 0.135 to 0.372 /ig/ml.
EXPERIMENT 2
METHOD
A match-to-sample discrimination task was used for the
baboon and the toxicants were gases administered through
inhalation in a flow-thru system.
The behavioral test chambers and gas exposure chambers
have been previously described [7]. Two large stainless steel
chambers equipped with a walk-in air lock were used to con-
duct the exposure. The chambers measured approximately
nine feet high and nine feet in diameter. Juvenile male ba-
10
to 8H
Q
O 6-
CJ
LU
CO
4-
2-
RESPONSE TIME
LEFT LEVER
» x CONTROL
• • PBB
10-
8-
co
Q
§
o
LU
CO 4.
2-
2 4 6 8 10 12 14 16 18
2 4 6 8 10 12 14 16 18
WEEKS
FIG. 1. Effect of polybrominated biphenyl on response time of the
rat.
boons approximately two years of age were housed in behav-
ioral test chambers which were maintained in the large expo-
sure chambers. The behavioral test chambers were designed
so that an intelligence panel could be slipped down between
the outside wall of the cage and the baboon. The intelligence
panel was equipped with a row of three translucent discs
which served as levers. Under the appropriate experimental
conditions, pressing either side disc produced a banana pel-
let reward. Experimental sessions of two-hour duration were
conducted on Monday through Friday of each week.
When the session timer was activated, a variable interval
(VI) tape was set in motion. The tape programmed the occur-
rence of a stimulus on the center lever on the average of once
every three minutes. The VI tape was inoperative during
each trial which began with the illumination of one of the
stimuli, the probe stimulus on the center lever. The stimulus
was terminated after a 30-sec period or by a response on the
lever. Termination of the stimulus activated a timer for a
two-minute delay interval. At the end of the delay interval,
stimuli appeared on both levers adjacent to the center lever!
The correct matching stimulus was varied between these two
levers in a mixed order. A response on the correct lever
where the stimulus matched the center lever stimulus termi-
nated the stimuli, activated the feeder and produced a
banana pellet reward. Responses on the incorrect lever sim
ply terminated the stimuli and again set the VI tape in m
tion.
-------
TOXICANTS AND BEHAVIOR
11
A record was kept of the number of probe stimuli pre-
sented during each 15 min segment of a two-hr session, the
numbers of correct matching responses on right and left lev-
ers and the number of incorrect responses. A record was also
kept of any extra responses that may have occurred on the
three levers when the stimuli were not activated or during
the delay interval. The time it took the animal to press the
lever after a stimulus was activated was also measured (re-
sponse time). After the baboons were trained to 90-100%
efficiency on the discrimination task, the exposure phase of
the study was begun.
The animals were exposed to methyl ethyl ketone (MEK)
or methyl isobutyl ketone (MIBK). Exposure was by means
of the vapor saturation technique [3], For the vapor satura-
tion method, air is bubbled through a gas washing bottle
containing the liquid to be vaporized. In passing through the
liquid, the air becomes saturated with vapor which is then
directed to the air intake ducts of the exposure chamber.
Changing the flowrate with the fine metering valve or chang-
ing the temperature of the constant temperature bath allows
one to produce a range of pollutant concentrations in the
exposure chamber. The technique is simple and works well
for substances that are liquids at room temperature. A
Hewlett-Packard gas chromatograph, modified for automatic
sampling was employed. This allowed for automatic sam-
pling, quantitation and recording of pollutant concentrations
in the exposure chamber.
The animals were also exposed to carbon monoxide (CO).
The exposure atmospheres for the carbon monoxide studies
were produced using compressed gas cylinders of CO ob-
tained in 99.5% purity. The correct amount of carbon
monoxide was introduced into the chamber by means of a
calibrated flowmeter and a fine metering valve. Samples of
chamber air were withdrawn with 1 ml gas tight syringe and
analyzed on a gas chromatograph. The GC uses the principle
of catalytic conversion to hydrogenate CO to methane which
is detected with a conventional flame ionization detector.
Samples were analyzed on the average of once every 10 min.
The concentration was determined by a comparison of the
detector response for a chamber air sample with the detector
response for a series of standard samples. With these tech-
niques, exposure chamber atmospheres were maintained
within 10% of the desired value.
Animals were exposed to the ketones for 24 hr per day
during a seven-day period. They were exposed to 100 ppm
MEK, 50 ppm MIBK or to a combination of MEK or MIBK
at the same concentrations. These concentrations are half
the established threshold limit values [5]. While two animals
in one of the chambers were being exposed to a contaminant
atmosphere, the animals in the other chamber served as con-
trols and were exposed to clean air during the same period.
Thus, not only did other animals serve as controls, but each
animal served as its own control in that exposure data could
be compared with data obtained pre- and post-exposure
time.
Animals were exposed to 25 ppm or 50 ppm CO for six hr
per day during a one-week period while the two-hr behav-
ioral test was conducted in the morning or afternoon. Thus,
behavioral testing took place during the first two hr of CO
exposure in the morning or in the afternoon after the animals
had already been exposed for four hr.
RESULTS
For the ketones, performance on the match-to-sample
IX 529
MEAN RESPONSE TIME IN SEC.
CONTROL-— EXPOSURE —
" MEK otIOOPPM
% 3 On
5 2.8-
LI 26:
5 2.4-
1-22-
i i?
a 14-
£ '2-3
1 1.0-
£ 08-
MIBK at 50 PPM
MEK otIOOPPM
and
MIBK at 50 PPM
2367
DAYS
236
DAYS
236
DAYS
RESPONSE DURING DELAY
MEK otIOOPPM
MIBK at 50 PPM
MEK at 100 PPM
MIBK at 50 PPM
! 300-
: 200-
236
DAYS
1X531
MEAN RESPONSE TIME IN SEC.
CONTROL EXPOSURE —
30-
28-
26-
24-
22-
20-
I 8-
I 6-
14-
1.2-
10-
08-
MEK at 100 PPM
MIBK at 50 PPM
MEK at 100 PPM
and
MIBK at 50 PPM
12367 I 2 3 6 7
DAYS DAYS DAYS
RESPONSES DURING DELAY
MEK at 100 PPM
and
MEK allOOPPM MIBK at 5OPPM MIBK at 50 PPM
> 4OO-
)
1 30O-
! 200-
t
! 100-
123671236712367
DAYS DAYS DAYS
FIG. 2. Effect of seven-day exposures to 100 ppm MEK or 50 ppm
MIBK, administered alone or in combination on match-to-sample
behavior of baboons. Control data are represented by broken lines
and exposure data by solid lines.
task was not impaired under any of the three experimental
conditions. However, response times or numbers of re-
sponses during the delay periods were affected by the gases
in the four test animals.
Figure 2 shows these effects in two baboons. The solid
lines represent exposure data and the broken lines, control
-------
12
GELLERCTAL.
MEAN RESPONSE TIME IN SECONDS
CONTROL! ) EXPOSURE! )
CO at 50 PPM
GOO-
„ 500-
| 400.
I
IE 300-
RESPONSES DURING DEL*
PM
2 3
OtftS
FIG. 3. Effects of six-hr daily exposure to carbon monoxide during a
five-day period. Broken lines represent control data while solid lines
represent exposure data.
data, obtained during a seven-day, pre-exposure period. For
Baboon 529, mean response time increased above control
levels under 100 ppm MEK during each of five behavioral
sessions. The same was true for MIBK at 50 ppm. However,
under a combination of MEK and MIBK at the same con-
centrations, the exposure data approximated that of the
pre-exposure control. Mean response time for Baboon 531
increased gradually under 100 ppm MEK during the first
three exposure days. On Days 6 and 7 the data were like that
of controls. Mean response time under 50 ppm MIBK in-
creased throughout the week of exposure. Again a combina-
tion of 100 ppm MEK and 50 ppm MIBK produced data
similar to controls.
Responses during the delay intervals were like that of
controls for Baboon 529, while for Baboon 531 there oc-
curred a large increase under MEK, little or no effect under
MIBK and an increase with the combined MEK, MIBK ex-
posure.
For CO, a slight impairment of discrimination occurred at
50 ppm; animals exposed to this concentration of CO occa-
sionally made a mistake.
Data typically obtained for response time latencies or re-
sponses during the delay are shown in Fig. 3; the broken
lines in the figure represent control data averaged for each of
five pre-exposure days and the solid lines represent data for
five exposure days. At 25 ppm, CO produced a slowing of
response time on Day 4 of the morning and afternoon expo-
sures. These differences between exposed and control ani-
mals were not significant. Responses during the delay inter-
vals did not change significantly under 25 ppm CO. The 50
ppm CO exposure produced a slowing of response time after
Day 1 which persisted throughout the five-day period. This
effect was significant for the afternoon animal who had al-
ready been exposed four hr each day when behavioral testing
began. Responses during the delay interval increased signifi-
cantly only for the afternoon animal during exposure to 50
ppm CO.
EXPERIMENTS 1 AND 2
DISCUSSION
Discrimination tasks have been used for the study of a
number of psychoactive agents [6, 9, 10], and several rat
studies have indicated that discrimination behavior may be
of value for the study of certain central nervous system
(CNS) active compounds [6,9]. It would appear that the ap-
plication of discrimination behavior in a sub-human primate
as well as in the rat should be a valuable technique of the
relatively new field of behavioral toxicology. We have de-
scribed here a simple discrimination task for the evaluation
of toxicants in rats and a match-to-simple discrimination task
for the evaluation of toxicants in young baboons. Applica-
tion of these tasks to the study of effects of PBBs in rats, and
to the study of effects of inhaled ketone vapors in baboons, is
illustrated.
Rats treated with 1 mg/kg PBB were like controls with
respect to accuracy on the discrimination tasks, however,
extra responses and response latencies generally increased,
thereby reducing the animal's efficiency.
Similarly, MEK or MIBK administered chronically at half
the TLV over a seven-day period did not impair the baboons'
ability to discriminate but did alter response latencies and
extra responses during the delay intervals. The combinations
of MEK and MIBK produced less effect on response laten-
cies than did either one of the individual gases. Since the
animals were esposed to the single gases prior to being ex-
posed to the mixtures, it is possible that monooxygenases
were induced in liver or extra-hepatic tissues that affected
metabolism of the compounds on subsequent exposure, or
that simultaneous inhalation of one compound affected the
metabolism of the other compound. A loss of effect on re-
sponse latency which occurred on the sixth and seventh day
for two animals exposed to MEK alone might also be ac-
counted for in terms of enzyme induction which increased
metabolism of the inhaled gas.
The effects noted here with MEK and MIBK and the
lessening of effect with a combination of the two vapors is of
special interest. Both MEK and MIBK have been considered
to be non-neurotoxic whereas methyl n-butyl ketone
(MnBK), through the action of its metabolites, has been
found to produce peripheral neuropathy, which is poten-
tiated by simultaneous inhalation of MEK [1, 11, 12]. How-
ever, the potential CNS toxicity of MnBK has not been con-
sidered. The potentiation of the peripheral neurotoxicity of
MnBK by MEK has also been associated with MEK-induced
stimulation of microsomal enzyme activities: this effect is
also manifested as a decrease in hexobarbital-induced sleep
times [4]. However, these studies involved much higher
vapor concentrations and longer exposure times than em-
ployed in the present studies.
The behavioral effects produced by MEK or MIBK at
half the TLV concentrations indicate that the solvents are
acting on the central nervous system. If the ketones rather
than their metabolites are the neuroactive forms, enhanced
metabolism might account for the observed loss of central
nervous system effects in these studies. The consequences
of human exposure to sub-TLV concentrations of these
ketones should be evaluated.
Carbon monoxide increased extra responses during the
delay interval while its greatest effect was to slow reaction
time. Theodore et al. [13] reported a similar slowing of reac-
tion times in monkeys exposed to almost twice the TLV of
-------
TOXICANTS AND BEHAVIOR
13
CO (90 ppm). Our observation of minimal effects on the
discrimination task itself are not in agreement with those of
Beard and Wertheim [2] who reported disruption of an audi-
tory discrimination task in humans exposed to CO at 50 ppm.
The findings with CO as well as with the ketones suggest
that the match-to-sample discrimination task in the baboon
provides a sensitive indicator of CNS effects of pollutants
and should be evaluated further.
REFERENCES
1. Allen, N., J. R. Mendell, D. J. Billmaier, R. E. Fontaine and J.
O'Neill. Toxic polyneuropathy due to methyl n-butyl ketone.
Arch Neural. 32: 209-222, 1975.
2. Beard, R. R. and G. A. Wertheim. Behavioral impairment as-
sociated with small doses of carbon monoxide. Am. J. publ.
Hltli 57: 2012-2022, 1967.
3. Cotabish, H. N., P. W. McConnaughey and H. C. Messer. Mak-
ing known concentrations for instrument calibration. Am. ind.
Hyg. Ass. J. 22: 392-402, 1961.
4. Court, D., L. B. Hetland, M. S. Abdel-Rahman and H. Weiss.
The influence of inhaled ketone solvent vapors on hepatic mi-
crosomal biotransformation activities. Toxicol. appl. Pharmac.
41: 285-289, 1977.
5. Documentation of Threshold Limit Values: Am. Conf. Gov. ind.
Hyg. 3rd Edition, 1971.
6. Geller, I., R. J. Hartmann and K. Blum. Effects of nicotine,
nicotine monomethiodide, lobeline, chlordiazepoxide and caf-
feine on a discrimination task in laboratory rats. Psychophar-
macologia 20: 355-365, 1971.
7. Geller, I., R. L. Martinez, R. J. Hartmann and H. L. Kaplan.
Effects of ketones on a match-to-sample task in the baboon.
Proc. West. Pharmac. Soc. 21: 439-442, 1978.
8. Hanninen, H. Psychological test methods sensitivity to long-
term chemical exposure at work. Neurobehav. Toxic. (IN
PRINT).
9. Hughes, F. W. and E. Kopmann. Influence of pentobarbital,
hydroxyzine, chlorpromazine, reserpine and meprobamate on
choice discrimination behavior in the rat. Arch. Int. Pharmac.
Ther. CXXVI: 158-170, 1960.
10. Poling, A., M. A. Simmons and J. B. Appel. Morphine and
shock detection: effects on shock intensity. Commun.
Psychopharmac. 2: 333-336, 1978.
11. Saida, K., J. R. Mendell and H. S. Weiss. Peripheral nerve
changes induced by methyl n-butyl ketone and potentiation by
methyl ethyl ketone. J. Neuropathol. Exp. Neural. XXXV:
207-225, 1976.
12. Spencer, P. S. and H. H. Schaumburg. Feline nervous system
response to chronic intoxication with commercial grades of
methyl n-butyl ketone, methyl isobutyl ketone and methyl ethyl
ketone. Toxic, appl. Pharmac. 37: 301-311, 1976.
13. Theodore, J., R. D. O'Donnell and K. C. Back. Toxicological
evaluation of carbon monoxide in humans and other mammalian
species./, occup. Meet. 13(5): 242-255, 1971.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 15-22. ANKHO International Inc., 1979.
Effects of Toxicants on Visual Systems1
WILLIAM H. MERIGAN
University of Rochester Department of Radiation Biology and Biophysics
School of Medicine and Dentistry, Rochester, NY 14642
MERIGAN, W. H. Effects of toxicants on visual systems. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 15-22, 1979.—
The analysis of visual toxicity is complicated by the heterogeneity of visual capacities in different regions of the visual field.
Since various toxicants may impair different functions allied to localized portions of the visual field, it is important to
explore the relationship of field defects to residual visual abilities. We have begun this exploration by studying methylmer-
cury poisoning in macaque monkeys. Extended exposure to this toxicant produces a marked concentric constriction of
visual fields, a result similar to that found in human victims. In addition, visual sensitivity is greatly reduced on those tests
in which the periphery of the visual field is more sensitive than the center. Our findings suggest simple but reliable clinical
tests for screening suspected victims of substances impairing peripheral vision.
Visual fields Visual thresholds Methylmercury Neurotoxicity
VISION, which in humans may be the most important of the
exteroceptive senses, is extremely sensitive to toxic insult.
Many toxicants, including methanol, carbon disulfide, some
arsenicals, methylmercury and pesticides, act directly on the
visual system. Numerous other toxic substances, such as
lead and carbon monoxide may cause visual impairment as a
result of less specific damage to the central nervous system.
While visual toxicity can be assessed with the techniques
of physiology and pathology, the testing of visual capacities
remains the most direct way to detect impaired vision. Vis-
ual testing is superior to other techniques in that it is non-
invasive, it does not provoke great anxiety in patients, and it
is suitable for field testing of industrial workers or suspected
victims of environmental pollution.
Unfortunately, the measurement of vision is complicated
by the inhomogeneity of visual capacities in different por-
tions of the visual field. For example, the portion of the
visual field located at the center of gaze, the fovea, has con-
siderably greater visual acuity than other regions of the vi-
sual field (Fig. 1). Thus, visual impairment confined to the
foveal region would cause a much more severe loss of acuity
than damage in extrafoveal regions.
These considerations are of particular importance in vis-
ual toxicity because of the marked specificity of many tox-
icants for certain portions of the visual field. Figure 2 shows
examples of substances which can cause localized damage of
the visual field. The polar plots represent the central 30° of
the visual field of each eye with the origin located at the
fovea. The optic disc or blind spot, which is not shown, is
located between about 13 and 17° eccentricity on the tem-
poral (outer) side of each visual field. The blackened area of
the plots represents a reduction of visual sensitivity
(scotoma) but not necessarily a complete loss of vision.
Consumption of methanol (methyl alcohol) produces a
marked edema of the optic disc and typically results in a
central or circumcentral loss of vision [9,17]. Prolonged mal-
nutrition (Fig. 2) frequently causes optic atrophy and small
bilateral central scotomas [8]. Harrington [8] has pointed out
that the common finding of central field defects in the so-
called "tobacco and alcohol amblyopias" may also be due to
a Vitamin B12 deficiency in some alcoholics and heavy
smokers. Lead, carbon disulfide, thallium, and several drugs
can also induce scotomas in the center of the visual field [7,
8, 18].
A loss of peripheral vision is characteristic of exposure to
other substances [7,8]. Figure 2 shows an almost complete
loss of vision following idiopathic quinine poisoning but with
some residual central vision. Salicylate poisoning is quite
rare but results in a similar field constriction. Peripheral field
loss can also be caused by chloroquine, pentavalent arseni-
cals, carbon monoxide and some drugs. Both central and
peripheral impairment of the visual field may follow expo-
sure to nitrobenzol, dinitrobenzene and methylbromide. An
example of such a loss is shown in Fig. 2 for nitrobenzol.
The locus of pathological changes in the disorders de-
scribed above is most frequently the retina or optic nerve.
Central visual defects could be due to selective toxicity of
cone photoreceptors which predominate near the fovea, but
the usual basis is damage to the papillo-macular bundle of
nerve fibers which serves the foveal region [8]. Peripheral
losses most often result from damage to ganglion cells or the
pigment epithelium [7]. However, since the visual field re-
mains topographically segregated throughout the more cen-
tral portions of the visual system (e.g., in the lateral genicu-
late, striate cortex and pre-striate areas) visual field defects
could be produced by lesions in any of these structures. The
'Supported in part by Grants ES-01885 and ES-01248 from the National Institute of Environmental Health Sciences, and by a contract with
the U.S. Department of Energy at the University of Rochester (Report No. UR-3490-1602).
15
-------
16
MERIGAN
w
Q
T
K
R H0
B
P
M
B
C
Y
FIG. 1. A chart which demonstrates the rate of decrease in acuity with distance from the fovea or center of gaze. When the center of the
chart is fixated each of the letters is equally discriminable (from Anstis [1]).
clearest example of such a central toxic effect is the loss of
peripheral vision in methylmercury poisoning [11]. Methyl-
mercury causes a progressive cell loss in deep fissures of the
cerebral cortex, but this is most marked in the anterior por-
tion of the calcarine fissure, the region which receives the
projection from the periphery of the visual field [21,22]. A
similar pattern of pathology is seen in macaque monkeys [6].
Figure 3 shows a moderate degree of peripheral loss in a
patient exposed to methylmercury [14]. The dashed line rep-
resents normal field boundaries, and the solid line, those of
the patient. Visual field constriction in methylmercury
poisoning is progressive, permanent, and is one of the
clearest diagonistic indices of intoxication [14].
The frequent occurrence of restricted field defects in
toxic neuropathy makes the assessment of residual vision
difficult. It is clear that visual loss will depend both on the
-------
TOXICANTS AND VISUAL SYSTEM
17
L.E.
R.E.
L.E.
R.E.
METHfiNOL
VITRNIN DEFICIENCY
L.E.
R.E.
L.E.
R.E.
QUININE
NITROBENZOL
FIG. 2. Visual field plots showing field defects typical of toxic syndromes. Each polar plot represents the central 30 ° of
vision for the right (R.E.) and left (L.E.) eyes. Blackened areas show regions of reduced visual sensitivity. (Adapted from
Harrington [8]).
LEFT EYE
120 105
240 255
RIGHT EYE
1Q5 90 75 60
270 285 300
FIG. 3. Visual field charts showing a moderate concentric constriction of visual fields in methylmercury poisoning [14]. The
dashed lines indicate normal visual fields and the solid line the field boundaries of a victim of methylmercury.
-------
18
MERIGAN
NORMRL VISURL FIELD
10CM
LJ
>
-------
TOXICANTS AND VISUAL SYSTEM
19
LEFT EYE
RIGHT EYE
I* "
* 605
CONTROL
LEFT EYE
RIGHT EYE
« 82
ME HG
FIG. 6. Visual field boundaries of a control and a methylmercury poisoned monkey.
3 HZ 10 HZ 30 HZ
2LH
L-
0-
50X
LJ 2LH
U
z
CE _
o-
WWWWWVA;
17X
2L-
L-
0-
o's
TINE (SEC)
0-5
TIME (SEC!
sx
0.5
TIME (SEC)
FIG. 7. Representation of the temporal luminance profile of representative flickering test stimuli. Each plot shows that the
target luminance was modulated in time around the mean luminance (L). Three modulation depths are shown for each of
three frequencies of flicker
-------
20
MERIGAN
FIG. 8. Technique used to measure flicker sensitivity of monkeys. The monkey is
responding correctly on the left pushbutton which corresponds to the location of the
flickering display. Juice reward is delivered through a stainless steel drinking spout.
a monkey which had received sufficient methylmercury to
produce a moderate constriction of visual fields. We were
especially interested in those measures of flicker sensitivity
which are most dependent on the periphery of the visual
field. The methodological details are more fully described
elsewhere [16].
Animals
Results will be reported for two macaque monkeys, one
of which served as a control (No. 605) while the other (No.
82) was treated with methylmercury. The exposure history
of Monkey 82 has been reported [16]. Monkey 82 showed a
marked myopia but was corrected during testing with
corneal contact lenses.
Peri me try
The technique used for visual field measures is shown
schematically in Fig. 5. The monkey sat in an acrylic test
chair facing the circular aperture which was used to deter-
mine the limits of the visual field. The visual subtense of this
aperture could be changed in increments of 10° from 10 to
70° by moving the restraining chair or by inserting panels
with smaller openings. The tester attracted the monkey's
gaze by placing a small piece of fruit or a marshmallow at the
fixation point. When the monkey fixated this central target, a
1 cm white marshmallow was moved inside the aperture at
one of the eight test points located at 45° intervals. If the
monkey glanced toward the test marshmallow, this detection
was marked at the appropriate location on the visual field
chart. A range of eccentricities was tested at each location to
determine the limits of the field. All testing was monocular
and the non-tested eye was occluded with an opaque contact
lens. To prevent habituation, the pieces of food used as fixa-
tion and test stimuli were frequently extended to allow the
monkey to eat them.
The visual field plots of the control (No. 605) and
poisoned (No. 82) monkeys are shown in Fig. 6. The results
for Monkey 605 are typical of normal monkeys tested in our
laboratory. The visual field extends to approximately 50° in
the nasal and superior directions and to at least 65° in the
temporal direction. Where the field extended beyond the 70°
limit of our perimeter, it was marked as 75°. Targets pre-
sented at more than 50° eccentricity in the inferior direction
were partially blocked by the acrylic chair.
The visual fields of Monkey 82 show a marked constric-
tion compared to those of the control monkey. The two plots
for the left eye of this monkey represent determinations
separated by several days. All measures indicated a concen-
tric constriction of visual fields to about 35 to 40° of eccen-
tricity. Such a pronounced field constriction should be corre-
lated with a decreased sensitivity to stimuli which depend
primarily on the periphery of the visual field for detection.
Thus, we tested Monkey 82 and the control monkey to explore
the feasibility of using such stimuli to detect visual loss in
methylmercury poisoning.
Flicker Sensitivity
The measures of flicker sensitivity will be described by
analogy to the similar but more familiar quantification of
audition by the audiogram. The stimuli we use are shown
schematically in Fig. 7. While the audiologist uses pure
(sinusoidal) tones, we use sinusoidal flicker of a luminous
target. The audiogram measures audibility of tones of 50 to
10,000 Hz while we test the visibility of flicker rates of 2 to
about 70 Hz. Figure 7 illustrates flicker test stimuli of 3, 10,
and 30 Hz. The measure of auditory sensitivity is threshold
loudness (sound pressure amplitude) while flicker sensitivity
threshold is expressed as relative luminance modulation
depth. Modulation depth is defined as LMAX - LMIN
LMAX + LMIN
where LMAX is the maximum and LMIN the minimum
luminance of the flickering stimulus. Flicker modulation
depths of 50, 17, and 5% are shown in Fig. 7. A final param-
eter of our measures is the mean or adapting luminance of
our stimuli (L in Fig. 7), since light or dark adaptation
changes visual sensitivity much as adapting sound level
changes auditory sensitivity. Flicker sensitivity was exam-
-------
TOXICANTS AND VISUAL SYSTEM
21
10-
HIGH LUMINflNCE
DO
Q
CO
CO
0
10
100
0:
IDE
20-
nEDIlin LUNINflNCE
CO
z
LJ
CO
r~n—
10
100
od
10 =
20-
LOUI LUniNPNCE
r — i — i i i i |
10
FLICKER FREQ (HZ
T
100
FIG. 9. Flicker visuogram of the methylmercury poisoned monkey.
The dashed line represents the modulation sensitivity of the control
monkey. Distance below the dashed line indicates the degree of
sensitivity loss. High luminance=5 ft L, medium luminance=0.05 ft
L, low luminance=0.0005 ft L.
ined at 3 luminances: one at which cone photoreceptors are
dominant (high luminance), one at which only rod photo-
receptors are active (low luminance), and finally at an inter-
mediate level (medium luminance) which is slightly above
the point of transition from rod to cone vision. Detection of
the low luminance stimuli is strongly dependent on the pe-
riphery of the visual field [10]. In addition, at the medium
and high luminances used in this study, the detection of high
rates of flicker depends on peripheral vision [12] (see Fig. 4
above).
The actual arrangement of the test apparatus is illustrated
in Fig. 8. The seated monkey faced two display oscilloscopes
at a distance of 1 m. The face of each display was evenly
illuminated and subtended 4.6x5.7° of visual angle. On each
trial the entire illuminated portion of one of the two displays
was flickered. The monkey indicated the flickering display
by pressing the corresponding pushbutton on the test panel.
Correct choices were rewarded with a small amount of fruit
juice.
The frequency and modulation depth of flicker were set
remotely by the on-line computer which controlled the ex-
periments. The mean luminance of the stimuli was varied by
placing neutral density filters over the face of each display.
In each session, sensitivity to a single flicker frequency at
one luminance level was tested. Modulation thresholds were
measured by presenting flicker at 5 modulation depths, cho-
sen to bracket the threshold value, in random order. The
threshold was defined by interpolation on the resulting
psychometric function as the modulation depth at which per-
formance fell to 75% correct.
The results in Fig. 9 show the sensitivity loss of the
poisoned monkey relative to the control monkey over the
range of flicker frequencies tested. The most striking finding
at high and medium luminance was a great loss of sensitivity
to high frequencies of flicker. The maximum resolvable
flicker frequency (CFF) at high luminance was only 45 Hz in
the poisoned monkey compared to 70 Hz in the control. The
role of visual field constriction in this reduced sensitivity to
high frequency flicker can easily be seen by comparing Figs.
4 and 6. Figure 6 shows that the visual fields of Monkey 82
reach only to 35 or 40° of eccentricity. In addition, detailed
studies of human methylmercury victims [14] indicate that
visual impairment extends well inside the limits of the visual
field. Thus, Monkey 82 has suffered a severe impairment of
the very regions of the visual field which can resolve the
highest rate of flicker (see Fig. 4).
The slight loss of sensitivity to low rates of flicker shown
in the upper part of Fig. 9 is somewhat more puzzling be-
cause the visibility of such stimuli is not thought to depend
on peripheral vision. This loss could be due to reduced
acuity, since the perception of low flicker frequencies is
greatly enhanced by the presence of sharply focussed edges
[13]. However, since Monkey 82 shows only a slight loss of
visual acuity (20% below the mean of normal monkeys we
have tested) it is unlikely that the low frequency flicker loss
is due to poor pattern vision. Further studies will be required
to determine if low frequency impairment is a consistent sign
of methylmercury intoxication.
The bottom panel in Fig. 9 shows a profound loss of sen-
sitivity for all frequencies of low luminance flicker. This re-
sult is consistent with the observation [10,15] that the pe-
riphery is more responsive to flicker at extremely low lumi-
nances than is the center of the visual field. Thus, low lumi-
nance as well as high frequency flicker sensitivity is greatly
reduced in methylmercury poisoning and this is correlated
with a constriction of visual fields.
Stimuli which are detected most easily with central vision
should prove poor indicators of methylmercury induced vi-
sual loss. Indeed, methylmercury victims in Iraq showed no
impairment of visual acuity [19]. A particularly dramatic
illustration of residual central vision is the report of Iwata
[14] that CFF was unimpaired in a large population of
methylmercury victims. At first glance this result appears
inconsistent with our findings. However, the results of
Hylkema [12] show that this apparent inconsistency is due to
the size of the test stimulus. For small flickering stimuli such
as those used by Iwata, CFF is maximal at the fovea: thus,
there is no impairment of methylmercury poisoning. How-
-------
22
MERIGAN
ever, for large targets, such as those used to test Monkey 82,
CFF increases with eccentricity (Fig. 4). Thus, CFF could be
used for detection of either central or peripheral field loss
depending on the size of the target.
The present findings as well as those of previous investi-
gations can be used to devise clinical tests for visual loss in
methylmercury poisoning. Measurements of the boundaries
or the sensitivity profiles across the visual field provide
detailed descriptions of early visual loss [14].
The major drawback of these techniques is that they require
a cooperative patient and a skilled examiner. The measure-
ment of low luminance vision [3,4] gives an early index of
intoxication which is applicable to studies of non-human
primates. Iwata [14] has achieved even greater sensitivity in
the detection of toxicity by blocking the central 40° of the
visual field and measuring low luminance vision with evoked
responses. Unfortunately, these techniques require a lengthy
dark adaptation period and a subject who is cooperative in
the dark. On the other hand, the data from Monkey 82
suggest that the measurement of CFF with a large flickering
target may provide the most convenient early index of
methylmercury intoxication. This test could be used for the
rapid screening of populations potentially intoxicated with
methylmercury. Studies are underway in our laboratory to
determine if this simple measure is as early and reliable an
index of poisoning as the more difficult examinations de-
scribed above.
REFERENCES
1. Anstis, S. M. A chart demonstrating variations in acuity with
retinal position. Vision Res. 14: 589-592, 1974.
2. Aulhorn, E. and H. Harms. Visual perimetry. In: Handbook of
Sensory Physiology, Vol. VII (4) Visual psychophyslcs, edited
by D. Jameson and L. Hurvich. Berlin: Springer, 1973.
3. Berlin, M., C. Grant, J. Hellberg, J. Hellstrom and A. Shutz.
Neurotoxicity of methylmercury in squirrel monkeys. Arc/is
envir. tilth. 30: 340-348, 1975.
4. Evans, H. L. Early methylmercury signs revealed in visual
tests. In: Proc. Intern. Conf. on Heavy Metals in the Environ-
ment, Vol. 3, edited by T. C. Hutchinson. Toronto: University
of Toronto, 1977, pp. 241-256.
5. Fick, A. E. Ueber Stabchensehscharfe und Zapfensehscharfe.
Albrecht von Grafe's Archiv fiir Ophthalmologie 45: 336-356,
1898.
6. Garman, R., B. Weiss and H. Evans. Alkylmercurial enceph-
alopathy in the monkey (Saimiri sciureus and Macaco arc-
toides). Acta Neuropath. 32: 61-74, 1975.
7. Grant, W. M. Toxicology of the Eye. Springfield, 111.: Charles C. .
Thomas, 1974.
8. Harrington, D. O. The Visual Fields. St. Louis: C. V. Mosby
Company, 1976.
9. Hayreh, M. S., S. S. Hayreh, G. Baumbach, P. Cancilla, G.
Martin-Amat, T. R. Tephly, K. E. McMartin and A. B. Makar.
Methanol poisoning: III. Ocular toxicity. Archs Ophthalmol. 95:
1851-1858, 1977.
10. Hecht, S. and C. D. Verrijp. Intermittent stimulation by light.
III. The relation between intensity and critical fusion frequency
for different retinal locations. J. gen. Pliysiol. 17: 251-265,
1933a.
11. Hunter, D., R. R. Bomford and D. S. Russell. Poisoning by
methylmercury compounds. Q. J. Med. 9: 193-213, 1940.
12. Hylkema, B. S. Examination of the visual field by determining
the fusion frequency. Acta opthal. (Kbh.) 20: 181-193, 1942.
13. Kelly, D. H. Flickering patterns and lateral inhibition. J. Opti-
cal Soc. Am. 59: 1361-1370, 1969.
14. Iwata, K. Neuroophthalmological findings and their transition
in "Minimata Disease" organic mercury poisoning in Agano
area of Niigata perfecture. Acta soc. ophthal. japonica. 77:
1788-1834, 1974.
15. Lloyd, V. V. A comparison of critical fusion frequency for
different areas in the fovea and the periphery. Am. J. Psychol.
65: 346-357, 1957.
16. Merigan, W. H. Visual fields and flicker thresholds in methyl-
mercury poisoned monkeys. In: Neurotoxicity of the Visual
System, edited by W. H. Merigan and B. Weiss. New York:
Raven Press, in press.
17. Potts, A. M. The visual toxicity of methanol: VI. The clinical
aspects of experimental methanol poisoning treated with base.
Am. J. Opthalmol. 39: 86-92, 1955.
18. Potts, A. M. and L. M. Gonasun. Toxicology of the eye. In:
Toxicology: The Basic Science of Poisons, edited by J. Casarett
and J. Doull. New York: McMillan, 1975.
19. Sabelaish, S. and G. Hilmi. Ocular manifestations of mercury
poisoning. Bull WHO Suppl. 53: 83-86, 1976.
20. Sperling, H. G. and Y. Hsia. Some comparisons among spectral
sensitivity data obtained in different retinal locations and with
two sizes of foveal stimulus./. Opt. Soc. Am. 47: 707-713, 1957.
21. Takeuchi, T. Neuropathology of Minimata disease in
Kumamoto: especially at the chronic stage. In: Neurotoxicol-
ogy, edited by L. Roizin, H. Shiraki and N. Grcevic. New York:
Raven Press, 1977.
22. Takeuchi, T., N. Morikawa, H. Matsumoto and Y. Shiraishi. A
pathological study of Minimata disease in Japan. Acta
Neuropathol. (Berl) 2: 40-57, 1962.
23. Tynan, P. D. Motion processing in peripheral vision: reaction
time, contrast threshold and perceived velocity. Unpublished
dissertation, Northwestern University, 1979.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 23-31. ANKHO International Inc., 1979.
Effects of Toxicants on the Somatosensory
System
JACQUES P. J. MAURISSEN
Department of Radiation Biology and Biophysics, University of Rochester, Rochester, NY 14642
MAURISSEN, J. P. J. Effects of toxicants on the somatosensory system. NEUROBEHAV. TOXICOL. 1: Suppl 1, 23-31,
1979.—A computerized system for an objective and accurate study of vibration sensitivity has been designed.
Sensitivity can be assessed in human as well as nonhuman primates. Its usefulness in the study of peripheral nerve
disorders induced by chemical exposure is emphasized.
Vibration sensitivity Sensory disturbances Neurotoxicants
NUMEROUS chemicals (drugs and toxic chemicals) are
known to produce peripheral sensory neuropathy. Mor-
phological, electrophysiological and clinical studies have at-
tempted to define the exact nature of these neurotoxic ef-
fects. Morphological studies have investigated the anatomi-
cal structure and ultrastructure of peripheral nerves. Elec-
trophysiological studies have emphasized the importance of
a functional approach to these disorders. However, elec-
trophysiological measures (nerve conduction velocity,
somatosensory evoked potentials) do not always reflect
neurological status. Clinical studies have aimed at detecting
the presence of frank disorders but do not have the refine-
ments necessary to detect early signs of peripheral nerve
damage, or to follow sensory recovery after exposure to a
neurotoxicant.
Still another approach to these problems can be taken.
Psychophysics is the scientific and objective study of the
relations between stimuli and resulting sensations. Operant
and Pavlovian conditioning offer the possibility of studying
sensory processes in animals psychophysically.
This paper will briefly review the methods used to study
the effects of neurotoxic chemicals on somesthetic sensitiv-
ity and will describe a technique for the study of vibration
sensitivity in human and nonhuman primates.
METHODS FOR STUDY OF SOMESTHETIC SENSITIVITY
Clinical examination is essentially the sole source of in-
formation about the effects of chemicals on sensory func-
tions. However, uncontrolled factors make it a crude indi-
cator of sensory status.
Scope and Limitations of Neurological Examination
Sensory testing is an important part of the neurological
examination. Two-point discrimination is generally tested
with a pair of calipers, pain sensitivity with the "neurologic
pin," thermal sensitivity with tubes containing hot or ice-
cold water, light touch with a cotton wisp, and vibration
sensitivity with a tuning fork. These techniques are designed
to detect frank sensory abnormalities. They are suitable
neither for detecting a partial sensory loss, nor for following
sensory recovery after peripheral nerve insult. The stimuli
used are also hardly reproducible. There is no standard way
to apply them and in no case is the velocity of impact on the
skin controlled. Furthermore, evaluation of the subject's
performance is highly subjective. For example, the examiner
strikes a tuning fork with an unknown and irreproducible
force, applies its base on a bony eminence with an unknown
pressure, and asks the subject to report when vibration
ceases to be felt. At that time, the tuning fork typically is
applied to a similar part of the examiner's body. If the exam-
iner still perceives vibration, sensitivity of the subject is di-
agnosed as impaired. Even if the sensitivity of the examiner
stays constant over time, it is not possible to compare data
obtained by different examiners.
Necessity for Investigative Techniques
Several pieces of equipment have been designed for a
more quantitative and objective evaluation of somesthetic
sensitivity. Unfortunately they are not being introduced in
clinical practice very eagerly. Temperature sensitivity can be
assessed with simple, inexpensive and commercially avail-
able thermal stimulators [33] or with more complex ther-
moelectric units whose temperature can be very accurately
controlled [39, 53, 64]. Electrical sensitivity has also been
used and the parameters influencing it have been thoroughly
studied by psychologists in the last decades [50,128]. Precise
control of stimulus intensity can be achieved easily, and
absolute thresholds of electrical stimulation determined.
Electricity has also been used in the study of pain sensitivity
[93]. For the quantitative study of light touch sensitivity, an
apparatus that overcomes the difficulties encountered with
Von Frey hairs has been devised [34], It consists of a probe
mounted on a motor driven by a function generator. A sys-
tem for the quantification of vibration sensitivity will be de-
scribed later and the advantages offered by vibratory stimula-
tion will be discussed.
23
-------
24
MAURISSEN
TABLE 1
Drugs
Vincristine (Oncovin)
Vinblastine (Velban)
Vindesine (NSC-245467)
CYs-diamminedichloro-
platinum (Platinal)
Maytansine (NSC-1 53858)
Methylhydrazine
(Procarbazine)
Podophyllum resin
(Podophyllin)
Misonidazole
(Ro-07-0582)
Metronidazole (Flagyl)
Nitrofurantoin
(Furadantin)
Nitrofurazone
(Furacin)
Furaltadone (Altafur)
Thiophenicol
(Thiomycetin)
Chloramphenicol
(Chloromycetin)
lodochlorhydroxyquin
(Entero-Vioform)
Demeclocycline
(Declomycin)
Doxycycline (Vibramycin)
Colistimethate (Colistin)
Dapsone (Avlosulfon)
Ammoniated mercury
ointment
Methylmercury thioacetamide
Gold thiomalate
Therapeutic Use
antitumor
antitumor
antitumor
antitumor
antitumor
antitumor
antitumor
hypoxic cell
radiosensitizer
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antimicrobial
antiseptic
fungicidal
antarthritic
Sensory Signs and Symptoms
impaired touch, pinprick, vibration sense,
paresthesia, numbness, tingling
paresthesia
paresthesia
impaired vibration, touch, pinprick sense,
paresthesia, numbness, tingling
paresthesia
paresthesia, sensorium changes
paresthesia, generalized hypesthesia
paresthesia, reduced pinprick, light touch,
vibration and temperature sensitivity, coldness
paresthesia, hyperesthesia, decreased touch,
pinprick and temperature sense
numbness, paresthesia, tingling, dysesthesia,
sensory loss (pinprick, pain vibration,
temperature, light touch)
numbness, paresthesia, impaired touch, pain and
vibration sense
loss of all sensory modalities
tingling, burning, tactile hypesthesia, dysesthesia,
decreased vibration sense
numbness, tingling, burning, decreased touch,
vibration sense
numbness, tingling, dysesthesia, impaired touch,
temperature, pain and vibration sense
paresthesia, tingling, burning sensation
paresthesia, tingling
paresthesia
numbness, decreased light touch, pain, two-point
discrimination, vibration sense
decreased pain and vibration sense, numbness
numbness, sensory disturbance
tingling, pricking, paresthesia, decreased touch,
References
13, 15, 27, 85, 88, 90,
103, 105, 108, 110, 136
115
138
10, 67
9, 95
109
18, 24
30, 130
28, 101, 107, 129
35, 52, 63, 80, 83
106, 124
26
26, 61
114
20, 65
117, 120, 126
37
38
69
36, 70
123
96
32, 78
(Myochrysin)
Phenytoin (Dilantin)
Nitrous oxide
Perhexiline maleate
(Pexid)
Disulfiram (Antabuse)
Isoniazid (Niconyl)
Ethionamide (Trecator)
Hydralazine (Apresoline)
Thalidomide (Distaval)
Furosemide (Lasix)
temperature, pain and vibration sense, hyperesthesia,
burning, pins and needles sensations
antiepileptic dysesthesia, numbness, tingling, decreased pain
and vibration sense
anesthetic numbness, paresthesia, dysesthesia, tingling,
impaired touch and vibration
antianginal dysesthesia, paresthesia, impaired deep sensitivity,
decreased touch, pain, temperature and vibration
sense
antialcohol tingling, numbness, dysesthesia, cold or burning
sensation, impaired pinprick, touch, vibration and
temperature sense
antituberculous tingling, numbness, pins and needles, electric
shock sensation, burning pain, dysesthesia, impaired
vibration, temperature, touch and two-point
discrimination, hyperalgesia
antituberculous numbness, impaired vibration sense, dysesthesia,
burning sensation, hyperalgesia, impaired
vibration sense
antihypertensive pins and needles, numbness
sedative paresthesia, numbness, hyperalgesia,
impaired light touch, vibration sense
diuretic burning paresthesia
31, 81
73, 74
2, 11, 75, 79
6, 12, 21, 44, 48, 89
8, 42, 94
100
102
40, 41
84
-------
NEUROTOXICANTS AND SOMATOSENSORY SYSTEM
25
TABLE 2
Chemicals
Signs and Symptoms
References
Acrylamide
Triorthocresyl phosphate
n-Hexane
Methyl n-butyl ketone
Carbon disulfide
Arsenic
Mercury
Thallium
Lead
Cyanide
Chlorobiphenyl
Methyl bromide
decreased vibration, temperature, touch
and pinprick, numbness, paresthesia,
coldness, tingling, pins and needles
paresthesia, numbness, formication, dysesthesia,
aching pain, impaired touch, temperature, pain
and vibration sense, coldness
paresthesia, numbness, burning sensation or
coldness, tingling, dysesthesia, decreased
pinprick, touch, temperature and vibration sense
numbness, impaired pain, light touch,
temperature and vibration sense
numbness, paresthesia, tingling, decreased
touch, temperature and pain sensitivity
numbness, tingling, burning sensation,
paresthesia, dysesthesia, decreased touch,
temperature and vibration sense, formication,
hyperalgesia, hyperesthesia
paresthesia, numbness, tingling, sensation
of needle-pricking, superficial and deep
sensory disturbance (touch, pain, two-point
discrimination, vibration)
paresthesia, dysesthesia, tingling, burning
pain, numbness, hyperalgesia, impaired light
touch, pinprick and vibration sense,
hyperesthesia
paresthesia
paresthesia, numbness, pins and needles,
burning, aching pain, feeling of heat and cold,
decreased two-point discrimination, pinprick,
light touch, temperature and vibration sense
numbness, decreased temperature,
pain and touch sensitivity
numbness, decreased vibration sense, impaired
superficial sensation in all modalities
17, 45, 49
7, 14, 46, 87,
116, 122, 134
23, 47, 56, 71,
99, 104, 111, 125
86, 137
68, 131
22, 43, 55, 57,
77, 112
1, 3, 54, 59,
60, 76, 82, 97,
127, 135
4, 16, 51, 62
29, 113
98
92
66
CHEMICALS AS ETIOLOGIC AGENTS IN SENSORY DISORDERS
A wide variety of drugs and toxic chemicals of unrelated
structures and modes of action can induce sensory disorders
of peripheral or central origin. Drugs with unrelated
therapeutic actions have been implicated in the development
of peripheral neuropathy mainly characterized by sensory
disorders. The main sensory signs and symptoms due to
selected drugs are shown in Table 1.
Several groups of anticancer drugs have neurotoxic
peripheral side effects which may be severe enough to limit
therapeutic treatment. Of the vinca alkaloids, this is par-
ticularly the case with vincristine. Hypoxic cell sensitizers,
such as misonidazole and metronidazole, which are also clin-
ical antiflagellates, are used as adjuvants to radiotherapy and
also produce neurotoxicity.
Derivatives of the furan family, which possess antibacte-
rial activity, are also radiosensitizers of hypoxic mammalian
cells [19]. Nitrofurantoin neurotoxicity is a problem espe-
cially in patients with severe impairment of renal function, or
when the drug is administered in high doses for prolonged
periods of time. Various antibiotics of the chloramphenicpl,
oxyquinoline, tetracycline and polymyxin families can in-
duce neurotoxic peripheral side effects in a small susceptible
fraction of patients, as is also the case with some drugs of the
hydantoin and sulfone families. Polyneuropathy with early
sensory complaints is seen in subjects exposed to nitrous
oxide and is also a known entity in chronic alcoholism.
However, there is evidence that hypovitaminosis, rather
than alcohol per se, is largely responsible for this effect.
Control of alcoholism with disulfiram has been associated
with peripheral neuropathy and optic neuritis although their
occurrence has not been widely recognized. More common
are the sensory side effects associated with antituberculous
therapy, especially with isoniazid. There have also been re-
ports of iatrogenic neuropathy induced by some antianginal
agents, fungicides, antiseptics, antarthritics, sedatives and
diuretics, which seem to produce fleeting peripheral pares-
thesias and other sensory changes.
Besides drugs, many industrial and environmental toxic
chemicals have contributed significantly to clinical
peripheral nerve dysfunction (Table 2). Here, too, chemicals
of completely unrelated structures and modes of action can
present similar neurological profiles. In acrylamide toxicity,
sensory signs and symptoms precede and predominate over
motor disturbances. Triorthocresyl phosphate, which is used
as a plasticizer, has been the origin of several repeated out-
breaks in Morocco, where thousands of people were affected
by adulterated cooking oil. India (with contaminated mus-
tard oil) and the United States (with Jamaica ginger) have not
been spared. More or less important sensory involvement
-------
MAURISSEN
FIG. I. Human subject and monkey during a testing session.
has characterized these outbreaks. Several organic solvents
and heavy metals share in common adverse effects on the
somatosensory system. Such effects are protean and range
from subjective numbness and tingling to objective cutane-
ous sensory loss. Other industrial chemicals have similar
effects. Dietary cyanide has instigated outbreaks of sensory
ataxia mainly in tropical and subtropical regions, where the
staple food is cassava, very rich in cyanogenetic glycosides.
VIBRATION SENSITIVITY
Anatomical ami Functional Substrates
Vibration sensitivity is mediated by a duplex mechanism.
Psychophysical evidence for this claim is based on the shape
of the function relating vibration frequency to the just de-
tectable vibration amplitude (absolute threshold). Below
about 50 Hz, threshold is independent of frequency. Above
this value, there is a frequency dependent decrease in
threshold with an optimal sensitivity around 250 Hz. Then.
the amplitude-frequency detection curve rises. It was con-
cluded, from this function, that temporal summation oc-
curs above 50 Hz, and is absent below this point. In-
creasing the size of the vibrating contactor results in a
lower threshold at higher frequencies while it does not affect
thresholds in the low frequency range. This is evidence for
a spatial summation selective for higher frequencies
[132,133].
Electrophysiological data also support the idea that vi-
bration sensitivity is mediated by a pluralistic mechanism
and depends on at least two sets of thick myelinated nerve
fibers and receptors. One set is thought to end in Pacinian
corpuscles and has its tuning points and absolute thresholds
of afferent discharge at frequencies higher than 50 Hz. An-
other set has a maximal sensitivity in the low frequency
range. Its anatomical substrate has not been convincingly
identified yet, although Meissner's corpuscles have been
proposed [911.
Correlation with Subjective Symptoms
With most of the compounds cited in Tables 1 and 2,
subjective sensory symptoms, such as distal paresthesia,
numbness and tingling, are often early manifestations of
neurotoxicity and are followed by objective sensory dys-
function. No systematic study has been done to correlate
subjective symptoms with objective signs. However, in sev-
eral clinical case reports, where vibration was measured with
quantitative and objective methods, it was mentioned that
numbness occurred concurrently with vibration sensitivity
impairment [25,72]. Paresthesias tended to confuse the re-
sults of the examination of vibratory sensitivity [5,121].
We do not believe, however, that every subjective com-
plaint must necessarily be accompanied by vibration sen-
sitivity impairment. Some chemicals can be more or less
selective for some types of nerve fibers. Thalidomide, for
example, preferentially affects large diameter nerve fibers
-------
NEUROTOXICANTS AND SOMATOSENSORY SYSTEM
27
PROBE TIP,
HEIGHT
ADJUSTMENT
FIG. 2. Schematic drawing of the vibration assembly.
[41]. This is also the case for acrylamide [58], which also
inactivates mechanoreceptive properties of Pacinian cor-
puscles before any alteration of ultrastructural features of
the nerve fiber or the sensory terminal [118,119]. Disulfiram
damages myelinated fibers and spares unmyelinated fibers
[89]. Vincristine causes loss of both small and large diameter
fibers [85]. A drug that would selectively affect unmyelinated
fibers or thin myelinated fibers would certainly have a mini-
mal effect on vibration sensitivity, which is mainly carried
through thick myelinated fibers. Pain and temperature sen-
sitivity, on the other hand, might be more significantly im-
paired.
Methods and Preliminary Data
Vibration sensitivity can be determined objectively and
accurately in human and nonhuman primates with essentially
the same methods and equipment. The monkey sits in a re-
straining chair in front of a small table (Fig. 1). The left hand
is immobilized in a plasticene mold. The tip of the middle
finger is placed on the vibrating probe, which protrudes
through a hole (Fig. 2). The position of the vibrator can be
adjusted with a precision gear apparatus so as to indent the
skin by a constant depth when the probe makes electrical
contact with the finger. The right hand is free and has access
to a telegraph key. A spout next to the monkey's mouth
delivers fruit juice when an electromagnetic valve is ac-
tivated. Two loudspeakers are located at the top of the re-
straining chair. The whole system is enclosed in a sound-
attenuated double-walled chamber. A white noise generator
is turned on during the testing session.
PDP-12
D/fl CONVERTER
(FREQUENCY)
D/fl CONVERTER
( flflPL I TUDE )
FUNCTION
GENERflTOR
flMPLIFIER
VIBRflTOR
flCCELER-
OnETER
CHRRGE
flnPLIFIER
COUNT
METER
DIGITflL
VOLTMETER
FIG. 3. Block diagram of arrangement for stimulus delivery and
measurement.
Similarly, the human subject sits on a chair facing a table.
Instruction is given to keep the left hand in a relaxed position
with the tip of the middle finger on the vibrating contactor.
The subject wears headphones during the session. The right
hand has access to a telegraph key. Performance feedback is
given through an intelligence panel. Verbal instructions are
given to the subject about the task requested.
Sinusoidal vibrations are generated through the system
illustrated in Fig. 3. The whole experiment, including
stimulus presentation and data collection, is under complete
control of an on-line PDP-12 computer (Digital Equipment
Corporation). Frequency and amplitudes are independently
controlled with a function generator whose output is
amplified and drives the vibrator. On the moving shaft is
mounted a piezoelectric accelerometer which permits accu-
rate measurement of vibration amplitude.
The testing session is divided into discrete trials. A tone is
turned on. The subject then presses the key, holds it down,
and, after a variable interval, vibration is delivered to the
finger. A key release during vibration is rewarded either by
fruit juice (monkey) or by addition of a point to a counter
(human). In order to get a quantitative estimate of guessing
bias, catch trials are randomly introduced during the session.
They differ from normal trials in that no vibration is pre-
sented. The trial starts with the tone onset, the subject de-
-------
28
MAURISSEN
10 -q
D_
CO
z
o
CX
O
1 -
0. 1 -
1000
10 100
FREQUENCIES CHZD
FIG. 4. Normal vibration thresholds in monkey and human subject.
Empty and closed circles show two independent determinations of
thresholds. Half closed circles represent the performance of a
human subject.
the subject cannot detect the stimulus any more, in which
case its amplitude is subsequently increased.
With the equipment and methods described, it has been
possible to measure vibration sensitivity at different fre-
quencies. Normative data for one monkey and one human
subject can be seen in Fig. 4. The curves with the open and
closed circles represent absolute thresholds of vibration
sensitivity obtained in one monkey about two months apart.
A remarkable stability characterizes the monkey's perform-
ance. Human data are given to show the similarities between
human and nonhuman primates.
CONCLUSIONS
Monkeys studied with the technique described here rep-
resent a very good model of human vibration sensitivity.
Anatomical and physiological similarities account for the
agreement observed in cutaneous sensitivity of human and
nonhuman primates. Application of this model to toxicity
studies is underway, and we believe that these studies are
likely to provide meaningful data applicable to the human
situation. This approach has also the advantage of being
comparable to audition and vision, where the frequency di-
mension has proven to be an important parameter in the
prediction of toxicity.
presses the key and holds it down. After a variable interval,
the tone is turned off. Key release at that time is also re-
warded.
Vibratory stimuli are presented according to the up-and-
down method. When the subject is able to detect vibration,
the amplitude of the next vibration is decreased in steps until
ACKNOWLEDGEMENTS
This work was supported in part by Grant MH-11752 from the
National Institute of Mental Health, by Grants ES-01247 and ES-
01248 from the National Institute of Environmental Health Sciences,
and in part by a contract with the U.S. Department of Energy and
has been assigned Report No. UR-3490-1600.
REFERENCES
1. Ahlmark, A. Poisoning by methyl mercury compounds. Br. J.
ind. Meet. 5: 117-119, 1948.
2. Augustin, P., M. Samson and L. Verdure. A propos d'une
observation de polyneuropathie due au maleate de perhexiline.
Rev. Oto-Neuro-Ophthal. 49: 133-139, 1977.
3. Bakir, F., S. F. Damluji, L. Amin-Zaki, M. Murtadha, A.
Khalidi, N. Y. Al-Rawi, S. Tikriti, H. I. Dhahir, T. W.
Clarkson, J. C. Smith and R. A. Doherty. Methylmercury
poisoning in Iraq. Science 181: 230-241, 1973.
4. Bank, W. J., D. E. Pleasure, K. Suzuki, M. Nigro and R. Katz.
Thallium poisoning. Archs Neural. 26: 456-464, 1972.
5. Barach, J. H. Test for quantitative vibratory sensation in diab-
etes, pernicious anemia and tabes dorsalis. Archs int. Med. 79:
602-613, 1947.
6. Barry, W. K. Peripheral neuritis following tetraethylthiuram
disulphide treatment. Lancet 2: 937, 1953.
7. Bennett, C. R. A group of patients suffering from paralysis due
to drinking Jamaica ginger. Sth med. J., Birmingham 23: 371-
375, 1930.
8. Biehl, J. P. and J. H. Skavlem. Toxicity of isoniazid. Am. Rev.
Tuberc. pulm. Dis. 68: 296-297, 1953.
9. Blum, R. H. and T. Kahlert. Maytansine: a phase I study of an
ansa macrolide with antitumor activity. Cancer Treat Rep. 62:
435-438, 1978.
10. Bonomi, P. D., R. E. Slayton and J. Wolter. Phase II trial of
adriamycin and c/s-dichlorodiammineplatinum (II) in squam-
ous cell, ovarian, and testicular carcinomas. Cancer Treat.
Rep. 62: 1211-1213, 1978.
11. Bourrat, C., J. J. Viala and J. P. Guastala. Neuropathie
peripherique apres absorption prolongee de maleate de
perhexiline. Nouvelle Presse medicate 4: 2528, 1975.
12. Bradley, W. G. and R. L. Hewer. Peripheral neuropathy due to
disulfiram. Br. med. J. 2: 449-450, 1966.
13. Bradley, W. G., L. P. Lassman, G. W. Pearce and J. N. Wal-
ton. The neuromyopathy of vincristine in man. Clinical, elec-
trophysiological and pathological studies. J. neural. Sci. 10:
107-131, 1970.
14. Burley, B. T. Polyneuritis from tricresyl phosphate. J. Am.
med. Ass. 98: 298-304, 1932.
15. Casey, E. B., A. M. Jeliffe, P. M. Le Quesne and Y. L. Millett.
Vincristine neuropathy. Clinical and electrophysiological ob-
servations. Brain 96: 69-86, 1973.
16. Cavanagh, J. B., N. H. Fuller, H. R. M. Johnson and P. Rudge.
The effects of thallium salts, with particular reference to the
nervous system changes. Q. J. Med. 43: 293-319, 1974.
17. Cavigneaux, A. and G. B. Cabasson. Intoxication par
1'acrylamide. Archs Mai. prof. 33: 115-116, 1972.
18. Chamberlain, M. J., A. L. Reynolds and W. B. Yeoman. Toxic
effect of podophyllum application in pregnancy. Br. med. J. 3:
391-392, 1972.
19. Chapman, J. D., A. P. Reuvers, J. Borsa, A. Petkau and D. R.
McCalla. Nitrofurans as radiosensitizers of hypoxic mamma-
lian cells. Cancer Res. 32: 2616-2624, 1972.
-------
NEUROTOXICANTS AND SOMATOSENSORY SYSTEM
29
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
Charache, S., D. Finkelstein, P. S. Lietman and J. Scott.
Peripheral and optic neuritis in a patient with hemoglobin SC
disease during treatment of Salmonella osteomyelitis with
chloramphenicol. Johns Hopkins med. J. 140: 121-124, 1977.
Charatan, F. B. Peripheral neuritis following tetraethylthiuram
disulphide treatment. Br. med. J. 2: 380, 1953.
Chhuttani, P. N., L. S. Chawla and T. D. Sharma. Arsenical
neuropathy. Neurology, Minneap. 17: 269-274, 1967.
Cianchetti, C., G. Abbritti, G. Perticoni, A. Siracusa and F.
Curradi. Toxic poly neuropathy of shoe-industry workers. A
study of 122 cases. J. Neurol. Neurosurg. Psychiat. 39: 1151-
1161, 1976.
Clark, A. N. G. and M. J. Parsonage. Case of podophyllum
poisoning with involvement of the nervous system. Br. med. J.
2: 1155-1157, 1957.
Collens, W. S., J. D. Zilinsky and L. C. Boas. Quantitative
estimation of vibratory sense as a guide for treatment of
peripheral neuritis in diabetes. Proc. Am. Diabetes Ass. 6:
459-468, 1946.
Collings; H. Polyneuropathy associated with nitrofuran
therapy. Archs Neurol., Chicago 3: 656-660, 1960.
Costa, G., M. M. Hreshchyshyn and J. F. Holland. Initial clin-
ical studies with vincristine. Cancer Chemother. Rep. 24:
39-44, 1962.
Coxon, A. and C. A. Pallis. Metronidazole neuropathy. J.
Neurol. Neiirosurg. Psychiat. 39: 403-405, 1976.
Dagnino, N. and R. Badino. Polineuropatia da saturnismo id-
rico. Sist. nerv. 20: 417-420, 1968.
Dische, S., M. I. Saunders, M. E. Lee, G. E. Adams and I. R.
Flockhart. Clinical testing of the radiosensitizer Ro 07-0582:
experience with multiple doses. Br. J. Cancer 35: 567-579,
1977.
Dobkin, B. H. Reversible subacute peripheral neuropathy in-
duced by phenytoin. Archs Neurol., Chicago 34: 189-190,
1977.
Doyle, J. B. and E. F. Cannon. Severe polyneuritis following
gold therapy for rheumatoid arthritis. Ann. int. Med. 33: 1468-
1472, 1950.
Dyck, P. J., D. J. Curtis, W. Bushek and K. Offord. Descrip-
tion of "Minnesota Thermal Disks" and normal values of
cutaneous thermal discrimination in man. Neurology, Min-
neap. 24: 325-330, 1974.
Dyck, P. J., E. H. Lambert and P. C. Nichols. Quantitative
measurement of sensation related to compound action potential
and number and sizes of myelinated and unmyelinated fibers of
sural nerve in health, Freidreich's ataxia, hereditary sensory
neuropathy, and tabes dorsalis. In: Handbook of Electroence-
phalography and Clinical Neurophysiology. Amsterdam:
Elsevier Publishing Co., Vol. 9, 1971, pp. 83-118.
Ellis, F. G. Acute polyneuritis after nitrofurantoin therapy.
Lancet 2: 1136-1138, 1962.
Epstein, F. W. Dapsone-induced peripheral neuropathy. Archs
Derm. 112: 1761-1762, 1976.
Frost, P., G. D. Weinstein and E. C. Gomez. Methacycline and
demeclocycline in relation to sunlight. J. Am. med. Ass. 216:
326-329, 1971.
Frost, P., G. D. Weinstein and E. C. Gomez. Phototoxic
potential of minocycline and doxycycline. Archs Derm. 105:
681-683, 1972.
Fruhstorfer, H., U. Lindblom and W. G. Schmidt. Method for
quantitative estimation of thermal thresholds in patients. J.
Neurol. Neurosurg. Psychiat. 39: 1071-1075, 1976.
Fullerton, P. M. and M. Kremer. Neuropathy after intake of
thalidomide (Distaval). Br. med. J. 2: 855-858, 1961.
Fullerton, P. M. and D. J. O'Sullivan. Thalidomide
neuropathy: a clinical, electrophysiological, and histological
follow-up study. J. Neurol. Neurosurg. Psychiat. 31: 543-551,
1968.
Gammon, G. D., F. W. Burge and G. King. Neural toxicity in
tuberculous patients treated with isoniazid (isonicotinic acid
hydrazide). Archs. Neurol. Psychiat., Chicago 70: 64-69, 1953.
43. Garb, L. G. and C. H. Hine. Arsenical neuropathy: residual
effects following acute industrial exposure. J. occup. Med. 19:
567-568, 1977.
44. Gardner-Thorpe, C. and S. Benjamin. Peripheral neuropathy
after disulfiram administration. J. Neurol. Neurosurg.
Psychiat. 34:253-259, 1971.
45. Garland, T. O. and M. W. H. Patterson. Six cases of ac-
rylamide poisoning. Br. med. J. 4: 134-138, 1967.
46. Geoffrey, H., A. Slomic, M. Benenadji and P. Pascal. Myelo-
polyn6vrites tri-cresyl phosphatees. Toxi-epidemie marocaine
de septembre-octobre 1959. Wld Neurol. 1: 294-315, 1960.
47. Goto, I., M. Matsumura, N. Inoue, Y. Murai, K. Shida, T.
Santa and Y. Kuroiwa. Toxic polyneuropathy due to glue snif-
fing. J. Neurol. Neurosurg. Psychiat. 37: 848-853, 1974.
48. Graveleau, J. Les neuropathies peripheriques du disulfirame
(Antabuse). Rev. neural. 126: 149-153, 1972.
49. Graveleau, J., P. Loirat and V. Nusinovici. Polynevrite par
1'acrylamide. Rev. neural. 123: 62-65, 1970.
50. Green, R. T. The absolute threshold of electric shock. Br. J.
Psychol. 53: 107-115, 1962.
51. Grunfeld, O. and G. Hinostroza. Thallium poisoning. Ac/is int.
Med. 114: 132-138, 1964.
52. Hakamies, L. and M. Mumenthaler. Particolarita della
polineuropatia da nitrofurantoina. Minerva med., Roma 64:
2624-2626, 1973.
53. Hamann, H. D., H. O. Handwerker and H. Assmus. Quantita-
tive assessment of altered thermal sensation in patients suffer-
ing from cutaneous nerve disorders. Neuroscience Lett. 9:
273-277, 1978.
54. Harada, M. Minamata disease. Chronology and medical report.
Bull. Inst. constitutional Med. Kumamoto 25: 1-60, 1976.
55. Hassin, G. Symptomatology of arsenical polyneuritis. J. nerv.
mem. Dis. 72: 628-636, 1930.
56. Herskowitz, A., N. Ishii and H. Schaumburg. n-Hexane
neuropathy. A syndrome occurring as a result of industrial ex-
posure. New Engl. J. Med. 285: 82-85, 1971.
57. Heyman, A., J. B. Pfeiffer, R. W. Willett and H. M. Taylor.
Peripheral neuropathy caused by arsenical intoxication. A
study of 41 cases with observations on the effect of BAL (2,3,
dimercapto-propanol). New Engl. J. Med. 254: 401-409, 1956.
58. Hopkins, A. P. and R. W. Gilliatt. Motor and sensory nerve
conduction velocity in the baboon: normal values and changes
during acrylamide neuropathy. J. Neural. Neurosurg.
Psychiat. 34: 415-426, 1971.
59. Hunter, D., R. R. Bomford and D. S. Russell. Poisoning by
methyl mercury compounds. Q. J. Med. 9: 193-213, 1940.
60. Hunter, D. and D. S. Russell. Focal cerebral and cerebellar
atrophy in a human subject due to organic mercury com-
pounds. J. Neural. Neurosurg. Psychiat. 17: 235-241, 1954.
61. Hussain, K. K. and H. Koilpillai. Toxic effects of furaltadone.
Lancet 2: 490, 1960.
62. Innis, R. and H. Moses. Thallium poisoning. Johns Hopkins
med. J. 142: 27-31, 1978.
63. Jacknowitz, A. I., J. L. Le Frock and R. A. Prince. Nitrofuran-
toin polyneuropathy: report of two cases. Am. J. Hasp.
Pharm. 34: 759-762, 1977.
64. Jones, N., P. A. Twelker and D. Singer. A waterless thermal
stimulator. Am. J. Psychol. 75: 147-149, 1962.
65. Joy, R. J. T., R. Scalettar and D. B. Sodee. Optic and
peripheral neuritis. Probable effects of prolonged chloram-
phenicol therapy. J. Am. med. Ass. 173: 1731-1734, 1960.
66. Kantarjian, A. D. and A. S. Shaheen. Methyl bromide poison-
ing with nervous system manifestations resembling
polyneuropathy. Neurology, Minneap. 13: 1054-1068, 1963.
67. Kedar, A., M. E. Cohen and A. I. Freeman. Peripheral
neuropathy as a complication of c/^-dichlorodiammineplatinum
(II) treatment: a case report. Cancer Treat. Rep. 62: 819-821,
1978.
68. Knave, B., B. Kolmodin-Hedman, H. E. Persson and J. M.
Goldberg. Chronic exposure to carbon disulfide: effects on oc-
cupationally exposed workers with special reference to the
nervous system. Wk-Environment-Hlth 11: 49-58, 1974.
-------
30
MAURISSEN
69. Koch-Weser, J., V. W. Sidel, E. B. Fereman, P. Kanarek, D.
C. Finer and A. E. Eaton. Adverse effects of sodium colis-
timethate. Manifestations and specific reaction rates during 317
courses of therapy. Ann. int. Med. 72: 857-868, 1970.
70. Koller, W. C., L. K. Gehlraann, F. D. Malkinson and F. A.
Davis. Dapsone-induced peripheral neuropathy. Archs
Neural., Chicago 34: 644-646, 1977.
71. Korobkin, R., A. K. Asbury, A. J. Sumner and S. L. Nielsen.
Glue-sniffing neuropathy. Archs Neural., Chicago 32: 158-162,
1975.
72. Laidlaw, R. W., M. A. Hamilton and R. M. Brickner. The
occurrence of dissociated disturbances of pallesthesia and
kinesthesia. Bull, neural. Inst. N.Y. 7: 303-320, 1938.
73. Layzer, R. B. Myeloneuropathy after prolonged exposure to
nitrous oxide. Lancet 2: 1227-1230, 1978.
74. Layzer, R. B., R. A. Fishman and J. A. Schafer. Neuropathy
following abuse of nitrous oxide. Neurology, Minneap. 28:
504-506, 1978.
75. Le Menn, G., D. Mabin and P. Penther. Regression lente et
incomplete d'une neuropathic peripherique due au maleate de
perhexiline. Ann. Cardiol. Angeiol. 26: 149-150, 1977.
76. Le Quesne, P. M., S. F. Damluji and H. Rustam. Elec-
trophysiological studies of peripheral nerves in patients with
organic mercury poisoning. J. Neural. Neurosurg. Psychiat.
37: 333-339, 1974.
77. Le Quesne, P. M. and J. G. MacLeod. Peripheral neuropathy
following a single exposure to arsenic. J. neural. Sci. 32: 437-
451, 1977.
78. Lescher, F. G. Nervous complication following treatment by
gold salts. Br. med. J. 2; 1303-1305, 1936.
79. Lhermitte, F., M. Fardeau, F. Chedru and J. Mallecourt.
Polyneuropathy after perhexiline maleate therapy. Br. med. J.
1: 1256, 1976.
80. Loughridge, L. W. Peripheral neuropathy due to nitrofuran-
toin. Lancet 2: 1133-1135, 1962.
81. Lovelace, R. E. and S. J. Horwitz. Peripheral neuropathy in
long-term diphenylhydantoin therapy. Archs Neural., Chicago
18: 69-77, 1968.
82. Maghazaji, H. I. Psychiatric aspects of methylmercury poison-
ing. J. Neural. Neurosurg. Psychiat. 37: 954-958, 1974.
83. Martin, W. J., K. B. Corbin and D. C. Utz. Paresthesias during
treatment with nitrofurantoin: report of a case. Proc. Mayo
Clin. 37: 288-292, 1962.
84. Materson, B. J. Generalized burning paresthesia due to
intravenous furosemide. J. Fla med. Ass. 58: 34-35, 1971.
85. McLeod, J. G. and R. Penny. Vincristine neuropathy: an elec-
trophysiological and histological study. J. Neural. Neurosurg.
Psychiat. 32: 297-304, 1969.
86. Mendell, J. R., K. Saida, M. F. Ganansia, D. B. Jackson, H.
Weiss, R. W. Gardier, C. Chrisman, N. Allen, D. Couri, J.
O'Neill, B. Marks and L. Hetland. Toxic polyneuropathy pro-
duced by methyl n-butyl ketone. Science 185: 787-789, 1974.
87. Merritt, H. H. and M. Moore. Peripheral neuritis associated
with ginger extract ingestion. New Engl. J. Med. 203: 4-12,
1930.
88. Mittelman, A., R. Grindberg and T. Dao. Clinical experience
with vincristine (NSC-67574) in advanced cancer of the breast.
Cancer Chemother. Rep. 34: 25-30, 1964.
89. Moddel, G., J. M. Bilbao, D. Payne and P. Ashby. Disulfiram
neuropathy. Archs Neural., Chicago 35: 658-660, 1978.
90. Moress, G. R., A. N. D'Agostino and L. W. Jarcho.
Neuropathy in lymphoblastic leukemia treated with vincristine.
Archs Neural., Chicago 16: 377-384, 1967.
91. Mountcastle, V. B., R. H. LaMotte and G. Carli. Detection
thresholds for stimuli in humans and monkeys: comparison
with threshold events in mechanoreceptive afferent nerve fi-
bers innervating the monkey hand. J. Neurophysiol. 35: 122-
136, 1972.
92. Murai, Y. and Y. Kuroiwa. Peripheral neuropathy in
chlorobiphenyl poisoning. Neurology, Minneap. 21: 1173-
1176, 1971.
93. Notermans, S. -L. H. Measurement of the pain threshold de-
termined by electrical stimulation and its clinical application.
Part II. Clinical application in neurological and neurosurgical
patients. Neurology, Minneap. 17: 58-73, 1967.
94. Ochoa, J. Isoniazid neuropathy in man: quantitative electron
microscope study. Brain 93: 831-850, 1970.
95. O'Connell, M. J., A. Shani, J. Rubin and C. G. Moertel. Phase
II trial of maytansine in patients with advanced colorectal car-
cinoma. Cancer Treat. Rep. 62: 1237-1238, 1978.
96. Okinaka, S., M. Yoshikawa, T. Mozai, Y. Mizuno, T. Terao,
H. Watanabe, K. Ogihara, S. Hirai, Y. Yoshino, T. Inose, S.
Anzai and M. Tsuda. Encephalomyelopathy due to an organic
mercury compound. Neurology 14: 69-76, 1964.
97. Ordonez, J. V., J. A. Carrillo, M. Miranda and J. L. Gale.
Estudio epidemiologico de una enfermedad considerada como
encefalitis en la region de los altos de Guatemala. Boln Of.
sanit. pan-am. 60: 510-519, 1966.
98. Osuntokun, B. O. An ataxic neuropathy in Nigeria. A clinical,
biochemical and electrophysiological study. Brain 91: 215-248,
1968.
99. Paulson, G. W. and G. W. Waylonis. Polyneuropathy due to
n-hexane. Archs int. Med. 136: 880-882, 1976.
100. Poole, G. W. and J. Schneeweiss. Peripheral neuropathy due to
ethionamide. Am. Rev. resp. Dis. 84: 890-892, 1961.
101. Ramsay, I. D. Endocrine ophthalmology. Br. med. J. 4: 706,
1968.
102. Raskin, N. H. and R. A. Fishman. Pyridoxine-deficiency
neuropathy due to hydralazine. New Engl. J. Med. 273: 1182-
1185, 1965.
103. Reitemeier, R. J., C. G. Moertel and C. M. Blackburn. Vin-
cristine (NSC-67574) therapy of adult patients with solid
tumors. Cancer Chemother. Rep. 34: 21-23, 1964.
104. Rizzuto, N., H. Terzian and S. Galiazzo-Rizzuto. Toxic
polyneuropathies in Italy due to leather cement poisoning in
shoe industries. A light- and electron-microscopic study. /.
neural. Sci. 31: 343-354, 1977.
105. Rosenthal, S. and S. Kaufman. Vincristine neurotoxicity. Ann.
int. Med. 80: 733-737, 1974.
106. Rubenstein, C. J. Peripheral polyneuropathy caused by nit-
rofurantoin. J. Am. med. Ass. 187: 647-649, 1964.
107. Said, G., J. Goasguen and C. Laverdant. Polynevrites au cours
des traitements prolonges par le metronidazole. Rev. neural.
134: 515-521, 1978.
108. Sakamoto, A. Physical activity: a possible determinant of vin-
cristine (NSC-67574) neuropathy. Cancer Chemother. Rep. 58:
413-415, 1974.
109. Samuel, M. L., W. V. Leary, R. Alexanian, C. D. Howe and
E. Frei III. Clinical trials with N-isopropyl-(2-methyl-
hydrazino)-p-toluamide hydrochloride in malignant dissemi-
nated neoplasia. Cancer 20: 1187-1194, 1967.
110. Sandier, S. G., W. Tobin and E. S. Henderson. Vincristine-
induced neuropathy. A clinical study of fifty leukemic patients.
Neurology, Minneap. 19: 367-374, 1969.
111. Scala, R. A. Hydrocarbon neuropathy. Ann. occup. Hyg. 19:
293-299, 1976.
112. Senanayake, N., W. A. S. de Silva and M. S. L. Salgado.
Arsenical neuropathy—A clinical study. Ceylon med. J. 17:
195-203, 1972.
113. Seppalainen, A. M. and S. Hernberg. Sensitive technique for
detecting subclinical lead neuropathy. Br. J. ind. Med. 29:
443-449, 1972.
114. Shinohara, Y., F. Yamaguchi and F. Gotoh. Toxic neuropathy
as a complication of thiophenicol therapy. Eur. Neural. 16:
161-164, 1977.
115. Smart, C. R., D. B. Rochlin, A. M. Nahum, A. Silva and D.
Wagner. Clinical experience with vinblastine sulfate (NSC-
49842) in squamous cell carcinoma and other malignancies.
Cancer Chemother. Rep. 34: 31-45, 1964.
116. Smith, H. V. and J. M. K. Spalding. Outbreak of paralysis in
Morocco due to ortho-cresyl phosphate. Lancet 1: 1019-1021,
1959.
-------
NEUROTOXICANTS AND SOMATOSENSORY SYSTEM
31
117. Sobue, I., K. Ando, M. lida, T. Takayanagi, Y. Yamamura and
Y. Matsuoka. Myeloneuropathy with abdominal disorders in
Japan. A clinical study of 752 cases. Neurology, Minneap. 21:
168-173, 1971.
118. Spencer, P. S., R. B. Hanna and G. D. Pappas. Acrylamide
intoxication of Pacinian corpuscle function: a freeze fracture
study. J. Neuropathol. exp. Neuml. 36: 631, 1977 (abstract).
119. Spencer, P. S., R. Hanna, M. Sussman and G. Pappas. Inac-
tivation of Pacinian corpuscle mechanosensitivity by ac-
rylamide. J. gen. Physiol. 70: 17a, 1977 (abstract).
120. Spillane, J. D. The geography of neurology. Br. med. J. 2:
506-512, 1972.
121. Steiness, I. Vibratory perception in non-diabetic subjects dur-
ing ischaemia with special reference to the conditions in
hyperglycaemia, after carbohydrate starvation and after cor-
tisone administration. Acta med. scand. 169: 17-26, 1961.
122. Svennilson, E. Studies of triorthocresyl phosphate neuropathy,
Morocco 1960. Acta psychiat. scand. Suppl. ISO: 334-336,
1960.
123. Swaiman, K. F. and D. Flagler. Mercury poisoning with cen-
tral and peripheral nervous system involvement treated with
penicillamine. Pediatrics 48: 639-642, 1971.
124. Toole, J. F. and M. L. Parrish. Nitrofurantoin polyneuropathy.
Neurology, Minneap. 23: 554-559, 1973.
125. Towfighi, J., N. K. Gonatas, D. Pleasure, H. S. Cooper and L.
McCree. Glue sniffer's neuropathy. Neurology, Minneap. 26:
238-243, 1976.
126. Tsubaki, T., Y. Honma and M. Hoshi. Neurological syndrome
associated with clioquinol. Lancet 1: 696-697, 1971.
127. Tsubaki, T., K. Shirakawa, K. Hirota and K. Kanbayashi.
Neurological aspects of methylmercury poisoning in Niigata.
In: Minamata Disease, edited by T. Tsubaki and K.
Irukayama. New York: Elsevier Publishing Company, 1977,
pp. 145-165.
128. Tursky, B. and P. D. Watson. Controlled physical and subjec-
tive intensities of electric shock. Psychophysiology 1: 151-162,
1964.
129. Ursing, B. and C. Kamme. Metronidazole for Crohn's disease.
Lancet 1: 775-777, 1975.
130. Urtasun, R. C., P. Band, J. D. Chapman, H. R. Rabin, A. E.
Wilson and C. G. Fryer. Clinical phase I study of the hypoxic
cell radiosensitizer Ro-07-0582, a 2-nitroimidazole derivative.
Radiology 122: 801-804, 1977.
131. Vasilescu, C. Sensory and motor conduction in chronic carbon
disulphide poisoning. Eur. Neural. 14: 447-457, 1976.
132. Verrillo, R. T. Effect of contactor area on the vibrotactile
threshold. J. acoust. Soc. Am. 35: 1962-1966, 1963.
133. Verrillo, R. T. Vibrotactile thresholds for hairy skin. J. exp.
Psychol. 72: 47-50, 1966.
134. Vora, D. D., D. K. Dastur, B. M. Braganca, L. M. Parihar, C.
G. S. Iyer, R. B. Fondekar and K. Prabhakaran. Toxic
polyneuritis in Bombay due to ortho-cresyl-phosphate poison-
ing. J. Neural. Neurosurg. Psychiat. 25: 234-242, 1962.
135. Wahlberg, P., E. Karppanen, K. Henrikson and D. Nyman.
Human exposure to mercury from goosander eggs containing
methyl mercury. Acta med. scand. 189: 235-239, 1971.
136. Whitelaw, D. M., D. H. Cowan, F. R. Cassidi and T. A. Patter-
son. Clinical experience with vincristine. Cancer Chemother.
Rep. 30: 13-20, 1963.
137. Wickersham, C. W. and E. J. Fredericks. Toxic
polyneuropathy secondary to methyl n-butyl ketone. Conn.
Med. 40: 311-312, 1976.
138. Wong, P. P., A. Yagoda, V. E. Currie and C. W. Young. Phase
II study of vindesine sulfate in the therapy for advanced renal
carcinoma. Cancer Chemother. Rep. 61: 1727-1729, 1977.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 33-44. ANKHO International Inc., 1979.
Comparative Behavioral Toxicology1
WILLIAM C. STEBBINS AND DAVID B. MOODY
Departments of Otorhinolaryngology and Psychology
and
Kresge Hearing Research Institute, University of Michigan, Ann Arbor, MI 48109
STEBBINS, W. C. AND D. B. MOODY. Comparative behavioral toxicology. NEUROBEHAV. TOXICOL. 1: Suppl. 1,
33-44, 1979.—Behavioral conditioning together with conventional sensory testing methods may be used in the evaluation
of toxic effects on sensory systems in experimental animal models. Such procedures yield precise quantitative estimates of
impairment in absolute and differential acuity and in sensory perception. Additionally, these behavioral changes can be
related to the presence of histopathology in peripheral sensory structures; this orderly relation between structure and
function may aid in our understanding of the basis for sensory coding in the normal end organ.
Behavioral toxicology Hearing loss Monkeys Guinea pigs Cats Chinchillas Drugs Noise
THE earliest behavioral indications of toxicity may often be
subtle, and, in the absence of sufficiently sensitive measur-
ing techniques, go undetected. The quest for methodology
capable of detecting these early signs and, too, of determin-
ing the manner and precise action of toxic agents has led us
to examine a variety of both methods and animal models.
The comparative method is thus illustrated along two di-
mensions and will, we hope, add generality to our findings in
order that we might more effectively extrapolate to man. It
has been known for some time that certain drugs such as the
aminoglycosidic antibiotics, and intense sound exposure can
cause permanent deafness with severe cochlear histopathol-
ogy in man. Our progress in achieving an understanding of
the process and of the mechanisms in man himself is limited
by our inability to control the relevant variables, such as the
influence of other toxins, and to evaluate the histopathology
immediately post mortem. It is for these reasons among
others that we rely on animal models for their value in ex-
trapolation to man.
First, in experiments on ototoxicity with the amino-
glycosidic antibiotics and noise we have employed as subjects
guinea pigs, chinchillas, cats, and several species of monkey.
Thus, in a biological sense the approach can be considered
comparative; in fact, we have discovered a striking example
of species-specific toxicity in the course of this research.
Second, in the study of hearing and hearing loss brought
about by these drugs and intense noise we have tried to
utilize a variety of methods for our evaluation of impaired
hearing. Hearing, like any other behavioral process, can be
examined in more ways than one. The threshold test is per-
haps the most widely used method of evaluation, and, in
previous research, threshold changes have been related to
histopathological changes occurring in the inner ear and cen-
tral nervous system after exposure to toxic agents [8, 18,231.
However, under normal lifelike conditions we seldom attend
to stimulus events at minimum detectable levels. When hear-
ing becomes impaired we undergo subtle changes in our
ability to accurately locate and discriminate between acous-
tic events at normal speaking and listening levels. Conse-
quently, it is important that we establish reliable and valid
procedures for the evaluation of these suprathreshold dis-
criminative functions.
This paper will consider some of the different char-
acteristics of hearing and how they are measured in experi-
mental animals before and after exposure to toxic agents.
Examples will be drawn from different species and compari-
sons made between them. In most instances the similarity in
toxic effects across species for a given agent is clear; how-
ever, there is at least one very interesting exception which,
we think, illustrates the importance of diversity in both ap-
proach and choice of subject, particularly if our goal is to
extend these findings to ourselves. Although our experience
has been in the auditory system and the specific toxicants to
which it is sensitive, there is every reason to believe that the
behavioral methods described herein are applicable to the
behavioral analysis of sensory impairment in general.
A significant outcome, even a gratuity, of these experi-
ments on ototoxicity is that they have provided us with in-
formation about how the ear works. The relation between
hearing loss and the changes occurring in the inner ear is a
very predictable one. Stimulus coding of frequency and in-
tensity is, in large measure, a function carried out by the
receptor cells in the inner ear. These inner and outer hair
cells are selectively destroyed by intense sound and certain
of the ototoxic drugs with specific consequences for hearing,
thus revealing certain key aspects of the peripheral coding
process [21]. This issue will also be considered in this paper.
In behavioral, anatomical, and physiological studies of
'The research described herein was supported by research and program grants from NINCDS (NS 05077, NS 05065 and NS 05785).
Histological procedures and evaluation were carried out by Drs. J. E. Hawkins, Jr. and L.-G. Johnsson.
33
-------
34
STEBBINS AND MOODY
the auditory system, guinea pigs and cats have long been the
subjects of choice. Although much of what we know about
structure, function, and toxicity in the auditory system is
based on findings from these animal models, they have often
been considered recalcitrant subjects for behavioral experi-
ments. However, we have found operant conditioning
methods with the appropriate positive reinforcers to be quite
effective [9, 10, 12], In more recent years chinchillas [1,5]
and monkeys [20] have been selected as subjects for auditory
research both because of the relative ease with which they
may be behaviorally trained for hearing testing and of the
close semblance of their hearing thresholds to man's.
TRAINING AND TESTING PROCEDURES
We have employed the same basic training procedure
with all of these animals. There are obvious and critical
differences in some of the details such as the selection of the
response device and the nature of the reinforcer, but a gen-
eral training and threshold testing paradigm has been found
effective for all. Further, only minor modifications in the
basic paradigm for audiometric or threshold testing render it
useful for a variety of discrimination procedures where the
acoustic stimuli are well above threshold levels; i.e., fre-
quency and intensity discrimination, loudness judgment, and
so on [8].
Under the training procedure the animals are food de-
prived and food is then used as a reinforcer to strengthen
desired behaviors. A stimulus light indicates the opportunity
to respond by manually contacting either a switch or
contact-sensitive plate. Contact, once initiated, must be
maintained until an acoustic stimulus (pure tone) is pre-
sented. The time interval between contact initiation and
stimulus onset is randomized on successive presentations. If
manual contact is then broken while the tone is on, the food
reinforcer is delivered immediately. The holding response is
critical; if contact is broken prematurely, i.e., before the
auditory stimulus is turned on, the trial is terminated without
food and the next tone presentation (trial) may be delayed.
Time intervals assigned are brief. The holding requirement
can vary between 1 and 9 seconds. The penalty or time out
for releasing the response device too soon may be a 6-10 sec
delay before the stimulus light is turned on indicating the
next trial. In addition, catch trials in which no stimulus is
presented are interspersed randomly with tone trials in order
to assess an animal's guess rate. In a more elaborate form of
the procedure the trial is announced by a flashing light which
becomes steady when the contact response has been effec-
tively initiated. The light is turned off when food is given or if
contact is prematurely broken.
Training begins with a conditioning session in which the
animal is shaped by successive approximations to touch and
maintain contact with the response device and continues
until the contingencies described above are in effect and the
discriminative behavior with regard to the auditory stimulus
is stable [8]. Criteria for stability require that responding to
catch trials is reduced to a low steady level and that the
holding response is maintained until the tone is presented at
which time contact is quickly broken. At this stage one of
two psychophysical procedures for threshold testing is put
into effect.
Under the constant stimulus method, tones at 5 intensities
(equally spaced and bracketing the presumed threshold) are
presented in a random order on successive trials. The fre-
FIG. 1. Use of the tracking method for audiometric testing of mon-
keys. Correct detections cause the tone to be attenuated in 5-dB
steps, while failure to hear produces a subsequent 5-dB increase in
tone intensity. The threshold is indicated by the horizontal dashed
line at 61 dB [18].
quency or percentage of times that the subject responds at
each intensity level is noted and the threshold for a given
frequency (pure tone) is that sound level to which the animal
responds 50 percent of the time. A more efficient
psychophysical method but also more difficult for the animal
is that of tracking. The tone intensity is initially well above
normal threshold. Each correct detection by the subject
leads to a decrease in tone intensity on the subsequent trial.
Failure to respond to the tone, conversely, is followed by an
increase in tone intensity. Under this procedure the stimulus
is quickly brought down by the subject to the threshold re-
gion and kept there (see Fig. 1). The difficulty lies in main-
taining behavior which is continually under the control of
barely audible stimuli. Threshold which is indicated by the
dashed horizontal line in Fig. 1 is based on the average tran-
sition value between correct detections and failure to report
the tone. Although the procedure has been used successfully
with monkeys, the constant stimulus method appears more
effective with guinea pigs. Thresholds are determined in this
manner at frequencies encompassing an animal's entire
audible range. Typical threshold functions are shown in Fig.
2 for four Old World monkeys (Macaco) and in Fig. 3 for two
guinea pigs. These contours represent the minimum detect-
able sound pressure levels at various frequencies to which
these animals can respond, and the functions for individual
animals form baselines against which the effects of agents
which damage the ear and impair hearing are measured. The
data provide stable and reproducible measures of acoustic
sensitivity in a variety of experimental animals; yet these
measures are quick to reflect minute changes in the auditory
system after insult with ototoxic compounds or exposure to
intense sound.
Absolute sensitivity is but one aspect of auditory func-
tion. An animal's survival and certainly man's ability to cope
effectively with his environment depend on the discrimina-
tion of minima] differences in the waveform of acoustic
stimulation. Such differential sensitivity may be evaluated
by determining the smallest difference between two fre-
quencies or two sound pressure levels to which an animal
can respond. The testing procedure is very similar to that
described above for the threshold test. The subject initiates a
holding response which immediately produces a repetitively
pulsed pure tone (standard frequency). After a brief but var-
-------
COMPARATIVE BEHAVIORAL TOXICOLOGY
35
100
80
(VI
O
60-
£
CD
40
20
•o
c
3
O
co o
-20
A
O
-A M-2
•• M-3
-A M-5
-o M-6
.06
.125
.25
.5 I
Frequency
2
(kHz)
15
30
60
FIG. 2. Normal auditory thresholds for 4 macaque monkeys [15],
ied time interval, if the animal has maintained the contact
response, a second pure tone (variable frequency) alternates
several times with the first. The subject responds to the
change in frequency by breaking contact and is then rein-
forced with food [8,16]. Following training the variable fre-
quency is brought closer to the standard until the limits of the
animal's resolution are reached and a differential threshold is
determined as that frequency difference to which the subject
responds 50% of the time. Differential thresholds for inten-
sity or sound pressure may be determined with the same
paradigm and are measured at sound pressure levels well
above minimum detectable levels and, in the same manner as
absolute threshold, across much of the ear's audible range.
The significance of differential acuity for responding to the
small nuances in the sounds of speech and communication,
for example, can hardly be overestimated. The clinical
audiometric examination includes similar tests in the diag-
nosis of hearing disorders.
FREQUENCY
FIG. 3. Normal auditory thresholds for 2 guinea pigs [12].
Beyond the threshold measures of hearing capacity are
the perceptual functions which assume importance at normal
sound levels. One of these is the judgment of loudness. The
absolute threshold function across frequency is the limiting
case of the equal loudness contour. Significant decisions re-
garding, for example, the distance or exact location of a
sound source are to a considerable extent based on loudness
judgments. Recruitment refers to a form of hearing impair-
ment in which our ability to discriminate loudness levels is
markedly affected. In our laboratory we have devised a
strategy for training experimental animals to make these
judgments [6,14]. The paradigm is based on the fact that an
animal's speed of reaction to the onset of an acoustic
stimulus varies in an orderly fashion with the intensity of that
stimulus [22]. A typical response latency-stimulus intensity
function is shown in Fig. 4. Such differential responding at
different stimulus levels provides us with a measure of loud-
ness judgment and has been shown by us and others [11] to
be quite compatible (in man) to loudness judgments made
under more subjective (verbally instructed) conditions.
Further, these latency-intensity functions give us an addi-
tional and different measure of auditory impairment. Not
uncommonly we may find elevation of the hearing threshold
but without significant hearing loss at typical conversational
levels [7].
Training of animals to produce these loudness (latency-
intensity) functions requires one simple elaboration on the
basic threshold paradigm. Animals learn to report acoustic
stimulation by breaking contact with a response device, but,
in addition, they must do so quickly. In the course of train-
ing, food reinforcement is contingent upon a rapid response
to tone onset. Such training ensures a relatively invariant
response topography, response latencies on the order of 200
msec with minimal variability at high stimulus intensities,
-------
36
STEBBINS AND MOODY
1 500
5
M-42
S "
•3 I0
00 200 300
Latency (msec.)
0 10 20 30 40 50 60 70 80 90
Sound Pressure (dB re 00002 dyne/cm?)
FIG. 4. Reaction time (medians as a function of stimulus intensity)
for one macaque monkey. Vertical lines indicate semi-interquartile
ranges. Inset, upper right, indicates frequency distribution of reac-
tion times at 65 dB SPL from the reaction time-stimulus intensity
function [17].
and a latency-intensity function which covers most of the
animal's dynamic range of hearing above absolute threshold.
There are yet other behavioral questions to be directed at
the normally functioning and subsequently impaired auditory
system. The acuity for the localization of sound in space is a
binaural phenomenon which depends on the integrity of both
ears and key portions of the central auditory system. Differ-
ential loss in one ear as, for example, with an eighth nerve
tumor, leads to a deficiency in this important function. Be-
havioral measures of frequency analysis or selectivity, such
as the critical band or psychophysical tuning curve, may
provide an evaluation of inner ear function since the cochlea
is the first and probably foremost frequency analyzer in the
auditory system.
Tuning curves are well-known properties of single audi-
tory neurons. More recently with behavioral techniques in-
volving tone-on-tone masking we have been able to obtain
psychophysical tuning curves in animals which closely re-
semble their electrophysiological counterparts [13]. One
such function taken from a monkey is presented in Fig. 5.
The test signal was a 2 kHz pure tone at 5 dB above
threshold. The subject was conditioned to respond to the
tone (as described above). Following correct detections the
test tone remained at the same level but the intensity of the
pure tone masker was raised. If the subject failed to report
the test tone the intensity of the masker was decreased. The
function in Fig. 5 represents the levels of various masker
frequencies at which the 2 kHz test tone was just detectable.
The discrimination measure of frequency selectivity is QlodB
which is the test tone frequency divided by the bandwidth of
the function 10 dB above the minimum. Human listeners
with sensorineural hearing loss and probable cochlear in-
volvement demonstrate pronounced changes in these func-
tions suggesting a serious disruption of frequency analytic
capabilities [4].
so
o.
in
m eo
50
30
M-102 LEFT
2 kHz TEST TONE
5dB SL
I 2
MASKER FREQUENCY (kHz)
FIG. 5. Psychophysical tuning curve for a macaque monkey. The
test tone was at 2 kHz, 5 dB above threshold.
EXPOSURE TO OTOTOXIC AGENTS
We have been working principally with the amino-
glycoside antibiotics kanamycin, neomycin, and dihydro-
streptomycin in monkeys, guinea pigs, and cats or with in-
tense sound exposure in monkeys and chinchillas. Drugs are
administered daily following hearing testing. Treatment is
continued until a given level of impairment has been pro-
duced. Animals are tested until their hearing loss has
stabilized for at least one month following treatment when
they are sacrificed for histologic evaluation. Sound expo-
sures are given in eight-hour daily sessions followed by hear-
ing testing. Animals are sacrificed one month or later after
the final exposure.
HISTOLOGIC EVALUATION
At the time of sacrifice the animal's temporal bones are
removed so that the cochlea of the inner ear may be prepared
for examination and evaluation by phase contrast micros-
copy [2]. More detailed scrutiny may also be carried out with
scanning or transmission electron microscopy. Under light
microscopy inner and outer hair cells still present are
counted and cytocochleograms, which display the number of
hair cells as a function of their position on the basilar mem-
brane, are constructed. Direct comparisons can then be
made between these cochleograms and behavioral
audiograms which represent the extent of the animal's hearing
across the audible frequency range.
OTOTOXIC EFFECTS
Daily hearing testing both during and following adminis-
tration of the toxicant reveals the progressive and orderly
-------
COMPARATIVE BEHAVIORAL TOXICOLOGY
37
-10
0
10
2O
"° 30
Z 40
I 5°
| 60
ai 70
SL
H 80
90
100
M-14 Left Ear
l.M. Kanamycin
100 mg/kg/day
for ISO days
60
FIG. 6. Progressive changes in threshold for a macaque monkey for different acoustic frequencies
during and following daily kanamycin treatment. The zero line represents normal hearing at all fre-
quencies prior to drug treatment [23].
T
-20
-10
$ 0
.£ 10
20
.
30
40
50
60
£70
80
90
100
1 T
1 T
T
"I T
M-CAT
I.M. Kanamycin
60mg/kg/day
for 44 day's
I
16
20
24
26 32 36
Days
40
44 48 52 56
FIG. 7. Progressive changes in threshold for a cat for different acoustic frequencies during and
following kanamycin treatment. The zero line represents normal hearing at all frequencies prior to drug
treatment.
nature of the hearing loss. In Fig. 6 threshold changes at
several frequencies are shown for a macaque monkey over a
six-month period during which kanamycin was given daily.
The results of hearing testing after cessation of treatment are
also shown. Typically with kanamycin as with the other
aminoglycosides hearing loss occurs first at the higher fre-
quencies and inevitably spreads in time to the lows. If the
antibiotic is continued, complete deafness ensues. The pro-
cess resembles presbycusis (hearing loss with aging) although
the time frame is considerably shorter. In Fig. 7 similar re-
sults are presented for a cat, and in Fig. 8 for a guinea pig.
The similarities are striking. In addition to the gradual high-
to-low frequency loss, the abruptness with which the loss
occurs at each frequency is worth noting.
In Fig. 9 the terminal audiogram or threshold test (lower
panel) for each ear before sacrifice is shown together with
the cytocochleogram (upper panel) for right and left
cochleas. The monkey is the same one described previously
(Fig. 6). Note that while no response to the most intense
acoustic stimuli could be obtained at the higher frequencies,
hearing was completely normal at the lower frequencies. Our
protocol requires testing until the degree of hearing loss has
shown no further change for at least one month following the
end of treatment; thus, there is every reason to consider the
final thresholds as stable. The corresponding cytocochleo-
gram is shown in the upper part of Fig. 9. While no inner or
outer hair cells remain in the lower half of the cochlea, both
kinds of cells are present and appear normal in the upper
half. The symmetry of the pattern of missing hair cells and of
the hearing loss in each ear is the usual finding in these
animals. The boundary between present and missing hair
cells at about 15 mm from the base resembles the sharp
transition between normal hearing and complete deafness in
the threshold function. The significance of this relation be-
-------
38
STEBBINS AND MOODY
-10
So
•E 10
'30
"«50
£
£60
(-
70
S-9
Sc Kanamyoin
200 mg/kg/day
for 23 days
I I I I I I I I | I
I I I I I I I I I I I I
12
16
20
24
28 32
Days
36 40
44
D-
8"
48
70
FIG. 8. Progressive changes in threshold for a guinea pig for different acoustic frequencies during and
following daily kanamycin treatment. The zero line represents normal hearing at all frequencies prior
to drug treatment.
LENGTH OF BASILAR MEMBRANE mm FROM APEX
100
g1 80
"c
'a
E
£
-------
COMPARATIVE BEHAVIORAL TOXICOLOGY
39
LENGTH OF BASILAR MEMBRANE ' mm FROM APEX
LENGTH OF BASILAR MEMBRANE :mm FROM BASAL END
25 20 15 10 5
~T
FREQUENCY (kHz)
FIG. 11. Auditory threshold shift (lower panel) for the guinea pig
represented in Fig. 8 measured 5 weeks after last treatment with
kanamycin and cytocochleogram (upper panel) [12].
differential receptor cell loss in the cochleas of the guinea
pigs. Many more outer than inner hair cells are missing from
the basal region of the cochlea. The possible significance of
these findings for stimulus intensity coding will be discussed
below.
One of our most interesting and serendipitous discoveries
was the finding of a form of species-specific toxicity in
nonhuman primates [3]. This is a finding with clear implica-
tions for the use of animal models in toxicologic research and
one which should be pursued. Macaque monkeys appear
relatively insensitive (with regard to their hearing) to dihy-
drostreptomycin at dose levels up to 100 mg/kg (5 times the
clinical dose in man) for daily injection periods of as long as
eight months. Yet patas monkeys suffer considerable ir-
reversible hearing impairment in a matter of weeks when
given the clinical dose. Examples with thresholds taken one
week after cessation of treatment and again when hearing
loss had stabilized just before sacrifice together with
cytocochleograms are shown in Figs. 12 and 13.
These data are very similar to those just described for the
kanamycin-treated guinea pigs. The hearing loss progressed
from the high frequency end of the audible range toward the
lows. Failure to hear at all at the highest frequencies tested, a
moderate to severe loss over the midrange, and normal or
nearly normal hearing at the lowest frequencies characterize
these monkeys. Correlated with the behavioral impairment
was a complete loss of all receptor cells in the extreme base
of the cochlea, a long stretch of membrane with normal ap-
pearing inner hair cells and no outer hair cells, and finally a
full or nearly full complement of both cell types in the apex
of the cochlea. While these effects were observed in the
patas monkeys, hearing in the treated macaques remained
normal and there were no significant alterations in their
cochlear morphology. Seven patas and six macaques were
examined in this experiment. With regard to sensitivity to
FREQUENCY (kHz}
FIG. 12. Auditory threshold shift (lower panel) and cytocochleo-
gram (upper panel) for a patas monkey treated with dihydrostrep-
tomycin. The dashed function represents hearing loss 1 week after
drug treatment was stopped; the solid function indicates hearing loss
12 weeks later [3].
LENGTH OF BASILAR MEMBRANE •• mm FROM BASAL END
25 20 15 10 5
~T
025 05 I 2 4
FREQUENCY (kHz)
FIG. 13. Auditory threshold shift (lower panel) and cytocochleo-
gram (upper panel) for a patas monkey treated with dihydrostrep-
tomycin. The dashed function represents hearing loss 1 week after
drug treatment was stopped; the solid function indicates hearing loss
17 weeks later [3].
-------
40
STEBBINS AND MOODY
dihydrostreptomycin, man is probably somewhere between
these two extremes; a small but significant number of hu-
mans treated with dihydrostreptomycin experience severe
hearing loss. A unique part of the effect for this drug in both
man and patas is its delayed toxic action (see Figs. 12 and
13). Months following the last treatment, the patas monkeys
continued to experience a deterioration in their hearing. We
rarely see this degree of delayed effect with the other antibi-
otics. The continued daily testing protocol brings it out very
clearly.
Though we tend to consider as toxicants only those liq-
uids, gases or solids which threaten our health or even our
life, intense sound as a severe and usually transitory mechan-
ical disturbance in air should qualify as a health-threatening
pollutant, and therefore the appropriate subject matter of
toxicology. Although different from other toxicants in that
no lingering residue remains in tissue or in the environment,
nonetheless intense sound leaves its permanent mark at least
on hearing and on the inner ear and undoubtedly on psycho-
logical functions less subject to precise measurement. Unlike
the antibiotics, intense sound or noise can cause temporary
hearing loss but depending on its intensity and duration it can
also produce permanent loss in the same manner as the anti-
biotics by destroying the receptor cells in the inner ear.
However, the pattern of destruction is often somewhat
different and depends to some extent on the spectral proper-
ties of the sound source.
Figure 14 illustrates an example of noise-induced perma-
nent hearing loss in a chinchilla exposed to band limited
noise (710 to 2800 Hz) at 123 dB SPL for only 15 min. A
moderate to severe loss is evident throughout the animal's
hearing range (lower panel). Outer hair cells have disap-
peared throughout the middle and basal regions of the coch-
lea, while considerably more than 50% of the inners remain.
The final audiogram was taken one month after the noise
exposure. In Fig. 15 we show the results for the left ear of a
monkey exposed on a working day schedule for three
months to the intense and complex noise recorded from an
automotive stamping plant and played back at the same level
(105 dBA with impulse peaks reaching 126 dB SPL). The
hearing loss though moderate is significant (see lower panel)
and is correlated with a pronounced loss of first row outer
hair cells. The differential sensitivity of inner and outer hair
cells to ototraumatic insult, whether by drugs or intense
sound, is readily apparent.
Measures of absolute threshold or the minimum detecta-
ble levels of acoustic energy by the ear provide an evaluation
of one important aspect of auditory function. Agents, toxic
at least to the peripheral auditory system, destroy the recep-
tor cells which lie along the basilar membrane of the cochlea
of the inner ear, and these events are followed by loss of
cochlear supporting cells and degeneration of the auditory
nerve and even cellular destruction at the level of the
cochlear nucleus in the brain stem. Behavioral threshold
testing either at high frequencies after antibiotic therapy or at
those frequencies which are most intense in a high-level ex-
posure sound or noise provide the earliest evidence of the
beginning of the toxic process. We consider this kind of early
warning system to be a most important goal for behavioral
toxicology.
Differential sensitivity to frequency or other dimensions
of acoustic stimulation may tell us something about an ani-
mal's ability to resolve those dimensions. In Fig. 16 the stand-
ard audiogram is plotted on the right for an animal treated
with kanamycin. The familiar shape of the post drug function
DISTANCE FROM BASAL END (mm)
188 15 10 5
IOO
090
§ SO
' tt 6O
: en 50
(r
< 2O
x
IO
O
I I I I i i i i I i i i i i i i i i i—n-
C-2 28 DAY RECOVERY
IHC
OHC I
710-2800 Hz BAND
123 dB SPL
15 MIN.
A
f
_l I I L
20
1248
FREQUENCY (KHz)
FIG. 14. Auditory threshold shift for a chinchilla after exposure to
band-limited noise (lower panel) and cytocochleogram (upper panel)
UJ.
LENGTH OF BASILAR MEMBRANE mm FROM BASAL END
25 20 15 10 5
"^ I I I—1—1—I—I—I—I—\—I—1—I—I—I—I—[—i—i—i—|—i—
3 20
I MONTH POST TREATMENT
M-7S L
M. RADIATA
INDUSTRIAL NOISE
105 dBA, 126 dB IMPULSE
8 HRS / DAY. x 5 DAYS / WK x 12 WKS
25 ,5 I 2
FREQUENCY (kHz)
FIG. 15. Auditory threshold shift (lower panel) and cytocochleo-
gram (upper panel) for a macaque monkey exposed to industrial
-------
COMPARATIVE BEHAVIORAL TOXICOLOGY
41
needs no further comment. The animal was trained and
tested on alternate days for his ability to discriminate
changes in frequency—the frequency difference threshold.
The data indicate that impairment in frequency discrimina-
tion precedes the shifts in absolute threshold. Although the
animal's ability to hear frequencies up to 4 kHz is normal, he
is significantly impaired at higher frequencies. Yet at 4 kHz
differential sensitivity to frequency is significantly affected.
Additional measures such as this provide alternate ways of
examining the effects of ototoxins.
The difficulties in training an animal on two different
tasks is outweighed by the information gained. For example,
in Fig. 17 audiometric (threshold) findings are presented for a
patas monkey treated with dihydrosteptomycin. After 75
days of treatment a moderate high frequency loss was ob-
served. For three months following treatment the hearing
loss progressed to the low frequencies. If we now examine
the animal's reaction time—stimulus intensity functions at a
single key frequency (8 kHz), the progressive nature of the
post-drug hearing loss becomes apparent (Fig. 18). The
greatest loss occurred in the first two weeks following treat-
ment, and particularly significant is the observation that the
loss was more severe at intensity levels near threshold. Note
that the functions come together at high sound levels
suggesting little change in sensitivity at these levels. The
effect, commonly noted with sensorineural hearing loss in
man, is referred to as loudness recruitment. The reaction
time procedure affords us an opportunity to study auditory
perception in experimental animal models and its deteriora-
tion after exposure to toxic substances. It also gives us an
added insight into the detailed behavioral changes occurring
after such exposure.
A somewhat similar pattern of loudness recruitment is
seen in Fig. 19 for a monkey following exposure to an octave
band of noise centered at 2 kHz. Recruitment tests were
carried out at 500, 2000, 4000, and 8000 Hz. Significant ef-
fects on hearing are seen only at 2000 and 4000 Hz, and,
similar to the drug-treated animal described above, these
effects are obvious at low stimulus levels near absolute
threshold. The effects of the same exposure level and dura-
tion are more pronounced for a second animal whose re-
cruitment functions are shown in Fig. 20. Sizeable shifts in
the latency-intensity functions occurred at all frequencies
close to and above the noise exposure band. This animal's
separately determined threshold function measured over the
same four test frequencies together with the cytocochleo-
gram is presented in Fig. 21. As seen before, significant high
frequency hearing loss is correlated with complete loss of
outer hair cells and some sporadic loss of inners in the basal
half of the cochlea. The recruitment functions indicate a
sizeable impairment in loudness discrimination near
threshold and at moderate sound levels perhaps 40 to 50 dB
above threshold; at higher levels still, hearing is relatively
normal although the animal's dynamic (intensity) hearing
range is considerably compressed.
SENSORY CODING IN THE EAR
The orderly and predictable nature of the deterioration in
hearing and the related disappearance of the receptor cells in
the inner ear following toxic insult may tell us something of
the function of these cells in normal hearing. The conceptual
notion of frequency coding according to place of stimulation
along the basilar membrane in the cochlea from base to apex
c c 2
M-17 Left Ear
Konamycin i.m.
I00mg/kg • 29 day*
O—a Frequency Different
O—O Absolute Threshold
20 .£
vz.
30 o«
Frequency (kHz)
FIG. 16. Changes in absolute auditory threshold and frequency
difference threshold in a macaque monkey treated with kanamycin
[18].
-20
-10
0
10
3 30
O
§40
60
70
M-II5 R
E. PATAS
DHSM
20 mq / kg /day x 75 days
.063 .125 .25
.5124
FREQUENCY (kHz)
FIG. 17. Auditory threshold shift for a patas monkey treated with
dihydrostreptomycin. The upper function represents the threshold
shift at the end of drug treatment; the lowerfunction was taken more
than 3 months following the end of drug treatment.
is widely held. If we consider other data similar to those
shown in Fig. 9 we can begin to generate an actual frequency
map of the cochlear spiral as presented in Fig. 22 based on
data from four drug-treated animals. The frequency scale for
threshold testing is aligned with the linear extent of the basi-
lar membrane according to certain rules which are consistent
between animals. The result is strongly supportive of a place
principle of frequency coding and yields a frequency map of
the monkey's cochlea.
Intensity coding in the mammalian inner ear is still poorly
understood. Some coding is surely done in single auditory
nerve fibers but this seems insufficient to account for the
enormous (greater than 100 dB) dynamic range of the mam-
malian ear. One reasonable possibility depends on a func-
tional difference between the inner and outer hair cells not
dissimilar to the differences observed for the rods and cones
of the vertebrate retina. The receptor cells in the ear may be
part of a two stage intensity code with the more delicate
outer hair cells subserving sound pressure levels from
threshold to about 50 dB above threshold and the inner hair
-------
42
STEBBINS AND MOODY
10 20 30 40 50 60 70 80 90
SOUND PRESSURE LEVEL (dB re ZO/iN/m2}
FIG. 18. Reaction time-stimulus intensity functions at 8 kHz for the
same monkey shown in Fig. 17 before and varying times after drug
treatment.
M NEMESTRINfl
2 kHz OCTAVE BAND El
120 dB SPL.B MRS
500 HI TEST TONE
8 kHz TEST TONE
10 20 30 40 50 60 70 80 90 100 MO IHO 10 20 30 40 50 60 70 80 90 100 110
SPL (dB re£O^N/m2)
FIG. 20. Reaction time-stimulus intensity functions for a macaque
monkey at 4 test tone frequencies before and after exposure to
noise. Arrows near abscissa indicate pre- and post-exposure
thresholds at those test tone frequencies.
IISEl POST NOISE
M NEMESTRINA
2kHi OCTAVE BANDE:
120 dB SPL.8 MRS
KHz TEST TONE
RE NOISE
32dB—
2 HHi TEST TONE
8hHi TEST TONE
10 20 30 40 30 SO 70 BO 90 IOO 110 120 10 20 30 40 50 60 70 80 90 100 110
SPL (dBreaOuN/m*)
FIG. 19. Reaction time-stimulus intensity functions for a macaque
monkey at 4 test tone frequencies before and after exposure to
noise. Arrows near abscissa indicate pre- and post-exposure
thresholds at those test tone frequencies.
LENGTH OF BASILAR MEMBRANE mm FROM BASAL END
25 20 15 10 5
"I 1 \ 1 T-
M-95 L
M RADIATA
2 kHz OCTAVE BAND
120 dB SPL
6 MRS.
.063 .125 .25 .5 I 2
FREQUENCY (kHzl
-10
0
10
20
30 £
10 «
50 o
z
60 <
70
80
90
100
FIG. 21. Auditory threshold shift (lower panel) and cytocochleo-
gram (upper panel) for the monkey whose reaction time data are
shown in Fig. 20.
-------
COMPARATIVE BEHAVIORAL TOXICOLOGY
43
FREQUENCY LOCALIZATION IN THE MONKEY
8000
8mm
M-13
15*000
-4mm
M-16
FIG. 22. Cochlear locations of the regions of threshold responses at
15, 8, 4 and 2 kHz in the monkey as determined in experiments with
kanamycin and neomycin [18].
cells taking over from 50 dB to the higher sound levels. Our
data form only one line of evidence in support of this notion.
Note for example in Figs. 10-13 in those basal portions of the
cochlea, where only inner hair cells remain, there is about a
50 dB hearing loss at the frequencies corresponding to those
locations on the basilar membrane. We suggest that the inner
hair cells are viable and therefore responsible for the residual
hearing in those ears. If the inner hair cells are also removed
as in Fig. 9 all traces of hearing disappear and the impairment
is complete. An intermediate example with hearing loss at
some frequencies greater than 50 dB accompanied by inner
hair cell loss is seen in Fig. 14.
SUMMARY
Behavioral methods for the evaluation of the effects of
toxins on hearing in experimental animal models have been
described. Measures of threshold and differential sensitivity,
of frequency selectivity, and of loudness discrimination were
discussed. Representative findings from guinea pig, chin-
chilla, cat, and monkey were shown following treatment with
one of the aminoglycosidic antibiotics or exposure to intense
sound. The behavioral methods should not be considered
limited to hearing testing for they have more general appli-
cation in the assessment of toxic effects on sensory systems
in a wide variety of mammalian species. Further, experi-
ments such as those described herein may yield significant
information concerning the basic mechanisms underlying
sensory processing in the unimpaired or normal animal.
REFERENCES
1. Clark, W. W., C. L. Clark, D. B. Moody and W. C. Stebbins.
Noise-induced hearing loss in the chinchilla determined by a
positive reinforcement technique. J. Ac-oust. Soc. Am. 56:
1202-1209, 1974.
2. Hawkins, J. E., Jr. and L.-G. Johnsson. Microdissection and
surface preparations of the inner ear. In: Handbook of Auditory
and Vestibitlar Research Methods, edited by C. Smith and J.
Vernon. Springfield, 111.: Charles C. Thomas, 1976, pp. 5-52.
3. Hawkins, J. E., Jr., W. C. Stebbins, L.-G. Johnsson, D. B.
Moody and A. Muraski. The patas monkey as a model for di-
hydrostreptomycin ototoxicity. Acta. Otolaryngol. 83: 123-129,
1977.
4. Hoekstra, A. and R. J. Ritsma. Perceptive hearing loss and
frequency selectivity. In: Psychophysics and Physiology of
Hearing, edited by E. F. Evans and J. P. Wilson. New York:
Academic Press, 1977, pp. 263-271.
5. Miller, J. D. Audibility curve of the chinchilla. J. Acoust. Soc.
Am. 48: 513-523, 1970.
6. Moody, D. B. Reaction time as an index of sensory function. In:
Animal Psychophysics: The Design and Conduct of Sensory
Experiments, edited by W. C. Stebbins. New York: Appleton-
Century-Crofts, 1970, pp. 277-303.
7. Moody, D. B. Behavioral studies of noise-induced hearing loss
in primates: loudness recruitment. In: Advances in
Otorhinolaryngology: Otophysiology, edited by J. E. Hawkins,
Jr. and M. Lawrence. Basel: Karger, 1973, pp. 82-101.
8. Moody, D. B., M. D. Beecher and W. C. Stebbins. Behavioral
methods in auditory research. In: Handbook of Auditory and
Vestibular Research Methods, edited by C. Smith and J. Ver-
non. Springfield. 111.: Charles C. Thomas, 1976, pp. 439-497.
9. Orr, J. L., D. B. Moody and W. C. Stebbins. Behavioral system
and apparatus for tone detection and choice reaction times in
the cat. J. Acoust. Soc. Am. 62: 1268-1272, 1977.
10
Petersen, M. R., C. A. Prosen, D. B. Moody and W. C. Steb-
bins. Operant conditioning in the guinea pig. J. exp. Analysis
Belwv. 27: 529-532, 1977.
Pfingst, B. E., R. Hienz, J. Kimm and J. Miller. Reaction-time
procedures for measurement of hearing. I. Suprathreshold
functions. J. Acoust. Soc. Am. 57: 421-430, 1975.
Prosen, C. A., M. R. Petersen, D. B. Moody and W. C. Steb-
bins. Auditory thresholds and kanamycin-induced hearing loss
in the guinea pig assessed by positive reinforcement procedures.
J. Acoust. Soc. Am. 63: 559-566, 1978.
Serafin, J. V., D. B. Moody and W. C. Stebbins. Psychophysi-
cal tuning curves in monkeys. J. Acoust. Soc. Am. 63: S30,
1978.
14. Stebbins, W. C. Auditory reaction time and the derivation of
equal loudness contours for the monkey. J. exp. Analysis Be-
hav. 9: 135-142, 1966.
Stebbins, W. C. Studies of hearing and hearing loss in the mon-
key. In: Animal Psychophysics: The Design and Conduct of
Sensory Experiments, edited by W. C. Stebbins. New York:
Appleton-Century-Crofts, 1970, pp. 41-66.
Stebbins, W. C. Hearing of Old World monkeys (Cer-
copithecinae). Am. J. Phys. Anthro. 38: 357-364, 1973.
Stebbins, W. C., C. H. Brown and M. R. Petersen. Sensory
processes in animals. In: Handbook of Physiology, Sensory
Processes, Vol. I, edited by I. Darian-Smith, J. Brookhart and
V. B. Mountcastle, section eds. Washington, D.C.: American
Physiological Society, 1979, in press.
Stebbins, W. C., W. W. Clark, R. D. Pearson and N. G. Wei-
land. Noise- and drug-induced hearing loss in monkeys. In: Ad-
vances in Otorhinolaryngology: Otophysiology, edited by J. E.
Hawkins, Jr. and M. Lawrence. Basel: Karger, 1973, pp. 42-63.
Stebbins, W. C. and S. Coombs. Behavioral assessment of
ototoxicity in nonhuman primates. In: Behavioral Toxicology.
New York: Plenum Press, 1976, pp. 401-427.
11
12
13
15
16
17
18
19
-------
44
STEBBINS AND MOODY
20. Stebbins, W. C., S. Green and F. L. Miller. Auditory sensitivity
of the monkey. Science 153: 1646-1647, 1966.
21. Stebbins, W. C., J. E. Hawkins, Jr., L.-G. Johnsson and D. B.
Moody. Hearing thresholds with outer and inner hair cell loss.
Am. J. Otolaryngol., 1979, in press.
22. Stebbins, W. C. and J. M. Miller. Reaction time as a function of
stimulus intensity for the monkey J exp. Analysis Behav. 7:
309-312, 1964.
23. Stebbins, W. C., J. M. Miller, L.-G. Johnsson and J. E. Haw-
kins, Jr. Ototoxic hearing loss and cochlear pathology in the
monkey. Trans. Am. Otol. Soc. 57: 110-128, 1969.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 45-52. ANKHO International Inc., 1979.
Trialwise Tracking Method for Measuring
Drug-Affected Sensory Threshold Changes in
Animals1
KIYOSHI ANDO2 AND KOHJI TAKADA
Department of Psychopharmacology, Preclinical Research Laboratories
Central Institute for Experimental Animals, 1433 Nogawa, Takatsu-ku, Kawasaki, Japan 213
ANDO, K. AND K. TAKADA. Trialwise tracking method for measuring drug-affected sensory threshold changes in
animals. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 45-52, 1979.—Rats and rhesus monkeys were trained under a
multiple schedule, the components of which were random ratio schedules for food presentation and for shock presentation.
The discriminative stimulus for the shock presentation component was a pure tone for the rats and a light for the rhesus
monkeys. In the test session under the extinction condition for the shock presentation component, the intensity of the
discriminative stimulus was successively either decreased by fixed units when the conditioned suppression was observed
or increased when the conditioned suppression was not observed. The levels finally oscillated within a narrow range around
the threshold. The auditory thresholds of rats were increased by intramuscular administration of quinidine at 20 mg/kg and
also by repeated intramuscular administration of kanamycin at 250 and 500 mg/kg/day. In rhesus monkeys, visual
thresholds were raised by application of pilocarpine at 0.02-0.16 mg/kg to the eyes and also by subcutaneous administration
of LSD-25 at 4-8 ^ig/kg in one monkey and at 20-30 /^.g/kg in another. The method used for tracking the animals' sensory
thresholds was sensitive enough to test the selective effect of the drugs and was also a relatively easy way to obtain a stable
behavioral baseline for experimental purposes.
Trialwise tracking method Auditory threshold Quinidine Kanamycin Rat Visual threshold
Pilocarpine LSD-25 Rhesus monkeys
ACCORDING to clinical reports, a number of drugs have
been found to affect sensory functions. For example,
aminoglycoside antibiotics (e.g., kanamycin), diuretics (e.g.,
ethacrynic acid), salicylates, quinidine and quinine, etc.,
have toxic effects on auditory function while cholinomimetic
drugs (e.g., pilocarpine), LSD-25, chloroquine (an antimala-
rial drug), ethanbutol (an antitubercular drug), etc., have var-
ious effects on visual function [10,15].
It is important to know whether such effects on sensory
functions as have been observed in man can be reproduced
in animal experiments. One must determine whether sensory
toxic effects similar to those observed in clinical situations
may be obtained in animal experiments, whether effects are
specific to a certain sensory function and are reversible,
what relationship exists between the effect and the condi-
tions of administration (e.g., dose, route, whether single or
repeated administration), and what the mechanism of the
action is. To answer these questions, there is an urgent need
to establish appropriate behavioral as well as physiological
and histopathological methodologies. A variety of behavioral
methods have been used in this area. Some examples
are the unconditioned pinna reflex elicited by tonal stimulus
in guinea pigs [1], the pole climbing response in rats [8], and
conditioned inhibition of cold shivering in refrigerated guinea
pigs [2,9], all of which have been used to test the ototoxic
effects of antibiotics. Operant procedures such as multiple
schedules [12, 20, 21], choice situations [4,23], an avoidance
situation [18] and a conditioned suppression paradigm [6]
have also been used to test drug effects on sensory functions.
However, these methods do not directly test the drug effects
on sensory thresholds. In contrast, there exists a more
sophisticated operant approach called the "tracking"
method (also known as the "stair case," "up and down" or
"titration" method) which does test drug-affected sensory
threshold changes directly. In this method, the stimulus pre-
sentation is essentially response-dependent. For example,
one type of response leads to increments in the intensity of a
stimulus while another leads to decrements, and the conse-
quence is that the increments and decrements gradually tend
to oscillate within a narrow range about the threshold value
'This research was supported by a special fund from the Science and Technology Agency of Japan for promoting multiministerial projects
1973-1975 concerning "Studies on the Regeneration of Nervous Tissue". The authors wish to thank Dr. Tomoji Yanagita for his comments on
the study and also Mr. Edgar M. Cooke for his suggestions in writing the manuscript.
2Requests for reprints should be sent to: Dr. K. Ando, Department of Psychopharmacology, Preclinical Research Laboratories, Central
Institute for Experimental Animals, 1433, Nogawa, Takatsuku, Kawasaki, Japan 213.
45
-------
46
ANDO AND TAKADA
[5,24]. In the application of this type of method in drug
studies, increased auditory threshold changes have been
successfully observed after repeated administration of
kanamycin in monkeys [26,27], and in rats [16], and in-
creased visual threshold changes have also been observed
after single administration of LSD-25 in pigeons [3] and after
repeated administration of trans ll-amino-10, ll-dihydro-5-
(3-dimethylaminopropyl)-5, 10-epoxy-5H-dibenzo[a,d]-cyclo-
heptene dihydrochloride in dogs [17]. Although fewer trials
are required with the tracking method because most stimuli
are presented near the threshold level, it usually requires a
long training period as well as a complex apparatus [5].
However, there have been some studies in which sensory
thresholds were tested in animals using tracking methods
adopting a conditioned suppression paradigm [16, 19, 22, 24,
25, 28]. In this so-called trialwise tracking approach [5], the
easy establishment and very effective stimulus control of the
conditioned suppression provide a considerable advantage.
The purpose of the present study was to test the effects of
quinidine and kanamycin on the auditory threshold in rats
and also the effects of pilocarpine and LSD-25 on the visual
threshold in rhesus monkeys by using a trialwise tracking
method adopting a conditioned suppression paradigm.
EXPERIMENT 1
The effect of a single administration of quinidine and of
repeated administration of kanamycin on auditory threshold
were tested in rats using a trialwise tracking method.
METHOD
Animals
Six experimentally naive, adult male Wistar rats were
used. They were maintained at approximately 80% of their
free-feeding body weights throughout the experiment.
Apparatus
Two operant-conditioning chambers (Lehigh Valley Elec-
tronics Inc., Model 143-22) mounted in sound-attenuating
cubicles (40x112x40 cm) were used. Each chamber had a
lever which was located 6 cm above the floor and 7.5 cm to
the right of the vertical midline of the front wall. A minimum
force of 20 g was required to operate the lever. A food pellet
weighing 45 mg was dispensed into a hopper from a dis-
penser (Lehigh Valley Electronics Inc., Model 114-20) lo-
cated behind the front wall of the chamber. An interrupted
pure tone (10 KHz) used as the test stimulus was generated
by an oscillator (Nagashima Medical Instruments Co.,
Model PA) through a speaker (Foster Electric Co., Model
UP-163). The speaker was located 60 cm away from the front
wall towards the rear. The rear wall was made of wire mesh
to allow the tone to pass through freely. The tone was pre-
sented for 1 sec intervals separated by 1 sec intervals of
silence. The intensity of the tone was adjusted manually by
turning the oscillator dial, which had been calibrated by a
sound level meter (Briiel and Kjaer Co., Model 2107) at-
tached to a condenser microphone (Briiel and Kjaer Co.,
Model 4144) which was located in the middle of the chamber
and was pointed toward the speaker. A green lamp (dia. 1.2
cm, 24 VDC) amounted 4 cm above the lever was used as a
control discriminative stimulus (the control light). Two mA
of electric shock were delivered to the grid floor of the
chamber via a shocker. A fluorescent lamp attached to the
sound-attenuating cubicle provided continuous illumination
in the chamber. Experimental contingencies and data record-
ing were arranged by a PDP 8/1 computer (Digital Equipment
Co.). Cumulative response recorders (Ralph Gerbrands Co.,
Model C-3) were also used. This controlling and recording
equipment was located in the adjacent room.
Procedure
Training under the multiple schedule. Rats were trained
to press a lever for a food pellet under a random ratio
schedule in which lever pressing was reinforced randomly on
the average of every 20 presses (RR 20 schedule). After re-
sponding by the rats stabilized under the RR 20 schedule for
food reinforcement, training under a multiple schedule [13]
was begun. A discriminative stimulus, either the pure tone
(frequency, 10 KHz; sound pressure level, 60 dB re 0.0002
dyne/cm2) or the control light was presented for 1 min every
10 reinforcements while the rats were responding for food
under the RR 20 schedule. During presentation of the dis-
criminative stimulus, responding was shocked under the RR
20 schedule instead of delivery of a food pellet (mult RR 20
[food] RR 20 [shock] schedule). One training session under
the multiple schedule included 10 to 15 food presentation
periods (food periods) and the same number of discrimina-
tive stimulus presentation periods with shock (stimulus
w/shock periods) occurring alternately and beginning with a
food period. Two of the stimulus w/shock periods were cho-
sen randomly to be governed by the control light and all the
other stimulus w/shock periods occurred in conjunction with
the tone. This multiple schedule training was held every day
except Sunday.
Threshold test. The auditory threshold test sessions were
given after a stable high rate of responding during the food
periods and suppression of responding in the stimulus
w/shock periods were observed. These test sessions were
generally given once a week avoiding Monday, the rest of the
week being occupied by training sessions under the multiple
schedule. The procedure for the test sessions was the same
as for training sessions except that responding by the rat
during presentation of the discriminative stimulus was not
shocked (stimulus periods), and the intensity of the test tone
in each stimulus period was changed according to the degree
of conditioned suppression recorded during each proceeding
tone presentation. The degree of the conditioned suppres-
sion was expressed by the suppression ratio (SR), defined as
follows:
A: Response rate during the last stimulus period.
B: Response rate during the food period between the last two
stimulus periods.
The sound pressure of the tone for each stimulus period was
decreased by 4, 6, or 10 dB steps when the SR was greater
than or equal to 0.90 and was increased by 4, 6, or 10 dB
steps when the SR was less than 0.90. The levels finally
oscillated within a narrow range around the threshold. The
intensity of the control light was kept constant in Experi-
ment 1.
Drug tests. After consistent and reproducible tracking
data were obtained in the threshold test, the drug test ses-
-------
TRIAL WISE TRACKING METHOD
47
sions were conducted in the same manner as the threshold
test sessions. In the drug test involving quinidine using 2
rats, vehicle only, carboxymethylcellulose in saline at 0.5%,
was administered at 1 ml/kg IM, 95 min before the first session.
A week later, quinidine sulfate (Hoei Yakko Co.) dissolved in
the same vehicle was administered at 20 mg/kg and tested
in the same manner. Between these two test sessions, the
rats were trained under the mult RR 20 [food] RR 20 [shock]
schedule every day except Sunday. The same two rats were
used again in the kanamycin test two months later, as no
residual effect of quinidine was observable.
The drug test sessions involving kanamycin were con-
ducted with three groups of two rats each. These included a
saline control group, a kanamycin low-dose group and a
kanamycin high-dose group. Saline at a volume of 1 ml/kg
was administered intramuscularly to the rats in the saline
control group at 10:00 a.m. every day except Sundays.
Kanamycin sulfate (Meiji) was administered to the rats in the
kanamycin low-dose group at 250 mg/kg in the same manner
as in the saline control group. The kanamycin high-dose
group received an additional dose of 250 mg/kg at 4:00 p.m.
The other experimental conditions were the same as in the
kanamycin low-dose group. For each rat the drug test ses-
sion was conducted once every 2 to 10 days at noon over
several weeks. Between test sessions the rats were trained
on the mult RR 20 [food] RR [shock] schedule every day
except Sunday. The threshold test session was conducted
again one month after the last saline or kanamycin adminis-
tration. Between the last administration and the following
threshold test session, rats were trained under the RR 20
schedule for food every day except Sunday.
RESULTS
After lever pressing under the continuous reinforcement
schedule had been established, 8-12 sessions were required
to obtain stable responding under the RR 20 schedule for
food. After 76-111 sessions of training under the mult RR 20
[food] RR 20 [shock] schedule, a stable and high rate of
responding (45.7-95.7 R's/min) during the food periods as
well as conditioned suppression during the stimulus w/shock
periods (SR&0.99) was observed in all rats. In the threshold
test, the intensity of the tone initially decreased from 60 dB
by several steps and then oscillated around certain values
(e.g., 20-30 dB with R 504, see Fig. 1). This stabilized track-
R504
20 dBSPL
CO
E
Ot o
!
3 01
C
m >>
OJ —
CO
oo
60
50
40
30
20
10
10
20
30 40
Time (min)
50
60
70
FIG 1 Sample cumulative response record (upper panel) and tracking of the sound pressure level of the stimulus tone
(lower panel) by Rat No. 504 in the threshold test. The sound pressure level, indicated under the offset of event pen in the
upper panel, was successively decreased when the conditioned suppression was observed, or increased when the con-
ditioned suppression was not observed. A light, indicated as L under the offset of event pen, was used as the control
stimulus. The data points in the lower panel correspond to the sound pressure levels directly above in the upper panel.
After several presentations of the tone, the levels oscillated between 20 and 30 dB.
-------
48
ANDO AND TAKADA
60
50
40
30
20
10
0
60
50
40
30
20
10
R504
o— Vehicle
o— Quinidine 20mg/kg, i.m.
R506
Vehicle
—o— Quinidine 20mg/kg, i.m.
100 120 140 160 180
Time after administration (min)
200
FIG. 2. Effect of intramuscular administration of quinidine at 20
mg/kg to rats on the tracking of the stimulus tone sound pressure
level. Compared to the vehicle, the levels in both rats increased with
quinidine.
ing within a range of less than 10 dB was observed 20-30 min
after the start of the session in all rats.
In all rats, neither the level nor the range of the stabilized
tracking were affected by administration of the vehicle. With
quinidine at 20 mg/kg, the level of the tracking was higher
than for the vehicle between 100 min and 160 min after ad-
ministration in R 504 and between 135 min and 160 min after
administration in R 506 (Fig. 2). In both rats, the level re-
covered to the vehicle level about 170 min after administra-
tion. The conditioned suppression with the control light was
not attenuated either after vehicle administration or after
quinidine administration in either rat (SRs=0.95).
The effects of repeated administration of kanamycin on
auditory threshold are presented in Fig. 3. The threshold was
herein defined as the minimum sound pressure level of the
tone after the 4th presentation of the tone in the session. The
threshold did not change after several weeks of repeated
administration of saline in either rat. However, the threshold
increased during test sessions on the 15th day of administra-
tions of 250 mg/kg/day IM of kanamycine in R 510 and on the
24th day in R 509. In the test session on the first day of
administration with the kanamycin high-dose group, rats
were tested at noon after the first dose of 250 mg/kg IM of
kanamycin and the second dose of 250 mg/kg was not given
until 4:00 p.m. R 506 failed to respond during the test session
after this first administration of kanamycin. With kanamycin
at 500 mg/kg/day IM, the threshold increased on the 13th day
in R 506 and on the 19th day in R 508. One month after the
last administration, the threshold was still at the same level
as that of the last administration for each rat. Throughout the
test sessions with all the rats in the kanamycin test, the con-
ditioned suppression was generally not attenuated by the
control light (SR&0.90). However, the SR for the control
light in one out of two stimulus periods in the test session
was sometimes less than 0.90 in all rats of both the kanamy-
cin low- and high-dose groups.
EXPERIMENT 2
The effects on visual threshold of the application of
pilocarpine to both eyes and of the subcutaneous adminis-
tration of LSD-25 were tested in rhesus monkeys using a
trialwise tracking method similar to that used in Experi-
ment 1.
METHOD
Subjects
The same two male rhesus monkeys were used in both the
pilocarpine test and the LSD-25 test. They had been used for
the conditioned emotional response experiment after Estes
and Skinner [11]. Their body weights at the start of the ex-
periment were 6.0 kg for M239 and 7.1 kg for M543. Both
monkeys were housed in individual living cages. They were
fed a total of 80 g of food per day throughout the experi-
ments.
Apparatus
During the experimental session, the monkeys were re-
strained in a primate chair in a darkened experimental room.
In front of the monkey was a panel which contained a lever,
a food tray, and a red lamp which was used as a stimulus
light. The lever was made of stainless steel (5x5x0.3 cm)
and required a minimum force of 50 g to operate. The food
tray was placed 7 cm under the lever and was dimly illu-
minated by a small lamp. The red lamp (dia. 1.5 cm) was lo-
cated 20 cm away from a point midway between the mon-
key's eyes. A pellet dispenser (Ralph Gerbrands Co., Model
A) was used to dispense a soybean (about 0.2 g) into the food
tray. The luminance of the stimulus light was changed by
adjusting the voltage supplied to the red lamp by manual
operation of the dial of a variable resistor. The luminance of
the red lamp at each voltage was calibrated by a Pritchard
photometer (Photo-Research Co., Spectra), the data values
being represented in millilamberts (mL). An interrupted pure
tone (4 KHz) as the control discriminative stimulus (the con-
trol tone) was generated by a home-made oscillator through
two speakers each located 30 cm away from the monkey's
ears. The tone was presented for 0.5 sec intervals separated
by 0.5 sec intervals of silence. Masking noise was presented
throughout the sessions through the same speakers. Neither
the intensity of the tone nor of the masking noise was cali-
brated; however, the tone could easily be detected through
the masking noise. Three mA of electric shock were deliv-
ered to the monkey's tail. The experimental contingencies
and data recording were arranged as in Experiment 1.
Procedure
Training under the multiple schedule. Monkeys were
trained under the same schedule as in Experiment 1 (i.e., the
mult RR 20 [food] RR 20 [shock] schedule). Each training
-------
TRIALWISE TRACKING METHOD
49
60
40
20
Saline Iml/kg/day, i.m.
—o— R 504
—• - R514
o o
03
T3
T3
*O
01
i.
60
40
20
Kanamycin 250mg/kg/day, i.m
R509
R510
Kanamycin 500mg/kg/day, i.m.
20
Pre 1
10 20
Days of administration
30
Post
FIG. 3. Effect of intramuscular administration of kanamycin on the auditory threshold in rats. The threshold was defined as
the minimum sound pressure level after the 4th presentation of the tone in the session. By repeated administration,
kanamycin at 250 mg/kg/day and at 500 mg/kg/day increased the threshold level while saline did not. One month after the
last administration of kanamycin, the increased threshold levels were still maintained (data points marked "Post").
-------
50
ANDO AND TAKADA
session included 24 stimulus w/shock periods. In the 3rd, the
11th, and the 19th stimulus w/shock periods, the control tone
was used while in the other periods, the light (a red lamp
supplied with 50 VAC) was used. In this training as well as in
the threshold test and the drug tests described below, the
monkey was kept in the primate chair and dark-adapted to
the experimental room for 20 min before the session. The
other procedures for the training under the multiple schedule
were the same as in Experiment 1.
Threshold test. After both a stable, high rate of respond-
ing during the food periods and suppression of responding in
the stimulus w/shock periods were observed, the visual
threshold test sessions were given in the same manner as in
Experiment 1 except that the intensity of the test light was
changed by adjusting the voltage to the lamp by either 1, 2.5,
5, or 10 V steps. The threshold for luminance was deter-
mined in the same manner as in Experiment 1. The intensity
of the tone was kept constant throughout Experiment 2.
Drug tests. After consistent and reproducible tracking
data were obtained in the threshold test, the effect of the
drug was tested. In the pilocarpine test, saline was applied to
both eyes at a volume of 0.15 ml per eye 25 min prior to the
test session. Following this, pilocarpine hydrochloride (Mo-
han Yakuhin Co.) dissolved in saline was tested at several
doses once a week in the same manner as in the saline test
session. The doses tested were 0.16 mg/kg in M239, and 0.02,
0.04, 0.08, and 0.16 mg/kg in M543. The pupil size of the left
eye was measured in both monkeys after each test session.
Between test sessions, monkeys were trained under the mult
RR 20 [food] RR 20 [shock] schedule every day except Sun-
day.
In LSD-25 test sessions, the vehicle, tartaric acid diluted
in saline at a concentration of 0.0048% was administered
subcutaneously at 0.5 ml/kg, 25 min prior to the test session.
Each dose of LSD-25 dissolved in the vehicle was tested
once a week in the same manner as the vehicle test session.
The doses tested were 10, 20 and 30 /ug/kg in M239, and 2, 4,
and 8 /u.g/kg in M 543. In the days following the LSD-25 test
sessions, the threshold test session was given again without
any administration to check the effect of LSD-25 after 24 hr.
Between drug test sessions, the monkeys were trained under
the mult RR 20 [food] RR 20 [shock] schedule every day
except Sunday. The interval between the pilocarpine test
and the LSD-25 test was 2 weeks or longer.
RESULTS
After 10 sessions of training under the RR 20 schedule for
food and 10-13 sessions of training under the mult RR 20
[food] RR 20 [shock] schedule, a stable and high rate of
responding (118.5-158.2 R's/min) during the food periods as
well as conditioned suppression during the stimulus w/shock
periods (SR&0.99) were observed in both monkeys. In the
threshold test, the luminance of the light initially decreased
from 0.64 log mL (50 VAC) by several steps and then oscil-
lated around certain values (e.g., -3.79 to -4.88 log mL in
M239 and -4.04 to -4.82 log mL in M543). The levels and
the range of the stabilized tracking were not affected by ad-
ministration of the vehicle in either the pilocarpine test or in
the LSD-25 test in either monkey.
Pilocarpine applied to the eyes decreased the pupil size.
The pupil sizes were 5.0 mm in M239 and 4.5 mm in M543
after saline administration and decreased to 1.0 mm at 0.16
mg/kg in M239, and for M543, 1.5 mm at 0.02 mg/kg, 1.3 mm
M239
-2
-6
Vehicle
Pilocarpine 0.16mg/kg/eye
M543
-2
-4
-6
- Vehicle
Pilocarpine 0.02mg/kg/eye
Pilocarpine O.OSmg/kg/eye
20 40 60 80 100
Time after administration
120
FIG. 4. Effect of application of pilocarpine to the eyes of rhesus
monkeys on the tracking of the luminance of the stimulus light.
Compared to the vehicle, pilocarpine slightly increased the levels in
both monkeys.
at 0.04 mg/kg and 1.0 mm at 0.8 and 1.6 mg/kg of pilocarpine.
As shown in Fig. 4, the levels of the tracking for the light
were slightly higher after pilocarpine administrations than
after saline administration. The levels of the tracking after
pilocarpine at 0.04 and at 0.16 mg/kg in M543 are not pre-
sented in Fig. 4 as they were almost identical to that of
pilocarpine at 0.08 mg/kg.
Subcutaneous administration of LSD-25 increased the
tracking levels at 20 and 30 ,u,g/kg in M239 and at 4 ,u,g/kg and
8 /xg/kg in M543 (Fig. 5). The increased levels at 4 and 8
iu,g/kg in M543 recovered to the vehicle level 24 hr after
administration. Although data are not shown, no remarkable
difference in the levels of the tracking compared with the
vehicle levels was observed at 10 /^g/kg in M239 or at 2 /ng/kg
in M543.
In both the pilocarpine test and the LSD-25 test with the
two monkeys, the conditioned suppression for the control
-------
TRIALWISE TRACKING METHOD
M239
51
01
o
-2
-4
-6
Vehicle
LSD-25 20pg/kg, s.c.
LSD-25 30pg/kg, s.c.
O)
u
c:
M543
-2
-4
-6
o— Vehicle
o— LSD-25 4yg/kg, s.c.
•— LSD-25 Syg/kg, s.c.
VWW
20
40 60 80 100 120 140
Time after administration (min)
160
180
FIG. 5. Effect of subcutaneous administration of LSD-25 on ths tracking of the luminance of the stimulus light in rhesus
monkeys. Compared to the vehicle, LSD-25 increased the levels in both monkeys.
tone was not attenuated either after vehicle administration or
after drug administration (SR^O.95 except in LSD-25 at 20
/Ag/kg in M239 where SR was 0.89).
DISCUSSION
One of the advantages of the tracking method lies in its
power to assure reliability with a relatively small set of data
points. However, the tracking method may require long
training sessions and complex apparatus. A combination of a
tracking procedure and a conditioned suppression paradigm
seems to produce some advantages both because con-
ditioned suppression may be well established and because
the baseline under this paradigm is stable within each session
and over a series of sessions. As the presentation of the
stimulus at a certain intensity is given trialwise in the present
tracking procedure, the experimenter has enough time to
adjust the dial of the stimulus generator manually. An appa-
ratus providing automatic variation of stimulus intensity is
not indispensable with trialwise tracking as it is with con-
tinuous tracking [5].
In the present experiments, the levels of tracking in the
threshold tests were stabilized within a narrow range after
several stimulus presentations. This tracking range was stable
-------
52
ANDO AND TAKADA
for each animal over all threshold test sessions (e.g., the
ranges of the stabilized tracking were 10 dB in Experiment 1
and 1 log mL in Experiment 2). To ensure that the present
method was specifically evaluating sensory threshold, the
effect of eardrum lesions was tested in rats. In the threshold
test just after the lesion, both a high rate of responding dur-
ing the tone presentation periods and complete conditioned
suppression during the light presentation periods were ob-
served, and the tracking levels in this test session were
above 50 dB. When the threshold test session was given in
the darkened room to a non-dark-adapted monkey, the levels
of tracking were higher than in the test sessions with dark
adaptation, while stabilized tracking and complete suppres-
sion to the tone were observed. This evidence with rats and
monkeys may indicate that the present method is valid
enough to measure sensory threshold specifically.
In the drug tests, both single administrations of quinidine
and repeated administrations of kanamycin increased audi-
tory threshold in rats, and single administrations of pilocar-
pine or of LSD-25 increased visual threshold in rhesus mon-
keys. As the conditioned suppression to the control stimulus
was not markedly attenuated in the present experiments, the
action of drugs tested were different from benzodiazepines
which have general attenuating effects on conditioned sup-
pression [7,14].
The auditory threshold change with quinidine observed in
rats was in agreement with the change previously observed
in a monkey [27]. While the effect of quinidine was revers-
ible, the effect with repeated administration of kanamycin
was irreversible, because hearing loss with kanamycin is
caused by the loss of receptor cells of the cochlea [26]. Audi-
tory threshold changes with kanamycin were detected on the
13th day at 500 mg/kg and the 20th day at 250 mg/kg.
Gourevitch [16] detected changes after several months of
repeated administration of kanamycin at 100 or 200
mg/kg/day in rats, and Stebbins [26] after two months of
repeated administration of kanamycin at 100 mg/kg/day to
monkeys. The visual threshold changes with LSD-25 in
monkeys were more marked than those produced by the
miotic agent, pilocarpine. Threshold increase with LSD-25
was also reported by Blough [3] for pigeons at 100 /ag/kg PO
or IP. Thus it can be seen by comparison of the above with
the results reported in the present experiments, that in the
kanamycin test our procedure effected a saving of time,
while in the LSD test the drug effect was observed at a
considerably lower dose, although in the latter case the spe-
cies difference of the test animals must also be considered.
The trialwise tracking method used was sensitive enough to
test the selective effect of drugs on sensory thresholds and
was also a relatively easy way to obtain a stable behavioral
baseline for experimental purposes.
REFERENCES
1. Akiyoshi, M. and K. Sato. Reevaluation of the pinna reflex test
as screening for ototoxicity of antibiotics. In: Proceedings of the
6th International Congress of Chemotherapy. Tokyo: Univer-
sity of Tokyo Press, 1970, pp. 621-627.
2. Anderson, H. and E. Wedenberg. A new method for hearing
tests in the guinea pig. Acta otolar. 60: 375-393, 1965.
3. Blough, D. S. Effect of lysergic acid diethylamide on absolute
visual threshold of the pigeon. Science 126: 304-305, 1957.
4. Blough, D. S. Some effects of drugs on visual discrimination in
the pigeon. Ann. N.Y. Acad. Sci. 66: 733-739, 1957.
5. Blough, D. and P. Blough. Animal psychophysics. In: Hand-
book ofOperant Behavior, edited by W. K. Honig and J. E. R.
Staddon. Englewood Cliffs, New Jersey: Prentice-Hall, Inc.,
1977, pp. 514-439.
6. Chiba, S. and K. Ando. Effects of chronic administration of
kanamycin on conditioned suppression to auditory stimulus in
rats. Jap. J. Pharmac. 26: 419-426, 1976.
7. Cook, L. and A. B. Davidson. Effects of behaviorally active
drugs in a conflict-punishment procedure in rats. In: The Ben-
zodiazepines, edited by S. Garattini, E. Mussini and L. O. Ran-
dall. New York: Raven Press, 1973, pp. 327-345.
8. Courvoisier, S. and O. Leau. Study of the onset of deafness in
rats treated with streptomycin, dihydrostreptomycin, and
neomycin. Antibiotics and Chemother. 6: 411-420, 1956.
9. Crifo, S. Shiver-audiometry in the conditioned guinea-pig
(Simplified Anderson-Wedenberg test). Acta otolar. 75: 38-44,
1973.
10. Darcy, P. F. and J. P. Griffin, latrogenic Disease. London:
Oxford University Press, 1972, pp. 151-164.
11. Estes, W. K. andB. F. Skinner. Some quantitative properties of
anxiety. J. exp. Psychol. 29: 390-400, 1941.
12. Evans, H. L. Scopolamine effects on visual discrimination:
Modifications related to stimulus control. J. Pharmac. exp.
Ther. 195: 105-113, 1975.
13. Ferster, C. B. and B. F. Skinner. Schedules of Reinforcement.
New York: Appleton-Century-Crofts, 1957, pp. 503-579.
14. Geller, I., J. T. Kulak and J. Seifter. The effects of chlor-
diazepoxide and chlorpromazine on a punishment discrimina-
tion. Psychopharmacologia 3: 374-385, 1962.
15. Goodman, L. S. and A. Oilman. The Pharmacological Basis of
Therapeutics, 5th edition. New York: The Macmillan Company,
1975.
16. Gourevitch, G. Detectability of tones in quiet and in noise by
rats and monkeys. In: Animal Psychophysics: The Design and
Conduct of Sensory Experiments, edited by W. C. Stebbins.
New York: Plenum Press, 1970, pp. 67-97.
17. Hanson, H. M. Psychophysical evaluation of toxic effects on
sensory systems. In: Behavioral Pharmacology: The Current
Status, edited by B. Weiss and V. G. Laties. New York: Plenum
Press, 1976, pp. 195-205.
18. Hearst, E. Drug effects on stimulus generalization gradients in
the monkey. Psychopharmacologia 6: 57-70, 1964.
19. Heffner, R., H. Heffner and B. Masterton. Behavioral mea-
surements of absolute and frequency-difference thresholds in
guinea pig. J. acoust. Soc. Amer. 49: 1888-1895, 1971.
20. Laties, V. G. and B. Weiss. Influence of drugs on behavior
controlled by internal and external stimuli. J. Pharmac. exp.
Ther. 152: 388-396, 1966.
21. Laties, V. G. The modification of drug effects on behavior by
external discriminative stimuli. J. Pharmac. exp. Ther. 183:
1-13, 1972.
22. Ray, A. R. Psychophysical testing of neurologic mutant mice.
In: Animal Psychophysics: The Design and Conduct of Sensory
Experiments, edited by W. C. Stebbins. New York: Plenum
Press, 1970, pp. 99-124.
23. Reiter, L. W., G. M. Talens and D. E. Woolley. Parathion
administration in the monkey: Time course of inhibition and
recovery of blood cholinesterases and visual discrimination per-
formance. Toxic, appl. Pharmac. 33: 1-13, 1975.
24. Rosenberger, P. B. Response-adjusting stimulus intensity. In:
Animal Psychophysics: The Design and Conduct of Sensory
Experiments, edited by W. C. Stebbins. New York: Plenum
Press, 1970, pp. 161-184.
25. Rosenberger, P. B. and J. T. Ernest. Behavioral assessment of
absolute visual thresholds in the albino rat. Vision Res. 11:
199-207, 1971.
26. Stebbins, W. C. Studies of hearing and hearing loss in the mon-
key. In: Animal Psychophysics: The Design and Conduct of
Sensory Experiments, edited by W. C. Stebbins. New York:
Plenum Press, 1970, pp. 41-66.
27. Stebbins, W. C., W. W. Clark, R. D. Pearson and N. G. Wei-
land. Noise- and drug-induced hearing loss in monkeys. Adv.
Otorhino-laiyng. 20: 42-63, 1973.
28. Sidman, M., B. A. Ray, R. L. Sidman and J. M. Klinger. Hear-
ing and vision in neurological mutant mice: A method for their
evaluation. Exp/ Neural. 16:377-402, 1966.
-------
est Methods for Definition of Effects of Toxic Substances on Behavior anil Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 53-66. ANKHO International Inc., 1979.
Motor Activity: A Survey of Methods with
Potential Use in Toxicity Testing
LAWRENCE W. REITER AND ROBERT C. MACPHAIL
Neurotoxicology Division, Health Effects Research Laboratory, U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
REITER, L. W. AND R. C. MACPHAIL. Motor activity: A survey of methods with potential use in toxicity testing.
NEUROBEHAV. TOXICOL. 1: Suppl. 1, 53-66, 1979.—Activity measurements are expected to have widespread use in
toxicity testing. The multifaceted nature of motor activity will directly influence the selection of a measurement technique
since the relative contribution of various motor acts to any particular measurement will depend upon the detection method.
Because of the apparatus-dependent nature of motor activity measurements, it is recommended that consideration be given
to how accurately the various devices measure locomotor activity. In the present paper, two types of body movement will
be considered as locomotor activity: ambulation (horizontally directed movement) and rearing (vertically directed move-
ment). Discussion focuses on the various methods currently used to record motor activity, the various components of
motor activity which are likely to be recorded, and the advantages and disadvantages of these techniques for the measure-
ment of locomotor activity. Finally, consideration is given to studies which have compared treatment effects on motor
activity derived from two or more measurement techniques.
Motor activity Toxicity testing Measurement technique
THE Toxic Substances Control Act specifically includes be-
havioral disorders as a health effect for which standard test
data are to be developed. Behavioral testing will be used to
evaluate both new substances and substances currently in
commerce. Although it is likely that the strategy for these
two test situations may be quite different, the actual test
systems employed may overlap. New substances need to be
tested for their neurotoxic properties. In this case, a series of
behavioral tests is needed which will provide a reliable fore-
cast of neurotoxicity. The greatest concern here is that a
neurotoxicant not go undetected. On the other hand, testing
of substances currently in use should focus on more clearly
defining the neurotoxicity and should, therefore, be con-
cerned with characterizing the types of behavioral changes
associated with exposure along with the factors which influ-
ence these changes. Nevertheless, in both situations, there is
a need for behavioral test systems which are reliable, sensi-
tive and cost-effective.
It is very likely that motor activity will be widely used in
evaluating toxic substances, see [44, 79, 88]. In the present
paper, discussion of locomotor activity will focus on two
types of body movement: ambulation (horizontal movement)
and rearing (vertical movement). Next we will consider meas-
urement techniques that are used with rodents including
both observational and automated techniques. Consideration
will be given to the behavioral components that contribute to
different measures of activity including the fidelity with
which the devices measure locomotor activity. Finally, we
will examine the issue of cross reliability by reviewing exper-
iments which compared two or more measures of motor ac-
tivity.
GENERAL MOTOR ACTIVITY VS LOCOMOTOR
ACTIVITY
Motor activity is not a unitary class of behavior. The
terms "general motor activity" and "spontaneous motor ac-
tivity" refer to the numerous motor acts, occurring either
alone or in combination, which constitute an animal's behav-
ioral repertoire. Quantitation of the general activity level of a
rodent requires measurement of the total frequency of acts,
such as walking, rearing, sniffing, grooming, etc. Because of
the heterogeneous nature of general activity, it is doubtful
that a single measure could ever be developed. Skinner
[105], recognizing this point, concluded that "no attempt is
made to distinguish between the various forms that the ac-
tivity may take. Since each form should have its own units, a
quantitative measure of activity as a whole is practically im-
possible."
Many investigators have utilized observational tech-
niques to quantitate general activity. Draper [24], for exam-
ple, developed an extensive list of descriptive behavioral
units which he used to record the home-cage activity of rats
(see Table 1). These behavioral units were divided into three
categories: ambulatory, non-ambulatory and inactivity. The
extensive (albeit incomplete) nature of this list should be
cause for concern when one considers that a toxicant could
differentially affect the various components of an animal's
general activity. Norton [70], for example, used time-lapse
photography to examine the effects of amphetamine on 15
components of general activity in rats. Her results demon-
strated (see Table 2) that amphetamine produced a differen-
tial effect on the frequency of occurrence of various motor
53
-------
54
REITER AND MACPHAIL
TABLE 1
BEHAVIORAL COMPONENTS OF HOME CAGE ACTIVITY OF RATS
Ambulatory Movements
(a) Vigorous
(1) Chase tail
(2) Climb
(3) Jump
(4) Roll over
(5) Run
(b) Non-Vigorous
(1) Circle
(2) Stand on cage
(3) Rear
(4) Stretch
(5) Walk
Inactivity
(a) "Relaxed" position
(1) Head under
(2) Side
(b) "Alert" position
(1) Stomach
(2) Stand still
Non-Ambulatory Movements*
(a) Groom
(1) Bite coat
(2) Lick coat
(3) Genitals
(4) Nails
(5) Wash face
(6) Scratch
(7) Chew tail
(b) "Scanning" movements
(1) Sniff
(2) Turn head
(c) Displaced Activity
(1) Chew cage
(2) Chew chain
(d) Miscellaneous movements
(1) Shake
(2) Twitch
(3) Sneeze
(4) Yawn
-Rapid circular movement with tail in or near mouth
-Movement across sides or roof of cage, all four feet off floor
-Quick upward jump, all feet leave floor
-Turn onto back or completely over
-Rapid movement across floor of cage
-Slow circular shifting of position similar to that seen in dogs or cats prior to sleep
-Front feet raised off floor, placed on sides of cage
-Front feet raised off floor or held close to body
-Back arched, front, rear legs extended
-Any leg movement that moves animal slowly across floor, roof or sides of cage
-Head tucked under front feet, forehead on floor
-Head on side on floor, body curled around
-Animal on stomach, head erect, immobile
-Animal stands, body off floor, immobile
-Quick bite into fur
-Smoothing fur with tongue
-Bite or lick genital area
-Chewing feet or nails
-Stroking whiskers and/or face with one or both front feet
-Quick scratch of ear or coat with hind leg
-Nose wrinkles, audible sniff
-Horizontal or vertical head movement
-Chew bars of cage
-Chew chain holding bottle on cage
-Quick shake of entire body
-Small movement of portion of body
*Ambulatory and non-ambulatory movements are not mutually exclusive.
From: Draper (Behaviour 28: 280-293, 1967)
components of activity. Whereas behaviors such as walking
and rearing increased following amphetamine, others such as
grooming and sitting decreased, and still others including
turning, looking (stationary head orientation) and pawing
showed no consistent alteration. Since psychoactive chemi-
cals may differentially alter the frequencies of the motor
items comprising general activity, any attempt to quantitate
this activity, short of recording all motor components, may
be "practically impossible". Regardless of the technique
which is employed to measure activity, it is essential to
understand which components of motor activity contribute
to that measure.
-------
MOTOR ACTIVITY AND TOXICITY TESTING
55
TABLE 2
TOTAL OCCURRENCES OF BEHAVIOR ACTS IN IS MIN
8 Saline Control Rats
per Group
8 Amphetamine Rats
per Group (mg/kg)
Acts Increased
Bobbing
Rearing
Turning
Walking
Acts Decreased
Eating
Grooming
Patting
Scratching
Sitting
Smelling
Standing
Washing Face
A
by Amphetamine
211
1554
153
517
by Amphetamine
239
148
87
26
893
2358
4097
219
Acts Not Consistently Altered by
Head Turning
Looking
Pawing
872
1425
60
B
222
1785
167
417
116
139
85
39
684
2368
1473
265
Amphetamine
904
1265
60
C
206
1932
117
384
210
141
110
45
786
2377
3936
338
824
1189
37
0.25
A
293*
1935*
273*
745*
46*
70*
44*
18
461*
2314
3833*
205
993
1416
45
0.5
B
356*
1848
344*
926*
28*
19*
21*
6*
423*
1756*
3771*
159*
972
1491*
45
1.0
C
248
2906*
365*
1139*
0*
4*
14*
7*
138*
1719*
2658*
36*
902
747
31
*Significant difference (p«0.05) using f-test to compare 8 control rats with 8 amphetamine-treated rats
observed during the same 15 min
From: Norton (Physiol. Behav. 11: 181-186, 1973)
The multifaceted nature of motor activity impacts directly
on the use of automated measurement techniques, since the
relative contribution of various motor acts to any particular
measurement of activity will depend upon the detection
method. In this regard, Draper [24] cautioned that the use of
automated techniques has, on occasion, caused confusion
since the concept of activity has become apparatus bound,
and hence provides no common unit for purposes of com-
parison.
One major aspect of general motor activity is locomotor
activity which, in the strictest sense, is defined as movement
from one place to another. This movement can be either
horizontally directed ambulation, including walking and
running, or vertically directed including rearing, climbing,
and jumping. In this discussion locomotor activity will in-
clude these two types of movement although in most exper-
imental situations vertically directed movement is restricted
to rearing.
METHODS USED TO MEASURE MOTOR ACTIVITY
A. OBSERVATIONAL TECHNIQUES
To some extent, the direct observation of behavior should
always be included in toxicity testing. There will always be a
need for information on the animal's general condition. The
initial use of direct observation may provide important in-
formation on the toxicological properties of a compound
which in turn increases the likelihood of detecting behavioral
changes that may otherwise go unnoticed [39]. Toman and
Everett [126], for example, noted the characteristic hunch-
back posture of reserpine-treated rats which would have
gone undetected by most automated activity monitoring de-
vices. The important question here, however, concerns the
degree to which observational techniques may be employed
in critically evaluating toxicant-induced changes in behavior.
Observational techniques employ either quantitative or
qualitative measurements (Table 3). With the quantitative
approach, the frequency, duration and/or sequencing of var-
ious motor components of behavior are measured. Norton,
Mullenix and Culver [72], for example, have used this tech-
nique to evaluate activity in rats that had sustained brain
damage from x-irradiation, carbon monoxide and pallidal le-
sions. In their study, the frequency and duration of 15 motor
acts and the sequencing of pairs of acts were determined. An
intriguing finding was that disruptions of motor sequencing
were directly correlated with increases in photocell-activity
counts, indicating that the increased activity measured in a
photocell cage reflected a loss of the normal sequencing of
behavior.
The qualitative approach is used to collect data on the
presence or absence of certain components of activity. This
approach is best exemplified by check lists and rating scales
which have been developed by various investigators
[45,135]. Use of these qualitative observational techniques
has been limited primarily to the study of the acute effects of
psychoactive drugs. They are well suited for identifying
-------
56
REITER AND MACPHAIL
TABLE 3
OBSERVATIONAL TECHNIQUES
A. Type of Measurement
1. Quantitative - measures the frequency, duration, and/or
sequencing of various motor components of behavior
Norton (1973), Norton et al. (1976)
Premack (1965)
2. Qualitative - measures the presence or absence of certain
motor components of behavior, i.e. check lists or rating
scales
Weismann et al. (1966)
Irwin (1968)
B. Type of Environment
1. Home cage
Draper (1967)
2. Novel environment, i.e. open field, Y-mazes, etc.
Hall (1934)
See Reviews by Walsh and Cummins (1976) and
Archer (1973)
gross changes in behavior such as tremors, ataxia, or
stereotypies.
However, the use of observational techniques to detect
subtle changes in behavior has three major limitations: (1)
The techniques require considerable commitments of time
and manpower for the data-collection process. Depending on
the precision to which general activity is characterized, more
than one observer may be required [103]. (2) Because of
possible subjective influences on the data-collection process,
a considerable degree of technical competency is required to
insure reliability. The problem of interobserver reliability is
especially important when interlaboratory comparisons are
to be made. In this case, a minimal requirement is that strict,
objective criteria are used to identify the various motor com-
ponents of activity. (3) When direct observation of behavior
is employed, subject-observer interaction may be an impor-
tant consideration. The mere presence of the observer may
modify the animal's behavior. For example, McCall et al.
[64] demonstrated that the distribution of a rat's activity in
an open field tended to concentrate on the side closest to the
observer.
The use of video-tape recordings or time-lapse photogra-
phy has increased considerably in observational analysis. In
most instances, this has minimized the problem of subject-
observer interaction and also has provided a permanent re-
cord of behavior which can be used for standardizing obser-
vations, both within and between laboratories. Recent ad-
vances in computer-automated pattern-recognition tech-
niques have also been applied to behavioral analysis [73].
Since rigidly defined criteria are required for computer-
automated identification of the various motor components of
behavior, this has alleviated somewhat the problems of sub-
jectivity and labor-intensive data collection discussed above.
Experiments using observational techniques to measure
motor activity frequently remove the animal from the home
cage and observe its behavior in a novel environment. The
complexity of this environment has ranged from a flat sur-
TABLE 4
CLASSIFICATION OF OPEN-FIELD-DEPENDENT
PARAMETERS
I. Behavior
A. Whole or Major Body Movement
1. Type of movement
a. Distance covered per unit time
b. Time spent in ambulation
c. Rearing frequency
d. Escape attempts
e. Latency (usually time taken to leave start area)
f. Time spent without movement
2. Locations
a. Field area visited (inner or peripheral areas,
corners, etc.)
b. Affiliation (distance from partner subject)
c. Stimulus interaction (e.g., distance from
stimulus object)
B. Part Body Movement
1. Manipulation of objects
2. Sniffing
3. Scratching
4. Digging
5. Teeth chattering
6. Grooming
7. Vocalization
8. Visual exploration
II. Autonomic Nervous System
A. Defecation
B. Digestive Transit Time
C. Urination
D. Heart Rate and Rhythm
E. Respiratory Rate
III. Adrenal Activity
A. Adrenal Ascorbic Acid
B. Serum Corticosteroids
IV. Electrophysiology
A. Hippocampal Theta Activity
B. Electromyogram Activity
From: Walsh and Cummins (Psychol. Bull. 83: 482-504, 1976).
Copyright (1976) by the American Psychological Association.
Reprinted by permission.
face to a simple cubicle (resembling the home cage) to com-
plex mazes consisting of interconnecting alleys.
The most extensively employed test environment has
been the open field (for review see [3,132]). Much of its
popularity is almost certainly due to the simplicity of the
apparatus. With few exceptions, an open field is defined as
an arena that is considerably larger than the cage used to
house the animal [91].
Table 4, taken from a review by Walsh and Cummins
[132], lists the dependent variables which have been meas-
ured in the open field. The most widely recorded classes of
behavior include ambulation (often distinguished on the
basis of whether ambulation occurs in the inner or peripheral
-------
MOTOR ACTIVITY AND TOXICITY TESTING
57
areas of the arena), frequency of rearing and grooming, and
latency to first crossover. Typically, the floor of the open
field is marked off into a grid arrangement and ambulation is
measured by the number of times an animal enters a new
square (crossovers).
The simplicity of design and ease of use of the open field
almost certainly insure its continued use in behavioral
testing. However, the numerous factors, both environmental
and organismic, which influence open field measurements
(e.g., size, illumination, species, sex and experience) require
that investigators extensively report the methods employed.
Surprisingly little consideration has been given toward stan-
dardization of the open field. Walsh and Cummins ([132], p.
493) summarized the problem as follows:
"Almost every physical characteristic of the apparatus, its
surroundings and every procedural step have been widely
varied, so that although standardization may have been es-
tablished within individual laboratories, there is a disturbing
lack of conformity in procedure and results within the litera-
ture as a whole.
In fact, it is hard to think of any facet which has not been
modified. This difficulty of standardization is compounded
by the extreme rarity of reports which cite details of more
than a small proportion of relevant procedural variables."
Observational techniques have also been used with ro-
dents tested in a complex environment such as a Y-maze.
The maze consists of three alleys which join to form a Y.
With this apparatus Steinberg and coworkers [92, 93, 111]
have evaluated the effects of psychoactive drugs and past
experience on ambulation (alley entrances) and rearing.
B. AUTOMATED TECHNIQUES
A variety of techniques have been employed to automati-
cally record motor activity in rodents, the major classes of
which are listed in Table 5. Photocell devices, field detectors
and touch plates all provide direct measurements of activity
because they measure the movement of the animal. The me-
chanical devices, on the other hand, provide an indirect
measurement because they detect movement of the cage.
/. Photocell Devices
Photocell devices provide direct, relatively unambiguous
measures of motor activity in which beams of light traverse a
cage and impinge on photoreceptors. Any movement of an
animal which interrupts a beam is recorded as an activity
count; in most cases, the process of detection provides no
feedback to the animal which may affect subsequent activity.
Photocell chambers are stationary and therefore are free of
momentum (carry-over) effects. The use of an infrared light
source allows recording during the nocturnal period without
detection by the animal. The greatest limitation of this tech-
nique is the possible contamination of the data with non-
locomotor activities (e.g., movements of the head and paws)
when these occur immediately in front of a photobeam.
Photocell devices may be categorized primarily in terms
of the complexity of the test environment and, secondarily,
in terms of the number and orientation of the photobeams.
a. Simple test environments are typically circular or
rectangular and generally measure activity with only one or a
few photobeams. Siegel [101] is credited with using the first
photocell cage to monitor activity. This device consisted of a
rectangular box with a single photobeam oriented along the
TABLE 5
AUTOMATED TECHNIQUES
A. Photocell Measurements
1. Simple environment
a. Horizontal (Siegel, 1946; Dews, 1953)
b. Vertical (Schnitzer and Ross, 1960)
2. Complex environment
Norton et al. (1975)
Ljungberg and Ungerstedt (1977)
B. Mechanical Measurements
1. Stabilimeter
a. Jiggle cage
Richter (1927)
b. Tilt cage
Bousfleld and Mote (1946)
Campbell et al. (1961)
c. Force platform
Denenberg et al. (1975)
Segal and Mandell (1974)
2. Wheels
Stewart (1898)
Skinner (1933)
C. Field Detectors
1. Ultrasonic
Peacock and Williams (1962)
2. Capacitance of a resonant circuit
Svensson and Thieme (1969)
D. Touch plates (proximity counter)
Silbergeld and Goldberg (1973)
long axis. In 1953, Dews [23] employed a similar device to
record the effects of pharmacological agents on the motor
activity of mice (see also [108, 109, 129]). Numerous experi-
ments have been conducted by Iversen and her colleagues on
the effects of various treatments on the activity of rats in
rectangular cages with parallel photobeams horizontally di-
rected along the long axis (e.g. [18, 19, 54]). Rectangular
chambers have also been employed in which one or two
photobeams bisect the chamber across the short axis (e.g.,
[25, 60, 69]) and in which photobeams intersect at right an-
gles (e.g., [41, 89, 127]).
Several simple circular photocell chambers have been
used with a variety of photobeam arrangements (e.g.,
[29,133]). One chamber measures approximately 60 cm in
diameter and is transected by two sets of three beams that
intersect at right angles. Separate counts are available for the
two banks of photobeams. The device has been used exten-
sively to assess the effects of several chemical and nonchem-
ical treatments on motor activity (e.g., [14, 38, 40, 110]),
-------
58
REITER AND MACPHAIL
although lately it has been supplanted by more sophisticated
photocell chambers (see below).
There is a recurring concern over the importance of
photocell placement [60,133]. Watzman et al. [133], for
example, compared the effects of chlorpromazine on activity
of mice tested with two different photobeam arrangements
and tested either individually or in groups of five. When
three parallel beams transected the chamber, more interrup-
tions were recorded under control conditions from each of
the two peripheral beams than from the center one. Also, the
number of center beam interruptions was less than that ob-
tained when the beam was oriented at a right angle with a
second beam. Chlorpromazine decreased activity in a dose-
dependent fashion, and this decrease depended on both the
group size and on the particular photobeam arrangement. At
low doses of chlorpromazine a significant dose-response was
found for the crisscross photocell arrangement but not for
the parallel arrangement. Therefore, with a given chamber
configuration, the photocell arrangement may affect the
sensitivity of a measurement technique.
In the pieces of apparatus described above, the orienta-
tion of the photobeams is along the horizontal plane. An
alternative approach is to direct the photobeams downward
to the sensors. One of the earliest devices that used this
arrangement was described by Schnitzer and Ross [95] in
which light from a single overhead source impinged on
photocells located below a translucent plastic floor. Al-
though the device has the advantage of using only one light
source, thereby eliminating problems with photocell align-
ment, a possible limitation is that with too dense a con-
centration of photocells, beam interruptions could conceiv-
ably be produced by non-locomotor activities such as head
turning, tail swishes, and defecation [55]. This limitation
highlights a continuing concern with the use of photocell
chambers: whereas too few photobeams would allow signifi-
cant amounts of locomotor activity to go undetected, too
many photobeams increase the likelihood that non-
locomotor activities will contribute to the measurement.
Almost without exception, most simple photocell devices
permit recording of horizontally-directed locomotor activity
(ambulation) and exclude measurement of vertically-directed
activity (rearing). In addition, very little information is col-
lected regarding the spatial distribution of locomotor activ-
ity. There is at least one commercial photocell device, how-
ever, which can be used to measure both rearing and the
spatial distribution of activity. The Electronic Motility Meter
manufactured by Motron Products consists of a rectangular
platform with a translucent plastic floor with 40 photosen-
sors situated underneath. Infrared light is provided by a
single overhead source. With an appropriate data-collection
device, information can be collected on the spatial distribu-
tion of locomotor activity by recording and storing individual
photobeam interruptions. In addition, the device has an ad-
justable array of photobeams and sensors for detecting rear-
ing. The Motron device has been used for investigating the
effects of drugs and other variables on locomotor activity
(e.g. [1, 2, 4, 16, 33, 35, 36, 65, 77, 128, 130, 131]). Very few
of these studies, however, have collected rearing data, and
to date there has been no published investigation of treat-
ment effects on the spatial distribution of locomotor activity.
Although the device has many advantages over the other
simple photocell test chambers currently in use, the same
concerns regarding density of photocells that were discussed
above apply here. Additionally, widespread use of the Mo-
tron chamber has been hampered by its high cost. At present,
however, there has been at least one attempt to develop a
similar but much less expensive prototype [80].
Adaptation to these simple test environments occurs
rapidly, with animals showing initial high levels of activity
that decrease to asymptotically low levels after several (e.g.
10-60) minutes. Downward trends in activity levels are also
evident with repeated testing. Since asymptotic levels are
characteristically low, simple test environments are well
suited for detection of chemically induced increases in
locomotor activity and/or retardation of the within-session
habituation of activity.
b. Complex test environments. We consider a photocell
chamber to be complex if it contains more than a single
boundary. For example, a variety of otherwise simple
photocell test chambers include an obstruction in the center
area. One device is doughnut-shaped with the light source
located in the center of the annulus and an array of six
photosensors spaced equidistantly on the outer periphery.
The device was first described by Wright et al. [136] and was
manufactured as the Woodard Actophotometer. It has been
used to investigate the effects of drugs and biochemical le-
sions of the central nervous system on motor activity (e.g.
[14, 21, 62, 125, 137]). A related device was used by Krsiak
et al. [57] in which light impinged on two photocells that
intersected at right angles at the center of a square testing
area. Access to the center of the chamber was precluded by a
square translucent column, which was included to prevent
excess activity counts by an animal simultaneously interrupt-
ing both photobeams at or near their point of intersection.
It is important to note that devices such as that used by
Krsiak et al. [57] and Wright et al. [136] may have unique
operating characteristics as a result of this central obstruc-
tion. It is frequently demonstrated, for example, that the
locomotor activity of rodents tends to concentrate at the
periphery of the test area (i.e., "wall-seeking" behavior or
thigmotaxis which has been described by Barnett [4]; see
also [59]). Since devices with central obstructions provide
two boundaries for wall-seeking to find expression, it is
likely that the relationship between locomotor activity and
an independent variable differs from that found with devices
that permit access to the center of the test area.
Support for this notion comes from an experiment by
Stewart [113] on the effect of environmental complexity on
locomotor activity in control and in scopolamine-treated
rats. Although direct observation and not a photocell device
was used to measure activity, the results illustrate how a
complex boundary can modify behavior. Figure 1 shows the
within-session time course of locomotor activity for control
rats and rats treated with scopolamine (0.25 mg/kg). In a
simple open field, scopolamine-treated rats had a higher
level of activity than controls. However, when the complex-
ity of the environment was increased by adding walls which
divided the field into either two (Cl) or four (C2) com-
partments, the scopolamine-treated animals were less active
than controls. This difference in the drug response was due
to both an increase in the rate of habituation of drug-treated
animals and to a decrease in the rate of habituation of the
controls. The rate of habituation, therefore, was a function
of the complexity of the environment as well as the drug.
Norton et al. [71] have described a more complex photo-
cell chamber with two continuous alleys that connect to a
small central area and that form a figure eight when placed
on its side. Two additional blind alleys extend from the cen-
tral area and may provide stimuli such as water or a nest box.
Locomotor activity in the figure-eight alleys is detected by
-------
MOTOR ACTIVITY AND TOXICITY TESTING
59
o
V-
D
CD
5
80
60
40
20
SI
C1
C2
I I I I I I I I I I I I I I I I I I
5 10 15 5 10 15 5 10 15
15
O
oc
<
111
cc 6
3 -
5 10 15 5 10 15 5 10 15
TIME (minutes)
FIG. 1. Mean squares entered (ambulation) and rearing responses in
saline- and scopolamine-treated (0.25 mg/kg) rats tested in three
36x36 in. environments differing in complexity (SI, simple; Cl, one
insert: Cl, two interlocking inserts, in the field) during 2.5 min inter-
vals of the 15 min test session. (Redrawn from: Stewart, Neurosci.
Lett. 1: 121-125, 1975.)
six pairs of photobeam and sensor and an additional beam
and sensor detects excursions into each of the blind alleys.
The device has been used extensively by Norton and her
colleagues in studies of the effects on locomotor activity of
time-ofjday and social variables, as well as exposure to
psychoactive drugs and environmental pollutants [71, 72, 85,
86, 87, 88].
In light of the above discussion of environmental com-
plexity, it is interesting to compare the rate of habituation in
this test chamber with habituation in the open field. Whereas
Stewart [113] reported that habituation occurred within 10
min in a simple open field, this process extends over hours in
the maze. These two test environments may, therefore, be
expected to differ in their sensitivity to detect chemically
induced alteration in the habituation process.
Ljungberg and Ungerstedt [61J have recently described a
device which purports to measure several key aspects of
motor activity. The outer boundary of the chamber is square
with a square obstruction in the center. Horizontally
oriented photobeams and sensors detect ambulation and
vertically oriented beams and sensors record entries into the
corners of the chamber. The floor is replete with holes below
which horizontally oriented beams and sensors detect head
(and other types of) entries. Finally, a microphone is used to
detect gnawing behavior. Preliminary data suggest that the
device is capable of recording several behavioral effects pro-
duced by drugs that alter transmission in dopamine-
containing neurons. Although the device can be used to
measure the spatial distribution of activity, it does not record
rearing behavior. At this time, it remains to be seen whether
conclusions regarding the effects of chemicals on locomotor
activity obtained with this device are in any way more pre-
cise than those obtained with simpler devices.
2. Mechnical Measurements
The motor activity recorded with these devices involves a
vertical or horizontal displacement of the chamber in re-
sponse to the animal's movement. The displacement is
transduced to a digital electronic signal which is counted. As
with any type of mechanical measurement device, care must
be taken to insure that activity counts are not confounded by
momentum. Displacement of the chamber will depend in part
on the age and weight of the animal. Additionally, it should
be noted that mechanical devices may provide propriocep-
tive feedback that is positively correlated with motion.
a. Stabilimeters. Stabilimeters are so named because
movements of the animal cause the chamber to be displaced
from its resting position. Displacement may be in either the
horizontal or vertical plane.
1. Jiggle cages. Richter [90] described a triangular activity
measurement device for rats. Movements displaced the
chamber vertically and this was transduced pneumatically
and recorded on a revolving smoked-drum kymograph. This
type of jiggle cage does not allow separate determination of
rearing and ambulation, nor does it as a rule allow determi-
nation of an animal's location within the chamber. Depend-
ing on the sensitivity, non-locomotor activities may or may
not be recorded.
Another type of jiggle cage, the Williamson jiggle cage,
confines the animal in a very restrictive environment (ap-
proximately 15 cm in all directions). The transduction of
displacements is characteristically complex [123], which
makes calibration and, consequently, standardization dif-
ficult. Moreover, the device routinely detects fine-grained
non-locomotor activities (e.g., tremor and face-washing) in
rodents. The device has had only limited employment in the
past two decades (e.g. [34, 120, 123, 124]).
Davis and Ellison [20] have described a device in which
the floor of the test chamber rests on ball bearings, and
movement causes the floor to displace laterally. Transduc-
tion is accomplished by closing a circuit between a metal
plumb bob, that is normally centered (but not touching)
within a hole in the floor just outside the chamber, and a thin
film of metal that is wrapped around the rim of the hole. This
-------
60
REITER AND MACPHAIL
device suffers from the same limitations (e.g., carry-over
effects and dependence of displacement on body weight) de-
scribed above for most of the mechanical measurement de-
vices. In addition, since the device records only lateral dis-
placements it has an undefined relationship to the more
common devices which record vertical displacements.
2. Tilt cage. Perhaps the most widely used of the tilt cages
was originally described by Bousfield and Mote [16] and sub-
sequently modified by Campbell et al. [12]. A wire-mesh
rectangular cage rests on a central fulcrum that runs across
the chamber's width. Ambulation from one side to the other
causes a slight vertical displacement of the cage which ac-
tivates a microswitch. This type of tilt cage has been used
extensively to characterize the effects of development, de-
privation, temperature, time of day, drugs, and biochemical,
electrolytic and mechanical lesions of the central nervous sys-
tem (e. g. [8,11, 27, 28, 30, 48, 49, 66, 67]). The most positive
feature of this device is its simplicity and, consequently, its
ease of construction and standardization. Moreover, in con-
trast to the jiggle cage, it is insensitive to relatively minor
non-locomotor activities such as grooming and tremor. A
significant limitation, however, is that some locomotor ac-
tivity may go unrecorded if the animal does not cross from
one side the fulcrum to the other. Also, the device does not
record rearing, and provides only a crude estimate of the
spatial locus of activity. Finally, although momentum
(carry-over) effects are negligible, the device does provide
proprioceptive stimuli as positive feedback for ambulation.
Circular tilt cages have also been described. Eayrs [26],
for example, described a doughnut-shaped wire-mesh
chamber that was balanced on a central pivot. Movement
about the cage was detected by making and breaking electri-
cal connections between pairs of brass contacts that were
placed at 90° intervals underneath the chamber. Interesting-
ly, in a counterbalanced crossover design, Eayrs [26] found
no correlation between daily levels of activity of female rats
measured in the circular tilt cage and those measured in a
running wheel. Campbell [8] has also described a circular tilt
cage, but it has not received widespread use.
3. Force platforms. Force platforms are those in which
ambulatory movements cause very small displacements of
the floor of the chamber. Movements are either digitized to
increment counters or in other cases an analog signal is the
output. Segal and Mandell [100], for example, described a
chamber in which horizontal locomotor activity was re-
corded as crossovers on a floor that was electronically divided
into quadrants. Rearing was recorded by metal touch plates
located on the walls approximately 12 cm above the floor
(rearing that did not include wall contact went unrecorded).
That the device was relatively insensitive to small non-
locomotor activities was shown in the following way. Large
doses of ^/-amphetamine had a triphasic effect on activity; an
initial increase in locomotor activity was followed by a
period in which the rats exhibited stereotyped sequences of
sniffing, grooming and gnawing. During this phase of am-
phetamine intoxication, ambulation and rearing were virtu-
ally absent and no activity counts were recorded. After ap-
proximately 1.5-2 hr, stereotyped behavior subsided and
was accompanied by a return of locomotor activity. Al-
though the device has many positive features, it has not been
extensively used [97-100].
A particularly elegant force platform has recently been
described by Denenberg and co-workers [22]. A large square
test arena is supported on four corner-mounted elec-
tromechanical force transducers that provide analog outputs
which are proportional to their respective supportive forces.
With this device, and an appropriate data-acquisition sys-
tem, it is possible to record all horizontally directed move-
ments of the animal, however small. It is also possible to
measure the spatial distribution of ambulation and, conceiv-
ably, the exact path the animal takes in moving about the
chamber. Measurements have been shown to correlate well
with ambulation measured by an observer [37]; it has an
advantage over the open field of recording ambulation that
does not result in a grid crossover [22]. It does not, however,
permit registration of rearing. To date, there are no pub-
lished data on the effects of chemical or non-chemical treat-
ments on ambulation with this device and, consequently, its
utility for toxicity testing remains to be determined.
b. Wheels. Running wheels are ordinarily designed so that
the wheel is mounted on a horizontal axle and the animal's
ambulation within the wheel causes it to rotate on its axle.
The direction of rotation of the wheel may be unidirectional
or bidirectional, and transduction may be accomplished in
any number of ways [94, 106, 107]. Prevention of rotation
with a braking device is not uncommon [78,94]. Running
wheels have been used in animal behavior for over three-
quarters of a century and were originally described by
Stewart [112] in 1898. Skinner [106] provides an excellent
description of the construction of a running wheel and the
variables that influence the measurement of ambulation.
It should be obvious that ambulation recorded in running
wheels involves considerable neuromuscular coordination.
Also, there is considerable stimulation in the form of prop-
rioceptive feedback, that correlates positively with the level
of ambulation. The specialized nature of this activity is
perhaps best reflected in the time-course for establishing
stable levels of activity. In rats, activity in the running wheel
typically increases with repeated testing [15, 26, 31, 91] even
though the within-session trend is downward. Once asymp-
totic levels are reached, however, they may be very stable
for long periods [15,26], although fluctuations about the
mean are apparent, for example, with estrus or time of day.
Large individual differences in baseline levels may be appar-
ent [47]. Unless special precautions are taken, momentum
may influence activity counts.
Running wheels have been used to study the effects of
food deprivation, water deprivation, estrus, ambient tem-
perature and time of day, lesions of the central nervous sys-
tem, and a wide array of drugs on locomotor activity (e.g. [8,
32, 45-47, 51, 52, 78, 94, 96, 118, 122, 123, 134, 138]).
3. Field Detectors
Many devices have been introduced recently that record
the disturbances that an organism creates in moving about a
prearranged field within the test cage. Both ultrasonic and
capacitive fields have been employed and each has the ad-
vantage of being stationary and, therefore, providing no
feedback stimulation for activity. Measurements of activity
are also unconfounded by momentum effects. The principle
limitations of these techniques, however, have to do with
both the degree to which locomotor activity counts may be
inflated with non-locomotor activity data and the inability to
adequately calibrate these devices.
a. Ultrasonic field detectors. These devices deliver an
acoustic signal to the test chamber that creates a three-
dimensional pattern of standing waves (e.g. [74,75]). Move-
ments of an organism produce changes in the pattern of en-
ergy which are ordinarily transmitted to a microphone, di-
-------
MOTOR ACTIVITY AND TOXICITY TESTING
61
gitized and counted. The degree of pattern interference is
correlated with the speed and extent of body movement.
Typically, the sensitivity of the devices can be varied over a
wide range, as can the size of the test chamber. Problems
arise, however, over uncertainties regarding the uniformity
of the propagated field. Operating characteristics may
change with fluctuations in temperature and humidity, or the
deposition of fecal material. Standardization of one or more
instruments is extremely tedious. Artifacts may arise
through electrical interference with the signal that is sent to
the counter; for this reason, coaxial cable is required. De-
spite the limitations described above, these types of field
detectors may be of use especially when a measure of gen-
eral motor activity is desired.
b. Capacitance-sensing devices. These devices utilize a
tuned oscillator circuit. In the simplest case, an adjustable
oscillator supplies high-frequency current to an input coil
which creates a field around the test chamber. Movement
within the chamber produces momentary changes in the
voltage that is induced in the output coil. This signal is then
amplified and an output pulse is counted. Devices of this
type (e.g., Animex and Varimex) have been used extensively
to evaluate drug effects (e.g. [5, 50, 52, 63, 68, 81, 104, 116,
119]). Svensson and Thieme [121] have described the uses
and limitations of the Animex for recording motor activity in
control and drug-treated mice. Although commercially avail-
able models are being continually refined, many of the same
uses and limitations that characterize ultrasonic field sensing
devices are also applicable.
4, Touch Plates
The devices in this category are included because they do
not fit handily into any of the above categories. Although
these devices all measure motor activity by recording con-
tacts of the animal with sections of the chamber floor, they
differ in the means used to detect these contacts.
a. Proximity counter. The floor of this device is com-
prised of a rectangular checkerboard arrangement of
copper-clad phenolic plates that are separated by insulating
material. A low-voltage oscillating current is passed be-
tween alternate plates and is adjusted so that ambulation
from a ground plate to an active plate produces changes in
the frequency of the oscillating current. The magnitude of
this change in current is a function of the animal's
capacitance, and it is this change that is typically amplified,
transduced and counted. Although the principle of proximity
counters is not new [139], there are very few investigations
which have used them. The most notable instances are the
experiments carried out by Silbergeld and her colleagues on
the effect of postnatal lead exposure on the motor activity of
mice (e.g. [102]). Silbergeld's device records primarily
large-scale ambulatory activity; rearing would not be re-
corded unless it occurred concomitantly with a crossover
from one to another floor plate. Although the results of these
experiments have stimulated much research, no apparent at-
tempt has been made to critically evaluate the relationship
between the activity that is measured in a proximity counter
and that measured in one of the more conventional devices.
b. Touch plates. These devices represent modifications of
the chamber originally described by Kissel [56] and by Eayrs
[26]. The basic chamber consists of a small circular alley that
creates a somewhat restrictive and complex test environ-
ment for monitoring motor activity. Because of its restrictive
interior, it is likely that the age and size of the rodent will
determine whether ambulation is largely unidirectional or
bidirectional. The original chambers were suspended on a
central pivot, and in some instances delicately balanced with
counter weights [26]. Movement about the chamber pro-
duced displacements that operated microswitches which in
turn were used to increment counters. Subsequent modifica-
tions replaced the pivot arrangement with small floor plates
that were spaced equidistantly about the cage and which
rested on microswitches [83]. Although this type of chamber
detects ambulation and is relatively unaffected by the occur-
rence of non-locomotor activities, it has no means for record-
ing rearing. More importantly, however, its complex config-
uration, together with its restrictive interior, raise the
possibility that the effect of a treatment may be much more
complex than that determined in a larger chamber which has
no central obstruction. A salient example of the complex
nature of the locomotor activity recorded by this device is
the dose-response curve for rf-amphetamine [116]. In rats,
increasing doses of ^-amphetamine produce progressively
greater increases in locomotor activity until a dosage of 2-4
mg/kg is reached. Further dosage increases up to about 12
mg/kg produce either no greater increase or a decrease in
locomotor activity. In most instances, however, dosages of
12 mg/kg and beyond produce further increases in activity.
At these heroic dosages, it is likely that a measure of ambu-
lation reflects more the occurrence of the repetitive behav-
ioral sequences that are produced by of-amphetamine than an
effect of the drug on ambulation per se. The device has been
used in several studies on the effects of psychoactive and
biogenic amine-depleting drugs (e.g. [82, 83, 114-116]).
CROSS-COMPARISONS OF MOTOR-ACTIVITY-
MEASUREMENT DEVICES
A limited number of studies has compared two or more
measures of motor activity either in nontreated animals or in
animals subjected to different types of treatments.
Tapp et al. [123] compared several different measures of
motor activity in rats. Their results showed a lack of correla-
tion between the various activity measures shown in Table 6.
When, however, two measures were taken within the same
apparatus (i.e., measures of activity within the circular
field), a significant correlation was found. Otherwise, there
was generally no correlation between relative levels of ac-
tivity in the different test environments.
Tapp [124] has also investigated the effect of food depri-
vation on several measures of activity. Food deprivation did
not affect activity in a Williamson jiggle cage, and decreased
activity in both a simple photocell cage and a complex
doughnut-shaped mechanical device (circular field). There
are additional data that support the conclusion that the ef-
fects of food deprivation are apparatus-dependent. Strong
[117], for example, studied the effect of food deprivation on
stabilimeter activity under two conditions of measurement
sensitivity. Food deprivation did not affect activity when
measured in a microswitch-operated stabilimeter and de-
creased it when measured in a more sensitive stabilimeter
device. On the other hand, Weasner et al. [134] obtained
increases in photocell-cage activity and activity in a running
wheel with food deprivation. Campbell [8] also obtained in-
creases in activity in a running wheel and in a stabilimeter
during food deprivation. The latter two studies both obtained
greater increases in activity in the wheels than with the other
types of measurement device. Finger [32] has also shown
-------
62
REITER AND MACPHAIL
TABLE 6
MATRIX OF INTERCORRELATIONS BETWEEN ACTIVITY MEASURES
Test
Number
1
2
3
4
5
Test 1
Williamson Cages
Photocell Cages
Activity Wheel
Circular Field (alternations)
Circular Field (total counts)
234
-0.03 0.11 0.18
-0.02 -0.01
0.15
5
0.19
0.11
0.19
0.90*
*p<0.01 Adapted from Tapp, et al. (Psychol. Rep. 23: 1047-1050, 1968)
that substantially greater changes in activity occur during the
estrus cycle of rats when measured in wheels than in a simple
photocell device.
Animals treated with various drugs or toxicants have also
been shown to respond differently when tested in differ-
ent activity devices. Reiter et al. [86] reported changes in
activity of male rats exposed to the pesticide Kepone which
were apparatus-dependent. Following three weeks of treat-
ment, rats showed a dose-related increase in locomotor ac-
tivity in a figure-eight maze but a dose-related decrease in
activity in an open field (Fig. 2).
The data of Krsiak et al. [57] illustrate further the limita-
tions of a single measure of motor activity (Fig. 3). They
examined motor activity in rats following administration of
either c/-amphetamine (dexamphetamine) or amobarbital
(amylobarbitone). Animals were placed in a complex test
environment (described earlier) and activity was determined
simultaneously by recording photocell interruptions and by
direct observation of walking, rearing and grooming. In
saline-treated animals, the combined frequency of walking
and rearing closely agreed with activity levels determined by
photocell interruptions. In these animals, photocell counts
were significantly correlated with both walking and rearing
(r=0.68 and 0.79, respectively). However, with increasing
doses of rf-amphetamine, photocell activity counts exceeded
the observationed activity. Following 1.0 nig/kg of
of-amphetamine, walking (r=0.82) but not rearing (r=0.11)
was correlated with photocell counts. On the other hand,
following administration of amobarbital, there was a close
agreement between photocell interruptions and observa-
tional scoring of walking and rearing.
There also appear to be discrepancies between the effects
of drugs on activity measured in photocell cages and those
measured in capacitance-sensing devices. Maj et al. [63]
found that the activity of rats was about equally reduced
following reserpine or a-methyHcwa-tyrosine in both the
Animex chamber and a simple photocell chamber. A large
dose of /-dopa substantially blocked the decrease produced
by reserpine and by a-methyl-/?ara-tyrosine in activity
measured in the Animex but not that measured in the photo-
cell cage. Amantadine produced a dose-related blockade of
the effect of reserpine on activity in the Animex chamber
whereas the largest dose produced only a slight blockade of
the effect in the photocell chamber. Similarly, Ljungberg [60]
found that whereas a small dose of apomorphine decreased
activity in both a simple photocell cage and an Animex
chamber, a large dose had no effect on activity in the former
"1300 - RESIDENTIAL MAZE
£:i?nn —
8
O 900
5 800 -
'• I
OPEN FIELD —
0 40 80 0 40 80
DIETARY LEVELS OF KEPONE, ppm
FIG. 2. Effects of dietary kepone exposure (3 weeks) on two meas-
ures of locomotor activity in the rat. Kepone exposure produced a
dose-related increase in locomotor activity in a figure-eight maze but
a decreased activity in an open field. (From: Reiter et al., Toxic.
appl. Pharmac. 41: 143, 1977.)
device but increased it in the latter device. The effects of
several other drug treatments showed no correlation be-
tween activity measurements in the two types of devices.
Experimentally induced brain damage is another manipu-
lation which is reported to produce apparatus-dependent
changes in activity. Capobianco and Hamilton [13], for
example, employed several activity measures with rats sus-
taining lesions of the fornix, diagonal band, and the medial
forebrain bundle. Lesions of the medial forebrain bundle and
diagonal band reliably increased activity in a revolving wheel
whereas lesions of the fornix had a much smaller effect.
Lesions of the medial forebrain bundle and the diagonal
band, on the other hand, either had no effect or decreased
activity in a stabilimeter whereas fornix lesions substantially
increased activity. Measures of both open-field ambulation
and rearing were unaffected by these lesions. This experi-
ment suffers, however, from one possible source of experi-
mental error which was recognized by the authors: stabilime-
ter testing always preceded open-field testing. It is possible
that an "order effect" of testing could account for the lack of
a treatment effect on activity in the open field. If, for exam-
ple, lesions of the diagonal band altered the animal's reactiv-
-------
MOTOR ACTIVITY AND TOXICITY TESTING
63
0 0-25 0-5 1-0 2-0
DEXAMPHETAMINE, mg/kg
3-75 7-5 15-0 30-0
AMYLOBARBITONE, mg/kg
FIG. 3. Ten groups of 8 rats were injected subcutaneously either
with saline or a dose of a drug 35 min before a 10 min trial in the
activity cage. After dexamphetamine, marked discrepancies oc-
curred when the results with photocells were compared with obser-
vation (Fig. 3.a), but agreement was close after amylobarbitone (Fig.
3b). (Adapted from: Krsiakef al., Psychopharmacology 17: 258-274,
1970.)
ity to a novel environment, then a "carry over" effect from
stabilimeter to open field would preclude detection of an
effect in the latter.
Table 7 represents data obtained in adult male rats tested
on successive days in a figure-eight (Residential) maze and in
an open field. The order of testing was counterbalanced so
that half the animals were tested first in each apparatus.
Order of testing had a significant effect on activity in the
open field when animals had prior exposure to the maze but
the converse was not true. This potential for a carry-over
effect in open-field testing has obvious implications for ex-
periments using multiple test systems.
CONCLUSIONS
A. Direct observations are indispensible in preliminary
toxicity testing since many overt neurotoxic symptoms (i.e.,
tremor or ataxia) are readily detected by direct observation.
The problem with this approach is that the data are not
amenable to quantitative analysis.
B. Quantitative characterization of a chemical effect on
motor activity can be carried out using any number of non-
automated and automated techniques. Quantitative obser-
vational techniques may be used to detect subtle changes,
but they require personnel commitments, not found with
TABLE 7
EFFECT OF ORDER OF TESTING ON RESIDENTIAL MAZE
ACTIVITY AND OPEN FIELD ACTIVITY IN ADULT RATS
Open Field
First
Residential Maze
First
Open Field
(5-minutes)
t = 4.378
43df
p<0.005
Residential Maze
(5-minutes)
t= 1.33
43df
NS
Residential Maze
(120-minutes)
t= 1.116
43df
NS
67.5 ± 4.8
40.3 ± 3.8
automated techniques, which ultimately limit the number of
animals which can be tested.
C. Many automated techniques are available for record-
ing motor activity. The particular components of motor ac-
tivity which are measured depend on the particular tech-
nique chosen. Numerous chemical and nonchemical treat-
ments have been shown to selectively affect various compo-
nents of motor activity. It is imperative, therefore, to know
which components of activity a particular device is likely to
detect.
D. Relationships between devices will be understood only
when they are compared by measuring one or a few compo-
nents of motor activity. We recommend that different de-
vices be compared on the basis of their ability to measure
locomotor activity since this represents a major component
of motor activity and since most devices are designed to
detect this movement. This restricted focus on locomotor
activity is likely to uncover certain fundamental relation-
ships and/or differences between the devices. We recognize,
however, one possible shortcoming of this approach: since
treatments may selectively affect various components of
motor activity, too great a restriction on the variables to be
measured may decrease sensitivity of the measure.
E. A basic requirement of activity measurements is that
test methods be standardized at least within a given labora-
tory and that experimental details be extensively reported.
F. It is premature to conclude which of the many avail-
able activity measuring techniques is the most appropriate
for behavioral toxicity testing. Appreciation of the relative
sensitivity of the different devices will come when extensive
data have been collected under standardized laboratory
conditions. Future studies should, therefore, compare
different measures of motor activity following exposure to
neurotoxic substances.
-------
64
REITER AND MACPHAIL
REFERENCES
1. Ahlenius, S., N.-E. Anden and J. Engel. Restoration of
locomotor activity in mice by low /-dopa doses after suppres-
sion by a-methyl-tyrosine but not by reserpine. Brain Res. 62:
189-199, 1973.
2. Anden, N.-E., U. Strombom and T. H. Svensson. Dopamine
and noradrenaline receptor stimulation: Reversal of reserpine-
induced suppression of motor activity. Psychopharmacology
29: 289-298, 1973.
3. Archer, J. Tests for emotionality in rats and mice: A review.
Anim. Behav. 21: 205-235, 1973.
4. Barnett, S. H. The Rat: A Study in Behavior. Chicago: Aldine
Publishing Co., 1963, pp. 31-32.
5. Benkert, O. Measurement of hyperactivity in rats in a dose-
response-curve after intrahypothalamic norepinephrine injec-
tion. Life Sci. 8: 943-948, 1969.
6. Bousfield, W. A. and F. A. Mote. The construction of a tilting
activity cage. J. exp. Psychol. 32: 450-451, 1943.
7. Erase, D. A., H. H. Loh and E. L. Way. Comparison of the
effects of morphine on locomotor activity, analgesia and pri-
mary and protracted physical dependence in six mouse strains.
J. Pharmac. exp. Ther. 201: 368-374, 1977.
8. Campbell, B. A. Theory and research on the effects of water
deprivation on random activity in the rat. In: Thirst. Proceed-
ings from the 1st International Symposium on Thirst in the
Regulation of Body Water. Oxford: Pergamon Press, 1964, pp.
317-334.
9. Campbell, B. A. and G. A. Cicala. Studies of water deprivation
in rats as a function of age. J. comp. physiol. Psychol. 55:
763-768, 1962.
10. Campbell, B. A. and P. D. Mabry. The role of catecholamines
in behavioral arousal during ontogenesis. Psychopharmacol-
ogy 31: 253-264, 1973.
11. Campbell, B. A., L. D. Lytle and H. C. Fibiger. Ontogeny of
adrenergic arousal and cholinergic inhibitory mechanisms in
the rat. Science 166: 637-638, 1969.
12. Campbell, B. A., R. Teghtsoonian and R. A. Williams. Activ-
ity, weight loss and survival time of food deprived rats as a
function of age. J. comp. physiol. Psychol. 54: 216-219, 1961.
13. Capobianco, S. and L. W. Hamilton. Effects of interruption of
limbic system pathways on different measures of activity.
Physiol. Behav. 17: 65-72, 1976.
14. Carey, R. J. and A. P. Salim. Changes in rf-amphetamine reac-
tivity resulting from septal forebrain injury. Physiol. Behav. 5:
133-136, 1970.
15. Cornish, E. R. and N. Mrosovsky. Activity during food depri-
vation and satiation of six species of rodent. Anim. Behav. 13:
242-248, 1965.
16. Costa, E., A. Groppetti and M. K. Naimzada. Effects of (+)-
amphetamine on the turnover rate of brain catecholamines and
motor activity. Br. J. Pharmac. 44: 742-751, 1972.
17. Costa, E., K. M. Naimzada and A. Revuelta. Phenmetrazine,
aminorex and ( ± )-p-chloroamphetamine, their effects on
motor activity and turnover rate of brain catecholamines. Br. J.
Pharmac. 43: 570-579, 1972.
18. Creese, I. and S. D. Iversen. Blockage of amphetamine-
induced motor stimulation and stereotypy in the adult rat fol-
lowing neonatal treatment with 6-hydroxydopamine. Brain
Res. 55: 369-382, 1973.
19. Creese, I. and S. D. Iversen. The pharmacological and anatom-
ical substrates of the amphetamine response in the rat. Brain
Res. 83: 419-436, 1975.
20. Davis, J. D. and G. D. Ellison. A general purpose activity
recorder with variable sensitivity. J. exp. Analysis Behav. 7:
117-118, 1964.
21. Davis, W. M., M. Babbini, S. F. Pong, W. T. King and C. L.
White. Motility of mice after amphetamine: Effects of strain,
aggregation and illumination. Pharmac. Biochem. Behav. 2:
803-809, 1974.
22. Denenberg, V. H., J. Gartner and M. Myers. Absolute meas-
urement of open field activity in mice. Physiol. Behav. 15:
505-509, 1975.
23. Dews, P. B. The measurement of the influence of drugs on
voluntary activity in mice. Br. J. Pharmac. Chemother. 8:
46-48, 1953.
24. Draper, W. A. A behavioral study of the home cage activity of
the white rat. Behaviour 28: 280-293, 1967.
25. Dyer, R. S., D. A. Weldon and A. H. Thoumaian. Task-
specific effects of (/-amphetamine and blindness of BALB/cj
mice. Physiol. Psychol. 4: 175-179, 1976.
26. Eayrs, J. T. Spontaneous activity in the rat. Br. J. Anim. Be-
hav. 2: 25-30, 1954.
27. Erinoff, L., R. C. MacPhail, A. Heller and L. S. Seiden. Age-
dependent effects of 6-hydroxydopamine on locomotor activity
in the rat. Brain Res. 64: 195-205, 1979.
28. Fibiger, H. C. and B. A. Campbell. The effect ofpara-chloro-
phenylalatiine on spontaneous locomotor activity in the rat.
Neuropharmacology 10: 25-32, 1971.
29. Fibiger, H. C., C. Trimbach and B. A. Campbell. Enhanced
stimulant properties of (+)-amphetamine after chronic reser-
pine treatment in the rat: Mediation by hypophagia and weight
loss. Neuropharmacology 11: 57-67, 1972.
30. Fibiger, H. C., L. D. Lytle and B. A. Campbell. Cholinergic
modulation of adrenergic arousal in the developing rat. J.
comp. physiol. Psychol. 72: 384-389, 1970.
31. Finger, F. W. Measuring behavioral activity. In: Methods in
Psychobiology, Vol. 2, edited by R. D. Myers. New York:
Academic Press, 1972, pp. 1-19.
32. Finger, F. W. Estrus activity as a function of measuring device.
J. comp. physiol. Psychol. 55: 100-102, 1961.
33. Foldes, A. and E. Costa. Relationship of brain monoamine and
locomotor activity in rats. Biochem. Pharmac. 24: 1617-1621,
1975.
34. Freeman, J. J. and F. Sulser. Iprindole-amphetamine interac-
tions in the rat: The role of aromatic hydroxylation of am-
phetamine in its mode of action. J. Pharmac. exp. Ther. 183:
307-315, 1972.
35. Fuentes, J. A. and N. H. Neff. Selective monoamine oxidase
inhibitor drugs as aids in evaluating the role of Type A and B
enzymes. Neuropharmacology 14: 819-825, 1975.
36. Fuentes, J. A., M. A. Oleshansky and N. H. Neff. Comparison
of the apparent antidepressant activity of (-)- and (+)-
tranylcypromine in an animal model. Biochem. Pharmac. 25:
801-804, 1976.
37. Gartner, J., M. Myers, V. H. Denenberg and S. Bhat. A
center-of-pressure force platform for open field activity tests.
Trans. Second New Engl. Bioengineer. Conf., 449-457, 1975.
38. Glick, S. D. and S. Millory. Rate-dependent effects of
(/-amphetamine on locomotor activity in mice: Possible rela-
tionship to paradoxical amphetamine sedation in minimal brain
dysfunction. Eur. J. Pharmac. 24: 266-268, 1973.
39. Goldstein, A., L. Aronow and S. M. Kalman. Principles of
Drug Action: The Basis of Pharmacology, 2nd edition. New
York: John Wiley and Sons, 1974, p. 735.
40. Gordon, J. H. and M. K. Schellenberger. Regional catechol-
amine content in the rat brain: Sex differences and correlation
with motor activity. Neuropharmacology 13: 129-137, 1974.
41. Grant, M. Cholinergic influences on habituation of exploratory
activity in mice. J. comp. physiol. Psychol. 86: 853-857, 1974.
42. Hollister, A. S., G. R. Breese and B. R. Cooper. Comparison
of tyrosine hydroxylase and dopamine-j8-hydroxylase inhibi-
tion with the effects of various 6-hydroxydopamine treatments
on (/-amphetamine-induced motor activity. Psychopharmacol-
ogy 36: 1-16, 1974.
43. Hollister, A. S., G. R. Breese, C. M. Kuhn, B. R. Cooper and
S. M. Schanberg. An inhibitory role for brain serotonin-
containing systems in the locomotor effects of (/-amphetamine.
J. Pharmac. exp. Ther. 198: 12-22, 1976.
-------
MOTOR ACTIVITY AND TOXICITY TESTING
65
44. Human Health and the Environment-Some Research Needs.
U.S. Department of Health, Education and Welfare, Public
Health Service/National Institutes of Health. Washington,
B.C.: U.S. Government Printing Office, 1977.
45. Irwin, S. Comprehensive observational assessment: la. A sys-
tematic, quantitative procedure for assessing the behavioral
and physiologic state of the mouse. Psychopharmacology 13:
222-257, 1968.
46. Irwin, S. The action of drugs on psychomotor activity. Rev.
Can. Biol. 20: 239-250, 1961.
47. Irwin, S., M. Slabok and G. Thomas. Individual differences: I.
Correlation between control locomotor activity and sensitivity
to stimulant and depressant drugs. J. Pliannac. exp Ther. 123:
206-211, 1958.
48. Jacobs, B. L., W. D. Wise and K. M. Taylor. Differential
behavioral and neurochemical effects following lesions of the
dorsal or median raphe nuclei in rats. Brain Res. 79: 353-361
1974.
49. Jacobs, B. L., C. Trimbach, E. E. Eubanks and M. Trulson.
Hippocampal mediation of raphe lesion and PCPA-induced
hyperactivity in the rat. Brain Res. 94: 253-261, 1975.
50. Jenner, P. and C. D. Marsden. The influence of piribedil
(ET495) on components of locomotor activity. Eur. J. Phar-
mac. 33:211-215, 1975.
51. Karczmar, A. G., C. L. Scudder and D. L. Richardson. Inter-
disciplinary approach to the study of behavior in related mice
types. In: Neurosciences Research, Vol. 5, Chemical Ap-
proaches to Brain Function, edited by I. Kopin and S. Ehren-
pries. New York: Academic Press, 1973, pp. 159-244.
52. Kavanau, J. L. and D. H. Brant. Wheel-running preferences of
Peromyscus. Nature 208: 597-598, 1965.
53. Kellogg, C. and P. Lundborg. Ontogenic variations in re-
sponses to /-dopa and monoamine receptor-stimulating agents.
Psychophannacology 23: 187-200, 1972.
54. Kelly, P. H., P. W. Seviour and S. D. Iversen. Amphetamine
and apomorphine responses in the rat following 6-OHDA le-
sions of the nucleus accumbens septi and corpus striatum.
Brain Res. 94: 507-522, 1975.
55. Kinnard, W. J. and N. Watzman. Techniques utilized in the
evaluation of psychotropic drugs on animal activity. J. Phar-
mac. Sci. 55: 995-1012, 1966.
56. Kissel, J. W. Nutating annular cage for measuring motor activ-
ity. Science 134: 1224-1225, 1963.
57. Krsiak, M., H. Steinberg and I. P. Stolerman. Uses and limita-
tions of photocell activity cages for assessing effects of drugs.
Psychopharmacology 17: 258-274, 1970.
58. Kuschinsky, K. Are cholinergic mechanisms involved in mor-
phine effects on motility? Naunyn-Schmiedeberg's Arch.
Pharmac. 281: 167-173, 1974.
59. Lat, J. The spontaneous exploratory reaction as a tool for
psychopharmacological studies. A contribution towards a
theory of contradictory results in psychopharmacology. In:
Pharmacology of Conditioning, Learning and Retention,
edited by M. Mikhel'son, V. G. Longo and Z. Votava. New
York: Pergamon Press, 1965, pp. 47-66.
60. Ljungberg, T. Reliability of two activity boxes commonly used
to assess drug-induced behavioral changes. Pharmac.
Biochem. Behav. 8: 191-195, 1978.
61. Ljungberg, T. and U. Ungerstedt. A method for simultaneous
recording of eight behavioral parameters related to monoamine
neurotransmission. Pharmac. Biochem. Behav. 8: 483-489,
1978.
62. Maickel, R. P., R. M. Levine and J. E. Zabik. Differential
effects of d- and /-amphetamine on spontaneous motor activity
in mice. Res. comnums. chem. pathol. Pharmac. 8: 711-714,
1974.
63. Maj, J., B. Durek and W. Palider. Differences m locomotor
activity'of mice as measured by an Animex and photoresistor
actometer. J. Pharmac. Pharmac. 23: 979-981, 1971.
64 McCall R. B., M. L. Lester and C. M. Corter. Caretaker ef-
fects in'rats. Devi. Psychol. 1: 771, 1969.
65 Modigh, K. Central and peripheral effects of
5-hydroxytryptophan on motor activity in mice. Psychophar-
macology 23: 48-54, 1972.
66. Moorcroft, W. H. Ontogeny of forebrain inhibition of behav-
ioral arousal in the rat. Brain Res. 35: 513-522, 1971.
67. Moorcroft, W. H., L. D. Lytle and B. A. Campbell. Ontogeny
of starvation-induced behavioral arousal in the rat. J. comp.
physiol. Psychol. 75: 59-67, 1971.
68. Mukherjee, B. P., P. T. Bailey and S. N. Pradhan. Correlation
of lithium effects on motor activity with its brain concentra-
tions in rats. Neuropharmacology 16: 241-244, 1977.
69. Nagy, Z. M. and M. Ritter. Ontogeny of behavioral arousal in
the mouse: Effects of prior testing upon age of peak activity.
Bull. Psychon. Soc. 7: 285-288, 1976.
70. Norton, S. Amphetamine as a model for hyperactivity in the
rat. Physiol. Behav. 11: 181-186, 1973.
71. Norton, S., B. Culver and P. Mullenix. Measurement of the
effects of drugs on activity of permanent groups of rats.
Psychopharm. Commitn. 1: 131-138, 1975.
72. Norton, S., P. Mullenix and B. Culver. 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, 1976.
73. Norton, S. and W. Servos. Pattern recognition of rat behavior
using a laboratory computer. 5th Int. Congress of Pharmacol-
ogy, July 1972.
74. Peacock, L. J. and M. Williams. An ultrasonic device for re-
cording activity. Am. J. Psychol. 75: 648-652, 1962.
75. Peacock, L. J., M. H. Hodge and R. K. Thomas. Ultrasonic
measurements and automatic analysis of general activity in the
rat. J. comp. physiol. Psychol. 62: 284-288, 1966.
76. Pirch, J. H. and R. H. Rech. Behavioral recovery in rats during
chronic reserpine treatment. Psychopharmacology 12: 115-
122, 1968.
77. Post, R. M. and H. Rose. Increasing effects of repetitive
cocaine administration in the rat. Nature 260: 731-732, 1976.
78. Premack, D. Towards empirical behavioral laws: I. Positive
reinforcement. Psychol. Rev. 66: 219-233, 1959.
79. Principles for Evaluating Chemicals in the Environment. Na-
tional Academy of Sciences, Washington, D.C., 1975.
80. Rafales, L. S., R. L. Bomschein, I. A. Michaelson, R. K. Loch
and G. F. Barker. Drug induced activity in lead-exposed mice.
Pharmac. Biochem. Behav. 10: 95-104, 1979.
81. Rastogi, R. B. and R. L. Singhal. Lithium: Modification of
behavioral activity and brain biogenic amines in developing
hyperthyroid rats. J. Pharmac. exp. Ther. 201: 92-102, 1977.
82. Rech, R. H. Antagonism of reserpine behavioral depression by
d-amphetamine. J. Pharmac. exp. Ther. 146: 369-376, 1964.
83. Rech,R. H.,L. A. CarrandK. E. Moore. Behavioral effects of
a-methyltyrosine after prior depletion of brain catecholamines.
J. Pharmac. exp. Ther. 160: 326-335, 1968.
84. Reed, J. D. Spontaneous activity of animals: A review of the
literature since 1929. Psychol. Bull. 44: 393-412, 1947.
85. Reiter, L. W., G. E. Anderson, J. W. Laskey and D. F. Cahill.
Developmental and behavioral changes in the rat during
chronic exposure to lead. Envir. Hlth Perspect. 12: 119-123,
1975.
86. Reiter, L. W., K. Kidd, G. Ledbetter, L. Gray and N. Cher-
noff. Comparative behavioral toxicology of mirex and Kepone
in the rat. Toxic, appl. Pharmac. 41: 143, 1977.
87. Reiter, L. W. Behavioral toxicology: Effects of early postnatal
exposure to neurotoxins on development of locomotor activity
in the rat. J. occup. Med. 19: 201-204, 1977.
88. Reiter, L. W. Use of activity measures in behavioral toxicol-
ogy. Envir. Hlth Perspect. 26: 9-20, 1978.
89. Remington, G. and H. Anisman. Genetic and ontogenetic var-
iations in locomotor activity following treatment with
scopolamine or rf-amphetamine. Devi. Psychobiol. 9: 579-585,
1976.
90. Richter, C. P. Animal behavior and internal drives. Q. Rev.
Biol. 2: 307-343, 1927.
91. Robbins, T. W. A critique of the methods available for the
measurement of spontaneous motor activity. In: Handbook of
Psychopharmacology, Vol. 7, edited by L. L. Iversen, S. D.
Iversen and S. H. Snyder. New York: Plenum Press, 1977. pp
37-82.
-------
66
REITER AND MACPHAIL
92. Rushton, R., H. Steinberg and C. Tinson. Effects of single
experience on subsequent reaction to drugs. Br. J. Pharmac.
20: 99-105, 1963.
93. Rushton, R., H. Steinberg and M. Tomkiewicz. Equivalence
and persistence of the effects of psychoactive drugs and past
experience. Nature 220: 885-889, 1968.
94. Schaeffer, R. W. A new device for programming contingencies
between drinking, running and level pressing. J. exp. Analysis
Behav. 9: 3-24, 1963.
95. Schnitzer, S. B. and S. Ross. Effects of physiological saline
injections on locomotor activity in C57BL/6 mice. Psychol.
Rep. 6: 351-354, 1960.
96. Searle, L. V. and C. W. Brown. The effect of subcutaneous
injections of benzedrine sulfate on the activity of white rats. J.
exp. Psychol. 22: 480-490, 1938.
97. Segal, D. S. Behavioral characterization of d- and /-am-
phetamine: Neurochemical implications. Science 190: 475-477,
1975.
98. Segal, D. S. Behavioral and neurochemical correlates of re-
peated" ^-amphetamine administration. In: Neurobiological
Mechanisms of Adaptation and Behavior, edited by A. J. Man-
del. New York: Raven Press, 1975, pp. 247-262.
99. Segal, D. S. Differential effects ofpora-chlorophenylalanine on
amphetamine-induced locomotion and stereotypy. Brain Res.
116: 267-276, 1976.
100. Segal, D. S. and A. J. Mandell. Long-term administration of
(/-amphetamine: Progressive augmentation of motor activity
and stereotypy. Pharmac. Bioc/iem. Behav. 2: 249-255, 1974.
101. Siegel, P. S. A simple electronic device for the measurement of
gross bodily activity of small animals. J. Psychol. 21: 227-236,
1946.
102. Silbergeld, E. K. and A. M. Goldberg. A lead-induced behav-
ioral disorder. Life Sci. 13: 1275-1283, 1973.
103. Silverman, A. P. and H. Williams. Behaviour of rats exposed
to trichloroethylene vapour. Br. J. Ind. Med. 32: 308-315,
1975.
104. Simpson, L. L. A study of the interaction between am-
phetamine and food deprivation. Psvchopharmacologv 38:
279-286, 1974.
105. Skinner, B. F. The Behavior of Organisms: An Experimental
Analysis. New York: Appleton-Century-Crofts, Inc., 1938, pp.
362-363.
106. Skinner, B. F. The measurement of spontaneous activity. J.
gen. Psychol. 9: 3-24, 1933.
107. Skinner, B. F. and W. H. Morse. Concurrent activity under
fixed-interval reinforcement. J. comp. physiol. Psvchol. 50:
279-281, 1957.
108. Smith, C. B. Enhancement by reserpine and a-methyl dopa of
the effects of rf-amphetamine upon locomotor activity of mice.
J. Pharmac. exp. Ther. 142: 343-350, 1963.
109. Smith, C. B. and P. B. Dews. Antagonism of locomotor sup-
pressant effects of reserpine in mice. Psvchopharmacology 3:
55-59, 1962.
110. Sobotka, T. J. and M. P. Cook. Postnatal lead acetate exposure
in rats: Possible relationship to minimal brain dysfunction. Am.
J. Ment. Defic. 79: 5-9, 1974.
111. Steinberg, H., R. Rushton and C. Tinson. Modification of the
effects of an amphetamine-barbiturate mixture by the past
experience of rats. Nature 192: 533-535, 1961.
112. Stewart, C. C. Variations in daily activity produced by alcohol
and by changes in barometric pressure and diet, with a de-
scription of recording methods. Am. J. Phvsiol. 1: 40-56, 1898.
113. Stewart, W. J. Environmental complexity does affect
scopolamine induced changes in activity. Neurosci. Let. 1:
121-125, 1975.
114. Stolk, J. M. and R. H. Rech. Enhanced stimulant effects of
rf-amphetamine on the spontaneous motor activity of rats
treated with reserpine. ./. Pharmac. exp. Ther. 158: 140-149,
1967.
115. Stolk, J. M. and R. H. Rech. Enhanced stimulant effects of
(/-amphetamine in rats treated chronically with reserpine. J.
Pharmac. exp. Ther. 163: 75-83, 1968.
116. Stolk, J. M. and R. H. Rech. Antagonism of d-amphetamine by
alpha-methyl-/-tyrosine: Behavioral evidence for the participa-
tion of catecholamine stores and synthesis in the amphetamine
stimulant response. Neuropharmacology 9: 249-263, 1970.
117. Strong, P. N. Activity in the white rat as a function of appara-
tus and hunger. /. comp. physiol. Psychol. 50: 596-600, 1957.
118. Strong, P. N. and W. J. Jackson. Effects of hippocampal le-
sions in rats on three measures of activity. J. comp. physiol.
Psychol. 70: 60-65, 1970.
119. Su, M.-Q. and G. T. Okita. Behavioral effects on the progeny
of mice treated with methylmercury. Toxic, appl. Pharmac. 38:
195-205, 1976.
120. Sulser, F., M. L. Owens, N. R. Norvich and J. V. Dingell. The
relative role of storage and synthesis of brain norepinephrine in
the psychomotor stimulation evoked by amphetamine or by
desipramine and tetrabenazine. Psvchopharmacology 12:
322-332, 1968.
121. Svensson, T. H. and G. Thieme. An investigation of a new
instrument to measure motor activity of small animals.
Psvchopharmacology 14: 157-163, 1969.
122. Tainter, M. L. Effects of certain analeptic drugs on spontane-
ous running activity in the white rat. J. comp. Psychol. 36:
143-155, 1943.
123. Tapp, J. T., R. S. Zimmerman and P. S. D'Encarnacao. Inter-
correlational analysis of some common measures of rat activ-
ity. Psycho/. Rep. 23: 1047-1050, 1968.
124. Tapp, J. T. Activity, reactivity and the behavior-directing
properties of stimuli. In: Reinforcement and Behavior, edited
by J. T. Tapp. New York: Academic Press, 1969, pp. 146-177.
125. Thornburg, J. E. and K. E. Moore. A comparison of effects of
apomorphine and ET495 on locomotor activity and circling be-
havior in mice. Neuropharmacology 13: 189-197, 1974.
126. Toman, J. E. P. and G. M. Everett. In: Psychopharmacology,
edited by H. Pennes. New York: Hoeber-Harper, 1958, p. 248.
127. van Rossum, J. M. and F. Simons. Locomotor activity and
anorexogenic action. Psychopharmacology 14: 248-254, 1969.
128. Vasko, M. R., M. P. Lutz and E. F. Domino. Structure-activity
relations of some indolealkylamines in comparison to
phenethylamines on motor activity and acquisition of
avoidance behavior. Psychopharmacology 36: 49-58, 1974.
129. Villareal, J. E., M. Guzman and C. B. Smith. A comparison of
the effects of (/-amphetamine and morphine upon the
locomotor activity of mice treated with drugs which alter brain
catecholamine content. J. Pharmac. exp. Ther. 187: 1-7, 1973.
130. Waldeck, B. Modification of caffeine-induced locomotor stimu-
lation by a cholinergic mechanism. J. Neural Transm. 35: 197-
205, 1974.
131. Waldeck, B. On the interaction between caffeine and barbi-
turates with respect to locomotor activity and brain
catecholamines. Acta pharmac. toxic. 36: 172-180, 1975.
132. Walsh, R. N. and R. A. Cummins. The open-field test: A criti-
cal review. Psychol. Bull. 83: 482-504, 1976.
133. Watzman, N.| H. Barry, III, W. J. Kinnard, Jr. and J. P.
Buckley. Comparison of different photobeam arrangements for
measuring spontaneous activity in mice. J. Pharmaceut. Sci.
55: 907-909, 1966.
134. Weasner, M. H., F. W. Finger and L. S. Reid. Activity
changes under food deprivation as a function of recording de-
vice. J. comp. physio/. Psychol. 53: 470-474, 1960.
135. Weissman, A., B. K. Koe and S. S. Tenen. Antiamphetamine
effects following inhibition of tyrosine hydroxylase. J. Phar-
mac. exp. Ther. 151: 339-352, 1966.
136. Wright, L. S., H. J. Horn, Jr. and G. Woodard. Activity pat-
terns in mice tested singly and in groups as a drug screening
tool. Fedn. Proc. 21: 420, 1962.
137. Zabik, J. E., R. M. Levine and R. P. Maickel. Drug interac-
tions with brain biogenic amines and the effects of am-
phetamine isomers on locomotor activity. Pharmac. Biochem.
Behav. 8: 429-435, 1978.
138. Zieve, L. Effect of benzedrine on activity. Psychol Rec. 1:
343-346, 1937.
139. Zucker, M. H. Electronic Circuits for the Behavioral and
Biomedical Sciences. San Francisco: W. H. Freeman, 1969,
Chapter 6.
-------
lest Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 67-72. ANKHO International Inc., 1979.
Reinforcing Properties of Inhaled Substances1
RONALD W. WOOD
Environmental Health Services Center, and Department of Radiation Biology and Biophysics
University of Rochester School of Medicine and Dentistry
Rochester, NY 14642
WOOD, R. W. Reinforcing properties of inhaled substances. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 67-72, 1979.—Inhaled
substances can support behavior by acting as reinforcing stimulus events. The deliberate inhalation of volatile materials is
attributable to the positively reinforcing properties of these substances and can induce profound toxicity. On the other
hand, inhaled substances can also be aversive, e.g., corrosives, certain solvents, and combustion products. Both positive
and negative reinforcing properties of inhaled materials can be used to support the behavior of laboratory animals. Several
rules of evidence should be met, however, to demonstrate conclusively that an inhalant has such properties. Such proper-
ties should be considered in industrial hygiene and environmental quality decisions.
Ammonia Sensory irritation Aversive atmosphere Escape behavior Mouse Irritants
Conditioning Behavior Toluene Nitrous oxide Self-administration Glue sniffing
Squirrel monkey Rhesus monkey Solvents Anesthetics Inhalant abuse
INHALED materials can affect behavior in a variety of
ways, the most obvious of which is direct toxic impairment
from acute or chronic exposure. Established techniques of
behavioral pharmacology can evaluate such effects, but not
without adequate attention to exposure techniques and
pharmacokinetics. Inhaled materials also can alter behavior
by acting as stimulus events. Such stimulus properties play a
role in industrial hygiene and environmental quality deci-
sions; laboratory techniques to study them have been re-
viewed [23]. This paper focuses on the positive and negative
reinforcing stimulus properties of inhaled materials, will de-
scribe control procedures that can determine if materials
have these properties, and will discuss their pertinence to
regulatory decision making.
INHALANTS AS POSITIVE REINFORCERS
Volatile materials are subject to abuse by inhalation (snif-
fing). Deliberate inhalation of these materials is a menacing
problem, especially among juveniles [18]. The materials sub-
ject to abuse encompass a wide variety of industrial and
consumer products (Table 1; [2]), falling under the jurisdic-
tion of several regulatory agencies. Establishing a uniform
regulatory stance towards this substance abuse practice is
much more complex than simply placing a new abused drug
on a schedule of controlled substances.
The abuse of these materials is not limited to children; it
also occurs on the job. As with any substance abuse prac-
tice, true incidence is difficult to assess, particularly inhalant
abuse on the job, where discovery of the practice could lead
to loss of employment. Access to the agents, of course, is a
precondition to their abuse; some employees apparently re-
gard access to these intoxicants as a job-related benefit. Oc-
cupational exposure to solvents has been associated with
chronic abuse [6, 11, 16, 20]. Anesthesiologists and hospital
technicians have been known to habitually self-administer
halothane, cyclopropane, ether, nitrous oxide, ethyl
chloride, and chloroform [12,19]. Even vinyl chloride has
been subject to deliberate inhalation, with workers becoming
like alcoholics [10].
TABLE 1
COMMERCIAL PRODUCTS SUBJECT TO ABUSE BY
INHALATION (ADAPTED FROM [2])
Volatile Products
Aerosols
contact cement and adhesives
paints, lacquers and thinners
dry cleaning fluids and spot removers
transmission and brake fluids
liquid waxes and wax strippers
shoe polishes
lighter fluids
nail polish remover
degreasers
refrigerants
nitrites (amyl, butyl, isopentyl)
fuels
cold weather car starters
air sanitizers
window cleaners
furniture polishes
insecticides
disinfectants
spray medications
deodorants and hair sprays
antiperspirants
vegetable oil sprays
spray paints and lacquers
'This work was supported by Grant DA-00623 from the National Institute on Drug Abuse, by Grant MH-11752 from the National Institute of
Mental Health, by Grant ES-01247 from the National Institute of Environmental Health Science, and in part under contract with the U. S.
Department of Energy at the University of Rochester, Department of Radiation Biology and Biophysics and has been assigned Report No.
UR-3490-1604.
67
-------
WOOD
FIG. I. Squirrel monkey seated in toluene self-administration apparatus. Subject is pushing button that
produces a 15 sec infusion of toluene vapor into the helmet resting on the neck restraint plate of the
chair.
It is important to be able to predict the abuse potential of
substances incorporated in consumer products or that em-
ployees are likely to be exposed to transiently at relatively
high concentrations. Examples are apparently innocuous
solvents, refrigerants, propellents, and non-irritating volatile
or gaseous substances. These materials may be toxic, or
serve as a reinforcing vehicle for the ingestion of more toxk
materials. Although non-human primates will self-administer
most CNS drugs abused by humans [3, 9, 17], the positivelj
reinforcing properties of inhaled substances have rarely beoi
studied in the laboratory. Recently developed laboratory
procedures demonstrate that volatile materials may act as
-------
INHALED SUBSTANCES
69
FR
FIG. 2. Cumulative records of fixed-ratio performance maintained by 15 sec deliveries of 60% nitrous oxide for subject 52. Session duration
was two hours. Increasing fixed-ratio size increased response rates. Deflection of stepping pen marks onset of reinforcement delivery; event
pen is deflected for duration of reinforcement.
positive reinforcers, and bear substantial abuse potential.
Several rules of evidence for such a demonstration are of-
fered below.
An agent is a positive reinforcer if a response that pro-
duces the agent increases in frequency. This is implicit in
any attempted demonstration and is subject to the qualifica-
tions listed below. Yanagitaef al. [28] reported that macaque
monkeys increased lever pressing when it produced
intranasal infusions of chloroform, lacquer thinner, or ether.
Wood et al. [26] demonstrated that squirrel monkeys will
respond to produce delivery of nitrous oxide. Grubman [7]
has demonstrated that the macaque also can be trained to
self-administer nitrous oxide.
Terminating response-contingent access to the agent
should eventually reduce the frequency of behavior (extinc-
tion). Ending nitrous oxide delivery lowered responding by 3
to 4 fold in squirrel monkeys [26]. Early in the monkeys'
self-administration history, withholding toluene vapor pro-
-------
70
WOOD
duced irregular responding that occurred in bursts [23,24].
The toluene self-administration preparation is shown in Fig.
1. Oxygen or nitrogen substitution for nitrous oxide did not
support the self-administration behavior of macaques [7].
Typical patterns of behavior should be generated by
schedules of reinforcement. Increasing the number of re-
sponses required for reinforcement increases response rate
with a variety of reinforcers, such as food [1] and cocaine [4,
5, 13]. This is also the case with nitrous oxide in the squirrel
monkey (see Fig. 2; [26]) and in the macaque [7]. Rhesus
monkeys also display typical patterns of responding with a
multiple fixed-ratio 30 fixed-interval 5 min schedule of rein-
forcement [7].
The frequency of responding maintained by the agent
should be a function of concentration. Response and rein-
forcement rates usually are related by an inverted U-shaped
function to the magnitude of the event maintaining behavior.
Examples are provided by heat in the cold [21], electrical
stimulation of the brain [14,15], food [4], and a wide variety
of drugs injected intravenously [4, 8, 15, 22, 27]. This is also
the case with toluene serf-administration by the squirrel
monkey [23,24] and nitrous oxide self-administration by the
macaque [7]. This relationship was not cleanly demonstrated
in squirrel monkeys that self-administered nitrous oxide [26].
Beyond the minimum concentration necessary to maintain
the behavior, the relationship between nitrous oxide con-
centration and reinforcement rate was flat (15-75%). How-
ever, when 20 responses were required for nitrous oxide
delivery, response and reinforcement rates were an increas-
ing function of concentration. Within the concentrations
employed, rate decrements were not produced by increasing
the nitrous oxide concentration, with the exception of one
transition in one subject.
The behavior must not result from a general rate-
increasing effect of the agent. Squirrel monkeys that self-
administered toluene were seated in a chair facing two but-
tons [23,24]. A response on one button diverted a portion of
the air stream through toluene in a gas washing bottle and
back into the helmet (Fig. 1). Responses on the second but-
ton had no effect. Animals responded at a higher rate on the
active button, and was not related to concentration, indicat-
ing that the performance did not result from a general in-
reversed. Responding on the active button was related to
concentration, as described above. Responding on the in-
active button occurred at a much lower rate than on the
active button, and was not related to concentration indicat-
ing that the performance did not result from a general in-
crease in response rate produced by toluene. Noncontingent
toluene delivery also reduced the frequency of self-
administration, further demonstrating that responding is
maintained by the reinforcing properties of toluene, and is
not merely a nonspecific increase in responding.
INHALANTS AS NEGATIVE REINFORCERS
Air pollutants and other inhaled materials can display
noxious biological effects ranging from unpleasant odor and
eye and upper airway irritation, to tissue destruction. Lab-
oratory studies of inhalants that examine physiological or
morphological changes address the behavioral significance
of exposure only indirectly. The aversiveness of an inhalant
stimulus could be assessed by determining the concentra-
tions at which an animal responds to turn it off. If the
stimulus is intense enough so that its termination supports
behavior, then it is an effective negative reinforcer. A tech-
FIG. 3. Irritant (ammonia) exposure chamber: (A) before irritant
delivery; (B) during irritant delivery. The chamber atmosphere was
introduced from the top of the chamber, where it struck a baffle to
ensure even mixing. The mouse stood on a perforated stainless steel
platform through which the atmosphere exhausted. The irritant was
added to the dilution air immediately above the chamber. The irri-
tant delivery could be terminated by the mouse interrupting a light
beam located in a conical recess in the wall. At any given time, only
one of these two sensors would terminate the irritant delivery. When
the irritant was shut off, either by a nose poke or at the end of 60 sec,
a 1 liter/min stream of clean humidified air was delivered through
each cone. This was done to minimize the delay of irritant termina-
tion after a response occurred.
u
2 50'
_ 40^
CL
QC
ID
Q
30-
20-
0-0 .05 .10
• 20
CONCENTRnTION ('/.),
FIG. 4. Average duration of ammonia delivery as a function of am-
monia concentration. These data represent the mean performance of
six mice given one session of 50 deliveries at each concentration,
except for the 0.1% level, which is the mean of the means of 4-10
sessions per mouse. Plotted are the means ± 1 SD
-------
INHALED SUBSTANCES
71
Q
LJ
o:
u
LJ
(J
cr
u
CL
100-
90-
80-
70-
60-
50-
40-
30-
20-
10-
0-
0.0 .05
.10
.20
CONCENTRflTION (X)
FIG. 5. Percentage of ammonia deliveries terminated by a response
as a function of ammonia concentration. Data are as described in
Fig. 4.
nique has recently been developed to investigate the deter-
minants of escape from an aversive atmosphere [25]. It is
illustrated in Fig. 3. Work done by this technique can serve
to illustrate the control procedures that must be performed in
order to demonstrate conclusively escape from aversive at-
mospheres. They include the following:
An inhalant is a negative reinforce/' if the duration of
exposure tolerated is inversely related to concentration.
When mice were given the opportunity to terminate am-
monia delivery, they tolerated a progressively shorter dura-
tion of ammonia as concentration increased (Fig. 4).
An inhalant is a negative reinforcer if the proportion of
deliveries terminated by a response is directly related to
concentration. The mice terminated a greater percentage of
the ammonia deliveries as the concentration increased to
0.1%; at this and a higher concentration (0.2%) virtually
every delivery was terminated by the mice (Fig. 5).
The 'behavior must not result from a general rate-
increasing effect of the agent. Responses per delivery cycle
(irritant onset to onset) were related to concentration on the
sensor that terminated delivery, while responses per cycle
on the inactive sensor were not so related. If the nose-poking
behavior in this preparation was generated by a general in-
crease in activity, a differential frequency of responding be-
tween the two identical sensors would not be expected.
The subject must discriminate between responses that
terminate inhalant delivery and those that do not. On ap-
proximately 85% of the trials terminated by a response, the
animals went directly to the appropriate sensor without mak-
ing a response on the inappropriate sensor. When the func-
tion of the two sensors was exchanged, the mice were behav-
ing appropriately by the third hour after the reversal.
APPLICATIONS AND IMPLICATIONS
Preparations that evaluate inhalants as positive or nega-
tive reinforcing stimuli address a variety of questions perti-
nent to regulatory decision making. Both classes of proce-
dures help determine the irritant potency and abuse potential
of materials. Time- or exposure-dependent changes arising
from adaptation, sensitization, or tolerance can also be
determined. Adaptation phenomena are of special importance
for aversive inhalants since the unpleasant properties of
materials are sometimes relied upon to limit worker expo-
sure in industrial settings. Many compounds could be
studied using such techniques, including corrosives,
solvents, and the combustion products of industry, the auto-
mobile, and the catastrophic conflagration.
Intoxicating substances pose a special hazard to workers
who develop a fondness for the materials with which they
work. These properties should, be taken into account when
short-term exposure limit values are set. The abuse potential
of inhalants can be assessed without resorting to experi-
mental human exposures. In addition, animal self-
administration preparations should be able to determine
self-administration limit values, i.e., exposure levels insuffi-
cient to maintain the abuse of these materials. It should be
remembered that the exposures generated during the abuse
of these materials are usually high concentration spikes, and
that these exposures are unlike those used in typical evalua-
tions of toxicity. The consistency and chronicity of these
spiked exposures can be remarkable; for example, one man
noticed that toluene produced euphoria while working with
pain thinner at an aircraft company. His habit continued over
a fourteen-year period, and escalated to deep inhalation of
concentrated vapors more than ten times per hour through-
out the day, including mealtimes [11]. The reinforcing prop-
erties of these materials are critical when these materials are
carcinogenic, have acute organ toxicity, or induce chronic
toxicity such as polyneuropathies.
REFERENCES
. Boren, J. J. Resistance to extinction as a function of the fixed
ratio. J. exp. Psychol. 61: 304-308, 1961.
Cohen, S. Inhalant abuse: an overview of the problem. In: Re-
view of Inhalants: Euphoria to Dysfunction, edited by C. W.
Sharp and M. L. Brehm. NIDA Research Monograph 15,
USDHEW Publication No. (ADM) 77-55 pp. 2-11, 1977.
Deneau, G. A., T. Yanagita and M. H. Seevers. Self-
administration of psychoactive substances by the monkey—a
measure of psychological dependence. Psychopharmacologia
16: 30-48, 1969.
4. Goldberg, S. R. Comparable behavior maintained under fixed-
ratio and second-order schedules of food presentation, cocaine
injection or d-amphetamine injection in the squirrel monkey. J.
Pharmac. exp. Ther. 186: 18-30, 1973.
5. Goldberg, S. R., F. Hoffmeister, U. U. Schlichting and W.
Wuttke. A comparison of pentobarbital and cocaine self-
administration in rhesus monkeys: effects of dose and fixed-
ratio parameter. J. Pharmac. exp. Ther. 179: 277-283, 1971.
6. Grabsky, D. A. Toluene sniffing producing cerebellar degener-
ation. Am. J. Psychiat. 118: 461, 1961.
-------
72
WOOD
7. Grubman, J. Behavior maintained by nitrous oxide reinforce-
ment in the rhesus monkey. Unpublished masters thesis, De-
partment of Pharmacology, University of Michigan, 1977.
8. Hoffmeister, F. and U. U. Schlichting. Reinforcing properties
of some opiates and opioids in rhesus monkeys with histories of
cocaine and codeine self-administration. Psychopharmacologia
23: 55-74, 1972.
9. Kelleher, R. T., S. R. Goldberg and N. A. Krasnegor. Sym-
posium on control of drug-taking behavior by schedules of rein-
forcement. Pharmac. Rev. 27: 287-555, 1975.
10. Klein, J. The plastic coffin of Charlie Arthur. Rolling Stone, pp.
43-47, January 15, 1976.
11. Knox, J. M. and J. R. Nelson. Permanent encephalopathy from
toluene inhalation. New Engl. J. Meet. 275: 1494-1496, 1966.
12. Nagle, D. R. Anesthetic addiction and drunkeness: a contempo-
rary and historical survey. Int. J. Addict. 3: 25-39, 1968.
13. Pickens, R. and T. Thompson. Cocaine-reinforced behavior in
rats: effects of reinforcement magnitude and fixed ratio size. J.
Pharmac. exp. Ther. 161: 122-129, 1968.
14. Plutchik, R., W. L. McFarland and B. W. Robinson. Relation-
ships between current intensity, self-stimulation rates, escape
latencies, and evoked behavior in rhesus monkeys. /. comp.
physiol. Psychol. 61: 181-188, 1966.
15. Reynolds, R. W. The relationship between stimulation voltage
and rate of hypothalamic serf-stimulation in the rat. J. comp.
physiol. Psychol. 51: 193-198, 1958.
16. Satran, R. and V. N. Dodson. Toluene habituation: report of a
case. New Engl. J. Med. 268: 719-721, 1963.
17. Schuster, C. R. and T. Thompson. Self-administration of and
behavioral dependence on drugs. Ann. Rev. Pharmac. 9: 483-
502, 1969.
18. Sharp, C. W. and M. L. Brehm. Review of Inhalants: Euphoria
to Dysfunction. NIDA Research Monograph 15, U.S.D.H.E.W.
Publication No. (ADM) 77-553, 1977.
19. Spencer, J. D., F. O. Raasch and F. A. Trefny. Halothane abuse
in hospital personnel. J. Am. Med. Ass. 235: 1034-1035, 1976.
20. Weisenberger, B. L. Toluene habituation. J. occup. Med. 19:
569-570, 1977.
21. Weiss, B. and V. G. Laties. Magnitude of reinforcement as a
variable in thermoregulatory behavior. J. comp. physiol.
Psychol. 53: 603-608, 1960.
22. Wilson, M. C., M. Hitomi and C. R. Schuster. Psychomotor
stimulant self-administration as a function of dosage per injec-
tion in the rhesus monkey. Psychopharmacologia 22: 271-281,
1971.
23. Wood, R. W. Stimulus properties of inhaled substances. Envir.
Hlth. Perspect. 26: 69-76, 1978.
24. Wood, R. W. Toluene self-administration by the squirrel mon-
key. Seventh Int. Cong. Pharmac., Paris, 1978, Abstract 1497.
25. Wood, R. W. Behavioral evaluation of sensory irritation evoked
by ammonia. Toxicol. Appl. Pharmac., in press, 1979.
26. Wood, R. W., J. Grubman and B. Weiss. Nitrous oxide self-
administration by the squirrel monkey. J. Pharmac. exp. Ther.
202: 491-499, 1977.
27. Woods, J. H. and C. R. Schuster. Reinforcement properties of
morphine, cocaine and SPA as a function of unit dose. Int. J.
Addict. 3: 231-237, 1968.
28. Yanagita, T., S. Takahashi, K. IshidaandH. Funamoto. Volun-
tary inhalation of volatile anesthetics and organic solvents by
monkeys. Jap. J. din. Pharmac. 1: 13-16, 1970.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral.Toxicology, Vol. 1, Suppl. 1, pp. 73-84. ANKHO International Inc., 1979.
Behavioral Assessment of Risk-Taking and
Psychophysical Functions in the Baboon
JOSEPH V. BRADY, L. DIANNE BRADFORD AND ROBERT D. HIENZ
Department of Psychiatry and Behavioral Sciences, The John Hopkins University School of Medicine,
Baltimore, MD 21205
BRADY, J. V., L. D. BRADFORD AND R. D. HIENZ. Behavioral assessment of risk-taking and psychophysical
functions in the baboon. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 73-84, 1979.—Laboratory procedures have been
developed for the experimental analysis of risk-taking and psychophysical functions in dog-faced baboons (Papio anubis).
In a procedure analogous to the traffic light situation, animals are rewarded with food pellets for completing a fixed ratio of
100 responses in the presence of a green light. Superimposed upon this baseline performance are 5-second presentations of
a yellow warning light terminated by a red light in the presence of which all responses are punished with electric shock.
When the yellow light is introduced late in the sequence (e.g., after 93 responses have been completed), response rates
increase and the 100-response ratio is completed before the 5-second yellow light times out. When the yellow light appears
early in the sequence (e.g., after 73 responses) a marked decrease in response rate is observed with cessation of responding
before onset of the red light. The sensitivity of components of this risk-taking performance to pharmacological toxicants is
reported and psychophysical assessment of relevant sensory-motor effects described.
Risk-taking Psychophysical functions Baboons Pentobarbital Diazepam d-Methylamphetamine
Chlordiazepoxide Visual thresholds Auditory thresholds Reaction time Pharmacological toxicants
Behavioral toxicology Amobarbital
THE DEVELOPMENT of test methods for defining the ef-
fects of toxic substances on behavior have for the most part
focused upon well-explored procedures of demonstrated
sensitivity, usually with pharmacological or physiological
parametric interactions as a reference base [7]. The scope of
such methodological approaches has included a broad range
of naturalistic (e.g., activity and feeding cycles) and learned
(e.g., schedule-controlled performances) behaviors [4], as
well as innovative and sophisticated combinations of the two
in the application of psychophysical procedures to basic sen-
sory [2,5] and motor [8] assessments. CharacteristicaEy,
these developments have reflected a shift from traditional
screening approaches involving limited observations on large
numbers of organisms to more precise assessment tech-
niques where the large N is provided by the number of ob-
servations (frequently of more than one behavior simulta-
neously) and the number of organisms is relatively small.
The obvious advantages of such precision evaluations in-
clude the ability to focus upon specific behavioral systems
under well-controlled baseline conditions which minimize
the likelihood that minor random perturbations will contrib-
ute significantly to the observed variance.
The extension of this behavioral toxicology assessment
approach to less well-analyzed affective repertoires would
seem to require a comparable shift from dependence upon
traditionally subjective psychological evaluations to per-
formance measures derived from applied behavior analysis
techniques [6]. Furthermore, experimental developments
emphasizing the analysis of such complex behavioral inter-
actions and psychophysical processes at the animal labora-
tory level should serve to counteract the tendency for per-
formance assessment methods and measures in behavioral
toxicology to remain rigid in response to the dictates of famil-
iarity. In this regard, recent efforts to extend the breadth and
sensitivity of animal behavior baselines for evaluating the
effects of pharmacological toxicants have explored experi-
mental procedures for the analysis of risk-taking perform-
ances in laboratory baboons [1].
One such procedure may be described as an analogue of
the traffic light situation commonly encountered in urban
ecologies. In the presence of a green light, the animal is
rewarded with food pellets for completing a fixed number of
responses (e.g., 100) on a spring-loaded lever switch (i.e.,
Lindsley manipulandum). An added counter is provided for
the animal in the form of 10 small white pilot lights illumi-
nated sequentially upon completion of each block of 10 re-
sponses, thus serving to indicate location and progression
through the 100-response ratio requirement. Superimposed
upon this baseline performance are presentations of a yellow
light stimulus (5 sec in duration) which replaces the green
light and can occur at any point in the fixed 100-response
ratio sequence. The 5-sec yellow light interval is terminated
by the appearance of a red light (60 sec in duration) in the
presence of which all lever responses are punished by an
electric shock administered through an electrode attached to
a shaved portion of the baboon's tail. Responses in the pres-
ence of both the yellow and red light serve to advance the
100-response count required for delivery of the food pellet
reward even though each response in the red light is accom-
panied by shock.
Initial explorations of this procedure have focused upon
the effects of the yellow warning light, with the not unex-
73
-------
74
BRADY, BRADFORD AND HIENZ
SALINE
3.2mgAg PENTOBARBITAL
10.0 mgAg PENTOBARBITAL
3.2 mq/kg DIAZEPAM
' rx r-i
iaOmg/(
-------
RISK-TAKING AND PSYCHOPHYSICAL FUNCTIONS
75
d-METHYLAMPHETAMINE
10 min.
FIG. 2. Baseline control performance (top record) and portions of two drug sessions following doses of
0.1 mg/kg (middle record) and 0.32 mg/kg (bottom record)
-------
76
BRADY, BRADFORD AND HIENZ
PENTOBARBITAL
5.6mg/kg, 0%
56mg/kg, 100%
IQOmg/kg, 100%
10 min.
FIG. 3. Cumulative records showing effects of pentobarbital (5.6 and 10.0 mg/kg) upon performance
under conditions of low shock risk (0% shock probability, top record) and high shock risk (100% shock
probability, middle and bottom records) for Baboon S-WI.
DIAZEPAM
5.6mg/kg, 0%
5.6mgAg, 100%
10.0 mgAg, 100%
10 min.
FIG. 4. Cumulative records showing effects of diazepam (5.6 and 10.0 mg/kg) upon performance under
conditions of low shock risk (0% shock probability, top record) and high shock risk (100% shock
probability, middle and bottom records) for Baboon S-WI.
-------
RISK-TAKING AND PSYCHOPHYSICAL FUNCTIONS
77
• -Early Yellow O-Early Gre
*-Mte Yallo. A-L.te Brto
High Shock Risk
FIG. 5. The effects of diazepam upon "risk taking" behavior and control rates under conditions of low
shock risk (0% shock probability, left panel) and high shock risk (100% shock probability, right panel)
for Baboon S-WI. Brackets enclosing control values indicate ± one standard deviation.
marks on baseline under top record). With the same dose of
pentobarbital (5.6 mg/kg) under the 100% shock probability
condition, however, this same animal (S-WI) generated the
performance shown in the middle record of Fig. 3. Except
for the moderate overall rate decrease, the 5.6 mg/kg dose of
pentobarbital under these conditions had little apparent ef-
fect upon risk-taking, and there was a striking similarity to
the baseline performance for this animal (i.e., S-WI) shown
in the top section of Fig. 2. With the higher dose of pen-
tobarbital (i.e., 10.0 mg/kg) however, shown in the bottom
record of Fig. 3, rather severe rate suppression and some
disruption of the risk-taking performance were apparent
even with the 100% shock probability condition in effect.
Figure 4 shows a similar progression of experiments with
diazepam for baboon S-WI under conditions of low and high
shock probability. Again, a marked increase in responding
during both the early yellow and red light intervals accom-
panied by a high incidence of earned but undelivered shocks
followed administration of 5.6 mg/kg diazepam under the 0%
shock probability condition (top record, Fig. 4) despite evi-
dent decreases in the overall response rate. And although
some normalization of the risk-taking performance was ap-
parent when the same diazepam dose (i.e., 5.6 mg/kg) pre-
ceded an experimental session programmed under the 100%
shock probability condition (middle record, Fig. 4), intermit-
tent disruptions (e.g., running through the early yellow-light;
responses initiated in the presence of the red light) continued
to occur and were observed with even higher frequency fol-
lowing 10.0 mg/kg diazepam (bottom record, Fig. 4). In this
regard, the apparent greater sensitivity of baboon S-WI (i.e.,
by comparison with baboon Q-74) to pentobarbital and
diazepam is reflected in both the overall rate and the risk-
taking performance changes following drug administration.
The extent to which these drug-related performance
changes can be characterized as effects upon risk-taking de-
pends, of course, upon a critical analysis of differential sen-
sitivity to such pharmacological toxicants of selected com-
ponents of the multi-operant repertoire shown in Figs. 1
through 4. Figure 5, for example, illustrates such a compo-
nent analysis and compares the effects of diazepam over the
dose range of 3.2 to 17.0 mg/kg under conditions of low (i.e.,
0% probability) and high (i.e., 100% probability) shock risk.
The right side of Fig. 5 shows the generally suppressing ef-
fect of diazepam at all four doses upon response rates in the
presence of the green light and the late (i.e., after the 93rd
response in the 100-response green light sequence) yellow
light while at the same time producing at least a moderate
increase in the early (i.e., after the 73rd response in the
100-response green light sequence) yellow response rate
(solid line connecting filled circles at bottom of graph). The
data points plotted as open circles and triangles represent
control response rates measured during green-light-only
trials over early and late time segments in the 100-response
sequence corresponding to the exact location (i.e., after the
73rd and 93rd responses, respectively) of those intervals dur-
ing yellow light trials. The orderly and systematic relation-
ships between these rates and those measured in the pres-
ence of the early and late yellow light (i.e., filled circles and
triangles, respectively) clearly differentiate the component
performances of this multi-operant risk-taking repertoire.
The late yellow rates (filled triangles) are slightly but consis-
tently higher than the late green rates (open triangles), while
the early yellow rates (filled circles) are consistently and
markedly lower than the early green rates (open circles).
And at least under the high shock risk condition (right side,
Fig. 5), these relationships are as stable across all indicated
doses of diazepam as they are during the no-drug control
sessions from which the data shown in the upper left hand
corner of each graph were derived.
Though the differential (and somewhat dose-dependent)
increase in early yellow response rates following diazepam
administration under the high shock risk condition shown on
the right side of Fig. 5 is suggestive of a pharmacological
toxicant effect upon selective aspects of risk-taking, the
-------
78
BRADY, BRADFORD AND HIENZ
well-established rate-dependency hypothesis (i.e., a wide
range of drugs are known to decrease high rates of respond-
ing and increase low rates) would seem to provide a more
parsimonious interpretive alternative. Such is not the case,
however, when the data from the low shock risk condition
shown on the left side of Fig. 5 is considered. Again, all four
doses of diazepam can be seen to suppress response rates in
the presence of the green light and the late yellow light. But
the dramatic (and clearly dose-dependent) increase in re-
sponding during the early yellow light (solid line connecting
filled circles) under this low shock risk condition (i.e., 0%
shock probability) can not be attributed to a drug effect upon
low response rates alone. Both the low and high shock risk
condition are characterized by equally low response rates in
the early yellow under control conditions. But only in the
low shock risk conditions did diazepam increase the early
yellow rate above the early and late green rates. It would
appear that diazepam has differential effects upon risk-taking
performances depending upon whether there is a high or low
shock risk.
There remain, of course, many unanswered questions
about the validity, reliability, selectivity, feasibility, and
economics of such an experimental approach to defining the
effects of toxic substances upon so-called risk-taking behav-
iors. Not the least among these concerns is the extent to
which effects rather glibly described in such interpretive lan-
guage can be accounted for in terms of the sensory and
motor processes upon which the performances involved ob-
viously have a basic dependence. To help define some of
these conceptual and methodological limits, a series of
studies was undertaken to assess the sensory and motor ef-
fects of those pharmacologic toxicants which appeared to
alter risk-taking. The psychophysical methodology [3] in-
volved the use of a reaction-time procedure which required
the baboons to press a lever and hold it depressed for varying
intervals until presentation of a light flash or tone burst sig-
nalled the availability of food reward following lever release.
Correct responses (i.e., lever releases occurring within 1.5
sec of signal onset) were rewarded with banana-flavored
food pellets, and detection thresholds were determined by
systematically varying stimulus intensity and recording the
frequency of correct and incorrect responses (i.e., lever re-
leases occurring later than 1.5 sec after stimulus onset). In
addition, response latencies (i.e., elapsed time between
signal onset and lever release) were recorded to the nearest
millisecond as a measure of reaction time.
The subjects in the studies completed to date were 4 dog-
faced baboons (Papio anubis), housed in individual cages
and maintained on a 22-hr restricted feeding schedule with
supplemental monkey chow and fresh fruit provided on a
daily basis after each experimental session. The testing
apparatus consisted of a modified baboon squeeze cage fitted
within a double-walled sound attenuating chamber. A 30x38
inch intelligence panel attached to one side of the cage con-
tained a microswitch actuating lever, a red LED cue light, a
one-inch circular visual stimulus patch, and a tube feeder for
delivery of banana pellets. With the animal positioned facing
the panel, the cue light and visual stimulus patch were at eye
level, the feeding tube at mouth level, and the response lever
at waist level in front of the right arm. Additionally, a wide-
range acoustic driver suspended outside the cage and located
directly over the animal's head approximately 8 inches
above ear level provided for the delivery of auditory signals.
The light source for the visual stimuli was provided by a
slide projector mounted on the outside of the chamber and
projecting white light on to the back of the one-inch stimulus
patch through an otherwise light-tight aperture in the
chamber wall. Stimulus intensity was varied by using neutral
density filters in the slide projector. Light intensities were
calibrated with a light meter. Acoustic signals were gener-
ated by a Krohn-Hite oscillator passed through an electronic
switch (20 msec rise and fall times), programmable at-
tenuator, amplifier, and the wide range acoustic drive (i.e.,
speaker). The system was calibrated with a General Radio
sound level meter, and a Bruel and Kjaer amplifier, and '/2
inch condenser microphone located at ear level facing the
speaker. Programming of the experiments was accomplished
with a solid-state control system. Data recording involved
the use of electromechanical counters and a microprocessor
interfaced to a video terminal which recorded all response
latencies and computed median latency and Q values.
Following initial shaping of lever pressing and discrimi-
nation of the holding and release components of the re-
sponse, all animals were introduced to the discrete trial
reaction time procedure. In the presence of a flashing red cue
light (5/sec), a lever press changed the flashing red light to a
steady red light which remained instated as feedback as long
as the animal held the lever switch in the closed position. At
varying intervals (range 0.5 to 6.5 sec) following initiation of
this maintained holding response, a test stimulus (white light
on the circular patch or tone burst through the speaker) was
presented for 1.5 sec. Release of the lever within the 1.5 sec
test stimulus interval delivered a single banana pellet and
initiated a 1 sec intertrial interval (ITI) during which no
stimuli were presented and lever responses re-initiated the
ITI. Incorrect responses (i.e., lever presses prior to test
stimulus onset or after the 1.5 sec test stimulus interval)
reinstated the 1 sec ITI without reinforcement. Following the
1 sec ITI, the flashing red cue light signalled initiation of the
next trial in the series of several hundred which comprised
each daily 2 to 3 hour experimental session. Asymptotic
levels of performance on this procedure typically required 2
to 3 months of such daily training sessions.
Auditory and visual thresholds were determined by ran-
domly varying (in accordance with the method of constant
stimuli) the intensity of the test stimuli from trial to trial and
examining detection frequencies (i.e., correct lever releases)
at each intensity. For the auditory modality, four intensity
levels (10 dB apart) of a 16.0 kHz pure tone were used, with
the lowest level set just below the animal's estimated
threshold. Interspersed among the test trials were a series of
catch trials during which no tone was presented to measure
the false alarm (i.e., guessing) rate. For the visual modality,
four intensity levels (0.5 log density units apart) of the white
light were used with the lowest level again set just below the
animal's estimated threshold. Again, catch trials with no
light were programmed to occur intermittently. In addition,
sessions involving visual threshold determinations were pre-
ceded by a 30-min dark adaptation period in the light-proof
chamber followed by a 5-min, 50 trial warm up with the
highest test intensity of white light.
For both the auditory and visual threshold determina-
tions, each test session was divided into four blocks of ap-
proximately 150 trials with each of the 4 intensity levels (plus
catch trials) presented randomly approximately 30 times dur-
ing each block. This provided 4 independent within-session
estimates of the sensory thresholds and functions relating
reaction time to intensity. Sensory thresholds were deter-
mined from percent correct detections at each intensity in-
terpolating to the intensity which produced a detection score
-------
RISK-TAKING AND PSYCHOPHYSICAL FUNCTIONS
79
halfway between the false alarm rate and 100%. Stable audi-
tory thresholds were based upon determinations from 3 suc-
cessive test sessions with estimates which varied by no more
than 4 dB. Stable visual thresholds were based upon deter-
minations from 3 successive test sessions with estimates
which varied by no more than 0.2 log density units. In both
cases, such a determination of threshold stability required a
false alarm rate below 30% and no systematic change trends
in the data. With regard to the response latency measure of
reaction time (typically skewed due to the physiological lim-
its on lever release time), the standard measure of central
tendency employed for such distribution was the median,
with variability reported in terms of the interquartile range.
Following stabilization of the threshold and reaction time
measures, preliminary studies were undertaken to explore
the validity, reliability, sensitivity, and specificity of these
psychophysical methods with respect particularly to the
evaluation of pharmacological toxicants. All drugs have been
administered intramuscularly at the beginning of each exper-
imental session, followed by a 30-min warm-up before formal
threshold determinations were begun. Saline control ses-
sions have been conducted between each drug session and
return-to-baseline performances were required during these
intervening saline control sessions before further drug ad-
ministrations were programmed.
Figure 6 illustrates the psychometric function relating
percent correct lever releases to intensity in decibels sound
pressure level (dB SPL) of a 16 kHz pure tone obtained in the
course of a psychophysical experiment with one baboon. As
stimulus intensity decreased, percent correct detection also
decreased, describing the typical S-shaped function. The
absolute threshold, as indicated on Fig. 6, is defined by the
stimulus intensity value at which the percent correct detec-
tion is halfway between the catch or guessing rate and 100%
correct detections. Similar relationships were observed
when percent correct detections were determined as a func-
tion of white light intensity.
When response latency was measured as a function of
stimulus intensity with the baboon performing in a reaction
time experiment, the psychophysical functions shown in Fig.
7 were obtained. With both a white light and a 16 kHz tone,
response latencies decreased (i.e., reaction time became fas-
ter) as stimulus intensity increased. This naturally-occurring
relationship between response latency and stimulus intensity
required no shaping or training and has been consistently
observed in all animals. Additionally, the data plot in Fig. 7
shows that not only do latencies decrease as intensity in-
creases, but the variability in the latency measure also de-
creases, with the result that trial to trial data replicability was
extremely good at the higher stimulus intensities. Finally,
Fig. 7 also shows that there was a consistent difference be-
tween the response latencies to auditory and visual stimuli.
Significantly, at the higher intensities where variability is
low, auditory reaction times were consistently found to be
approximately 100 msec faster than visual reaction time.
That this basic stimulus intensity-response-latency func-
tion may provide a sensitive measure of toxic effects is
suggested by the data plot in Fig. 8 which shows the orderly
effects of increasing doses of amobarbital sodium (i.e., 3.2,
10.0, 17.0 mg/kg IM) upon reaction time as a function of
white light stimulus intensity in the baboon. The latency-
intensity functions were recorded 1 to 2 hr after drug admin-
istration (i.e., peak action time) and show the systematic
relationship between drug dose and response latency at all
but the lowest (i.e., threshold) intensity level. Even at the
100-
STIMULUS INTENSITY (dB SPL)
FIG. 6. A psychometric function showing percent correct lever re-
leases to a 16.0 kHz pure tone as a function of stimulus intensity for
Baboon S-PE. C=catch trial rate. The indicated threshold is cor-
rected for this catch trial rate (see text).
o
A A WHITE LIGHT
• « 16 KHz TONE
f- «--_
(nlitln dntllt/) 1.0 t.l 1.3 !.» 1.7 1.3 l.t
INCREASING STIMULUS INTENSITY »
FIG. 7. A latency-intensity function showing median reaction time
to either white light (solid lines) or a 16.0 kHz pure tone (dotted
lines) as a function of stimulus intensity for Baboon S-PE. Inter-
quartile ranges are bracketed for each point.
-------
80
BRADY, BRADFORD AND HIENZ
900-
800-
o
O>
CO
E
~ 700-
LJJ
5
I—
O 600-
i-
o
111
(T
500-
400-
AMOBARBITAL SODIUM
o
D a -17 m g / k g
Q A -10 mg/kg
o -o -3.2 mg/kg
-Saline Control
3.5 3.0 2.5
INCREASING LIGHT INTENSITY
(relative density )
2.0
FIG. 8. Latency-intensity functions showing median reaction time to
white light as a function of stimulus intensity, with dose level of
amobarbital sodium as the parameter. Each curve was obtained at
the estimated peak action time of the drug. All data are from one
Baboon, S-IK.
highest intensity levels where response variability is minimal
(see Fig. 7), the orderly progression of increasing latencies
with increasing doses is apparent. The basic psychophysical
functions illustrated in Figs. 6 and 7 have been determined
and replicated in each of the 4 baboons participating to date
in the experiments described below.
Against the background of these basic psychophysical re-
lationships, a series of studies was undertaken to deter-
mine the differential sensitivity of the methodology to exper-
imental interventions with predictable effects. Figures 9 and
10, for example, contrast the effects of noise (i.e., 100 dB
SPL for 45 min) and pentobarbital (17 mg/kg) upon selective
aspects of the animal's psychophysical performance. Expo-
sure to the two independent variables was programmed to
occur immediately before the separate experimental sessions
scheduled for each intervention, and effects were measured
as changes in auditory sensitivity and/or reaction time. Pre-
dictably, the data plots in Fig. 9 show an elevation in the
threshold sensitivity required for detection of the 16 kHz
pure tone during the first 1 to 2 hours following noise expo-
sure. No such auditory threshold shifts were observed fol-
lowing either pentobarbital administration or the non-
intervention control procedure. Significantly, experimental
sessions which involved visual reaction time performances
showed no changes in visual threshold following noise expo-
— 20-
l—
u_
X
cn 15-
o
O
X 1
cu
cr
I 5-
i-
i o-
H
D
16 KHz TONE
X X NOISE (100dB/45')
» * PENTOBARBITAL (17mg/kg)
0...-0 CONTROL
60
120 180 240 300
TIME (min)
24 hrs
FIG. 9. Changes in the auditory threshold to a 16.0 kHz pure tone as
a function of time after 3 different experimental manipulations. (1)
Exposure to broadband noise of 100 db for 45 min. (2) I.M. injection
of 17.0 mg/kg pentobarbital sodium. (3) No intervention.
E 800-
zaoo
o
o '
<
16 KHz TONE
* X NOISE MOOdB/45'l
. * P£NTOBARBITAL(17mg/kg)
a- • • • -a CONTROL
0 10 20 30
STIMULUS INTENSITY (dB)
FIG. 10. Latency-intensity functions showing median reaction time
to a 16.0 kHz pure tone as a function of stimulus intensity for the
same 3 experimental conditions shown in Fig. 9. Each curve was
taken from that point in the session showing the peak effect of each
of the respective experimental manipulations.
sure. When the performance measure examined following
noise exposure and pentobarbital administration focused
upon reaction time (i.e., response latency) however, the ef-
fects were quite different. Figure 10 shows the latency in-
creases recorded for each tone intensity during the 1 to 2 hr
period following drug administration (i.e., peak action time)
as compared to the stable, unchanged reaction times rec-
orded over the same intensity range and time period follow-
ing noise exposure and the non-intervention control proce-
dure. (The threshold shift following noise exposure made the
0 dB intensity value inaudible in the course of these determi-
nations with the result that no latencies were obtained at this
intensity and no point is plotted.)
-------
RISK-TAKING AND PSYCHOPHYSICAL FUNCTIONS
81
-VISUAL THRESHOLD, PENTOBARBITAL (17 mg/kg)
-VISUAL THRESHOLD, NO DRUG
•- — — -. -AUDITORY THRESHOLD, PENTOBARBITAL (17 mg/kg)
AUDITORY THRESHOLD, NO DRUG
_ 2
— 2
120
T I M
(mi
180
n )
FIG. 11. Changes in the visual threshold to white light and the auditory threshold to a 16.0 kHz pure
tone as a function of time after IM administration of 17.0 mg/kg pentobaribtal sodium in Baboon S-PE.
Also shown are saline control sessions for both auditory and visual threshold changes. Visual
threshold scale is to the left, auditory threshold scale is to the right.
Although no changes in visual sensitivity were observed
following noise exposure, pentobarbital administration did
significantly elevate visual thresholds. Figure 11 shows the
striking elevation in visual threshold during the first 1 to 2 hr
of an experimental session following pentobarbital adminis-
tration, and contrasts these changes with the unaltered audi-
tory thresholds recorded following identical drug exposure in
the same animal. Control values recorded during experi-
mental sessions without drug are also plotted for compari-
son. These preliminary findings support the differential sen-
sitivity of the psychophysical methodology to modality-
specific effects and the relative independence and selectivity
of the response latency and stimulus threshold measurement
components of the procedure.
Figure 12 summarizes the effects of pentobarbital sodium
as a function of dose (0, 1.0, 3.2, 10.0, 17.0 mg/kg) upon
response latencies and both visual and auditory thresholds in
one of two animals showing similarly toxic effects. The indi-
cated reaction time determinations were made with tone and
light stimuli at the high end of the intensity distribution
where response variability is minimized (see Fig. 7), and
during the peak action time of the drug (i.e., 1-2 hr after IM
administration). The range of control values obtained during
experimental sessions following saline administration is
bracketed in each graph by the broken lines labeled saline
control range. There were clear effects upon reaction time
and visual threshold as a function of dose with both animals,
and no change was observed in auditory thresholds over the
range of doses studied. For baboon S-PE (Fig. 12), doses of
1.0 and 3.2 mg/kg pentobarbital produced no change in any
of the measured functions, though 10.0 and 17.0 mg/kg pen-
tobarbital can be seen to increase response latencies and
visual threshold in a dose-dependent manner. No such
changes in auditory threshold occurred at any of the indi-
cated doses.
Figure 13 illustrates the effects of diazepam as a function
of dose (0, 0.32,1.0, 3.2,10.0 mg/kg) upon response latencies
and sensory thresholds determined with the same two ba-
boons used in the pentobarbital experiment under the same
stimulus conditions and peak drug action time (i.e., 1-2 hr
following IM administration) conditions. The range of saline
control values is again represented within the broken-line
bracketed portion of each graph. With both animals, clear
effects upon reaction time and the visual threshold were ob-
served as a function of dose, and auditory thresholds were
similarly affected. For the most part, all three effects—
increased response latency, visual threshold elevations, and
auditory threshold increases—appeared in a dose-dependent
manner with progressively greater decrements through the
range from 1.0 to 10.0 mg/kg.
The comparative data plots shown in Fig. 14 for baboon
S-PE illustrate the typically-observed differences in duration
of action between pentobarbital and diazepam for a 10 mg/kg
dose of each drug. The previously described effect of both
drugs upon reaction time is represented in terms of response
latency increases over the course of successive trial blocks
(150 trials/block) during a 2-3 hr test session (Day 2) for
pentobarbital on the left and diazepam on the right. Within
24 hr, response latencies had recovered to control levels for
pentobarbital, as shown by the Day 3 values on the left side
of Fig. 14. In contrast, slowed reaction time (i.e., response
latency increases) persisted well beyond 48 hr following
diazepam administration before recovery of control levels
was approximated (Day 5, left side of Fig. 14).
The results of preliminary studies with two additional
compounds are summarized in Figs. 15 and 16. Interestingly,
observations thus far completed with chlordiazepoxide (Fig.
15) over the same dose range (1.0 to 17.0 mg/kg) used in the
pentobarbital (Fig. 12) and diazepam (Fig. 13) experiments,
reveal little or no effect of this benzodiazepine compound
upon sensory thresholds or reaction time.
£/-Methylamphetamine, on the other hand, appears to pro-
duce a selective decrement in visual threshold at the highest
dose studied (0.32 mg/kg) with little or no effect on reaction
time or auditory threshold (Fig. 16).
The conclusions which can be drawn from these two sets
-------
82
BRADY, BRADFORD AND HIENZ
LU
65CH
550-
U «450-
I- E
350-
cc
WHITE LIGHT
S-PE
SALINE CONTROL
RA_NOE_
Saline 1.0 3.2 10.017.0
PE-NTOBARBITAL DOSE (mg/kg)
Q _ 3.2-1
I'm 3.0-
CO c
^2.8-
< 03
CO ~ 2.4-
2.2.
WHITE LIGHT
• • •
| .- ^, ^ SALINE CONTROL
1 A. " RANGE
\
\;
'
Saline 1.0 3.2 10.017.0
PENTOBARBITAL DOSE (mg/kgi
O
i
CO
GO
I
o-
4-
O
Q
Z>
12
16KHz TONE
SALINE CONTROL
RANGE
Saline 1.0 3.2 10.0 17.0
PENTOBARBITAL DOSE (mg/kg)
FIG. 12. A three-part figure showing the dose-related effects of pen-
tobarbital sodium on median reaction time to white light (top), visual
threshold to white light (middle), and auditory threshold to a 16.0
kHz pure tone (bottom). In each case, the dashed lines bracket the
total range in saline control values. All data points were obtained at
the peak action time of the drug, and are from the same Baboon,
S-PE. Visual reaction time data are medians from the highest
stimulus intensity employed.
650-
550-
O
cr
350-
WHITE LIGHT
S-PE
§AL~TNE~CONTROL
RANGE
Saline 0.32 i.o 3.2 10.0
DIAZEPAM DOSE (mg/kg)
3.4 -t
coc
LU 0)
CO
3.0^
;. 2.6-
WHITE LIGHT
SALINE CONTROL
RANGE
Saline 0.32 1.0 3.2 10.0
DIAZEPAM DOSE (mg/kg)
O
I
CO
LU
rr'
Q.
CO
CD
rr.
O
-8-
-4-
0-
- —• —
4 1 —
' 8-
12-
16 KHz TONE
SALINE CONTROL
Saline 0.10 0.32 i.o 3.2 10.0
DIAZEPAM DOSE (mg/kg)
FIG. 13. The dose-related effects of diazepam on median reaction
time to white light (top), visual threshold to white light (middle), and
auditory threshold to a. 16.0 kHz pure tone (bottom) in Baboon
S-PE. Further description as in Fig. 12.
-------
RISK-TAKING AND PSYCHOPHYSICAL FUNCTIONS
83
16 KHz TONE
S-PE
600-
"a500"1
i
1400-
300
200
Sal ine
Pentobarbital,
10 mg/kg
Saline
Diazepam, 10 mg/kg
r
Sal i-ne
.-.-/"—'
Saline
Saline /s
. / Saline
DAY 1 DAY 2 DAY 3 DAY 1 DAY 2 DAY 3
SUCCESSIVE BLOCKS OF TRIALS
DAY 4
FIG. 14. Changes in median reaction time to a 16.0 kHz pure tone for one Baboon, S-PE, showing the
time course of changes in reaction time following IM administration of pentobarbital sodium (10.0
mg/kg) and diazepam (10.0 mg/kg).
of data with respect to the independence of the behavioral
processes involved in the risk-taking and psychophysical
assessment procedures are, of course, limited. In the first
instance, it would appear that overlapping functions may be
represented by the response rate and reaction time measures
which characterize the effects of toxic substances upon the
two assessment methods. The decreases in response rate on
the risk-taking procedure and the dose-dependent increases
in response latency on the psychophysical procedure follow-
ing both barbiturate (Figs. 1, 3, 8, 10, 12, 14) and ben-
zodiazepine (Figs. 4, 5, 13, 14) administration are certainly
consistent with such a relationship. Close inspection of Fig.
16 also suggests a slight decrease in response latencies at
0.32 mg/kg d-methylamphetamine, a finding which accords
well with the modest response rate increases observed under
similar conditions during risk-taking performance (e.g., Fig.
2). By the same token, it seems unlikely that all of the ob-
served pharmacological toxicant effects can be accounted
for by reductionistic appeals to common processes in the
two methodological approaches. The selective changes in
visual threshold following 0.32 mg/kg d-
methylamphetamine, (Fig. 16), for example, occurred in the
absence of any effects upon yellow or red light response
probabilities in the risk-taking procedure under the same
drug condition (Fig. 2). And the differential effects of pen-
tobarbital and diazepam upon early and late yellow respond-
ing (Fig. 1) and upon low and high shock risk performances
(Figs. 3, 4, 5) occurred under conditions where unaffected
components of the multi-operant baseline controlled, to
some extent at least, for sensory threshold changes. The
early and late yellow light stimuli were the same in the exper-
iment shown in Fig. 1 even though performance change oc-
curred only during the late yellow. And the same yellow light
stimulus was used for both the low and high shock risk con-
ditions shown in Figs. 3, 4, and 5 even though the perform-
ance changes were restricted to the early yellow response
rates during low shock risk. Of course, the possibility of
complex interaction effects (e.g., between shock probability
and visual threshold under drug conditions) can not be ruled
out, but it would seem reasonable to conclude on the basis of
available data that the risk-taking procedure may provide an
approach to assessing the effects of toxic substances upon
aspects of a behavioral repertoire which are not necessarily
coextensive with sensory-motor processes.
ACKNOWLEDGEMENT
The research reported in this paper is supported in part by a grant
from the National Institute on Drug Abuse No. DA-00018 and DBA
Contract DEA78-9.
-------
84
BRADY, BRADFORD AND HIENZ
650-,
550-
^ 0>
O tn450.
o"
LU
CC
350-
WHITE LIGHT
S-PE
SALINE
CONTROL RANGE,
Saline 1.0 3.2 10.0 17.0
CHLORDIAZEPOXIDE DOSE (mg/kg)
Q S.BT
WHITE LIGHT
LU o>
CC-o 3.2--
I
!_ 0)
> g
CD
CT ~o
o ~
4- .
• \ OWIN i n^/i
12-
r>
<
20
16 KHz TONE
SALINE
CONTROL RANGE
Saline 1.0 3.2 10.0 17.0
CHLORDIAZEPOXIDE DOSE
(mg/kg)
FIG. 15. The dose-related effects of chlordiazepoxide on median
reaction time to white light (top), visual threshold to white light
(middle), and auditory threshold to a 16.0 kHz pure tone (bottom) in
one Baboon, S-PE. Further description as in Fig. 12.
650-1
550-
O
< 350-
UJ
DC
WHITE LIGHT
S-PE
SALINE
CONTROL RANGEv
Saline 0.01 0.032 0.10 0.32
d-METHYLAMPHETAMINE DOSE (mg/kg)
Q
3.4-1
I'm
CO c
LU CD
DC -a
IE.
3.2-
3.0-
,= 2.8-
CO —
>
2.6-
WHITE LIGHT
SALINE
CONTROL RANGE'
Saline 0.01 0.032 0.10 0.32
d-METHYLAMPHETAMINE DOSE (mg/kg)
O
I
co
LU ^
X Q_
4-
8-
>- CO
rr -a
0-12H
Q
Z)
16
16 KHz TONE
SALINE
CONTROL RANGE
< Saline 0.01 0.032 0.10 0.32
d-METHYLAMPHETAMINE DOSE (mg/kg)
FIG. 16. The dose-related effects of rf-methylamphetamine on me-
dian reaction time to white light (top), visual threshold to white light
(middle), and auditory threshold to a 16.0 kHz pure tone (bottom) in
Baboon S-PE. Further description as in Fig. 12.
REFERENCES
Bradford, L. D., F. Grollman and J. V. Brady. The effects of
pentobarbital, diazepam, and J-methylamphetamine on con-
ditioned and unconditioned punishment. Presented at Eastern
Psychological Association, 1979.
Evans, H. L. Early methylmercury signs revealed in visual tests.
In: Proceedings of International Congress on Heavy Metals in
the Environment, Vol. 3, edited by T. C. Hutchinson. Toronto,
Canada: Institute for Environmental Studies, 1977.
Hienz, R. D., F. Grollman and J. V. Brady. Psychophysical
assessment of drug effects on sensory and motor processes in the
baboon. Presented at Eastern Psychological Association, 1979.
Laties, V. G., P. B. Dews, D. E. McMillan and S. E. Norton.
Behavioral toxicity tests. In: Principles and Procedures for
Evaluating the Toxicity of Household Substances. Washington,
D.C.: National Academy of Science, 1977, p. 111.
Stebbins, W. C. and S. Coombs. Behavioral assessment of
ototoxicity in nonhuman primates. In: Behavioral Toxicology,
edited by B. Weiss and V. G. Laties. New York: Plenum Publ.
Corp., 1975, p. 401.
Weiss, B. The behavioral toxicology of metals. Fedn Proc. 37:
22-27, 1978.
Weiss, B. and V. G. Laties. Behavioral Toxicology. New York:
Plenum Publ. Corp., 1975.
Wood, R. W., A. B. Weiss and B. Weiss. Hand tremor induced
by industrial exposure to inorganic mercury. Archs Envir. Hlth.
26: 249-252, 1973.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 85-92. ANKHO International Inc., 1979.
Operant Conditioning of Infant Monkeys
(Macaca fascicularis) for Toxicity Testing
DEBORAH C. RICE
Toxicology Research Division, Health Protection Branch, Food Directorate
Bureau of Chemical Safety, Health and Welfare Canada, Tunney's Pasture
Ottawa, Ontario, Canada
RICE, D. C. Operant conditioning of infant monkeys (Macaca fascicularis) for toxicity testing. NEUROBEHAV.
TOXICOL. 1: Suppl. 1. 85-92, 1979.—A technique has been developed that allows infant monkeys to perform on an operant
schedule as soon as they are able to self-feed. Behavior is shaped in small increments through a series of operants; sensory
and motor systems as well as performance on schedules using intermittent reinforcement may be tested as early as 3-4
weeks of age. This is accomplished by exposing the infant to the operant situation almost continuously, and allowing the
infant to feed only by operantly responding. Infants exposed to lead post-natally differed from controls in pattern of fixed
ratio responding, "activity" as measured by pattern of responding over the course of the session, and on a two-choice form
discrimination reversal learning set paradigm. This technique allows rapid accumulation of large amounts of data without
experimenter intervention.
Operant conditioning Infant monkey Visual discrimination Discrimination reversal Fixed ratio Ac-
tivity Lead Neonatal exposure
THE developmental study of perception and learning in
newborn animals is of interest to investigators in many areas
of research, including behavioral toxicology. Even though in
many cases the primate offers the best model for understand-
ing the functioning of the human nervous system, there have
been only a few studies using operant responding to study
development of infant primates. Reasons for the lack of in-
formation on young primates include the poor motor coordi-
nation in the very young primate as well as the necessity of
monitoring the infant in its home environment under unre-
stricted conditions.
Recently an operant technique that allows collection of
large amounts of data on infant rhesus or pigtail macaques
performing in their home cage under ad lib motivation con-
ditions has been used to study the development of the visual
system [1,4]. This paper describes modifications to this basic
technique that allow a relatively large number of infant mon-
keys to be studied up until one year of age or longer with a
minimum amount of electronic and mechanical maintenance.
METHOD
Apparatus
Monkeys were housed from Day 1 of life in stainless steel
cages (38.1x48.3x48.3 cm), with a wire mesh floor and pan
underneath to collect urine and feces (Fig. 1). Pairs of cages
were supported by an angle iron frame with wheels. Each
cage opened from the top. The center of the front of the cage
was cut away, surrounded by slides so that "fronts" of
different types could be attached easily and adjusted to any
height. Fronts were of two types: 0.63 cm stainless steel bars
spaced 3.8 cm apart, and a clear Plexiglas sheet containing a
"mask" and touch-bar attachment (Fig. 2). The mask con-
sisted of a hemisphere with eye and mouth holes, with arm
holes beneath, made from mouth guard material. The first
mold was made by inserting a rubber ball halfway into Jel-
trate caulk; a positive was then made from dental cement to
be used as the template for the mouth guard material, which
was melted over the cast in an oven.
The behavioral apparatus was mounted on a frame made
from aluminum rods held together by Flexiframe clamps
(Fisher Scientific Co., Pittsburgh, PA), so that all levers,
feeders, etc., were independently adjustable in three dimen-
sions. This frame was also on wheels and was attached to the
cages by a rod fitting into a holder at either end of a pair of
cages. The frame containing the electrical equipment there-
fore could be disconnected easily from the housing units.
With the exception of the Plexiglas fronts, the cages were
washed by the automatic cage-washing equipment used by
the primate facility.
The feeding apparatus for infants under 30 days of age
consisted of a 110 V AC pull-type solenoid to which 0.32 cm
stiff Teflon tubing with a rubber nipple at one end could be
attached with clamps (Fig. 3). This tubing rested inside a
Plexiglas tube which fit up against the mask, thus protecting
the Teflon tubing from being grabbed and pulled into the
mouth hole by the monkey. The Teflon tubing was attached
by flexible tubing to a reservoir which sat below the level of
the nipple. Operation of the solenoid moved the nipple into
the mouth hole, thus allowing the monkey to suck. Backflow
into the reservoir was prevented by a one-way glass valve in
the line.
Infants over 30 days of age were reinforced with a small
85
-------
RICE
FIG. I. Pair of stainless steel units for operant conditioning of infant monkeys, with steel bar fronts and feeders with steel drinking times.
This configuration is for older infants responding on push-buttons.
amount of formula delivered by a 110 V AC Skinner (Skinner
Precision Industries, Inc., New Britain, CT) two-way
solenoid valve connected to a steel drinking tube which prot-
ruded about 2.5 cm into the cage.
Control of experimental procedures and data collection
were performed using the state notation language SKED [3]
operating on a Nova 3 minicomputer (Data General Cor-
poration, Southboro, MA). Data were collected by recording
every interevent time to the nearest 10 msec; thus, a session
could be precisely reconstructed from the raw data.
Experimental Procedure
The floor of the cage was covered with paper diapers for
young infants, and all infants had a surrogate attached to the
side of the cage and a paper diaper free in the cage. A heating
pad was used for the first few days of life.
Infants were housed in the cages 24 hours a day with
access to infant formula via a rubber nipple until they learned
to self-feed. Nursery personnel held them to the nipple on a
fixed schedule until they learned to find it themselves. After
they had been self-feeding for three days, they were exposed
to the following sequence of schedules:
(1) The infant interrupted a photocell beam by placing id
face in the mask in order to gain access to the nipple for II
sec. The beam had to remain interrupted for 0.2 sec (2 s»
sions), 0.5 sec (2 sessions), or 1 sec (2 sessions). If the beaa
were broken when the 10 sec feed time ended, the infant
could simply keep it broken for the specified time in order to
gain further access to the nipple.
(2) The infant touched a "touch bar" (contact sensw)
directly outside the arm holes while the mask photocell wa
interrupted. The infant could use either hand, and left and
right touches were recorded separately. Initially contact witk
the touch bar was made accidentally when the infant grabbed
the arm holes for support while feeding. The touch bar WM
moved further away from the Plexiglas front on successive
days so that the monkey had to reach progressively further
to respond. Infants that would be required to use both handl
had one arm hole blocked every session in an alternatin|
fashion once they had used their preferred hand for 2 ses-
sions. The touch bar was moved 1.26 cm every other night;
they were required to respond with their preferred and the«
non-preferred hand at each bar placement. Infants that
would respond with one hand (such as on a fixed ratio
schedule) were allowed to use only their preferred hand.
(3) At 30-40 days of age, the rubber nipple was replaced
-------
OPERANT CONDITIONING OF INFANT MONKEYS
87
FIG. 2. Close-up of Plexiglas front with mask, touch bar attachment, and rubber nipple feeder. The arm hole is on the side not visible.
with a drinking tube, and the photocell was no longer used;
0.2 ml of formula was delivered at each reinforcement.
(4) When the infant worked with its preferred hand or
either hand with the bar 5 cm from the front of the cage, a
push button with a colored light behind it was substituted for
the bar on the side of the preferred hand. Infants typically
pushed it spontaneously with no shaping necessary. Infants
performed on the push button for two sessions using the
preferred hand or two sessions with each hand.
(5) Once the infant responded on the push button, it was
put on one of two schedules. One of these was a non-spatial
discrimination, in which the final form discrimination was
shaped by a series of easier tasks. Infants were first intro-
duced to a red-dark discrimination. When they performed at
85% or above correct for five sessions, a green light was
introduced as the negative stimulus. When the same criterion
was met (85% correct or better for five sessions), a cross was
superimposed on the positive stimulus and a triangle on the
negative. When criterion was met, the colors were removed.
A series of up to 10 reversals were then performed, the cri-
terion for reversal again being 85% correct or better for five
sessions.
The other schedule was a fixed ratio (FR) 1 with the infant
using its preferred hand, followed by FR10, 20, 30 and 40.
Each FR value was tested for 14 sessions. Infants were
moved from FR1 to FR10 using intermediate FR values over
the course of three sessions.
The eating pattern (or response pattern) of each infant
throughout the course of the session was also determined.
Until the infants were 45 days old, sessions were 21 hours
long, from 3 p.m. to 12 noon. Sessions were then 16 hours
long, from 3 p.m. to 7 a.m. All infants received all their
formula during the session; there were no supplemental feed-
ings. Infants also received fresh fruit and primate diet as per
standard nursery procedure.
When infants were not performing on an operant
schedule, they were housed in standard nursery clear
polycarbonate housing units. Peer socialization began at 20
days of age; infants were placed in pairs in these housing
units for one hour each day. They were gradually introduced
in small groups to the large exercise cages, and allowed to
remain for increasing periods of time. At 45 days of age,
when they were in the operant cages for 16 hours overnight
only, they were housed in the large exercise cages with mon-
keys around their own age for approximately five hours per
day two or three times a week.
RESULTS
Once the infant was self-feeding, there was a minimum of
-------
88
RICE
.„***•**
FIG. 3. Rubber nipple feeder with pull-type solenoid. Solenoid activation pulls nipple forward: spring returns it to original position. Note
one-way glass valve to prevent backflow of formula.
TABLE 1
NUMBER OF SESSIONS (DAYS) FOR EACH INFANT TO
PROGRESS THROUGH SERIES OF OPERANTS REQUIRED TO
LEARN TO PRESS THE BUTTON. THE NUMBER IN THE
"SELF-FEED" ROW REFERS TO THE AGE OF THE INFANT
WHEN IT WAS ABLE TO FIND THE NIPPLE BY ITSELF. THE
"MASK" PROGRAMS REQUIRED THE INFANT TO INTER-
RUPT A PHOTOCELL BEAM ACROSS THE FACE MASK IN
ORDER TO GAIN ACCESS TO THE NIPPLE; THE "TOUCH"
PROGRAMS REQUIRED THE MONKEY TO TOUCH A CON-
TACT SENSOR WHICH WAS MOVED PROGRESSIVELY
FARTHER FROM THE CAGE FRONT. THE "PUSH-BUTTON"
PROGRAMS REQUIRED THE INFANT TO PRESS A BUTTON
WITH A RED LIGHT BEHIND IT, FOR TWO SESSIONS WITH
EACH HAND. THE NUMBER IN THE "MINIMUM REQUIRED"
COLUMN REFERS TO THE MINIMUM NUMBER OF SESSIONS
POSSIBLE TO PROGRESS THROUGH THE SERIES.
Minimum
Schedule Required
No. 86
No. 88 No. 89
Self-Feed
Mask
Touch Bar
Push Button
6
16
4
Day 3
6
8 (preferred)
17 (both)
4
Day 7
7
19
4
Day 5
6
16
4
26 (both hands) or 16 (one hand) sessions required until the
push button response was established. Of the first three
monkeys exposed to the procedure requiring responding
with both hands, one completed the series in the minimum
number of sessions and another required four extra sessions
(Table 1). These results are typical of infants exposed to this
regimen (now numbering 26 infants). The third infant was
allowed to use only its preferred hand until it acquired the
push button response; an unsuccessful attempt was then
made to transfer this response to the non-preferred hand.
The touch bar was then introduced on the non-preferred
side, and the sequence for two hands was initiated.
This method of shaping the behavior of the infants in
small increments insured that the formula consumption of
the infants did not suffer. This is clear from a comparison of
the three pilot infants with the three non-operant infants
closest in age (Fig. 4). In fact, infants required to make an
operant response in order to receive food actually consumed
more formula than those fed ad lib. Ad lib control monkeys
were offered a total of 240 ml of formula per day, while the
amount of formula consumed by the infants on the operant
study was not limited. This could artificially lower the con-
sumption for the ad lib control monkeys. Maximum formula
consumption occurred in 4% of feedings for one monkey,
10% for another, and 18% for the third. This was not a factor,
however, until at least 90 days of age.
After infants learned the push-button response with either
hand, they responded readily on the first session of the
light-dark discrimination. Infants learned each discrimina-
tion in an orderly fashion with progressive improvement in
performance across sessions, as demonstrated by the three
pilot infants (Fig. 5). These untreated infants also tended to
have fewer errors on the first session of successive reversals.
-------
OPERANT CONDITIONING OF INFANT MONKEYS
89
300 r
200
„ 100
1
O
5 0
-F--f-----
40 L
FIG. 5. Progression on the two-choice non-spatial discrimination for three untreated infants. Filled triangles represent light (red)—dark
discrimination: X's , red-green discrimination: and open triangles, red plus cross-green plus triangle. Filled circles represent the cross-
triangle discrimination and series of discrimination reversals. Breaks in the line connecting the circles represent points at which a reversal
was instituted. The break in the lower two graphs represents 19 sessions of lost data, to reversal 9 for the moddle graph and reversal 6 for the
bottom.
-------
90
RICE
50-1
40-
o
030H
10
o •
o •
•f
o
X
1 -.1
Of +
o
*
o
o +
123456789 0123456789
REVERSAL NUMBER
FIG. 6. Number of errors after the first for the second block of 50 trials in the first session of a reversal. Left, control
monkeys: right, treated. Each symbol type represents a different animal.
SUBJECT NO. 92
FIG. 7. Histograms of FR run times for the 14 sessions of FRIOfora
control monkey. Sessions are ordered from left to right: run times
from bottom to top. Z axis represents 35 sec in 0.5 sec bins.
SUBJECT HO. 93
FIG. 8. Histograms of FR run times for the FR10 sessions for a
lead-treated monkey. Axes as in Fig. 7.
This developmental pattern is quite consistent among un-
treated infants examined thus far.
The activity pattern over the course of the sessions in
which the infants were responding on an FR40 (approx-
imately 120 days of age) for a control and lead-treated mon-
key are in Figs. 10 and 11. The control monkey exhibits the
typical pattern of four to six eating bouts per session, while
the lead-treated monkey feeds more often and drinks less per
bout. Nor was this likely a function of delayed development,
as this pattern of responding persisted until these monkeys
were taken off operant testing at 270 days of age.
DISCUSSION
This paper describes a method for operantly conditioning
infant primates starting as early as the first week of life. The
procedure has several advantages:
(1) The cage design allows the electrical equipment to be
detached from the housing unit, and the cage itself is easily
cleaned by automatic equipment using high temperatures.
Maintenance of a clean environment is also facilitated by all
cabling being away from the floor or walls.
(2) The use of the touch bar allows shaping of the
pushbutton response without experimenter intervention.
This not only saves time, but also insures that each infant is
shaped by the same procedure, thus eliminating uncontrolled
variables introduced by experimenter-subject interaction.
The small increments of the shaping procedure also insure
that the infant will progress through the series without undue
stress or weight loss.
(3) The fact that the infants receive all formula in the
experimental chamber over a long (at least 16 hr) session
obviates the need for restriction of food or water intake. This
is especially important in the young animal, since early nu-
tritional deprivation may have detrimental effects on intelli-
gence in humans [2,5] and primates [6].
Sessions were run during the night because experience in
-------
OPERANT CONDITIONING OF INFANT MONKEYS
91
15
30
* 1
I*
2O(.
.J
8 12
TIME (HOURS)
16
FIG. 9. Response distribution pattern over the course of a session
for an untreated infant 15, 90, 150 and 200 days of age. Points on the
ordinate represent successive 15 minute segments of the session: the
abscissa is number of responses per 15 min segment. All responses
exceeding 60 in a 15 min segment are excluded.
FIG. 10. Histograms of activity across the FR40 sessions for a con-
trol monkey. Z-axis represents time into the session in 15 min in-
crements, Y-axis is number of reinforcements.
SUBJECT HO. 93
FIG. 11. Histograms of activity across the FR40 sessions for a lead-
treated monkey. Axes as in Fig. 10.
nursery rearing infants has demonstrated that this is the
period in which they consume the major portion of their daily
formula intake. This procedure has the added advantage of
freeing the infants during the day for peer socialization, ob-
servation, other sorts of toxicity testing, etc. Cleaning and
maintenance of equipment can also be done without interfer-
ing with the operant sessions.
This study demonstrates that infant macaques may be
operantly conditioned as soon as they can self-feed. The
method described allows the study of perceptual and learn-
ing functions very early in the neonatal primate's life, while
the fact that the infant performs in its home cage under ad lib
motivational conditions eliminates stress due to food depri-
vation, strange surroundings, etc. This is especially relevant
for perinatal exposure to pharmacologic or toxicologic
agents, as these sorts of stress may interact with the toxic
effects of the agent.
REFERENCES
1. Booth, R., D. Teller and G. Sackett. Trichromacy in normally
reared and light deprived infants (Macaco nemislrina). Vis. Res.
15: 1187-1191, 1975.
2. Craviato, J. and B. Robles. Evolution of adaptive and motor
behavior during rehabilitation from kwashiorkor. Am. J Orthop-
sychiat. 35: 449, 1965.
-------
92
RICE
3. Gilbert, S. and D. Rice, Nova SKED: A behavioral notation
language for Data General minicomputers. Behav. Res. Metho.
Instnim. 10: 705-709, 1978.
4. Sackett, G., R. Tripp, C. Milbrath and J. Gluck. A method for
studying visually guided perception and learning in newborn
macaques. Behav. Res. Meth. Instnim. 3: 233-236, 1971.
Stoch, M. and P. Smythe. Does undernutrition during infancy
inhibit brain growth and subsequent intellectual development?
Arc/i. Dis. Childhood 38: 546, 1963.
Zimmerman, R. and D. Geist. Attention deficiencies in mal-
nourished monkeys. In: Symposia of the Swedish Nutritional
Foundation. XII., edited by J. Cravioto, L. Hambraeus and B.
Vahlquist. Uppsula: Almqvist and Wiksell, 1974, pp. 115-126.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 93-103. ANKHO International Inc., 1979.
Effects of Pre- and Post-Natal Lead on Affective
Behavior and Learning in the Rat12
JOHN C. FLYNN, ELEANOR R. FLYNN
Biopsychology Laboratories, Department of Psychology, Baylor University
Waco, TX 76703
AND
JIM H. PATTON
Behavioral Science Laboratory and Division of Biochemistry, University of Texas Medical Branch
Galveston, TX
FLYNN, J. C., E. R. FLYNN AND J. H. PATTON. Effects ofpre- and post-natal lead on affective behavior and learning
in the rat. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 93-103, 1979.—Literature relevant to the relationship between early
ingestion of inorganic lead and subsequent hyperactivity in rodents is discussed. Original research in the area is presented.
Rats so exposed were not hyperactive in any of the situations investigated or under any of the dosage regimens employed.
They did show hypoactivity in the open field when dosed over a prolonged period. Using a new behavior measure,
lead-treated rats were found to be less active than controls in the passive avoidance situation. The possible utility of this
new measure for behavioral and developmental toxicology is discussed. It is concluded that the available evidence does not
support the contention that a meaningful relationship exists between early lead ingestion and hyperactive behavior. It is
suggested that future research may more profitably be directed to assessing the effects of lead ingestion on behavior in
stressful or fear provoking situations.
Lead Hyperactivity Affective behavior Passive avoidance learning Social interaction
Behavioral toxicology
THE demonstration that a relatively low body burden of Much of the work on lead and hyperactivity comes from
inorganic lead may be of etiological significance in childhood two laboratories. This work will be reviewed in some detail.
hyperactivity [3,4] is a matter of considerable interest to tox- It has served as a prototype for much of the later work in the
icologists. This is especially true in view of a number of field, and it has provided a research paradigm that is of con-
recent reports [22,34] indicating that hyperactivity can be siderable interest to developmental and behavioral toxicol-
induced in rats and mice as aresult of the post-natal adminis- ogy. Specifically, this work involves the exposure of devel-
tration of inorganic lead, in amounts less than those which oping rats or mice to lead via the maternal milk supply. Lac-
lead to obvious symptoms of lead intoxication. Whether tating dams are provided with inorganic lead salts in their
such reported hyperactivity is indeed analogous to that diet or drinking water. Nursing neonates are thereby ex-
demonstrated by the hyperkinetic child or the child diag- posed to relatively low concentrations of lead by virtue of
nosed as exhibiting minimal brain dysfunction (MBD) is a suckling these dams. Throughout our review of these studies
matter of considerable scientific and clinical import. If true, we will direct attention to two major methodological points.
an animal model (or at least a first approximation thereof) of The first of these is the necessity, in this research paradigm,
a perplexing clinical entity will have been found. Such a of pair-watering or pair-feeding the control animals, as a min-
possibility has been suggested [35,42]. The research reported imal control for the effects of malnutrition or dehydration.
in the present paper suggests that the presumed relationship When lactating dams are presented with adulterated food or
between early lead ingestion and subsequent hyperactivity in water they eat or drink less than control animals [30]. Their
rats is not a simple one, and is a weak relationship, at best. body weights decline, as does the body weight of their off-
When proper controls are imposed upon the experiment, spring [30]. Malnutrition has been shown to produce
either no relationship between early lead ingestion and later hyperactivity in rats [39,40].
hyperactivity is found, or the relationship is so small as to be The second methodological point stressed is the necessity
of little predictive utility. At times, an inverse relationship is of removing from the dependent variable, by either experi-
found, in which lead exposed animals are less active than mental or statistical means, apparent treatment effects which
control animals. are in reality effects that result from running intact groups.
'Supported in part by grants to John C. Flynn by the Baylor University Research Committee.
2The authors are grateful to Sterling-Winthrop Research Institute for supplying the MMTA.
93
-------
94
FLYNN, FLYNN AND PATTON
That is, it is necessary to control for pre-existing litter differ-
ences in this kind of developmental study. The necessity for
this type of control is well known in behavioral research and
has been forcefully stated [6]. In some of the studies re-
viewed, reporting of experimental procedures and statistical
analysis is fragmentary. It is, therefore, sometimes difficult
to be certain that controls for litter effects were employed.
In one series of experiments the focus has been on the
role of lead in producing hyperactivity in developing mice. In
the first of these studies [34] suckling mice were raised by
dams whose drinking water contained lead acetate, Pb(Ac)2,
in concentrations of 2 mg/ml, 5 mg/ml, or 10 mg/ml. Lead
treated offspring were reported to be three times more active
than the offspring of dams who drank a sodium acetate
(NaAc) solution. Control dams in this study were not pair-
watered. Development was retarded in the lead-treated ani-
mals, and some of them showed ataxia and disturbed gait. It
does not appear that litter effects were controlled in this
study.
In a similar study [35] where experimental dams received
Pb(Ac)2 while controls received NaAc, ataxia and disordered
gait are again reported in the lead-treated offspring. In fact,
while an increase in activity is reported for the lead-treated
mice, the authors state that activity was measured in animals
chosen for an absence of motor symptoms. Such non-
random selection of subjects provides additional opportunity
of observing spurious treatment effects. Lead animals are
reported to respond "paradoxically" to d-amphetamine and
phenobarbital in a manner consistent with the presumed re-
sponse of hyperactive children to these drugs [17,43]. No
growth data are presented for the developing mice. Pair-
watering was not utilized, and litter effects apparently were
uncontrolled.
In a subsequent study [36] these authors report on several
neurochemical studies of the lead exposed mouse. Effects
on the high affinity uptake of certain neurochemicals are
reported. Also, lead-treated animals are reported to differ
from controls in their response to a variety of pharmacologi-
cal agents. We will return to biochemical and pharmacologi-
cal considerations below.
In another series of studies the focus has been on the
relationship between lead and hyperactivity in rats. In the
first report [32] lactating rat dams were fed powdered labora-
tory chow containing 4.0% lead carbonate, Pb(CO3)2. In this
study, control dams were pair-fed. The activity of entire
groups of offspring was measured, rather than individual ac-
tivity. The offspring of lead mothers were said to show a40%
to 90% increase in activity when compared to controls.
While pair-feeding was used in this experiment, it is probable
that litter differences were not controlled. Lead animals also
were reported to show a 20% decrease in brain dopamine
(DA) relative to controls, but were not different from con-
trols in brain norepinephrine (NE). In a similar study [22]
these authors report on the activity of the offspring of dams
consuming powdered chow containing 5.0% Pb(Ac)2.
Group measures of activity were again used, and lead ani-
mals were found to be more active than controls. Pair-
feeding was again employed. The necessity of so doing was
demonstrated in this study by comparative growth curves.
Unfortunately, there is no evidence of any attempt to control
for litter effects. In a subsequent experiment [16] lead ace-
tate was administered to dams, as above, and the behavior of
the offspring was compared to the behavior of rat pups re-
ceiving either Pb(Ac)2 or NaAc via oral administration. The
growth of the offspring in the first group was markedly de-
pressed when compared to the growth of the other two
groups, suggesting some methodological advantages to be
found in the direct administration of lead to pups as opposed
to their dams. In 24 hr measures of group activity, the
Pb(Ac)2 animals were more active than the NaAc animals.
The authors were unable to replicate their previous findings
concerning NE and DA concentration. Again, there was no
control for the effects of litter differences. Finally, from the
same laboratory, is a report [18] of the effects of lead deliv-
ered to dams in drinking water as Pb(Ac)2. Animals were not
pair-watered. Twelve experimental and 12 control animals
were used in individual activity measures. Since animals
were run individually rather than in intact litters, litter ef-
fects may have been controlled. This cannot be determined
from the research report, since the manner of selection of the
12 animals is not specified. At any rate, no effects of lead on
activity were found, leading the authors to speculate that the
effects of lead on activity may be ". . minimal or evanes-
cent in nature."
The work reported above is frequently cited as evidence
of the link between lead and hyperactivity in rodents and as
being of relevance to the etiology of hyperactivity in children
[24,44]. Reports from other laboratories using a variety of
methods to expose developing rat pups to lead have yielded
varying and conflicting results. Thus, when male rats were
given daily oral doses of lead acetate during their first three
weeks of life, they did not differ from control rats in spon-
taneous activity measured in a photoactometer [42]. (How-
ever, on the basis of peformance in a shuttle task and the
response of the animals to amphetamine, these authors
suggest that lead may be of etiological significance in MBD.
Unfortunately, the data presented in this report are insuffi-
cient to evaluate this suggestion. While the authors ran
enough litters per condition to evaluate this source of vari-
ance statistically, this evaluation was not done. Student's t
was employed instead of the appropriate analysis of vari-
ance.)
Other investigators have examined the lead-behavior re-
lationship in a variety of situations, following a variety of
exposure routes. Thus, rat pups have been dosed
intraperitoneally [1], dams have been injected during gesta-
tion and following parturition [41], and dams have been fed
lead in diets, as described above [1], or in drinking water or
by gavage [1]. The behavior of offspring has been investi-
gated in the open field and in a perceptual discrimination task
[45], in a modified Hebb-Williams maze [41], and in a
T-maze. No clear pattern of lead related effects is evident in
these studies.
The methodological difficulties in much of the work
above have been recognized and succinctly stated by one
group of workers [20]. This group has conducted one of the
few methodologically sound pieces of research to be found in
this area. Pair-fed control animals were used, litter effects
were controlled, and appropriate statistical analyses were
conducted. In contrast to many previous reports, these
authors present sufficient data and statistical analyses to
permit the reader to assess the appropriateness of their con-
clusions. These authors were unable to demonstrate any lead
induced hyperactivity in their rats. Their results are at vari-
ance with the literature cited above. They question both the
purported relationship between lead and hyperactivity in rats
and the relevance of this animal model to childhood
hyperactivity. Of some interest is the report by these authors
that the lead-via-dam's food route of lead administration was
not satisfactory. Dams refused to eat the adulterated food.
-------
LEAD EFFECTS ON HYPERACTIVITY AND LEARNING
95
The authors were thus forced to intubation of dams as a
technique for delivery of lead. We had similar problems in
our laboratory at the outset of our work, observing refusal of
adulterated food and water and some related instances of
cannibalizing of litters by experimental dams.
The lack of relationship between lead exposure and
hyperactivity observed by these investigators is consistent
with findings in our laboratory. We have raised approx-
imately 100 litters of Long-Evans rats in this effort. We have
experimentally treated and/or dosed well over 550 of these
rats in various studies of the effects of pre- and post-natal
lead exposure on behavior and brain chemistry. When the
appropriate experimental controls are instituted, we are un-
able to demonstrate a convincing relationship between such
lead exposure and hyperactivity. Neither are we able to
demonstrate a consistent relationship between lead and be-
havioral variables thought to reflect learning and/or affective
states related to hyperactivity.
Some of these results have been reported elsewhere [13,
29, 30] and will be briefly summarized here. In the first phase
of this work lactating dams were fed 2.0% Pb(Ac)2 in pow-
dered chow. The effects of this administration on the open
field activity of the offspring and on the response of these
animals to amphetamine and phenobarbital were examined.
The open field was identical to that described below under
Methods, except that it was painted white. Intact litters were
not run. Litter effects were controlled by randomly assigning
animals to drug treatments within both lead and control
groups. No lead related differences in activity were found,
nor were there any differences in the amount and pattern of
rearing and grooming. Lead animals and control animals did
not respond differentially to drug treatment. No "paradoxi-
cal" effect of amphetamine was found.
In this study we did not employ pair feeding. A compari-
son of pup weight, dam weight, and dam food consump-
tion provided striking evidence for the necessity of employ-
ing pair feeding procedures when lead is administered in this
way. The weight of experimental dams was significantly
below that of controls throughout the pre-weaning period.
Experimental dams show drastic reductions in food intake,
virtually ceasing to eat upon the introduction of food contain-
ing 2.0% Pb(Ac)2. As might be expected, weights of experi-
mental pups lag behind those of controls.
In the second phase of this study, a dose-response inves-
tigation was carried out relating oral dosage of neonatal pups
to brain lead concentration, activity and other behaviors in
an open field, and passive avoidance behavior. Oral adminis-
tration was used in an effort to increase the reliability of lead
delivery to pups. This strategy was apparently effective.
Over a seven dose range, from 0 mg/kg to 220 mg/kg, brain
lead concentration was shown to increase monotonically
with oral dose. Brain weight of pups, on the other hand, was
shown to decrease monotonically with oral dose. Oral dosing
of neonates proved to be a reliable means of delivering lead
to experimental animals. Many of the difficulties of feeding
adulterated chow are circumvented in this way.
Results of the behavioral analysis in this study are an
object lesson in the need to control for litter effects. When
litters are ignored as a source of variance, a significant
analysis of variance treatment effect of lead is found for
number of squares crossed, number of center squares
crossed, and time spent in rearing. In each case the lead
treated animals differ from controls in the expected direc-
tion. However, when the data are treated in a hierarchical
design in which litters are treated as a source of variance,
only the number of squares crossed shows a statistically reli-
able effect of lead. And, the linear relationship between lead
and number of squares crossed is weak, lead treatment ac-
counting for only four percent of the variance in squares
crossed. Finally, there are reliable effects of litter on brain
lead concentration. Thus, even though lead is directly ad-
ministered to the offspring, litters are apt to differ in the
amount of lead which is found in brain following a given
dose.
In addition to an interest in behavioral concerns, some
investigators have examined the possible relationship be-
tween lead exposure and brain chemistry. Much of this in-
terest stems from the suggestion that dysfunction in brain
amines might be involved in hyperkinetic or MBD children
[43]. Conflicting reports of the effect of lead on DA and NE
have been cited. Other relationships between lead and brain
chemistry have been reported. Lead treated and control
mice are reported to differ in their response to a variety of
pharmacological compounds [36]. Significant differences in
high affinity synaptosomal transport of choline, dopamine,
and tyrosine have been reported by the same investiga-
tors. Other investigators have examined the effects of post-
natal lead on brain RNA, DNA, DA, NE, 5-HT, and GABA
(y-aminobutyric acid), finding either no effect, or inconsis-
tent and conflicting effects in the case of brain
catecholamines [16, 21, 32].
We now report on three additional studies done in our
laboratories. These involve both pre-natal, and pre-natal
combined with post-natal administration of lead. In addition
to measures of general activity, we present data on behaviors
reflecting learning and affective states presumed relevant
either to the hyperactive child syndrome or to some
presumed effect of lead on the CNS. Finally, we present the
effects of chronic lead administration on certain aspects of in
vivo and in vitro brain chemistry.
GENERAL METHOD
Some general statements are relevant to all studies con-
ducted in the Baylor laboratories. Animals were maintained
on a 12 hr light-dark cycle, lights on at 15:00 hr (see excep-
tion under Shuttle Avoidance, Study 3, below). All behav-
ioral testing was done between 09:00 hr and 14:00 hr during
the animals' dark, or active cycle. Behavioral testing was
conducted under low level, red light illumination, except in
those instances where level of illumination was an indepen-
dent variable. Background masking noise was provided for
all behavioral tests, and, with the exception of a large appa-
ratus such as the radial arm maze described below, each
behavioral apparatus was located within an isolation cubicle.
Feeding, watering, weighing, dosing, etc. of animals was
done between 15:00 hr and 18:00 hr. Lead contaminated
animals were housed separately from controls, in a nega-
tively pressurized room. Cross contamination of the control
colony from the atmosphere of the lead colony was therefore
avoided. All animals were bred in our colony, rather than
being purchased when pregnant. The desirability of this ap-
proach to developmental studies has been forcefully stated
[6].
STUDY 1
Animals
All animals were Long-Evans rats bred in the Baylor ani-
mal colony. Two females and one male were housed in a
-------
96
FLYNN, FLYNN AND PATTON
plastic breeding cage for four days. On the fifth day the
females were housed separately and randomly assigned to
either experimental or control groups. There were 16 exper-
imental animals and 16 control animals.
Treatment
Upon being housed separately, animals were maintained
on Purina Rat Chow. The experimental females were placed
on a 0.5% solution of Pb(Ac)2 in lieu of normal drinking
water. Control females were pair-watered each day by giving
them a volume of tap water equal to the consumption of the
paired experimental female on the previous day. This proce-
dure was continued until the pups were weaned on Day 22
following parturition, parturition being designated as Day 0.
Litters were culled to 8 pups (5 male and 3 female where
possible) on the second day following parturition. Not all
females conceived, and there were thus 10 experimental
dams and 9 control dams to provide offspring for the experi-
ments. One male pup from each of 10 experimental and one
male pup from each of 8 control litters were used in behav-
ioral testing.
Chemical Analyses
Whole brain lead was assayed when pups were 31-34
days old, using an organic solvent extraction modified in our
laboratory [30] from a procedure used to assay lead in blood
[46]. Brains were homogenized in 0.5% HNO3, lead was che-
lated with 1-pyrrolidinecarbodithioic acid (APDC) and then
extracted into methyl isobutyl ketone (MIBK) for assay via a
Perkin-Elmer Model 403 Atomic Absorption Spec-
trophotometer. Lead determinations were carried out on 9
experimental pups from 5 litters and on 11 control pups from
6 litters.
Brain calcium in hippocampus, striatum, and residual
brain was analyzed by atomic emission spectro-
photometry using the Perkin-Elmer Model 403. Brain dis-
section was by standard methods [15]. Tissue from hip-
pocampus and from the striatum was placed in ignition tubes
to which 0.2 ml concentrated nitric acid was added. The
tubes were then heated on a hot plate, at a temperature not
exceeding 110°C. This process of adding concentrated HNO;)
was repeated until a white ash was obtained. That is, if the
initial 0.2 ml concentrated nitric acid was not sufficient to
produce a white ash, another 0.2 ml was added. The residual
brain was homogenized in 5.0 ml of 0.5% (v/v) HNO3. Two
1.0 ml aliquots of the homogenate were then dried and di-
gested as above. When the tissue was reduced to a white ash
it was taken up with a solution composed of equal volumes of
0.1 N HC1 and 2.0% (w/v) potassium chloride. This proce-
dure is essentially the same as that used by other inves-
tigators [2]. The solutions so obtained were then analyzed for
calcium using a Perkin-Elmer Model 403 with a nitrous
oxide-acetylene flame.
Brain lead determinations at two days of age were ac-
complished on animals raised similarly to those above.
Breeding was continued for seven days rather than four.
Following breeding, experimental animals were given 0.5%
Pb(Ac)2 for drinking water and controls were pair-watered.
On Day 3 following parturition, 6 experimental animals (2
males from each of 3 litters) and 6 control animals (2 males
from each of 3 litters) were sacrificed by chloroform asphyx-
iation and their brains excised for lead analysis. The tissue
was wet ashed in warm, concentrated HNO:!. The white ash
was taken up in 0.5 ml of 0.1 N HNO3 and neutralized with
0.5 ml of 0.1 N NaOH. Five successive 100 ml aliquots of
this solution were dried in Delves Cups [5] and analyzed by
atomic absorption spectrophotometry.
Radial Arm Maze
A radial arm maze was constructed which was a modifi-
cation of one described earlier [27,28]. In the present case
the maze contained eight arms of length 86 cm, enclosed with
particle board floor 7 cm wide and particle board walls 4 cm
high. The tops of the arms were made of hardware cloth.
These arms were attached to an enclosed center octagon of
37 cm diameter made of particle board and fitted with a
hinged top of Plexiglas. The center octagon and each arm
had removable Plexiglas floors. During 12 days of testing in
the radial arm maze the following procedure was used:
Days I through 5: The rat was placed in the center octa-
gon and remained in the maze for at least 5 min. At the end of
the 5 min period it was removed, provided it had entered
each of the 8 arms at least once. The number of arm entries
during this 5-min period provided a measure of activity in a
novel, nonstressful situation. The number of different arms
entered in the first 8 entries provided a measure of spontane-
ous alternation under conditions of no reward and no depri-
vation. Spontaneous alternation has been associated with
hippocampal function [28], and lead has been shown to ac-
cumulate in the hippocampus [14]. If the animal had not
entered all 8 arms at the end of the 5-min period, it was
permitted to remain in the maze until it had done so or until
an additional 5 min had elapsed.
On Day 5, after testing, water deprivation was begun.
Both control and experimental animals were given water for
only one hr per day, between 15:00 and 16:00 hr. Effectively,
this resulted in a mean deprivation of 17 hr at the time of
testing.
Days 6 through 10: The testing procedure described
above, including deprivation, was continued except that a
large drop of water was placed in the drinking cup at the end
of each arm. In this way the effect of deprivation and reward
on activity levels and spontaneous alternation performance
could be observed.
Days 11 and 12: The procedure described above was con-
tinued except that animals were removed from the maze as
soon as they had entered all 8 arms. All animals had done so
within the first 5 min.
STUDY 2
Animals
All animals were Long-Evans rats bred in the Baylor ani-
mal colony. Two females and one male were housed in a
breeding cage. On the fifth day of breeding cage residence,
the females were removed and randomly assigned to the lead
or the control group. There were 6 lead females and 6 control
females. Two male offspring of each of these dams were
randomly sampled for inclusion in behavioral testing. Brain
chemistry was done on both male and female brains.
Treatment
Upon being housed individually, pregnant animals in the
experimental group were placed on a water supply that con-
tained 0.2% Pb, as lead acetate. The volume of water each
experimental female consumed each day was determined
-------
LEAD EFFECTS ON HYPERACTIVITY AND LEARNING
97
and this volume of water was given as the next day's ration
to a pair-watered control female. This procedure was contin-
ued until parturition. Beginning at birth, each litter was
weighed daily and a mean pup weight determined for each
litter. Lead animals were then individually administered lead
acetate via the oral cavity. Dose was 225 mg of lead per
kilogram of body weight (i.e. mean pup weight of the litter)
administered daily in a constant volume of 0.0025 ml per
gram of body weight. A 1.0 ml tuberculin syringe with needle
ground smooth was used to administer the lead solution.
Paired control litters were dosed in similar fashion with dis-
tilled water. At weaning, lead animals were housed as litters
and given drinking water containing 0.25% lead, as lead ace-
tate. Control litters were similarly housed and were provided
with tap water in a volume determined each day by the con-
sumption of their paired lead litter. Animals were separated
by sex on Day 25 and pair-watering was continued through-
out the experiment. Thus, these animals received lead pre-
natally and for the duration of their lives. Behavioral tasks
were accomplished in the order listed in this section. All
males to be used in behavioral testing were individually
housed on Day 30. Behavioral testing was carried out from
Day 49 to Day 58.
Biochemical Determinations
Animals were sacrificed by chloroform asphyxiation.
Brain lead was assayed as in the first study by atomic ab-
sorption spectrophotometry using an organic solvent ex-
traction for whole brain. Butanol extraction [19] was em-
ployed in the fluorometric assay of 5-Hydroxy-tryptamine
(5-HT). Brains were dissected [15] into hippocampus,
striatum, brain stem, and cortex, and 5-HT assayed sepa-
rately in these regions. The uptake by striatal tissue minces
of a-methyl-meta-tyramine (MMTA) was investigated.
Minces were incubated for 30 min in a 0.2 mg/ml concentra-
tion of MMTA. Following arrest of metabolic processes at
the end of this period by application of ice cold Krebs'-
Ringer wash, tissue was homogenized in 0.4 N perchloric
acid and extracted with a butanol:heptane mixture. The in-
doleamine was reacted with o-pthalaldehyde and assayed
fluorometrically. The complete procedure, composition of
Krebs'-Ringer, etc. is described elsewhere [7, 8, 33].
Radial Arm Maze
The radial arm maze described above was used. In Study
2 animals were run following the procedure of the previous
study, except that the animals were run only one day for a
five-min period. Measures of activity and of spontaneous
alternation were taken as described above.
Open Field
The open field was a cube with 60 cm sides, painted flat
black, and isolated from ambient noise. Animals were ob-
served for 10 min and recording of activity was taken in
Minutes 0-2, 3-5, and 6-10. Previous investigators [45] had
reported differences in the open field during the first two
min. This apparent effect in a novel situation was similar to a
trend we had observed in the radial arm maze. The floor of
the open field was inscribed with lines so as to divide it into
four equal quadrants. A circle of 16 cm diameter was in-
scribed with center at the intersection of the quadrant lines.
This circle served as a means for determining activity in the
center of the open field as opposed to activity toward the
periphery. Measures taken in this situation included number
of squares crossed, number of center squares crossed, and
total squares crossed. Ratings of behavior in this situation
have reliability coefficients in the 0.90's [30]. After each
animal was run, the glass floor of the maze was cleaned with
a commercial isopropyl alcohol solution. Animals were run
in the open field on two successive days.
Social Interaction
The amount and kind of social interaction engaged in by
pairs of animals was determined in two situations which have
been described as being sensitive to anxiety [10,11]. The
open field box was used as one of these situations, while a
similarly constructed box was used for the second. The first
is designated the "familiar" box; the second is designated
the "unfamiliar" box. The familiar box was painted flat
black and was always used in red light illumination. The un-
familiar box was painted white and was always used in bright
light illumination. The differential response of pairs of ani-
mals to these situations is well known [10, 11, 12].
The two open field trials served to familiarize each of the
animals with that box. At the time of the social interaction
test, each animal was placed in a box with a strange (non-
litter mate) animal. Pairs of lead animals so constituted were
compared with pairs of control animals. Pairs of animals
were randomly assigned to either a familiar box placed in a
red light illumination or to an unfamiliar box placed in a
bright illumination. The amount of time these animals spent
in social interaction [10,11] was scored, as were the in-
stances of aggressive behaviors defined as wrestling, boxing,
and biting.
The observer scored social interaction by activating a
running-time meter whenever the animals interacted so-
cially. The frequency of wrestling, boxing, and biting was
tabulated and summed and constituted the aggression score.
Passive Avoidance
Behavior in a passive avoidance situation was investi-
gated because of its presumed similarity to the inability of
the hyperactive child to inhibit responding. The apparatus
was a modification of one described elsewhere [30]. Essen-
tially it consisted of a two-compartment box made of Plexi-
glas, the compartments being separated by a guillotine door
that was manipulated by the experimenter. The com-
partments measured 42 cm by 20 cm, one being painted
white, the other flat black. The floor was a grid connected to
a shock source. A photoelectric beam was placed 4 mm
above the floor of the border of the two compartments.
Breaking this beam served to activate electronic equipment
that counted the number of interruptions and recorded their
duration.
Animals were placed in the white compartment facing
away from the closed guillotine door. After 10 sec, the door
was raised. When the animal crossed to the black side the
door was dropped and a shock of 0.5 mA delivered for 1.0
sec.
The animal was removed to a holding cage for a one-min
intertrial interval. This procedure was repeated until the
animal did not cross to the black box for eight min following
the raising of the door or until five trials had taken place.
Only one animal, a control, required five trials to meet the
-------
98
FLYNN, FLYNN AND PATTON
eight-min criterion. Latency to cross was recorded on each
trial.
The recording of incursions that break the photoelectric
beam was done on an Esterline Angus Model AW recorder,
using a paper speed of 7.6 cm/min. An incursion that occurs
prior to the animal's making full entry into the black box is
termed an Abortive Incursion (AI). AI's of less than 2 sec
duration (brief AI's) were simply counted, to obviate ob-
server variation in judgement of very brief time intervals. All
other AI's were both counted and timed for duration of in-
cursion. Thus, each animal yielded four measures of activity
at the photoelectric beam on each trial: number of untimed
incursions (brief AI's), number of timed incursions (timed
AI's), total number of incursions, and mean duration of
timed incursions.
STUDY 3
Animals
The animals in this study were housed and run in a sepa-
rate laboratory facility. As a result, some rearing parameters
were slightly changed from the procedure described earlier.
The animals were maintained on a 12-hr light-dark cycle with
lights coming on at 22:00 hr and going off at 10:00 hr. All
behavioral testing was done during the animals' dark period,
between the hours of 10:00 and 22:00. All litters were re-
duced to 8 pups on Day 2. Twenty male animals from 4 litters
were used in behavioral testing.
Treatment
Animals were treated through parturition in the same
fashion as animals in Study 2. That is, dams were maintained
on lead adulterated drinking water during pregnancy (0.2%
lead). Neonates were orally dosed with lead acetate at a lead
dose of 90 mg/kg, from Day 1 until 21 days. At weaning on
Day 21, offspring were placed on a 0.25% lead as Pb(Ac),
solution and were maintained on this solution until 33 days of
age. Thereafter they received tap water ad lib. Pair-
watering/pair-dosing were carried out as in Study 2.
Shuttle Activity
A standard Lehigh Valley shuttle avoidance box was
placed in a sound attenuating chamber. Lighting was pro-
vided by a 7.5 W red light bulb mounted in the ceiling of the
sound attenuating chamber. Background noise was provided
by the blower which provided for air exchange within the
chamber.
Each rat was placed in the apparatus for 15 min, and
crossings across the center line were recorded for 3 success-
ive five-min periods. Activity was measured between 30 and
33 days of age.
Shuttle Avoidance
A standard Lehigh Valley shuttle avoidance box with a
2.5 cm barrier installed was housed within a sound attenuat-
ing chamber. Scrambled shock of 0.6 mA constant current
was the unconditioned stimulus (US). The conditioned
stimulus (CS) was a 28 V bulb mounted in the end of the box,
plus a sound (Sonalert mounted in ceiling of box). The CS
came on for 5 sec, at which time the US was delivered. The
CS and US then continued paired for an additional 10 sec at
which time the US and CS were terminated. The animal
could avoid the shock by jumping to the other side of the
apparatus. The intertrial interval was 1 min. The chamber
was dark, the only illumination being provided by the CS.
Animals were run at age 58-60 days for 100 massed trials.
RESULTS
STUDY 1
Dam weight pre- and post-partum is shown in Fig. 1.
There are no statistically reliable differences between the
lead and control groups. There were no significant differ-
ences between lead and control groups in number of litters
delivered, number of pups per litter, or day of eye opening of
pups. There were no lead-control differences in pup weight
from birth to weaning.
/***%
$*
0
DAYS
10 20
POSTPABTUM
FIG. 1.
Pre- and post-partum weights of lead and pair-fed control
dams.
These results suggest that pair-watering is an effective
control in this paradigm. There were, however, differences
in brain weight. The brains of lead animals were consistently
lighter than the brains of control animals. This difference is
significant at Days 31-34. At this time the means and stan-
dard deviations for brain weight are, Lead: mean =
1.58 ± 0.06; Control: mean = 1.68 ± 0.06, for N of 9 and 11,
respectively. This decline in brain weight consequent to lead
exposure is consistent with our previous finding [13].
Chemical Analysis
At three days of age the lead animals have significantly
increased concentrations of brain lead when compared to
controls. There is no overlap between these groups at this
age. By 3 days of age, lead animals have a mean brain lead
concentration of 0.174 /u,g/gm of wet tissue weight while the
brain lead concentration of controls is essentially zero. This
difference is significant (f=14, p<0.001). At 30-34 days
differences in brain lead concentration are not statistically
significant. This phenomenon of disappearance of brain lead
-------
LEAD EFFECTS ON HYPERACTIVITY AND LEARNING
99
z
5 2
<"
I/}
Lead
Cont rol
Waler deprivation
and reward introduced
at -This point
DAYS
FIG. 2. Mean number of arms entered in the first five min in a radial
arm maze for twelve consecutive days of testing.
following cessation of lead ingestion has been reported by
others [18].
There were no differences in regional brain calcium be-
tween experimental and control groups at Days 40-45. Mean
brain calcium concentrations plus and minus one SD for lead
and control groups were: Hippocampus: Lead mean =
56.88 ± 6.37, Control mean =54.56 ± 6.13; Striatum: Lead
mean = 53.55 ± 6.61, Control mean = 52.92 ± 4.22 and
Residual Brain: Lead mean = 52.45 ± 5.13, Control mean =
52.16 + 3.01.
Radial Arm Maze
Figure 2 shows the mean number of arm entries in the
radial arm maze in the first 5 min as a function of days.
Although the experimental group was consistently more
active than the control group, the overall difference did not
attain statistical significance in a repeated measures analysis
of variance. The apparent first day difference between lead
and control animals was not replicated in our subsequent
work reported below. Insofar as the number of arm entries is
a measure of general activity, lead does not lead to hyperac-
tivity in this situation.
A measure of spontaneous alternation taken in the radial
arm maze failed to differentiate between lead and control
animals. Mean spontaneous alternation scores as measured
by the number of different arms entered in the first 8 choices
plus and minus one SD are Lead mean = 5.6 ± 0.97, and
Control mean = 5.63 ± 1.19 on Day 1. On Day 12 of testing,
Lead mean = 6.9 ± 0.63 and Control mean = 7.13 ± 0.35.
STUDY 2
Mean pup weights from Day 3 until Day 30 following
parturition are shown in Fig. 3. The groups do not differ from
one another during this period. Similarly, experimental and
control animals did not differ in body weight at time of test-
ing. Lead animals were slightly lighter than control animals,
but this difference is not statistically significant (Lead mean
= 221.75 ± 18.57; Control mean = 226.33 ± 24.56).
Chemical Analyses
Brain lead concentrations for experimental and control
offspring are shown in Table 1. These determinations are
based on 6 experimental and 6 control males assayed at Days
75-76. The difference in brain lead is significant, r(10) =
21.23, p<0.001.
Experimental and control brains did not differ in 5-HT
concentrations when assayed in hippocampus, striatum,
brain stem, and cortex.
MMTA uptake by striatal tissue minces did not differ in
the experimental and control groups. Mean MMTA concen-
trations in /ug/gm of tissue plus and minus one SD following a
en
E
EC
o
Q.
3
o.
z
<
UJ
90
80
70
60
50
40
30
20
10
Control n = 6 litters
Lead n = 6 litters
J L
11
15
19
23
27
31
DAYS POSTPARTUM
FIG. 3. Mean pup weights from birth to weaning for n=6 control and n=6 lead litters.
-------
100
FLYNN, FLYNN AND PATTON
TABLE 1
MEAN BRAIN LEAD CONCENTRATION IN ^G/GM WET TIS-
SUE WEIGHT AT DAYS 75-76 POSTPARTUM (STUDY 2)
Lead
Control
mean
SD
N
1.85
0.173
6
0.13
0.082
6
30-min incubation in MMTA in a concentration of 0.2 /u.g/ml
were 0.99 ± 0.57 for experimental animals and 1.77 ± 0.38
for control animals. These means are based on 4 experi-
mental and 3 control samples, resulting from the pooling and
mincing of 3 experimental and 3 control striata.
These differences are not statistically significant,
r(5)=1.73.
Radial Arm Maze
Lead rats did not differ from control rats in either activity
or in spontaneous alternation. The mean and standard de-
viation for number of arm entries in a 5-min period was mean
= 8.17 ± 4.62 for the lead animals and mean = 8.58 ± 4.52
for the control animals. The mean and standard deviation for
number of spontaneous alternations was mean = 4.67 ± 1.80
and mean = 4.83 ± 2.23 for the lead and control groups,
respectively.
Social Interaction
As expected, the familiar and the unfamiliar situations
differ in the amount of social interaction they elicit. Pairs of
rats interact more in the familiar situation than in the un-
familiar situation. There are, however, no statistically sig-
nificant differences in those situations between lead and con-
trol groups. The means and SD for time spent in social in-
teraction by the lead and control groups are Lead: mean=
362.23 ± 87.85 in the familiar situation, 194 ± 64.59 in the
unfamiliar situation; Control: mean = 371.23 ± 59.78 in the
familiar situation, 194.67 ± 28.84 in the unfamiliar situation.
Previous work has shown this test to be sensitive to anxiety
or fear. The above data suggest that the mean response to
this situation is not different in lead and control animals.
Open Field
Figure 4 presents the number of center arid peripheral
squares crossed by lead and control animals on Day 1. The
lead animals are reliably different from the controls with
regard to peripheral squares crossed when analyzed by a
2x3 factorial analysis of variance with repeated measures on
one factor. F(l,21)=4.29, p<0.05 for the main (lead) effect.
F(2,42)=5.49, /?<0.008 for the treatment by time-period in-
teraction effect. Examination of Fig. 4 reveals that there are
no differences between lead and control animals in Period 1
but there are differences in Periods 2 and 3. Note that the
lead animals in this study are less active than the control
animals, contrary to expectation.
Passive Avoidance
There were no differences between lead and control
30
O 24
g 21
16
12
0 ° Control periphery n = 12
* * Lead periphery n=ll
Control Center
MINUTES IN OPEN FIELD DAY 1
FIG. 4. Mean number of center and peripheral squares crossed by
lead and control animals on Day one in the open field for Minutes
1-2, 3-5, and 6-10.
18
z 16
O
i14
o
I 12
IU
p 10
1 8
I 6
Q
Z
o—o control
*—* Lead
2
TRIALS
FIG. 5.
Mean duration of timed incursions over trials in passive
avoidance for lead and control animals.
groups in trials to criterion (Lead: mean = 2.83 ± 0.687;
Control: mean = 3.08 ± 0.862). On the measures of abortive
entries into the dark side, however, there are interesting
differences between the experimental and control groups.
While lead and control animals did not differ in the mean
duration of timed incursions, they differed in the number of
incursions, both timed and untimed, made across trials. Fig-
ure 5 presents the mean duration of the timed incursions.
Figure 6 presents the number of untimed incursions. When
the number of untimed incursions is placed in ratio to the
total number of incursions, the difference between experi-
-------
LEAD EFFECTS ON HYPERACTIVITY AND LEARNING
101
18
16
14
12
10
8
6
4
2
O—O Control
*—A Lead
2
TRIALS
FIG. 6. Number of brief AI's over trials in passive avoidance for
lead and control animals.
mental and control animals is easily seen. These data are
shown in Fig. 7. (n changes in these graphs because some
animals reach criterion and/or show no AI's on the criterion
trial). The mean values of this ratio are presented in Table 2.
A repeated measures analysis of variance on these data show
TABLE 2
MEAN VALUES OF THE RATIO: BRIEF AI -f TOTAL AI
OVER TRIALS FOR LEAD AND CONTROL ANIMALS
Lead
Control
mean
SD
N
mean
SD
N
1
0.5
0.289
11
0.717
0.157
12
Trials
2
0.67
0.208
11
0.556
0.211
12
3
0.633
0.215
5
0.499
0.085
8
a significant treatment x trials effect, F(2,32)=4.98,
p<0.025.
STUDY 3
There were no significant differences in weight between
lead and control groups at 30-33 days when activity testing
was conducted. The mean weights plus and minus one SD at
this time are Lead: mean = 161.60 ± 10.70; Control: mean =
160.70 ± 7.6).
Shuttle Activity
The mean activity scores of the lead and control animals
0.7
O °-6
CD
Z
111
£
0.5
n=11
A A
Control
Lead
n=5
n=8
FIG. 7. Mean value of brief AI's H-
TRIALS
total AI's over trials in passive avoidance for lead and control
animals.
-------
102
FLYNN, FLYNN AND PATTON
were not statistically different when analyzed in a repeated
measures analysis of variance. However, the activity scores
of the lead animals were somewhat lower than those of con-
trols, a result consistent with the open field findings reported
in Study 2.
Shuttle Avoidance
There were no differences between lead and control ani-
mals in number of escape or number of avoidance responses.
DISCUSSION
Inorganic lead, in the dose range considered in this paper,
is not reliably associated with hyperactivity in laboratory
rats. Previous reports which have suggested that lead admin-
istration early in development is productive of later
hyperactivity should be re-evaluated. Methodological
shortcomings and inadequate or inappropriate statistical
analyses call this relationship into question. When necessary
methodological constraints are imposed upon the experi-
ment, lead is seen to bear little, if any, relationship to
hyperactivity as measured in a variety of situations. In fact,
lead in low doses may be unrelated to hyperactivity while
higher doses (which are still small enough not to be produc-
tive of obvious symptoms of toxicity) may be productive of
hypoactivity.
The behavior-related effects of lead exposure on
neurochemistry remain obscure. Other investigators have
reported conflicting results with respect to lead effects on
catecholamines [16,30]. We have found no effect of lead on
steady state DA concentrations, and no effect on regional
concentrations of 5-HT. The early hopes to find a simple
lead-neurotransmitter-behavior link have not been fulfilled.
Research accomplished at this point in time suggests that
such simple relationships are not likely to be found.
How, then, does the behavioral toxicologist proceed in
the search for the effects of lead in these experimental
paradigms? It will be necessary to concede at the outset that
the task is not an easy one. The neurochemical organization
of the brain is enormously complex, and the relationship
between this complex structure and the infinite variations
possible in behavior are only beginning to be understood.
Recognition of the difficult nature of the tasks suggests two
general strategies for further work in this area.
At the biochemical/neurochemical level, it will probably
be most productive to pursue investigation into the cellular
and molecular effects of lead. The effects of lead at this level
are being investigated, in some cases in situations directly
related to neurotransmitter systems presumed relevant to
hyperactivity [38]. In this regard, attention should be di-
rected to dynamic rather than static aspects of neurochemis-
try. Thus, we have found no differences in steady state cal-
cium concentrations. Studies of uptake, release, and turn-
over of neurotransmitters are likely to be more informative
than studies of steady state concentrations. Similarly,
studies of relevant enzyme systems are indicated. Thus, lead
has been shown to affect adenyl cyclase activity in vitro [24].
Our preliminary work along these lines has examined the
effects of chronic lead exposure on striatal uptake of
MMTA. We find no differences in striatal accumulation of
MMTA by experimental and control animals. MMTA is
taken up by striatal dopamine sensitive neurons, and the lack
of effect of lead on this process argues against any very
simple involvement of this system as mediator of behavioral
effects of chronic exposure to lead.
At the behavioral level, it seems unlikely that a relation-
ship is to be found between lead exposure and simple
measures of activity. It is preferable to direct attention to
and capitalize upon behaviors which exhibit more subtle var-
iation than does amount of activity. One promising line of
approach is that which examines qualitative activity differ-
ences in behaviors in which a large reactive component is
present. The characterization of the effects of pharmacologi-
cal agents on punished behavior is a good example of the
fruitfulness of this approach.
We have presented evidence above of a new means of
assessing such behaviors. In the repeated trials Passive
Avoidance we are able to measure two behaviors that serve
to quantify interesting qualitative differences between lead-
exposed and control animals. The mean duration of AI's is
not different in experimental and control groups. However,
the lead animals show significantly fewer of these AI's than
did controls. Furthermore, the number of brief AI's relative
to the total number of AI's increased over trials in the lead
group and decreased over trials in the control group. The
picture these data present is that of the lead-exposed animals
spending less time than the control in behaviors directed
toward the black side of the apparatus. When the lead-
exposed animal does engage in such behavior, it is increas-
ingly tentative and increasingly associated with retreat.
One possible interpretation of these observations, of
course, is one made in terms of fear, and/or punishment. It is
possible that the lead-exposed aniamls are more fearful than
are control animals. Or, punishment may be more disruptive
of ongoing behavior in the lead-exposed animals, more
readily bringing about an effect like conditioned suppression.
We are currently gathering additional data on these
phenomena. If these types of behavior are indeed stable
under appropriate conditions, a useful measure will be avail-
able to the behavioral toxicologist or pharmacologist.
We note finally that this passive avoidance behavior is not
consistent with a lead-based animal model of childhood
hyperkinesis. We expected lead-exposed animals to be less
able than controls to inhibit responding. They inhibit more.
And, to the extent that an interpretation based on punish-
ment and/or fear is valid, these lead-exposed animals are
unlike the hyperactive child. The latter is frequently char-
acterized as exhibiting little fear and as being unresponsive
to punishment. We suggest that early exposure of rats to
lead, as reviewed here, does not produce a reasonable ani-
mal model of childhood hyperkinesis.
ADDENDUM
The unexpected hypoactivity exhibited by the rats in
Study 2 was surprising. A further study was undertaken in an
effort to determine the underlying causes of the phenom-
enon. Long-Evans rats were treated as in Study 2 except that
all animals, both lead and control, were given food and water
ad lib one week before behavioral testing. Under these con-
ditions the lead-control differences seen earlier failed to
materialize. Mean activity scores consistently fell between
those of controls and experimentals in Study 2, suggesting
that two factors were operating to produce the hypoactivity
observed in that study. The pair watering procedure used in
Study 2 resulted in control animals who were acutely as well
as chronically water deprived at the time of testing. Evi-
-------
LEAD EFFECTS ON HYPERACTIVTTY AND LEARNING
103
dently this acute water deprivation was sufficient to elevate
activity levels. The lead treated animals exhibited hypoac-
tivity but only while they were actually ingesting lead. When
lead administration was stopped, activity levels returned to
normal.
REFERENCES
1. Brown, D. Neonatal lead exposure in the rat: decreased, learning
as a function of age and blood lead concentrations. Toxic, appl.
Pharmac. 32: 628-637, 1975.
2. Cardenas, H. L. and D. H. Ross. Morphine induced calcium
depletion in discrete regions of rat brain. J. Neurochem. 24:
487-^93, 1975.
3. David, O., J. Clark and K. Voeller. Lead and Tiyperactivity.
Lancet. 2: 900-903, 1972.
4. David, O. J., S. P. Hoffman, J. Sverd, J. Clark and K. Voeller.
Lead and hyperactivity—behavioral response to chelation—
pilot study. Am. J. Psychiat. 133: 1155-1158, 1976.
5. Delves, H. T. A micro-sampling method for the rapid determi-
nation of lead in blood by atomic absorption spec-
trophotometry. Analyst 95: 431-438, 1970.
6. Dennenberg, V. H. Assessing the effects of early experience.
In: Methods in Psychobiology, edited by R. E. Meyers. New
York: Academic Press, 1977.
7. Dorris, R. L. and P. A. Shore. Localization and persistence of
meteraminol and a-methyl-meta-tyramine in rat and rabbit
brain. J. Pharmac. expl. Ther. 179: 10, 1971.
8. Dorris, R. L. and P. A. Shore. Amine uptake and storage mech-
anisms in corpus striatum of rat and rabbit. J. Pharmac. exp.
Ther. 179: 15, 1971.
9. Douglas, R. J. The development of hippocampal function: im-
plications for theory and for therapy. In: The Hippocampus:
Volume 2: Neuropsysiology and Behavior, edited by R. I. Isaac-
son and K. H. Pribram. New York: Plenum Press, 1975.
10. File, S. E. and J. R. G. Hyde. The effects of
p-chlorophenylalinine and ethanolamine-O-sulphate in an ani-
mal test of anxiety. J. Pharm. Pharmac. 29: 735-738, 1977.
11. File, S. E. and R. R. G. Hyde. Can social interaction be used to
measure anxiety? Br. J. Pharmac. 62: 19-24, 1978.
12. File, S. E. Anxiety, ACTH, and 5-HT. Trends Neurosc. 1: 9-11,
1978.
13. Flynn, J. C., J. H. Patton and E. R. Flynn. Oral dosage and
brain lead concentration in neonatal rats: some implications for
animal models of the hyperactive child syndrome. Paper pre-
sented at the meeting of the Southwestern Psychological Asso-
ciation, New Orleans, April 1978.
14. Fjerdingstad, E. J., G. Danscher and E. Fjerdingstad. Hip-
pocampus: selective concentration of lead in the normal rat
brain. Brain Res. 80: 350-354, 1974.
15. Glowinski, J. and L. L. Iverson. Regional studies of
catecholamines in the rat brain. J. Neurochem. 13: 655-669,
1966.
16. Goiter, M. and O. Michaelson. Growth, behavior, and brain
catecholamines in lead-exposed neonatal rats: a reappraisal.
Science 187: 359-361, 1975.
17. Gross, M. and W. Wilson. Minimal Brain Dysfunction. New
York: Brunner/Mazel, 1974.
18. Hastings, L., G. P. Cooper, R. L. Bornschein and I. A.
Michaelson. Behavioral effects of low level neonatal lead expo-
sure. Pharmac. Biochem. Behav. 7: 37-42, 1977.
19. Jacobowitz, D. M. and J. S. Richardson. Method for the rapid
determination of norepinephrine, dopamine, and serotonin in
the same brain region. Pharmac. Biochem. Behav. 8: 515-519,
1978.
20. Krehbiel, D., G. A. Davis, L. M. LeRoy and R. E. Bowman.
Absence of hyperactivity in lead-exposed developing rats.
Envir. tilth. Perspec. 18: 147-157, 1976.
21. Michaelson, I. Effects of inorganic lead on RNA, DNA, and
protein content in the developing neonatal rat brain. Toxic.
appl. Pharmac. 26: 539-548, 1973.
22. Michaelson, I. and M. Sauerhoff. An improved model of lead-
induced brain dysfunction in the suckling rat. Toxic, appl.
Pharmac. 28: 88-96, 1974.
23. Michaelson, I. and M. Sauerhoff. Animal models of human dis-
ease: severe and mild lead encephalopathy in the neonatal rat.
Envir. Hlth. Perspec. 28: 88-96, 1974.
24. Nathanson, J. A. and F. E. Bloom. Lead induced inhibition of
brain adenyl cyclase. Nature 255: 419-420, 1975.
25. Niklowitz, W. J. and D. W. Yeager. Interference of Pb with
essential .brain tissue Cu, Fe, and Zn as main determinant in
experimental tetraethylead encephalopathy. Life Sci. 13: 897-
905, 1973.
26. Niklowitz, W. J. Subcellular mechanisms in lead toxicity:
significance in childhood encephalopathy, neurological
sequelae, and late dementias. In: Neurotoxicology, edited by
Roizin, H. Shiraki and N. Grevic. New York: Raven Press,
1977.
27. Olton, D. S. and R. J. Samuelson. Remembrance of places past:
spatial memory in rats. J. exp. Psycho/.: Animal Behav. Proc. 2:
97-116, 1976.
28. Olton, D. Spatial memory. Sclent. Am. 236: 82-98, 1977.
29. Patton, J. H., J. C. Flynn and E. R. Flynn. The effect of inor-
ganic lead (Pb) on the behavioral activity of young rats in open
field and passive avoidance situations. Paper presented at the
meeting of the Southwestern Psychological Association, New
Orleans, April 1978.
30. Patton, J. H. The behavior and physiological effects of adminis-
tering lead to neonatal rats: a test of the model of childhood
hyperactivity. Unpublished doctoral dissertation, Baylor Uni-
versity, 1978.
31. Pihl, R. O. and M. Parkes. Hair element content in learning
disabled children. Science 198: 204-206, 1977.
32. Sauerhoff, M. W. and I. Michaelson. Hyperactivity and brain
catecholamines in lead-exposed developing rats. Science 182:
1022-1024, 1973.
33. Shore, P. A. and H. S. Alpers. Fluorometric estimation of
meteraminal and related compounds. Life Sci. 3: 551-554, 1964.
34. Silbergeld, E. and A. Goldberg. A lead induced behavior disor-
der. Life Sci. 13: 1275-1283, 1973.
35. Silbergeld, E. and A. Goldberg. A lead-induced behavioral dys-
function: an animal model of hyperactivity. Expl Neural. 42:
146-157, 1974.
36. Silbergeld, E. K. and A. M. Goldberg. Pharmacological and
neurochemical investigations of lead-induced hyperactivity.
Neuropharmacology 14: 431-444, 1975.
37. Silbergeld, E. K. Interactions of lead and calcium on synap-
tosomal uptake of choline and dopamine. Life Sci. 20: 309-318,
1977.
38. Silbergeld, E. K. and J. J. Chisolm. Lead poisoning: altered
urinary catecholamine metabolites as indicators of intoxication
in mice and children. Science 192: 153-155, 1976.
39. Slob, A., C. Snow and E. de Natris-Mathot. Absence of behav-
ioral deficits following neonatal undernutrition in the rat. Devi
Psychobiol. 6: 177-186, 1973.
40. Smart, J. Activity and exploratory behavior of adult offspring of
undernourished rats. Devi Psychobiol. 7: 315-321, 1974.
41. Snowdon, C. Learning deficits in lead-injected rats. Pharmac.
Biochem. Behav. 1: 599-603, 1973.
42. Sobotka, T. J. and M. Cook. Postnatal lead acetate exposure in
rats: possible relationship to minimal brain dysfunction. Am. J.
ment. Defic. 79: 5-9, 1974.
43. Wender, P. H. Minimal Brain Dysfunction in Children. New
York: Wiley, 1971.
44. Wessel, M. A. and A. Dominski. Our children's daily lead. Am.
Scientist 65: 294-298, 1977.
45. Winneke, G., A. Brockhaus and R. Baltissen. Neurobehavioral
and systemic effects of longterm blood lead elevation in rats.
Archs Toxic. 37: 247-263, 1977.
46. Zinterhoffer, L., P. Jatlow and A. Fappiano. Atomic absorption
determination of lead in blood and urine in the presence of
EDTA. J. Lab. din. Med. 78: 664-674, 1971.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 105-111. ANKHO International Inc., 1979.
Performance and Acquisition of Serial Position
Sequences by Pigeons as Measures of
Behavioral Toxicity
D. E. MCMILLAN
Department of Pharmacology, School of Medicine, University of Arkansas for Medical Sciences
MCMILLAN, D. E. Performance and acquisition of serial position sequences by pigeons as measures of behavioral
toxicity. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 105-111, 1979.—A procedure has been developed to measure the
repeated acquisition of serial position sequences and to study the effects of drugs and toxic chemicals on the behavior
generated by the procedure. Thus far experiments using the procedure have shown: (1) Performance schedules generate
lower error rates than corresponding acquisition schedules: (2) Addition of a reset contingency further decreases errors
under both performance and acquisition schedules: (3) Chained acquisition and performance schedules generate lower
error rates than corresponding tandem acquisition and performance schedules: (4) Chained acquisition and performance
schedules produce behavior that usually is more sensitive to drugs than corresponding tandem acquisition and performance
schedules: (5) Acquisition schedules produce behavior that usually is more sensitive to drugs than corresponding perform-
ance schedules: and (6) Lead is an exception in that it produced clearer effects under a chained performance schedule with
a reset contingency than under a corresponding acquisition schedule. The greater sensitivity to drug effects of behavior
under acquisition schedules than behavior under performance schedules and of behavior under chained schedules may be a
function of the baseline error rates, rather than the behavioral processes of acquisition, performance, and stimulus control.
Behavioral toxicity Serial position sequences Acquisition Performance
EFFECTS OF CHEMICALS ON ACQUISITION
AND PERFORMANCE
There is a growing recognition that a complete assess-
ment of the possible toxic effects of a chemical must include
a determination of the effects of that chemical on behavior in
addition to the more traditional indices of toxicity [7, 11, 12,
13, 15]. One important aspect of behavior that might be af-
fected by a toxic chemical is the acquisition of new behavior.
Unfortunately, it is very difficult to study the effects of
chemicals on the acquisition of new behaviors because prob-
lems arise that are usually not encountered during studies of
the effects of chemicals on well established performance.
In order to study the effects of a chemical on established
performance, it has been customary to develop stable
baseline patterns of responding. Frequently, stable baselines
have been developed by food depriving the animal and then
conditioning the animal to make a specified response, such
as a lever press or a key peck, in order to produce food.
Once behavior has stabilized under a particular schedule de-
scribing the relationship between food delivery and respond-
ing, toxic chemicals are administered, either on an acute or a
chronic basis. This procedure allows the investigator to
study the effects of chemicals on the performance of indi-
vidual animals because the behavior is reproducible from
day to day. and changes in behavior after administration of
the chemical can be assumed to have been caused by the
chemical. In addition, intersubject comparisions usually can
be made, since all subjects have had similar training experi-
ences and their stable baseline performances are similar al-
though not identical. Some examples of the use of this ap-
proach to study the effects of toxic chemicals on behavior
would include the work of our group with lead (Barthalmus
et al., [2]) and the work of the Rochester group with mercury
(Evans et al., [6]).
It is much more difficult to study the effects of toxic
chemicals on the acquisition of new behaviors. The behav-
ioral history of some organisms may favor the acquisition of
new behaviors more than does the behavioral history of
other organisms, so that all organisms do not begin to ac-
quire new behaviors from a common baseline. Furthermore,
as acquisition proceeds, the behavior baseline is constantly
changing, perhaps at different rates in different organisms
and perhaps irreversibly. These factors make it almost im-
possible to make repeated measurements of the effects of
chemicals on the acquisition of behavior in individual sub-
jects and intersubject comparisions are often limited because
of the high degree of variability across subjects. Under such
circumstances, it may be impossible to detect subtle effects
of toxic chemicals.
REPEATED ACQUISITION OF SERIAL POSITION
SEQUENCES
Some of these problems have been solved by the devel-
opment of a technique for measuring the repeated acquisition
of serial position sequences [3, 4, 14]. In these experiments
monkeys were trained to respond on different levers in a
predetermined sequence to obtain food. Responses on a
105
-------
106
MCMILLAN
lever that were out of sequence (errors) produced a timeout
from reinforcement. Completion of the correct sequence of
responses on the different levers produced food. By present-
ing the monkey with a new sequence to be acquired each
session, it was possible to measure acquisition repeatedly in
a single subject.
Thompson [16,17] modified this procedure to study the
repeated acquisition of position sequences by pigeons.
Under a chained schedule, pigeons responded in a chamber
with three response keys, all of which were illuminated at the
same time with one of four different colors. Pecking the cor-
rect key changed the color of all three keys to another color.
In contrast, under a tandem schedule, the keys were always
illuminated with one color which did not change after a cor-
rect response. Under both chained and tandem schedules,
incorrect responses (responses out of sequence, or errors)
produced a timeout period of total darkness in the chamber
during which responses had no programmed consequences.
The completion of the four-response sequence operated the
feeder for a period that was too short to allow the pigeon
time to eat, but after the response sequence had been com-
pleted five times, the feeder operated for a period long
enough to allow the pigeon to eat. If the same problem is
presented to the pigeon each day, the bird's behavior may be
considered as a performance baseline, but if the sequence is
changed every day, acquisition can be measured repeatedly
in a single subject. Thompson found that after a few months
of training, the pigeon learned a new sequence each day
without wide variation in the number of errors.
In our laboratory, we have modified Thompson's proce-
dure by introducing a contigency whereby each peck on an
incorrect key resets the sequence to the beginning step [8].
Pigeons were trained in a chamber containing three response
keys to peck the center response key which was transillumi-
nated with a blue light. The two side keys remained dark and
responses on these keys had no programmed consequences.
Once responding on the center key was established, the key
transilluminated with the blue light changed position after
each food delivery. Only a response on a lighted key pro-
duced food (stage 1, Fig. 1). Subsequently, all three keys
were transilluminated with a red light, and a response on any
one of the three red keys changed the color to blue, but only
a peck on one predetermined blue key produced food (stage
2, Fig. 1). Pecking on either of the other blue keys produced
a timeout. If a reset contingency was to be used, it was
introduced during the same session as the timeout. The chain
was extended in a similar manner to include a green light
(stage 3, Fig. 1). Responses on one of the green keys changed
the color of all three keys to red, while responses on the
other green keys produced a timeout. If the correct green
key was pecked and all three keys became red, a response on
one red key changed the color of all three keys to blue, while
responses on the other red key produced a timeout (and reset
the sequence to the first step where all keys were green if a
reset contingency was in effect). If the correct red key was
pecked and all three keys became blue, a response on one
blue key produced food, while responses on the other blue
keys produced timeout (and reset the sequences to the first
step where all keys were green if a reset contingency was in
effect). Finally, the chain was extended in a similar manner
to include a yellow key light (stage 4, Fig. 1). After about
25-40 sessions, some of the birds were switched to a tandem
schedule where the key colors did not change following cor-
rect responses.
The repeated acquisition of serial position sequences in
Stage I
Stage 2
ORC
Stage 3
ORC
I TO TO
4, 4, I
-TO —TO 4,
4> I +
TO 4, TO-
ORC
FOOD
Stage 4
ORC
i*© © ©<
4, I 4,
TO 4> TO
I 4, 4r
i TO—TOJ
-TO—TO I
ORC
TO
TO
FOOD
FIG. 1. Diagram of method for training pigeons to perform serial
position sequences, with or without a reset contingency. Abbre-
viations are as follows: O, dark key; ®, blue key; ®, red key; ©,
green key; ®, yellow key; TO, timeout; ORC, optional reset
contingency. Other details are in the text.
these birds under chained and tandem schedules with a reset
contingency was compared to the repeated acquisition of
birds under chained and tandem schedules without a reset
contingency. Addition of the reset contingency reduced the
total number of errors made during a session under both
chained and tandem schedules of repeated acquisition as
shown in Fig. 2. Furthermore, errors were lower under the
chained schedules than under the corresponding tandem
schedules.
EFFECTS OF DRUGS ON RESPONDING UNDER ACQUISITION AND
PERFORMANCE SCHEDULES
Thompson [16-22] has used chained and tandem acquisi-
tion and performance schedules (without a reset contin-
-------
PIGEON SERIAL POSITION SEQUENCES
107
1000-
800-
600-
400-
200-
0-
CHAINED TANDEM
UJ
CO
UJ
a:
ir
h-
UJ
CO
Ul
cc
FIG. 2. Differences in total number of errors under chained and
tandem schedules with and without reset contingencies. All bars are
based on five sessions for three pigeons after several months of
training. (Redrawn from Harting and McMillan [8])
gency) to study the effects of drugs. Phenobarbital, chlor-
diazepoxide, (/-amphetamine, cocaine, imipramine, and
methylphenidate all increased errors under these schedules,
but fenfluramine and chlorpromazine did not increase errors
even at doses that decreased rates of key pecking [18-22].
The chained acquisition schedule was more sensitive to the
effects of drugs than was the corresponding tandem acquisi-
tion schedule [22]. Furthermore, behavior under the chained
acquisition schedule frequently was affected by lower doses
of drugs than was behavior under the chained performance
schedule [22]. Thus, the chained acquisition schedule ap-
peared to be the most sensitive schedule that Thompson
used to measure drug effects.
In our laboratory we have found that errors under the
chained acquisition schedule with a reset contingency are
increased to a greater extent than errors under this schedule
without a reset contingency [9]. Pigeons trained under the
chained acquisition schedule either with or without a reset
contingency were given various doses of pentobarbital. The
effects of pentobarbital on behavior under these schedules is
shown in Fig. 3. In Fig. 3 errors have been plotted both as
absolute number of errors and as a percentage of the control
errors so that both absolute and relative error increases can
be compared across the schedules.
Fig. 3 shows that pentobarbital produced greater increases
in the total errors per session under the chained schedule
when the reset contingency was in effect, both when errors
are considered in terms of absolute increases (top row) and
in terms of a percentage of control errors (middle row). These
increases in errors were independent of effects on rate of
responding (bottom row). Both with and without the reset
contingency in effect, the largest increases in errors occurred
after the 5.6 and 10 mg/kg doses, yet average rates of key
pecking were little affected at these doses.
Thompson [22] also has found behavior under chained
acquisition schedules to be affected by lower doses of drugs
CHAINED RESET
CHAINED NO-RESET
(0
1000-
800-
600-
400-
200-
0-
240-
200-
160-
120-
80-
40-
0-
0.8-
0.6-
0.4-
0.2-
0-
1 J T T
T T
C 0 3 5.6 10 17.5 C 0 3 5.6 10 17.5
MG/KG PENTOBARBITAL
FTG. 3. Effects of pentobarbital on total errors (top row), errors as a
percentage of the control errors (middle row), and mean rate of
responding (bottom row) for birds under the chained acquisition
schedule with a reset contingency and for birds under the chained
acquisition schedule without a reset contingency. The brackets at C
show ±2 standard errors of the mean for 6 control sessions in each
of 3 birds. The points at zero are duplicate observations of the ef-
fects of saline in these birds. Points were not plotted after 17.5 mg/kg
pentobarbital in some instances because this dose almost completely
eliminated responding. (Redrawn from Harting and McMillan [9])
than behavior under chained performance schedules where
the pigeons are exposed to the same serial position sequence
every day. Our data [1] suggest that behavior under a
chained acquisition schedule with a reset contingency also is
affected by lower doses of drugs than behavior under a per-
formance schedule with a reset contingency. Figure 4 shows
the effects of ethanol and diazepam on responding under
these acquisition and performance schedules. Again, in-
creases in errors have been plotted both on an absolute scale
and as a percentage of control errors.
Figure 4 shows that diazepam increased errors under the
acquisition schedule (marked A in the figure) at a dose of
1 mg/kg, but 3 mg/kg of diazepam were required to increase
errors under the performance schedule (marked P in the fig-
ure). Furthermore, absolute increases in errors after di-
azepam were much larger under the acquisition schedule.
Similarly, ethanol increased errors at lower doses and by a
larger absolute amount under the acquisiton schedule than
-------
108
MCMILLAN
DIAZEPAM
ETHANOL
400-
c/> 350-
o:
o 300-
IT
£ 250-
_i 200-
<
£ 150-
*~ too-
50-
_i 0-
AP
i
i
I
/ AA
P ' \
T •
J.
5* P-
!• • > •* // t— r"^
T T T T 1 1 jf T T 1
o
rr
h-
§ 500-
o
^ 400-
(/> 300-
w 200-
o 100-
QL
Jp
.
^
A j>
a" -/
-7
u CO til O3 i 3 " 0.2505 I
MG/KG G/KG
40n BIRD
9847
20-
o: C 1 10 20 30 40 50 60 C
DC 401
a:
uj 20-
_i 0-
BIRD
2066
I *•
< C 1
1 6°1
40-
?O-
n-
BIRD
3696
I"'
10 20 30 40 50 60 C
9
• \ *
• •
* .\ . • •
% • * •
« ••• •• •• •
C I 10 20 30 40 50 60 C
DAYS
FIG. 5. Total errors for individual birds under the chained perform-
ance schedule with a reset contingency. Brackets at C show ±2
standard deviations around the control mean based on at least 15
sessions before the beginning of lead administration. (Redrawn from
Diet? and McMillan [51)
FIG. 4. Effects of diazepam and ethanol on responding under per-
formance (P) and acquisition (A) schedules with a reset contingency.
The top row shows total errors and the bottom row shows total
errors as a percentage of control errors. Brackets at C show the
range around the mean for at least 7 control sessions. Points at zero
show the effects of vehicle administration. Diazepam was adminis-
tered intramuscularly and ethanol by oral intubation. The dotted
lines and unfilled circles marked A show data under the acquisition
schedule and the solid lines and filled circles marked P show data
under the performance schedule. (Redrawn from Barthalmus et al.
[21)
under the performance schedule. Although both drugs in-
creased errors under the acquistion schedule at lower doses
than under the performance schedule, the relative increases
in errors (errors as percent of control errors, bottom row)
were as great or greater under the performance schedule.
EFFECTS OF LEAD ON RESPONDING UNDER ACQUISITION AND
PERFORMANCE SCHEDULES
Recently we have used chained acquisition and perform-
ance schedules with a reset contingency to study the effects
of the chronic administration of lead in pigeons [5]. In pre-
vious experiments in our laboratory [2] we had found that
daily oral doses of lead acetate as low as 12.5 mg/kg had
changed rates of responding in adult pigeons whose perform-
ance was maintained under a multiple fixed-ratio fixed-
interval schedule of food presentation: however, a 6.25
mg/kg dose produced little or no effect. Therefore, the 6.25
mg/kg dose was administered chronically to pigeons under
the performance and acquisition schedules in order to de-
termine if these schedules could detect behavioral effects of
lead.
Figure 5 show the effects of chronic oral doses of 6.25
mg/kg of lead on the total number of errors made during
approximately two months of lead administration under the
chain performance schedule with a reset contingency. Bird
9847 shows no effect for about 5 weeks of lead administra-
tion after which a clear increase in the number of perform-
ance errors occurred. Bird 2066 showed a marginal increase
in the number of errors, beginning at about the same time. In
contrast, lead produced an almost immediate increase in
error rate in bird 3696 which had peaked by about the third
week of lead administration. Thus, chronic lead administra-
ion clearly increased errors under the performance schedule.
In general, these increases in errors occurred without
changes in the rate of key pecking. Following the discon-
tinuation of lead administration, the error rate gradually re-
turned toward the baseline error rate over a period of 6 to 10
weeks.
Figure 6 shows the effects of the same 6.25 mg/kg oral
dose of lead on the total number of errors made by three
birds under the chain acquisition schedule with a reset
contingency during approximately two months of lead ad-
ministration. None of the birds showed consistent error in-
creases. Bird 6885 did show increases in errors on days 29
and 49. Interestingly, during both of these sessions the cor-
rect sequence was center key, right key, left key, right key.
-------
PIGEON SERIAL POSITION SEQUENCES
109
BIRD •
400-1
300-
200-
100-
o-
6885
•
i
• •
A
• • • •
/ .* • • •
* • . • * .**•
•• •« •
• • • •
• ^»
• • •
C 1 10 20 30 40 50
c 300 -i
o
i 200-
LJ
100-
«-r r»-
BIRD
372
*• . •
.W-. .;.-^
• (
•
*
% ^
1 T
60 C
*.
>
PERFORMANCE
ACQUISITION
BIRD
6885
400-
300-
200-
100-
0-
C I 10 20 30 40 50 60 C
T BIRD 2781
T I 1 1 1 1 1 1 T
C I 10 20 30 40 50 60C
DAYS
FIG. 6. Total errors for individual birds under the chained acquisition
schedule with a reset contingency. Brackets at C show ±2 standard
deviations around the control mean based on at least 15 sessions
before the beginning of lead administration. (Redrawn from Dietz
and McMillan [51)
Prior to exposure to lead, this particular sequence did not
generate an especially large number of errors. Why this par-
ticular problem apparently interacted with chronic lead ad-
ministration to increase errors while other sequences did
not, is unclear. At any rate, the chained acquisition with a
reset contingency appeared to be much less sensitive to the
effects of a 6.25 mg/kg dose of inorganic lead than did the
corresponding performance schedule.
The finding that behavior under the performance schedule
was affected at a lower dose of lead than behavior under the
acquisition schedule contrasts sharply with the findings of
Thompson and with previous observations in our own lab-
oratory. Behavior generated by acquisition schedules,
whether or not a reset contingency was involved, has been
affected at lower doses of all drugs studied than has behavior
generated by peformance schedules. Lead appears to have
exactly the opposite effect, that is it increases error under
the chained peformance schedule with a reset contigency at a
dose that has little effect on errors under the corresponding
acquisition schedule.
The data in Fig. 5 and 6 were derived from pigeons whose
123456 123456
BLOCKS
FIG. 7. Within session error elimination for individual birds under the
chained performance (column 1) and chained acquisition (column 2)
schedules with a reset contingency. The abscissa is blocks of 100
correct responses and the ordinate is the number of errors in each
block. The brackets show ±2 standard deviations around a control
mean (mean not shown) based on the last 5 sessions prior to lead
administration. The sessions plotted are two successive sessions
after about 7 weeks of lead administration. (Redrawn from Dietz and
McMillan [5])
behavioral session terminated after 600 correct responses
had been made. This made it possible to study the within-
session error pattern by dividing each session into six blocks
of 100 correct responses and plotting the number of errors
made while completing each block of 100 correct responses.
Figure 7 shows that under the performance schedule (column
1) the number of errors made during each block was rela-
tively constant prior to lead exposure. Increases in errors
produced by lead tended to be concentrated at the beginning
of the session, especially for bird 9847. During the final block
of 100 correct responses, neither bird made more errors
under lead than under control conditions. Thus, for both
birds the increase in errors produced by lead had been elimi-
nated by the end of the session. Whether this elimination of
errors by the end of these sessions represents a behavioral
adaptation to the effect of lead, or is merely an artifact pro-
duced by the increase in control variability toward the end of
control sessions prior to lead exposure, cannot be answered.
Figure 7 also shows the within-session error elimination
-------
no
MCMILLAN
before and during lead administration for the chain acquisi-
tion schedule (column 2) with a reset contingency. For bird
6885, session 48 was chosen (filled points) because of the
unusually high error rate during this session. Figure 7 shows
that almost all of this error increase occurred during the first
block of 100 correct responses, where most of the acquisition
errors were occurring. During the next session, errors were
at a normal level in every block. The within-session pattern
of error reduction was normal during both lead sessions that
are shown for bird 2781.
The control variability was very high for bird 2781 in the
first block of 100 correct responses, after which the number
of errors decreased and became much less variable. The
variability in day to day errors, especially during the first
block of 100 correct responses under the chained acquisition
schedule with a reset contingency, is probably largely a
function of sequence difficulty. The acquisition of some se-
quences appears to be more difficult for the birds than is the
acquisition of other sequences. For example, the number of
errors under control conditions is outlined in Table 1 for
selected sequences for some of the birds studied by Dietz
and McMillan [5], It is obvious that the right, left, center,
right, and the right, center, left, center sequences consis-
tently generated a much higher number of errors than did the
center, left, right, center, and the left, center, left, right se-
quences. Although variability due to differences in problem
difficulty may mask subtle effects of lead, it should be em-
phasized that this chained acquisiton schedule with a reset
contingency produced behavior that was sensitive to the ef-
fects of a variety of drugs despite the baseline variability
(Fig. 3 and 4). In fact, this acquisition schedule produces
behavior that is generally sensitive to lower doses of drugs
than is behavior under the corresponding performance
schedule.
TABLE 1
AVERAGE NUMBER OF ERRORS FOR ALL BIRDS
DURING REPEATED ACQUISITION OF FOUR DIFFERENT
SEQUENCES*
TABLE 2
ERRORS BEFORE, DURING, AND AFTER LEAD ADMINIS-
TRATION IN BIRD 2654 UNDER THE CHAINED ACQUISI-
TION SCHEDULE WITH A RESET CONTINGENCY*
Sequence
RLCR
RCLC
CLRC
LCLR
Mean
Total
Errors
285
226
104
127
Standard Error
35
39
14
20
Number of Times
the Sequence
Was Studied
10
9
11
9
R = right; C = center; L = left
*Data from unpublished observations by Dietz and McMillan
Thompson (1978) has emphasized that after extended
training under acquisiton schedules without reset contin-
gency, a steady state is reached so that the total number of
errors per session stays relatively constant. Although most
of our early data supported the idea that a steady state was
achieved after extended training under acquisition schedules
with a reset contingency, a problem arose in data interpreta-
tion when the effects of a prolonged period of lead adminis-
tration were studied on behavior under this schedule. This
problem is illustrated in Table 2 where data are shown for
bird 2654 after almost two years of training under the chained
Condition
Control
Lead
Control
Lead
Control
Sessions
28
32
23
35
71
Errors
(Mean ± SEM)
230.2 ± 18.8
183.8 ± 15.0
170.3 ± 13.9
131.4 ± 7.9
112.3 ± 4.3
% Errors
in 1st Bin
73.1
63.9
55.3
52.6
57.5
*Data from unpublished observations by Dietz and McMillan.
acquisition schedule with a reset contingency. When the first
28 control sessions are compared to the 32 sessions under
lead, the data suggest that lead administration might be de-
creasing the error rate during acquisition, since there was an
apparent reduction in errors during chronic lead administra-
tion. Fortunately, lead administration was discontinued in
this pigeon, reinstituted and then discontinued again. During
each of these manipulations, the average number of errors
continued to decrease further. Thus it appears that even
after years of extended training under this schedule, the
baseline is very slowly changing in the direction of further
error elimination. This change is so slow that during the
determination of drug dose-effect curves over a few weeks
time, the baseline probably can be considered to represent a
steady state, but during long term chronic administration of
drugs or toxic chemicals these slow baseline changes could
lead to misinterpretations of the data, especially if the effects
of the toxic chemical are irreversible. Of course, if the be-
havioral effects of the chemical are large with respect to the
slow change in baseline, this will not be a serious problem.
In the study by Dietz and McMillan [5] it is possible that
the very gradual decrease in the error baseline over long
periods of time might be exerting an influence in a direction
opposite to an increase in errors produced by lead, so that
any small increases in errors produced by lead during re-
peated acquisition would be cancelled. We attempted to
evaluate this possibility by using a statistic called the split
middle method of trend estimation [10]. This statistic pro-
vides an estimate of trend over time. In the case of our
experiments, the null hypothesis is that there is no change in
the pre-lead baseline error trend during the period of lead
administration. Using the split middle trend analysis, we
were able to show that errors by some pigeons under the
acquisition schedule were significantly increased above the
number predicted by the trend of pre-exposure errors.
Whether this really is an effect of lead, or is an effect caused
by some other factor controlling the trend of error elimina-
tion, remains uncertain.
SUMMARY AND CONCLUSION
It is tempting to suggest that the behavioral processes
underlying repeated acquisition are more sensitive to drug
effects than are the behavioral processes underlying per-
formance, while the reverse is true for lead. However, it is
dangerous to interpret these data in terms of differential
-------
PIGEON SERIAL POSITION SEQUENCES
111
sensitivity of performance and acquisition. As we have dis-
cussed previously [5], it may be the baseline error rate that
determines the sensitivity of the behavioral baseline to chem-
icals. The performance schedules generally result in fewer
errors than do similar acquisition schedules. If a very simple
acquisition schedule is used, so that few errors are gener-
ated, and a complex performance schedule is used, so that
many errors are generated, the effects of chemical interven-
tion may be quite different from those recorded thus far.
Until experiments are performed to determine the effects of
drugs and toxic chemicals on acquisition and performance
baselines that have been equalized with respect to error
rates, it will be inappropriate to make general statements
about the relative sensitivities of these behavioral processes.
Nevertheless, lead does appear to be unique among chem-
icals that have been studied with present methods. All other
chemicals that have produced increases in total errors under
these acquisition and performance baselines have done so at
lower doses under the higher error-rate acquisition baselines
than under the lower error-rate performance baselines. In
contrast, a chronic dose of 6.25 mg/kg/day of lead produced
clear increases in errors under the performance schedule,
but it was difficult to show that this dose of lead was increas-
ing errors under the acquisition schedule.
REFERENCES
1. Barthalmus, G. T., J. D. Leander and D. E. McMillan. Com-
bined effects of ethanol and diazepam on performance and ac-
quisition of serial position sequences by pigeons. Psychophar-
macohgy 59: 101-102, 1978.
2. Barthalmus, G. T., J. D. Leander, D. E. McMillan, P. Mushak
and M. R. Krigman. Chronic effects of lead on schedule-
controlled pigeon behavior. Toxic, appl. Pharmac. 42: 271-284,
1977.
3. Boren, J. J. Repeated acquisition of behavioral chains. Am.
Psychol. 17: 421, 1963 (abstract).
4. Boren, J. J. and D. D. Devine. The repeated acquisition of
behavioral chains. J. exp. analysis behav. 11: 651-660, 1968.
5. Dietz, D. D. and D. E. McMillan. Effects of chronic lead admin-
istration on acquisition and performance of serial position se-
quences by pigeons. Toxic appl. Pharmac. 47: 377-384, 1979.
6. Evans, H. L., V. G. Laties and B. Weiss. Behavioral effects of
mercury and methylmercury. Jn:Contemporary Research in
Behavioral Pharmacology,edited by D. E. Blackman and D. J.
Sanger. New York: Plenum Press, 1978, pp. 207-224.
7. Evans, H. L. and B. Weiss. Behavioral toxicology. In: Contem-
porary Research in Behavioral Pharmacology, edited by D. E.
Blackman and D. J. Sanger. New York: Plenum Press, 1978, pp.
449-487.
8. Halting J. and D. E. McMillan. Repeated acquisition of re-
sponse sequences by pigeons under chained and tandem
schedules with reset and non-reset contingencies. Psychol. Rec.
26: 361-367, 1976.
9. Halting, J. and D. E. McMillan. Effects of pentobarbital and
rf-amphetamine on the repeated acquisition of response se-
quences by pigeons. Psychopharmacology 49: 245-248, 1976.
10. Kayden, A. E. Statistical analyses for single-case experimental
designs. In: Single-case Experimental Designs: Strategies for
Studying Behavior Change, edited by M. Hersen and D. H.
Barlow, New York: Pergamon Press, 1976, pp. 265-316.
11. Laties, V. G., P. B. Dews, D. E. McMillan and S. E. Norton.
Behavioral toxicity tests. In: Principles and Procedures for
Evaluating the Toxicity of Household Substances. Washington:
National Academy of Sciences, 1977, pp. 111-118.
12. Mello, N. K. Behavioral toxicology: A developing discipline.
In: Behavioral Pharmacology, edited by B. Weiss and V. G.
Laties. New York: Plenum Press, 1976, pp. 155-160.
13. NAS Report. Principles for evaluating chemicals in the en-
vironment. Washington: National Academy of Sciences, 1975,
pp. 198-216.
14. Sidman, M. and P. B. Rosenberger. Several methods for teach-
ing serial position sequences to monkeys. J. exp. analysis Be-
hav. 10: 467-478, 1967.
15. Spyker, J. M. Assessing the impact of low level chemicals on
development: Behavioral and latent effects. In: Behavioral
Pharmacology, edited by B. Weiss and V. G. Laties. New
York: Plenum Press, 1976, pp. 161-180.
16, Thompson, D. M. Repeated acquisition as a behavioral baseline
for studying drug effects. J. Pharmac. exp. Ther. 184: 506-514,
1973.
17. Thompson, D. M. Repeated acquisition of response sequences:
Effects of rf-amphetamine and chlorpromazine. Pharmac.
Biochem. Behav. 2: 741-746, 1974.
18. Thompson, D. M. Repeated acquisition of behavioral chains
under chronic drug conditions. J. Pharmac. exp. Ther. 188:
700-713, 1974.
19. Thompson, D. M. Repeated acquisition of response sequences:
Stimulus control and drugs. J. exp. analysis Behav. 23:429-436,
1975.
20. Thompson, D. M. Repeated acquisition of behavioral chains:
Effects of methylphenidate and imipramine. Pharmac. Bio-
chem. Behav. 4: 671-677, 1976.
21. Thompson, D. M. Development of tolerance to the disruptive
effects of cocaine on repeated acquisition and performance of
response sequences./. Pharmac. exp. Ther. 203: 294-302, 1977.
22. Thompson, D. M. Stimulus control and drug effects. In: Con-
temporary Research in Behavioral Pharmacology, edited by D.
E. Blackman and D. J. Sanger. New York: Plenum Press, 1978,
pp. 159-207.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 113-118. ANKHO International Inc., 1979.
Effects of Solvents on Schedule-Controlled
Behavior1
VICTOR A. COLOTLA AND SAMUEL BAUTISTA
National School of Professional Studies Iztacala and Faculty of Psychology
National University of Mexico, Mexico 20, D.F., Mexico
AND
MARTE LORENZANA-JIMENEZ AND RODOLFO RODRIGUEZ
Department of Pharmacology, Faculty of Medicine, National University of Mexico
Mexico 20, D.F., Mexico
COLOTLA, V. A., S. BAUTISTA, M. LORENZANA-JIMENEZ AND R. RODRIGUEZ. Effects of solvents on
schedule-controlled behavior. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 113-118, 1979.—Operant conditioning techniques
have been shown to be sensitive to the acute effects of industrial solvents. In the first experiment, five rats trained in a
multiple schedule with a fixed-ratio (FR) 10 component and a differential reinforcement of low rates (DRL) 20-sec compo-
nent, with a time out 60-sec between reinforcement periods, were exposed to 0.25, 0.50, 1 and 2 ml of toluene in the
experimental chamber. The effects were dose-dependent, with an increase in rate in the DRL component and a decrease in
FR responding. A second experiment assessing the effects of chronic exposure to thinner in the acquisition of a timing
behavior in rats showed an impairment in DRL learning after 4, 8 or 16 weeks of exposure to the solvent: however, rats
having a resting period did not differ from control animals. Whereas this finding suggests a reversible impairment in the
acquisition of a complex behavior, further research is needed to achieve more definitive conclusions.
Operant behavior Toluene Thinner Multiple schedule Rate-dependent effects Chronic solvent inha-
lation Temporal discrimination Rats
RECENT interest in behavioral pharmacology and
neurotoxicology has focused upon the health hazards of en-
vironmental contaminants (e.g. [16]), and on the industrial
substances that are being abused for deliberate self-
intoxication [13]. This concern for the effects of chemicals
on behavior and bodily health and development seems to
have arised from two sources of information: (a) recent re-
ports of mass poisoning due to undetected food and water
contamination [1] and of widespread abuse of solvents
among youths in several countries (e.g. [4,10]): and (b) in-
formation about the pathological, long-lasting effects in cen-
tral nervous system functioning due to prolonged exposure
to the contaminants and industrial inhalants [11]. During the
last few years, research in our laboratories has been aimed at
establishing the usefulness of the operant conditioning meth-
odology in the assessment of the behavioral effects of indus-
trial solvents. In the first experiment we present data on the
acute effects of exposure to toluene, and the second reports
on the chronic effects of paint thinner inhalation in the ac-
quisition of a temporal discrimination in laboratory animals.
EXPERIMENT 1
ACUTE EFFECTS OF TOLUENE ON OPERANT
BEHAVIOR
Previous experiments [5,6] have shown the sensitivity of
operant methods in the behavioral analysis of the effects of
industrial solvents. It was found, for instance, that the ef-
fects of paint thinner appear to be schedule-dependent, in
that responding under a fixed-ratio (FR) schedule was more
sensitive than performance under a differential reinforce-
ment of low rates (DRL) schedule, when rats trained in a
Mult FR9 DRL20 schedule were exposed to different doses
of the solvent [6], and that the effects of the same mixture of
solvents appeared to be rate-dependent when rats trained in
a fixed-interval (FI) schedule [5] were exposed to the same
doses of thinner as in the other study. The present experi-
ment extended the above findings to the main component of
paint thinner, toluene.
'We wish to thank Octavio Torres Chazaro for his help in the programming of the first experiment, and Jorge Rosas and Manuel Gonzalez
for their help in the conduct of the second experiment. Address reprint requests to: V. A. Colotla, apartado postal 69-716, Mexico 21, D.F.,
Mexico.
113
-------
114
COLOTLA ET AL.
METHOD
Animals
Five Wistar albino rats were used. They were approx-
imately three months old at the beginning of the study, and
were experimentally naive: the animals were kept in indi-
vidual home cages and placed in a 23-hr water deprivation
regime.
Apparatus
A standard BRS-Foringer operant conditioning chamber,
model RG-028 was 25 cm long, 22.5 cm wide, and 20 cm in
height. Both, the ceiling and the lateral walls were of trans-
parent Plexiglas, and the anterior and posterior walls, like
the 14 bars that formed the floor, were of stainless steel. In
the middle of the anterior wall there was a water magazine,
and 4 cm above and 4 cm to the left and right of the magazine
there were two retractile levers. Only the righthand side
lever was employed throughout the experiment; the left lever
was kept in a retractile position.
The experimental chamber was located inside a sound-
proof BRS-Foringer isolation cubicle, equipped with a fan
and an air extraction system. The programming and record-
ing of events was controlled by solid state equipment. A
Gerbrands cumulative recorder was also employed.
Solvent
Pure toluene (C|iH5CH3) was obtained from a commercial
chemical supplier (Merck) and was kept at environmental
temperature in a cool place. Toluene is the main ingredient of
commercial liquid paint thinner in Mexico, and is a volatile
flammable liquid, with a benzene-like odour [3].
Procedure
Each animal was trained to approach the water magazine,
and shaped to press the lever for water reinforcement by the
method of successive approximations. They then received a
session in which every response was reinforced, and were
subsequently trained to obtain reinforcements with an in-
creasing number of responses, until stable FR10 perform-
ance was achieved. Next, a multiple FR10 DRL20 TO60
schedule was instituted, where a 60-sec time out (TO) period
was followed by a 2-min period under the FR10 schedule,
followed by another TO period, and an 8-min interval with
the DRL20 component into effect. Thus, FR and DRL
periods alternated with a TO between them. A tone present
during the DRL components served as the discriminative
stimulus distinguishing both reinforcement components of
the multiple schedule. These sessions were 36-min long.
Three stability criteria were established before beginning
with exposure to toluene: (a) an appropriate stimulus con-
trol, in that performance under each component was char-
acteristic for that schedule: (b) a variability of no more than
10% from the average of the last three sessions in reinforce-
ment rate (number of reinforcements per minute) under the
DRL schedule: and (c) no ascending or descending trend in
DRL reinforcement rate. When an animal's performance
achieved the above criteria a dose of toluene was randomly
selected and employed until all animals were exposed to all
doses. If an animal's performance did not return to baseline
level in the day following a session with the solvent, no
further doses were given until the stability criteria were again
satisfied. In addition, every experimental session with tol-
uene was immediately followed by an additional session
without the solvent to assess behavioral recovery. The doses
employed were 0.25, 0.50, 1 and 2 ml, which given the vol-
ume of the experimental chamber (100 1), corresponded to
2.0, 4.0, 8.0 and 16.0 mg/1, or 574, 1148, 2296, and 4595 ppra,
respectively.
The experimental procedure for the administration of the
solvent was as follows: the dose of toluene was placed in a
glass dish plate which was in turn placed underneath the grill
floor of the chamber, directly below the animal. The fan and
exhaust holes were covered with plastic, the animal was in-
troduced to the chamber and the session with the multiple
schedule was begun. As a control procedure, each animal
was submitted to sessions in which the above procedure was
repeated, but substituting the solvent by a few ml of water.
These are called control sessions in the analyses of the data.
The animals had 30 min of free access to water at the end
of every session, which was conducted seven days a week.
Data registered included number of responses and rein-
forcements for each schedule component, interresponse
times (IRTs) in blocks of three seconds, and response
cumulative records.
RESULTS
Figure 1 shows the results obtained with one of the rats
employed in the experiment. The upper record is from one of
the baseline sessions and shows stable performance in the
multiple schedule, with responding typical of the component
schedules employed: a fast-rate FR responding and a low-
rate, paced responding in the DRL schedule. The other re-
cords show the effects of the different doses of toluene, a
decrease in responding in the ratio component and an in-
crease in response rate during the DRL periods. There was
also an increase in response frequency during the TO periods
with the highest doses employed. The records for the other
animals were similar, although in some of them FR perform-
ance was more affected than in the animal shown in this
figure. There was no significant change in performance dur-
ing the control sessions.
The analysis of reinforcement rate in both components is
shown in Fig. 2, expressed as a percent of baseline perform-
ance. There was a decrease in reinforcement rate for both
schedules under exposure to toluene, an effect that was more
evident with increases in dose. FR performance was more
affected than DRL responding, but the difference does not
appear significant, except probably with 1 ml of toluene.
Figure 3 depicts the changes in response rate in the two
schedules during the control sessions and the experimental
sessions with the solvent. A differential effect is evident for
the two schedules, an increase in response rate in DRL per-
formance and a decrease in FR responding with exposure to
toluene. Behavioral recovery data are still being analysed
and are not reported here.
EXPERIMENT 2
CHRONIC EFFECTS OF THINNER ON OPERANT
BEHAVIOR
Pathological clinical findings have been consistently re-
ported in the literature of chronic solvent abuse [9, 11, 14].
Thus, a second experiment was conducted to evaluate the
effects of chronic solvent inhalation on the acquisition of a
complex behavior in laboratory rats. Since the industrial
substance more widely used by solvent abusers in Mexico
-------
EFFECTS OF SOLVENTS ON OPERANT BEHAVIOR
115
baseline
I SR6
0.25 ml
0.5ml
42ml
Smin
FIG. 1. Representative results obtained when rats trained in a multiple schedule were exposed
to the doses of toluene indicated in the cumulative records. See text for details.
City is the commercial paint thinner easily obtained in hard-
ware stores, the solvent employed here was the same thinner
mixture (see composition below), and the behavior studied
was the acquisition of a temporal discrimination as evi-
denced in DRL performance.
METHOD
Animals
Twenty-four Wistar albino rats were randomly assigned
to one of three groups labeled A, B and C. They were all
male and aged one day old at the beginning of the study. The
rats remained with their mothers until weaning, which was
21 days post-partum, and were then kept in group cages with
free access to food and water. Each group was further di-
vided into two subgroups of eight animals each, an experi-
mental (E) and a control (C) subgroup. The only restriction
in the assignment to the different groups was that every ex-
perimental animal had a control counterpart from the same
litter. After the chronic inhalation procedure described
below all animals were transferred to individual home cages
and kept in a 23-hr water deprivation regime during the oper-
ant conditioning experiment.
Apparatus
Four glass inhalation chambers, with a 3-liter volume
each, were used during the inhalation treatment, maintained
at a constant temperature of 25 ± 1°C. For the operant train-
ing, a BRS-Foringer operant conditioning chamber as de-
scribed for Experiment 1 was employed.
Solvent
Commercial liquid paint thinner of low quality was em-
ployed. Figure 4 shows the composition of the thinner mix-
ture by gas chromatography and the insert details some of
the technical information for this analysis, and the percent-
age of each component of the sample.
Procedure
Inhalation treatment. All experimental animals were ex-
posed to 131.4 mg of thinner (50,000 ppm) in the inhalation
chambers during ten-min periods, twice daily, five days a
week during four (Group A), eight (Group B) or sixteen
(Group C) weeks. Each control rat was also placed in the
inhalation chamber under the same procedure as its experi-
mental counterpart, except that no thinner was injected into
the chamber.
Operant training. At the end of the inhalation treatment
for Group C (16 weeks) each subgroup of animals was further
divided into two (Phase I and Phase II) to proceed with the
training in the experimental operant conditioning chamber.
-------
116
COLOTLA£TAL
UJ
z
_l
UJ
10
<
m
o
o*
^e
UJ
(—
<
cr
t—
z
UJ
Z
UJ
o
or
o
u.
UJ
cc
180-
160-
140-
120—
100-
80-
60—
40—
20-
-DRL
^ 200-
UJ
5 180-
S 160-
<
CD 140—i
S 120-
n i i i i
C 0.25 0.50 1.0 2.0
o
«M*
UJ
UJ
\n
•z.
o
Q_
I/)
LU
o:
100—
80-
60-
40-
20-
o-
-ORL
I I I I I
C 0.25 0.50 1.0 2.0
TOLUENE (ml.).
FIG. 2. Reinforcement rate as a percent of baseline in the two
schedule components of the multiple schedule, under control and
experimental (toluene) conditions.
TOLUENE (ml.).
FIG. 3. Response rate as a percent of baseline in the two schedule
components of the multiple schedule, under control and experi-
mental (toluene) conditions.
Due to equipment limitations, Phase I animals were first
trained and then the Phase II set of rats. The experimental
procedure, however, was the same for all the animals.
Each rat was adapted to a water deprivation regime and
was then subjected to the following procedure: the rat was
exposed to an autoshaping situation [2], consisting of trials of
a 5-sec presentation of a light above the lever followed by the
free delivery of a drop of water in the magazine, every 50
sec. As expected, the pairing of light-above-the-lever and
water delivery led the animal to press the lever and obtain
the water reinforcement within two sessions. The rat then
received two sessions in which every response was rein-
forced, and was then placed in a DRL20 schedule, in which a
response was reinforced only if it was emitted at least 20 sec
after the preceding response. Each animal received 30 ses-
sions, each lasting 30 min, and conducted at the same time
every day. Data registered were number of responses and
reinforcements per session, and response cumulative re-
cords.
RESULTS
The main datum of interest was the percentage of rein-
forced responses for each subgroup of animals, taken as a
measure of efficiency in DRL responding, and thus of the
extent of the temporal discrimination achieved. Statistical
analyses of the data showed a nonsignificant difference,
F(l,6)=0.73, p>0.01, between the experimental animals in
Phase I (average of 13.4 reinforced responses) and Phase II
(11.6 mean reinforced responses). However, the average
number of reinforced responses for the control animals was
28.0 in Phase I and 15.1 in Phase II. This difference was
statistically significant, F(l,6) = 15.16,p<0.01, and thus data
from the two phases were not pooled. The results are shown
for the two phases separately.
Figure 5 shows the main finding of this experiment. First,
from the results of Phase I it is evident that the rats subjected
to chronic inhalation of thinner show an impairment in the
acquisition of the temporal discrimination, as compared to
their control counterparts, impairment that is more marked
as the inhalation treatment was extended. The rats in Phase
II, however, did not show a clear differential performance
with respect to the treatment of interest. Only the animals
subjected to 16 weeks of thinner inhalation appeared to be
more impaired than their controls.
GENERAL DISCUSSION
With respect to the first experiment, there appears to be
only one report in the literature [8] which may be contrasted
with the results obtained here. In that one experiment, Geller
et al. found an increase in barpressing rate of rats trained
under a variable-interval schedule when exposed to the
solvent ketone, and reported a long-lasting effect in that the
increased rate persisted for several days after initial expo-
sure to the solvent. In our first experiment, however, no
-------
EFFECTS OF SOLVENTS ON OPERANT BEHAVIOR
117
CO
to oo
X X
X
o
X
8
COLUMN a
TEMPERATURES:
FLOW
5* OF "CANDELILLA" MAX
CHROMOSORB W-AW 80/100 MESH
10 ft 1/8" STAINLESS STEEL
OVEN 95° C
DETECTOR 125'C
INYECTOR 150*C
HELIUM 40 ml/rain
THERMIC CONDUCTIVITY DETECTOR: FILAMENTS
CURRENT 150 mA.
SIGNALS
1
2
3
4
5
6
7
8
9
10
11
12
COMPOUND
METHANOL
ACETONE
NOT IDENTIFIED
HEXANE
HEPTANE ?
OCTANE ?
NOT IDENTIFIED
NOT IDENTIFIED
BENZENE
METHYL ISO-BUTYL KETONE
TOLUENE
XYLENE
TOTAL
PER CENT
24.605
1.966
1.154
8.787
4.799
4.456
0.594
0.274
0.486
10.545
41.888
0.439
99.999
16
22m in
FIG. 4. Gas chromatogram of the thinner mixture employed. See text for details.
iccn 30
O LJ
u- tn
20-
LJ
O
OL
UJ
Q.
10-
0J
FIRST PHASE
SECOND PHASE
CONTROL
"THINNER"
16
16
WEEKS OF INHALATION
FIG. 5. Percent of reinforced responses in DRL acquisition of exper-
imental (thinner) and control rats as a function of length of treatment
(weeks of inhalation), for Phase I and Phase II animals.
-------
118
COLOTLA ET AL.
persistance was found in the changes observed in response
rates. Nevertheless, the main differences between the
studies contrasted reside in the solvent employed and the
reinforcement schedule utilized as a behavioral baseline, and
thus the results should await further research for interpreta-
tion.
Furthermore, the main finding of the acute effects of the
solvent on schedule-controlled behavior seems to be that
there is a differential effect of toluene in the two schedules
employed, which might be interpreted as a rate-dependency
similar to that consistently reported for the amphetamines [7]
and other behaviorally active compounds [12], a finding
which would place toluene among the many other substances
that interact with the behavioral rate employed as a baseline.
The second experiment was concerned with the chronic
effects of exposure to commercial paint thinner on the ac-
quisition of a temporal discrimination, as evidenced in DRL
performance. The experiment followed the design employed
in a study which assessed the effect of prolonged alcohol
consumption in timing behavior [15], and the results suggest
that persistent inhalation of thinner vapors causes an im-
pairment in DRL acquisition, regardless of the length of
treatment, when the animals are tested within a relatively
short time after the inhalation procedure. However, provid-
ing a longer resting period after the exposure to the solvent
results in a lack of difference between experimental and con-
trol animals, except probably those exposed for a longer time
interval (i.e., 16 weeks). Whereas this finding might be inter-
preted as suggesting a reversibility of the brain dysfunction
responsible for the impairment of the timing behavior, it
should be pointed out that the control rats in Phase II had a
poor performance as compared to Phase I controls, a fact
that should be examined more closely in further experi-
mentation. It appears then, that chronic exposure to thinner
may cause an impairment in the acquisition of a complex
behavior in laboratory animals.
REFERENCES
1. Bakir, F., S. F. Damluji, L. Arnin-Zeki, M. Murtadha, A.
Khalidi, S. Tikriti, H. I. Dhakir, T. W. Clarkson, F. C. Smith
and R. A. Doherty. Methylmercury poisoning in Iraq: An inter-
university report. Science 181: 230-233, 1973.
2. Brown, P. L. and H. M. Jenkins. Autoshaping of the pigeon's
keypeck. J. exp. analysis Behav. 11: 1-8, 1968.
3. Bruckner, J. V. and R. G. Peterson. Review of the aliphatic and
aromatic hydrocarbons. In: Review of Inhalants: Euphoria to
Dysfunction, N1DA Research Monograph 15, edited by C. W.
Sharp and M. L. Brehm. Rockville, Maryland: National Insti-
tute of Drug Abuse, 1977, pp. 124-163.
4. Carroll, E. Notes in the epidemiology of inhalants. In: Review of
Inhalants: Euphoria to Dysfunction, NIDA Research Mono-
graph 15, edited by C. W. Sharp and M. L. Brehm. Rockville,
Maryland: National Institute of Drug Abuse, 1977, pp. 14-17.
5. Colotla, V. A., B. E. Z. Jacobo and M. M. G. Moctezuma.
Efectos agudos del "thinner" en la ejecucion de ratas en un
programa de intervalo fijo. Rev. Mex. Andlisis Cond. 4: 1978, in
press.
6. Colotla, V. A., M. Lorenzana-Jimenez, J. Echavarria Luna and
R. Rodriguez. Evaluacion de los efectos conductuales del tiner
con la metodologia operante. Cuad. dent. CEMESAM 8: 203-
220, 1978.
7. Dews, P. B. and G. R. Wenger. Rate-dependency of the behav-
ioral effects of amphetamines. In: Advances in Behavioral
Pharmacology, edited by T. Thompson and P. B. Dews. New
York: Academic Press, 1977, pp. 167-227.
8. Geller, I., J. R. Rowlands and H. L. Kaplan. Efectos de las
cetonas en la conducta operante en animales de laboratorio. In:
Inhalacion Voluntaria de Disolventes Industriales, edited by C.
M. Contreras Perez. Mexico: Editorial Trillas, 1977, pp. 125-
138.
9. Grabski, D. A. Toluene sniffing producing cerebellar degenera-
tion. Am. J. Psychiat. 118: 461-462, 1961.
10. Natera, G. Estudio sobre la incidencia del consume de disol-
ventes volatiles, en 27 centres de la Republica Mexicana. In:
Inhalacion Voluntaria de Disolventes Industriales, edited by C.
M. Contreras Perez. Mexico: Editorial Trillas, 1977, pp. 329-
351.
11. Prockop, L. and D. Couri. Nervous system damage from mixed
organic solvents. In: Review of Inhalants: Euphoria to Dys-
function, NIDA Research Monograph 15, edited by C. W. Sharp
and M. L. Brehm. Rockville, Maryland: National Institute of
Drug Abuse, 1977, pp. 185-198.
12. Sanger, D. J. and D. E. Blackman. Rate-dependent effects of
drugs: A review of the literature. Pharmac. Biochem. Behav. 4:
73-83, 1976.
13. Sharp, C. W. and M. L. Brehm (eds.) Review of Inhalants:
Euphoria to Dysfunction, NIDA Research Monograph 15.
Rockville, Maryland: National Institute on Drug Abuse, 1977.
14. Tsushima, W. T. and W. S. Towne. Effects of paint sniffing on
neuropsychological test performance. J. abnorm. Psychol. 86:
402-407, 1977.
15. Walker, D. W. and G. Freund. Impairment of timing behavior
after prolonged alcohol consumption in rats. Science 182: 597-
599, 1973.
16. Weiss, B. and V. G. Laties. Behavioral Toxicology. New York:
Plenum Press, 1975.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 119-127. ANKHO International Inc., 1979.
Testing for Behavioral Effects of Agents
P. B. DEWS AND G. R. WENGER1
Laboratory of Psychobiology, Harvard Medical School, 25 Shattuck Street, Boston, MA 02115
DEWS, P. B. AND G. R. WENGER. Testing for behavioral effects of agents. NEUROBEHAV. TOXICOL. 1: Suppl. 1,
119-127. 1979.—In the present state of science no morphological or chemical changes may be detectable at a time when
behavior is profoundly disturbed, as in schizophrenia. Until we are reassured to the contrary, we must assume that
exogenetic intoxication can produce changes detectable only as behavioral changes. Therefore behavioral toxicology must
be studied. In contrast to toxic manifestations such as lethality or carcinogenicity, which tend to be unequivocal and
irreversible, behavioral changes are like physiological changes in that they are quantitative, changing in time, and relate to
variables with a considerable range of normal variability. An experiment on behavioral teratology in mice is described and
the results used to illustrate the limits of the possible in behavioral toxicology. From reported and observed variability it is
surmised that changes that occur in as many as 1 per 100 of the population or average as large as a 10% decrement will still
be too small to be detected by direct experiment. Such risks are frequently unacceptable. Reasons are given for hoping that
epidemiological studies may be able to supplement experimental lexicological studies to provide a better assessment of risk
of small impairments or rare susceptibility.
Mice Caffeine Atropine Behavioral toxicity Behavioral teratology Safety-testing MultFR30FI
600 sec Schedule-controlled responding Spontaneous motor activity
THERE are many reasons for studying the behavioral effects
of toxic influences but one reason is fundamental: in the
present state of science no morphological or chemical
changes may be detectable at a time when behavior is pro-
foundly disturbed. The most dramatic example is acute
schizophrenia when almost any psychological or psychiatric
test will show gross abnormality while there are no detecta-
ble morphological changes and all chemical deviations found
to date are variable, not diagnostic, and can be related to
changes in dietary and other life patterns consequent on
schizophrenia. It is not necessary to assume that schizo-
phrenia is an intoxication for the importance of behavioral
studies to be emphasized by the example. If the undis-
covered influence in schizophrenia can cause such profound
changes in behavior then it must be assumed that exogenetic
intoxication can do the same, until we are reassured to the
contrary. There is insufficient evidence for us to be generally
reassured at the present time, although it appears likely that
profound behavioral disturbances not accompanied or even
preceded by other easily detectable evidences of intoxication
occurs with very few agents. The analogy to schizophrenia
may be taken one step further. Almost any behavioral test
will detect abnormality in a schizophrenic. Perhaps at the
present stage of behavioral toxicology we should concen-
trate on tests that are quick, efficient and informative and not
be too concerned about finding tests whose results have
plausible theoretical interpretations.
Other reasons for studying behavioral toxicology are as
follows. Behavioral tests are non-destructive so a subject
can be studied repeatedly. Thus the development of and
perhaps recovery from toxicity can be followed in a manner
not possible for histological and many chemical assays.
There is a widespread expectation that behavioral changes
may be an early manifestation of toxicity and that therefore
behavioral tests will provide very sensitive indices of toxic-
ity. Many people have gone so far as to suggest that the
reason for studying behavioral toxicity is that it might provide
the most sensitive tests, and that if behavioral tests are not
generally the most sensitive then behavioral toxicity is not
worth studying. As indicated above, the latter premise is unac-
ceptable. Whether behavioral tests are usually more sensitive is
a question to be answered by experiments. There is relatively
little in the way of systematic comparisons in the Western
literature. There is a formidable body of opinion in the USSR
that for a great many agents behavioral tests are the most
sensitive indications of toxicity, and maximum allowable
concentrations in the USSR based on behavioral tests are
much lower than in the US where they are usually based on
other criteria [5], Another reason for studying behavioral
toxicity is the hope that the effects of poisons will help the
analysis of behavioral phenomena in a way similar to the
large role that poisons played in the analysis of physiological
phenomena, from Claude Bernard and his curare to tet-
rodotoxin, bungarotoxin and actinomycin in our own time.
Finally, some people study behavioral toxicity simulta-
neously with neurological and neurochemical toxicities in
the hope of gaining insight into the normal relations among
behavioral, neurological and neurochemical phenomena.
The discussion of "test methods for the definition of ef-
fects of toxic substances on behavior" will be approached
through the description of a simple pilot experiment. The
pilot experiment was actually an experiment in behavioral
'Present address: Department of Pharmacology, University of Arkansas Medical Sciences, West Markham, Little Rock, AK.
119
-------
120
DEWS AND WENGER
teratology, but serves to illustrate problems general to be-
havioral toxicology and indeed to toxicology and safety-
testing in general. Teratological effects have the advantage
for behavioral studies that they may be permanent and rela-
tively stable.
When pregnant subjects are exposed to a teratogen, typi-
cally only occasional offspring are terata, the remaining
members of the litters being indistinguishable from normal.
The frequency of terata increases with increasing exposure
to the teratogen at the critical periods during pregnancy.
Teratology has been primarily a morphological science, and
terata have been recognized by their anatomical deviances.
(Indeed, Webster's dictionary defines teratology as con-
cerned exclusively with deviations from normal structure.
The legitimacy of behavioral teratology as a designation
would appear to derive from equating terata with monsters;
monsters may be monstrous by reasons of their behavior,
again according to Webster. It would have been better to
invent another term to cover the behavioral consequences of
antenatal influences but it is too late now.) It is only in recent
years that there has been much interest in the possibility that
permanently deviant behavior may result from exposure to
particular chemical substances during development, in the
absence of currently recognizable morphological or physi-
ological deviance. It is recognized that permanent behavioral
deviance must be the result of permanent physiological
changes; it must also be recognized, however, that the phys-
iological changes may not be indicated by any available
technique other than behavioral assessment.
That permanent changes in behavior may result from in
utero exposure to teratogens is undoubted. What is much
less clear, however, is whether permanent changes in what
the organizers of the workshop have chosen to call cognitive
behavior may result from exposures that do not have other,
more easily detectable, physical effects such as interference
with pregnancy itself, reduced survival of litters, lower
weight gain in offspring, and increased incidence of familiar
morphological terata or grossly obvious behavioral incompe-
tences. Most, if not all, examples of behavioral teratology in
the literature have employed exposures that have had clear
effects detectable by conventional teratologic techniques or
by such minor modifications as letting the offspring grow up
and looking at them and weighing them.
A fundamental consideration in behavioral teratology
does not seem to have been addressed definitively. If behav-
ioral terata occur as do anatomical terata, then most prenatal-
ly exposed offspring should be normal with only occasional
individuals showing gross behavioral changes that should be
fairly easily detected by simple general tests even if the
changes are specific. To detect such teratogenesis simple
and speedy tests should be used on a prolific species so that
as many subjects as possible can be examined. Alternatively
it may be suggested that during central nervous system mat-
uration so many different pathways and connections must
develop in exactly the right timing and sequence that most
subjects exposed to a sufficient noxious influence would
have some one or other process more or less perverted.
Higher behavioral activities require a great deal of sequential
processing in the CNS so that there is a high likelihood that if
deviances exist they will be encountered in the processing.
The results might be any one or more of a variety of de-
viances, perhaps slight or subtle, but occurring in most ex-
posed subjects. Such impairments are most likely to be de-
tected by intensive study of a few subjects. The alternative
suggestions thus point to different strategies for safety test-
ing. Our pilot experiment addressed, but did not solve, this
issue.
PILOT EXPERIMENT
The effects of atropine or caffeine throughout the period
of development of the CNS were studied in mice. Atropine
was selected because its high activity in antagonizing mus-
carinic effects of acetylcholine suggested it might interfere
with cholinergic function in the developing nervous system,
with permanent sequelae. Caffeine was chosen as an agent of
wide distribution that can have behavioral effects and in
massive doses is conventionally teratogenic [7]. It was of
interest to see whether smaller doses produce behavioral
teratology.
METHOD
Female mice in cages with males were subjected to a
12-hr light/12-hr dark diurnal cycle and examined each morn-
ing for vaginal plugs. When a plug was observed, the mouse
was randomly assigned to one of eight groups and housed
individually (Table 1). The mice were exposed to the various
solutions or water as sole drinking fluid throughout preg-
nancy. Litters were culled to 6 on the day of birth, and
exposure to the solutions continued.
Preliminary experiments suggested that mice restricted to
0.3 ml/g/day of drinking fluid would essentially always drink
all the fluid within 24 hr, so that dosages of agents could be
known, and that this fluid intake would support normal
growth. Accordingly, all groups of mice were fluid restricted
except one control group given unlimited water. All re-
stricted mice essentially always drank all their fluid, and
maintained normal growth rates except, perhaps, for some
slight slow-down in growth of pups just before weaning. At
21 days, the litters were separated from the dam, and sepa-
rated by sex, but they continued to receive only the solutions
as drinking water until 45 days of age. All litters were then
given unlimited water for the rest of the study. Testing was
started at 60 days. No attempt was made to conduct simulta-
neously classical teratological assessments (disssections and
histology). Three behavioral tests were used. Test 1 was a
simple assessment of spontaneous locomotor activity
(SMA). A transparent plastic mouse cage was traversed by a
light beam. A single mouse was put in the cage and the
number of times the light beam was broken during a period of
one half hour was counted. Test 2 was a motor activity test
rather similar to the first but with an additional feature. The
mouse was studied in a circular cage with 3 light beams and 3
photocells (Fig. 1). The beams (1, 2 and 3) made it possible to
know when the mouse was continuing round and round in
the same direction rather than backing and hauling. The total
number of times any light beam was broken and the number
of times the mouse broke the beams in the sequence 1 then 2
then 3 (rotation) were counted. The same ten mice of each
sex from each treatment group were studied in Tests 1 and 2
for 3 consecutive days starting shortly after the 60th day of
age. Test 3 was an assessment of schedule-controlled re-
sponding. Mice broke repeatedly a beam of light falling onto
a photocell by moving their noses in and out of the beam and
received milk under a mult FR30 FI 600 sec schedule [9].
One male and one female mouse from each of 3 litters in each
of the 8 groups was trained. Each mouse was studied daily 5
days/week for 5-6 weeks (which is why only 3 males and 3
females per group could be studied). The mice were exposed
-------
BEHAVIORAL EFFECTS OF AGENTS
121
TABLE 1
Regimen
Designation
Concentration
Volume
ad lib water
restricted water
10 mg/kg/day
caffeine
30 mg/kg/day
caffeine
100 mg/kg/day
caffeine
0.3 mg/kg/day
atropine
3.0 mg/kg/day
atropine
30 mg/kg/day
atropine
ad lib
R-H2O
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
0
0
0.033 mg/ml
0.10 mg/ml
0.33 mg/ml
0.001 mg/ml
0.01 mg/ml
0.10 mg/ml
ad lib
0.3 ml/g/day
0.3 ml/g/day
0.3 ml/g/day
0.3 ml/g/day
0.3 ml/g/day
0.3 ml/g/day
0.3 ml/g/day
to a standard training sequence, much as previously de-
scribed [9]. The final schedule was continued for 13 sessions,
and responses in last 3 sessions averaged for tabulation. The
number of times during training that a session was not com-
pleted in a nominal time (30 min) was also noted.
10 cm.
FIG. 1. Dotted lines show beams of light crossing circular track in
which mouse is placed for 30 min.
Comments
The simple assessments of spontaneous motor activity,
Tests 1 and 2, permitted a reasonably large number of mice
to be studied. Test 3, mult FRFI, represents a higher behav-
ioral activity, suitable for intensive appraisal of a few mice.
The schedule is well established as sensitive and productive
of valuable information in behavioral pharmacology [4] and
has already been used in behavioral toxicology [6].
RESULTS
Test 1
The average total counts per mouse in three sessions on
three consecutive days for the different treatment groups are
shown in Table 2. All the counts were between 540 and 647
with SE between 32 and 87 save for one group, the females
that had been raised on 100 mg/kg/day caffeine. This group
had a mean of 838 and SE of 273. The high SE suggests that
the high count may be due to a few high values. In fact it was
due to a single mouse that had a count of 2723. The next
highest count among the remaining 159 mice tested was 1362
and only two other mice had as many as 1000. None of the
mice in the other treatment groups differed from one another
or from those receiving free water: notably, the large dose of
30 mg/kg/day of caffeine and all doses of atropine were with-
out recognizable effects.
Test 2
The total number of breaks of the three light beams are
shown in Table 3. The results were unremarkable except for
the high variance in the females of the 100 mg/kg/day caffeine
group, which turned out to be due to the same mouse as the
high counts in Test 1. The numbers of complete clockwise
circumnavigations of the cage (rotations) distinguishes the
female of the 100 mg/kg/day group even more clearly. The
high variance of this group is seen in Table 4. The single
abnormal mouse in relation to the other 159 mice studied is
identified in the frequency-distribution of Fig. 2; the score of
-------
122
DEWS AND WENGER
80
60
I
UJ
ID
a
20
J I I I I L
100
300 500
ROTATIONS
700
OC78M6
FIG. 2. Frequency-distribution of different numbers of rotations in
three sessions of 160 mice.
TABLE 2
ACTIVITY 1/90 MIN
ad lib
RH.,0
10 CAF
30 CAP
100 CAF
0.3 AT
3 AT
30 AT
M
m
604
544
602
502
557
565
540
588
SE
32
47
56
56
73
36
51
66
m
647
581
608
593
838
532
606
509
F
SE
87
49
36
50
273
37
34
64
M&F
m
626
562
605
548
698
548
573
548
this one mouse is 3 SD's to the right of the mean. High
activity can arise from a variety of abnormalities unrelated to
caffeine so, obviously, no conclusion can be drawn from a
single mouse.
Test 3
All the mice trained under mult FR30 FI300 sec gave
similar performances without regard to early treatment, with
again, a single exception. No specific change in any aspect of
multi FRFI responding in most mice exposed to early treat-
ment with caffeine or atropine has been identified.
Table 5 summarizes information of FR. The mean rate of
responding was about 1 response per second, females a
shade lower than males, but not significantly so. Rates for
TABLE 3
ACTIVITY 2/90 MIN
ad lib
RH20
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
ad lib
RH20
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
M
m
3913
3540
3472
2987
3297
4046
3485
3878
ACTIVITY:
M
m
164
108
116
101
94
117
137
137
SE
253
161
129
271
143
279
248
242
TABLE 4
m
4252
4954
4114
4308
3987
4202
3989
3828
ROTATIONS/90
SE
22
9
7
15
11
14
21
25
m
130
147
103
116
174
96
113
128
F
SE
414
463
268
296
591
268
270
242
MIN
F
SE
22
44
11
11
79
9
16
12
M&F
m
4082
4247
3793
3648
3642
4124
3737
3853
M&F
m
147
139
110
109
134
106
126
133
100 mg/kg/day caffeine group were low but not lowest for
either males or females although for the average of both they
were, in fact, the lowest. The main reason for this turned out
again to be a single mouse that had a very low FR rate. After
responding normally for some sessions the FR rate declined
to very low levels in this mouse. Meanwhile, FI responding
remained substantially normal.
Spontaneous loss of FR pattern of responding is known to
occur not infrequently. It has been seen many times in a
variety of species down the years when no caffeine was in-
volved. FR responding has a positive feedback feature. The
faster the subject responds, the sooner the food is delivered;
and it has been suggested that this feature favors faster re-
sponding, which in turn brings food still closer which further
favors still faster responding and so on until the subject ap-
proaches its physiological limit of rate of responding. In-
deed, the fastest rates of responding that have ever been
recorded have been under these FR type schedules. One of
the lines of evidence for the positive feedback feature is that
it also seems to work in the opposite direction. If rate of
responding is slowed, food comes slower which favors
slower responding and so on until responding becomes de-
sultory. As an isolated finding, little significance can be at-
tached to a not uncommon phenomenon occurring in one
mouse.
Table 6 shows data for FI. There were no consistent con-
vincing differences between mice of any of the treatment
-------
BEHAVIORAL EFFECTS OF AGENTS
123
TABLE 5
FR (R/SEC)
ad lib
RH20
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
ad lib
RH2O
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
m
1.01
1.23
1.72
1.00
0.77
0.73
1.20
1.59
m
0.39
0.34
0.44
0.35
0.28
0.27
0.32
0.66
M
SE
0.46
0.44
0.48
0.12
0.34
0.11
0.18
0.18
TABLE 6
FI (R/SEC)
M
SE
0.12
0.02
0.08
0.15
0.11
0.04
0.07
0.09
m
0.63
0.87
1.45
0.84
0.56
0.87
0.55
0.86
m
0.29
0.33
0.52
0.31
0.47
0.42
0.17
0.32
F
SE
0.08
0.21
0.54
0.13
0.36
0.11
0.43
0.24
F
SE
0.10
0.12
0.07
0.03
0.23
0.12
0.08
0.09
M&F
m
0.82
1.05
1.58
0.92
0.66
0.80
0.88
1.22
M&F
m
0.34
0.34
0.48
0.33
0.38
0.34
0.24
0.40
groups under FI. The variability was again highest in the
females that had received 100 mg/kg/day caffine.
Training
In the training sessions for mult FRFI, the numbers of
times there was a failure to complete a session in 30 min (see
Method) in the 6 mice in each of the various groups was as
follows: ad lib, 4; restricted water, 7; 10 mg/kg/day caffeine,
7; 30 mg/kg/day caffeine, 4; 100 mg/kg/day caffeine, 10; 0.3
mg/kg/day atropine, 1; 3 mg/kg/day atropine, 1; 30 mg/kg/day
atropine, 6. Once again it is the 100 mg/kg/day caffeine group
that gives the extreme value although again, not statistically
significantly so, and dose-effect relations are not apparent,
casting serious doubt on the biological significance.
Comments
Three independent suggestions of behavioral abnormality
in the 100 mg/kg/day caffeine treated offspring were found.
None of them is, in itself, at all convincing, but the coinci-
dence of the 3 extreme values occurring in the same one of
the 8 groups is not easily dismissed ((1/8)3=0.002). Even
more important is that other workers have found behavioral
teratology at these sorts of levels of exposure in rats
(Sobotka, private communication).
If it is accepted that the tests showed the occurrence of
real behavioral teratology, what light is thrown on the issue
ad lib
RH2O
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
TABLE 7
PREGNANCIES
Plugs
6
8
21
8
24
5
5
5
Pregs
6
6
17
8
16
4
5
5
Non-preg
Plug
0
0.25
0.19
0
0.33
0.20
0
0
posed at the beginning: is behavioral teratology a phenom-
enon of sporadic occurrence among exposed offspring or is it
a matter of graded impairment of many? At face value, the
evidence certainly suggests sporadic occurrence of big de-
viances with the large dose of caffeine. The evidence is com-
patible with the nature of the deviance being of very different
type from pup to pup, just as a single teratogen may produce
very different types of morphological terata. No evidence of
impairment of most of the pups raised on 100 mg/kg/day was
detected. But the fact is inescapable that small changes could
not have been established. For example, the low rate of re-
sponding under FI of the females raised on 3 mg/kg/day at-
ropine was not statistically significant. In all the tests, coef-
ficients of variation were in the vicinity of 0.3, and some of
this variation is non-random. The implications of high varia-
bility with a component not due to sampling error has been
discussed previously [2] and will be elaborated later.
Other Effects
Although the pilot study was concerned with behavioral
toxicology, some routine biological information was col-
lected.
(1) As indicated in Method, mice were allocated at ran-
dom to the various groups after diagnosis of a vaginal plug.
The incidence of pregnancies is shown in Table 7. The inci-
dence was lowest in the 100 mg/kg/day caffeine group. The
other groups showing a less than 100% pregnancy rate were
the restricted water and the low doses of both of the agents.
The apparent paradox may be related to the relative con-
sumption of the various solutions. With water and the low
concentrations of agents, all fluid tended to be drunk in a
limited period following its presentation each day, leaving
the subjects without fluid for a substantial fraction of each
day, while with the higher concentrations, fluid was taken
more slowly through most of the 24 hours. It may be that
periods without drinking fluid are deleterious to pregnancy.
(2) Mean litter size was not affected by caffeine (Table 8)
in agreement with other reports [7].
(3) Two pups out of 30 on 100 mg/kg/day caffeine failed to
survive to 21 days while only 1 of the 180 in all the other
groups died in the first 21 days. This difference has less than
a 1 in 10 chance of occurring as a sampling error.
(4) At 60 days, there was no difference in mean body
weights between the groups (Table 9).
-------
124
DEWS AND WENGER
TABLE 8
MEAN LITTER SIZE
= log 0.01
log 0.99999
= 460, 515
m
SD
ad lib
RH20
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
8.8
11.6
10.1
9.8
9.6
10.0
7.8
11.4
2.4
1.0
2.1
2.3
2.6
1.6
2.8
0.9
TABLE 9
WEIGHTS AT 60 DAYS (g)
M
M&F
ad lib
RH20
10 CAP
30 CAP
100 CAP
0.3 AT
3 AT
30 AT
27.6
25.3
29.6
27.7
30.1
28.4
28.1
27.4
23.1
21.8
23.9
20.4
22.1
21.8
21.2
22.3
25.4
23.6
26.8
24.0
26.2
25.1
24.6
24.8
Comments
Caffeine at 100 mg/kg/day slightly reduced the viability of
fetuses and infants. No behavioral effects were detected at
exposures of less than 100 mg/kg/day. Thus it was only at an
exposure that reduced viability that behavioral changes were
detected. Is this assurance that caffeine is not a behavioral
teratogen at doses less than 100 mg/kg/day? In itself it is very
little assurance. If the incidence of detectable behavioral
terata were as high as 1% there is only a 20% chance that an
instance would occur in a sample of 20 mice. (With incidence
of 1% of terata there is a probability of 0.99 that a subject will
be normal. The probability of all of a sample of 20 being
normal is (0.99)20=0.818, so there is only about a 20% chance
of one or more abnormalities occurring in the sample of 20.)
To have a less than 1% chance of missing an incidence as
horrendously high as 1% would require over 450 subjects.
(For a 99% chance of having at least one abnormality in the
sample, the sample size n would have to be such that. .
(0.99)n=0.01
. or n log 0.99=log 0.01
n - log 0.01 = -2.00
log 0.99 -0.00436
=458
For similar assurance of detecting an abnormality with the
still unacceptably high incidence of 1 in 100,000 would re-
quire nearly half a million subjects. (If the incidence is 1 in
100,000 the corresponding n is given by. . .
Further, because of the high variability encountered in
assessment, even an abnormality that occurred in most sub-
jects would have to be large to be detected. With a coeffi-
cient of variation of 0.3, achange of 20% would be necessary
to attain statistical significance on 20 subjects. (With a coef-
ficient of variation of 0.3, a change averaging 20% in 20 sub-
jects would give a t value of 0.2 + (0.3/^20)=2.98. The
p<0.01 value oft for 19 d.f. is 2.86). As some of the variance
is not due to random sampling errors [2,3], it is likely that
less than, say, a 10% change could not be detected, no matter
how many subjects were studied. A decline of 10% in any
productive behavioral function is quite unacceptable. Final-
ly, only the possibility of detecting change has been consid-
ered. Changes in themselves do not prove serious toxicity as
they may be transient or even sought after, as with
therapeutic agents, so more testing would be indicated when
changes were detected.
Some of these limitations arise from the exigencies of
behavioral teratology where after long and laborious prep-
aration, each subject can only contribute one datum. In be-
havioral pharmacology in which each subject can function as
its own control in split sessions on each day, coefficients of
variation of mice under FR schedules of less than 0.05 have
been attained. But there will always be variability in test
measures, not all of it due to sampling errors, and infrequent
occurrences or small changes will remain undetectable
whether the changes are in behavioral functions or in physi-
ological functions that require time to measure. Consider, for
example, how hard it would be to detect a long term decline
of 10% in cardiac output, ventilatory capacity or red cell
count and to attribute the change to a specific agent.
It should be remarked parenthetically that assurance of
the safety of caffeine does not depend on studies such as the
one described but rather on the vast amount of scientific
work, laboratory as well as epidemiological, that has been
performed on caffeine attesting to its harmlessness as used
by sane human beings.
But what about new agents that have been studied little or
not at all? What follows addresses the problems faced by the
Office of Toxic Substances (O.T.S.) in meeting their man-
dated concern for behavioral testing. Many of the general
points apply to testing for other types of toxicity and for
other classes of agents, such as those regulated by the Food
and Drug Administration.
First, safety-testing, in seeking to prove the Null Hypoth-
esis that an agent causes no harmful effects, starts from a
logically unsound position, and the sooner we educate the
people and the Congress that they must accept this unpalat-
able truth, the better it will be for consumers, producers, and
testers. Risks can be assessed but it is fundamental that to
assess a risk one must have an effect, and further, for quan-
titative risk assessment not only an effect but a dose-
dependent effect is needed so that the slope of the dose-
dependent curve can be estimated. Such information may be
inadequate for realistic assessment of risks of very low expo-
sures such as human populations might receive, but it is
surely necessary as a starting point.
Second, while serious toxic effects in 1 in 100,000 or even
1 in 1,000,000 of exposed people or toxic declines in func-
tions of even a few percent are unacceptable, it is hard in
-------
BEHAVIORAL EFFECTS OF AGENTS
125
blind testing to establish in animals incidences as high as 1 in
100 or declines as great as 10%. By "blind" testing is meant
testing when there is no indication of what to look for in the
way of effects and therefore the attempt is made to detect
any and every deviance that might occur. The problem of
extrapolation to low incidence has been discussed most in
regard to carcinogens; paradoxically, not because it is more
formidable with respect to carcinogens but rather because it
appears to be so much easier that some have been deluded
into believing it possible. Cancer is able to be diagnosed with
greater certainty than almost anything else, except death,
and it persists for confirmation and can be documented per-
manently on a microscope slide with almost zero variance.
Compare that with a behavioral or physiological assessment.
Even a grossly deviant behavioral or physiological finding in
a single mouse carries little conviction of significance. An
intercurrent infection unrelated to the agent can cause such
deviance. What we need for conviction are consistent findings
in a number of mice systematically related to load of agent in
the form of a dose-effect relationship. The exploration of
such a relationship even in the 10~2 range involves im-
possibly large numbers of subjects. The same applies to be-
havioral changes less than about 10% (10"1) even if they occur
in most subjects. If a dose-effect curve is sigmoid, and,
properly plotted, essentially all of them are, then in seeking a
minimum effective dose we are working with a part of the
curve where the slope approaches zero as an asymptote,
where a large difference in dose makes almost no difference
to the effect and so, conversely, a small error in measure-
ment of the effect makes an enormous difference to the esti-
mate of minimum effective dose. If we are attempting to
determine the dose that kills less than 1 per 1000 mice and
one love-lorn mouse dies of a broken heart or because an
animal handler dropped a cage lid on him or her yesterday,
then that little mishap vitiates results on thousands of mice,
or worse, leads us to a wildly wrong answer.
The general approach to safety-testing seems to be to try
to administer an agent to a reasonably large number of rats
for as long as reasonably possible, 18 months to 2 years, or
even multigeneration, in as high a dosage as feasible, the
maximum that is tolerated without debility (M.T.D.). A rea-
sonably large number is frequently hundreds. The tests are
geared to measurements of body weight, hematology (be-
cause blood is easily accessible) and to organ weights and
histology post mortem. It is pathology oriented rather than
pharmacology oriented.
Safety-testing has another feature. In safety-testing, all
the immediate pressures are not to find anything. The manu-
facturer of an agent wants the testing to show that it is safe,
that is, for the studies to find nothing. Positive findings can
really be quite traumatic to the discoverer of them. As most
people most of the time are successfully finding nothing, the
unfortunate who makes a discovery is likely to have his
company besmirched on evening TV and in the morning
paper. Regulatory agencies also have some reason to prefer
negative findings. Positive findings generate all sorts of polit-
ical pressures and require hard decisions. Life is simpler and
pleasanter if nothing is found. Regulatory agencies try to
protect themselves by insisting on MTD's and GLP's but
both bring their own train of problems. The regulatory agen-
cies are starting to build their own in-house research groups,
who are already invaluable as eagerly-looking for positive
effects, but the groups will remain small in relation to the
size of the tasks for the foreseeable future. Very few
academic scientists work in the field. Consumer activists do
not have the money for testing and in any case they are sure
they know the answer: that nothing is safe. No one is being
accused of dishonesty nor even lack of dedication. Indeed,
most negative findings reflect that there are no effects of the
agents as tested and assessed so the results are replicable
and correct as presented. Doing test after test with negative
results is not, however, the best way to maintain alert inter-
est to recognize the occasional test when something unex-
pected happens. Further, the design, conduct and interpre-
tation of a test may be subtly altered when the best result is
no result, especially when we are dealing with low risks.
Schneiderman, Mantel and Brown [8] say: ". . . laboratory
errors, considered as noise, could tend to overwhelm any
signal present in the data, making it still more difficult to
differentiate between treated and control animals. The
statistical remedy is to enlarge the experiment by yet another
large factor increasing the risk of further blunders that could
wipe out the gains of the increased experiment size. In this
kind of experiment if one wanted to show 'safety' of a mate-
rial, quite possibly pressures would exist to do a poor job.
The more errors made, the less likely it is to show a differ-
ence between treated and controls." Conscientious clini-
cians have accepted the necessity for so-called "double-
blind" designs in the assessment of therapeutic procedures. I
am not advocating "double-blind" procedures for safety
testing except perhaps in a few special cases. The example of
clinicians is used to indicate that one does not have to accept
that one's honor is impugned to recognize the possibility of
bias.
Routine testing is inevitably directed toward preventing
previous disasters from happening again. Nothing could be
more destructive of imaginative approaches to devising
means of anticipating and preventing future disasters than
massive routine testing programs with preoccupation with
trivial procedural details imposed by GLP leaving no time,
inclination, or incentive to think. Routine blind subacute and
chronic toxicity testing of the limitless series of agents to
which we are exposed cannot provide the information
needed and is enormously expensive. It should be a last
resort.
The law of diminishing returns seem to operate with un-
usual force in toxicology. One can get a rough estimate of
LD50 with as few as a dozen or so mice (an assessment that
would probably have been sufficient to prevent the Elixir or
Sulfanilamide tragedy), and a similar number could indicate
where the dose-effect curve lay for a behavioral effect such
as change in spontaneous motor activity (SMA). A very large
number of agents could be studied in these two simple as-
says. To add a clinical assessment of the mice would in-
crease the cost manyfold. It remains to be shown that a
clinical examination can provide reliable evidence of abnor-
mality in mice that show normal SMA, or that a far more
elaborate screen provides more than a meagre increase in
information at a disproportionately large increase in demand
on resources. Evidence that an agent accumulates and there-
fore kills when given subacutely in a fraction of the acute
LD50 can be obtained on a very few mice. Again to get much
more information involves an enormous increase in invest-
ment in subacute and chronic testing. A much better use of
resources than massive blind safety testing is to study the
pharmacology of agents that do have effects. If we know
what an agent does and how it does it we are in a much better
position to determine whether it is affecting exposed hu-
mans. Agents showing no selective pharmacological effects
at any dose may be thereby eligible for low priority in further
-------
126
DEWS AND WENGER
testing. An obvious exception to this rule is when large num-
bers of people are exposed to substantial loads of a sub-
stance, especially when the exposure is involuntary. For
such an agent intensive safety testing must be performed
even if it has to be blind. As effects of more and more agents
are described, each new description will occur in a context
that will permit it to be seen in perspective and will avoid the
furor that currently follows each discovery of toxic effects of
a familiar chemical. In a word, any effect will not necessarily
be news. Decisions on what actions to take can then be made
in a cooler and more rational way. We are seeing perspective
develop in the field of carcinogens as a result of the large
amount of information being generated by the National
Cancer Institute. As more and more substances are found to
have some carcinogenic activity, objective appraisal of con-
sequences becomes easier. The National Cancer Institute's
results are probably more likely to kill the Delaney Amend-
ment than the Delaney Amendment is to kill saccharine.
The appropriate strategy of OTS would therefore be:
(1) to obtain information on LD50 and on a simple behav-
ioral test (e.g., SMA) in a few mice on as many of the agents to
which man will be exposed as possible. It is better to make a
few observations of this type on a large number of sub-
stances than more extensive observation on a few. Agents
with potential hazard in affecting behavior should show a
large difference between behavior-affecting doses and lethal
doses. Such agents should be referred for more extensive
behavioral assessment.
(2) to encourage pharmacological research on potentially
toxic agents. Such research should include, in particular, the
continued development of more discriminating behavioral
tests as illustrated in several papers in this symposium for
secondary and tertiary testing and elucidation of mech-
anisms.
(3) to develop an early-warning reporting system. We
should recognize frankly that no laboratory testing system
can assure safety even at much less stringent levels than
anyone would find acceptable. Therefore, damage must be
minimized when a toxic agent is loosed. In retrospect, the
number of cases of phocomelia following thalidomide could
have been drastically reduced by a relatively simple report-
ing system that is well within our capabilities.
It is recognized that giving huge doses by unnatural routes
may produce effects that need further study before interpre-
tation, but one can work from real effects toward real in-
terpretations while if one is forever groping in the muddy
noise of no effect, no real studies are possible. From real
effects one can also work toward mechanisms. It is only
through understanding mechanisms that we shall be able to
have confident predictions of safety. There are too many
chemicals to be tested exhaustively individually: we have got
to learn enough about molecular toxicology to be able to
identify from their chemistry which agents require individual
study. Progress in predictive ability is already encouraging
[1].
It will be noticed that nothing has been said about the
development of short-term in vitro tests for behavioral tox-
icity, comparable to the mutagenic assays for carcinogenic-
ity. That is because there is little information at pres-
ent. Insofar as behavior is an emergent function of whole
organisms the approach is perhaps less propitious than for
other types of function, but the possibility should certainly
not be rejected before extensive research on the effects of
agents on the behavior of a wide variety of species has been
studied, and it is entirely possible that, for example, species
of Arthropoda or Mollusca will provide useful screens for
particular types of effects. For neurological toxicity, in vitro
assays on cultured neural tissue deserves the interest it is
receiving, but, of course, lack of discernable effect of an
agent on neural tissue gives no assurance of its innocuous-
ness to behavioral functions.
The Role of Epidemiology
Epidemiology can give us direct fixes on risks at far lower
levels than can be attained in the laboratory. Through the
study of human populations we may be able to get informa-
tion on the characteristics of the dose-effect curve at very
low levels of effect.
What we need is an agent with a number of special prop-
erties. (1) Widely distributed in the human population and at
a considerable range of loads, say 100 to 1 or 1000 to 1
differences between individuals. (2) Causing a unique, un-
equivocal and measurable effect in man and other animal
species. (3) A single chemical entity, either not metabolized
or with a single metabolic path with inert intermediates and
similar in all species. (This requirement rules out cigarettes,
which have many desirable features for epidemiological
studies.) (4) Easily assayed and having kinetic features such
that measurements on urine or on salive faithfully reflect the
pharmacologically active load.
If we had such an agent, epidemiological estimates of its
effects in the human population should be compared with the
dose-effect curve of the agent in two or more species brac-
keting human sensitivity: that is, in species at least one of
which is more sensitive than humans and at least one less
sensitive. We can then study the rules for "interpolation
from animals to other animals such as man," instead of de-
bating "extrapolation from animals to man" like pre-
Darwinian mystics.
Now, of course, at the end of all this labor, all we would
have is information on the probable shape of the low end of
the dose-effect curve in man and its relation to the middle of
the curve in other species for one agent, and perhaps not
even an important agent as it will have been chosen for other
reasons than its lexicological importance. Is it worth it?
Consider the astronomical analogy. It was not until 1837 that
the distance to any star was measured. The star in question
had very special properties (only one of which was relative
nearness) that permitted its distance to be measured. In the
ensuing few years, the distances of a few other special stars
were measured, then, almost suddenly, when the scale of the
universe had been established, other means of determining
distances of stars were invented and now the distance of
everything is measured almost as a matter of routine.
Perhaps we can learn to do the same for astronomically small
risks in toxicology.
To return to the laboratory testing of behavioral toxicity
the immediate task is to generate more quantitative informa-
tion on dose-effect curves. Until we have a reasonable body
of comparable information we cannot hope to develop gen-
eral principles of optimum testing, which leads to the last
point. The word comparable means being able to be brought
together and compared for similarities and differences, the
similarities cumulating to a coherent body of information,
the differences being further studied. As in most new fields
without established practices, the handful of workers in be-
havioral toxicology have tended to go all their own ways,
choosing their agents and methods according to their indi-
vidual nature and nurture and education and prejudices and
-------
BEHAVIORAL EFFECTS OF AGENTS
127
funding. The result is that while many different approaches
are probed, very little comparable information is accumu-
lated. At this point, as has happened before in new disciplin-
es, people start calling for standardization of methods so
more comparable information can be collected, a laudable
goal. Again the experience of pharmacology is useful. At-
tempts to standardize methods have generally not been fruit-
ful, except in rather general terms. Assays in pharmacology
are invariably performed with respect to a standard agent
rather than by a standard method. There were international
standards for drugs such as insulin before active insulin
could be reliably purified chemically, the standard being a
stable powder with constant insulin activity. Toxicology
should follow the same principle. We should standardize
with respect to agents. If we agree on agents and loads, then
let us each go about assays with all our ingenuity—"let a
hundred flowers blossom, let a hundred schools of thought
contend" (Mao, 1956)—in the assurance that results will be
comparable because of the invariance of agent and load.
Thus methods can be compared, better ones selected and
improved, poorer ones discarded, all by reference to a con-
stant objective standard.
In conclusion, work establishing dose-effect curves
should concentrate on the measurable part of the curve from
about 0.1 to 0.9, attempts at direct extrapolation to very low
risk should be abandoned as fundamentally unsound and
strenuous attempts should be made to link with epidemiol-
ogy. Many of the points apply to areas of toxicology addi-
tional to behavioral toxicology.
We are not ready to recommend batteries of methods and
standards in behavioral toxicology to regulating agencies.
Indeed, detailed regulations at present would generally be
counter-productive. The requirements would be chosen with
a high degree of arbitrariness. Most regulatees would happily
do the tests, stop thinking or working on the problems, and
still more funds would be spent in pursuit of nothing, and ten
years from now we would simply know more of what does
not happen. With funds devoted to the pursuit of good sci-
ence, we shall have increasing knowledge with, surely, con-
sequent increase in ability to promote safety. If it is deemed
necessary to require some behavioral testing, the require-
ment should not go beyond the study of a few mice in an
objective test such as SMA until we know more. Resources
should be dedicated to research, not to routine blind testing.
ACKNOWLEDGEMENTS
The work described was supported by grants from the USPHS
(OH-00706 and MH-02094) and by the International Life Sciences
Institute. We would like to thank J. J. Holland for making the appa-
ratus, L. King for help with the experiments, J. Katz for statistical
checks, L. Levy for the MS and W. H. Morse for suggestions.
REFERENCES
1. Cramer, G. M. and R. A. Ford. Estimation of toxic hazard—a
decision tree approach. Food and Cosmetic Toxic. 16: 225-276,
1978.
2. Dews, P. B. Epistemology of screening for behavioral toxicity.
Environ, tilth Perspect. 26: 37-42, 1978.
3. Dews, P. B. and J. Berkson. On the error of bio-assay with
quantal response. In: Statistics and Mathematics in Biology,
edited by O. Kempthorne, T. A. Bancroft, J. W. Gowen and J.
L. Lush. Ames, Iowa: Iowa State College Press, 1954, Chapter
27, pp. 361-370.
4. Dews, P. B. and J. DeWeese. Schedules of reinforcement. In:
Handbook of Psychopharmacology, edited by L. L. Iversen, S.
D. Iverson and S. H. Snyder. New York: Plenum Press, 1977,
Vol. 7, pp. 107-150.
5. Ekel, G. J. and W. H. Teichner. An analysis and critique of
behavioral toxicology in the USSR. Washington, D.C.: Depart-
ment of Health, Education and Welfare, 1976.
6. Levine, T. E. Effects of carbon disulfide and FLA-63 on operant
behavior in pigeons. J. Pharmac, exp. ther. 199: 669-678, 1976.
7. Palm, P. E., E. P. Arnold, P. C. Rachwall, J. C. Leyczek, K. W.
Teague and C. Kensler. Evaluation of teratogenic potential of
fresh-brewed coffee and caffeine in the rat. Toxic, appl. Phar-
mac. 44: 1-16, 1978.
8. Schneiderman, M. A., N. Mantel and C. C. Brown. From mouse
to man—or how to get from the laboratory to Park Avenue and
59th Street. Ann. N.Y. Acad. Sci. 246: 237-250, 1975.
9. Wenger, G. R. and P. B. Dews. The effects of phencyclidine,
ketamine, rf-amphetamine and pentobarbital on schedule-
controlled behavior in the mouse. J. Pharmac. exp. Ther. 196:
616-624, 1976.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 129-135. ANKHO International Inc., 1979.
Some Problems in Interpreting the Behavioral
Effects of Lead and Methylmercury1
VICTOR G. LATIES AND DEBORAH A. CORY-SLECHTA2
Environmental Health Sciences Center
and
Department of Radiation Biology and Biophysics
School of Medicine and Dentistry
University of Rochester
Rochester NY 14642
LATIES, V. G. AND D. A. CORY-SLECHTA. Some problems in interpreting the behavioral effects of lead and methylmer-
cury. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 129-135, 1979.—Two sets of observations are reported as illustrations of
problems encountered in behavioral toxicology? First, in an attempt to determine the contribution of methylmercury-
induced ataxia to behavioral changes observed on the fixed-consecutive-number schedule, some ancillary control experi-
ments were undertaken. Neither pharmacologically-produced incoordination (ethanol) nor mechanically-produced incoor-
dination (foot taping) led to behavioral changes similar to those seen after exposure to methylmercury. Second, total crop
impaction in a pigeon that died during a behavioral experiment on lead suggested some further work. Lead-induced crop
stasis in pigeons was measured by x-raying the passage of force-fed stainless steel ball bearings through the crop. This
retardation of motility reliably preceded signs of overt toxicity. These results suggest that the behavioral changes in the
pigeon noted by us and reported by other investigators cannot be attributed to CNS dysfunction alone, but more likely arise
from starvation, or from combined CNS damage and starvation. In addition, these results demonstrate that the appearance
of behavioral effects prior to overt toxicity does not necessarily reflect CNS damage.
Methylmercury
Behavioral toxicology
Lead Pigeon
Crop dysfunction
Fixed consecutive number schedule
Behavior Urecholine
Ethanol
Crop stasis
THIS article will describe two sets of observations that illus-
trate problems that can be encountered by people working in
behavioral toxicology. The first concerns one of the side
effects of methylmercury in the pigeon, an effect that com-
plicates interpretation of changes in schedule-controlled be-
havior. The second concerns a serendipitous discovery of
one confounding factor that arises during exposure of pi-
geons to lead.
METHYLMERCURY ATAXIA AND PERFORMANCE ON THE FIXED
CONSECUTIVE NUMBER SCHEDULE
A pigeon can be trained to peck on one key a certain
number of times and then move over to a second key, peck
once, and get food. When the requirement on the first key is
specified as no fewer than 8 nor more than 9 pecks, the bird
will, indeed, learn to do what is required. Sequences of, for
example, 7 or 10 followed by a peck on the second key would
not be reinforced, and would lead to a resetting of the re-
sponse requirement. The bird would then have to start at
zero in making its next series of responses. No external
stimulus change informs the bird that the requirement has
been met, making it a tandem rather than a chain schedule
[5]. The sequence of pecks on the first key before the switch
to the second is called a run. Figure 1 (top left) shows that
run lengths of 8 and 9 come to predominate when only those
runs lead to reinforcement for the peck on the second key.
Figure 1 (bottom left) shows that this particular bird's run
length distribution was changed by exposure to methylmer-
cury. Details of exposure are not important for our purposes
and are given briefly in the figure legend and in more detail
elsewhere (Laties, V. G. and H. L. Evans, in preparation).
Suffice it to say here that these are effects of a chronic expo-
sure regimen in which the substance is given in such an
amount as to produce a marked effect on behavior in about
one or two months, the hope being that we then would be
able to follow recovery from this exposure (Fig. 1, top right).
The point in the present context concerns the way the run
length distribution for this bird shifts to the left, with shorter
runs now being emitted before the switch to the second key.
About the time the bird started showing these changes in
schedule-controlled behavior, it also started showing some
'This work was supported in part by grants ES-01247 and ES-01248 from the National Institute of Environmental Health Sciences and in
part by a contract with the U.S. Department of Energy at the University of Rochester Department of Radiation Biology and Biophysics and
has been assigned Report No. UR-3490-1611.
"Currently supported by a Jr. Staff Fellowship from the National Center for Toxicological Research, Jefferson, Arkansas.
129
-------
130
LATIES AND CORY-SLECHTA
BIRD 12
METHYLMERCURY
PRE
AFTER RECOVERY
FOOT TAPED
20 28 I
RUN LENGTH
10
20
28
FIG. I. Performance of a pigeon on the fixed consecutive number schedule of reinforcement. The bird was given 2 mg/kg
methylmercury daily Monday through Friday by intubation into the crop several hours after the sessions. Run lengths of 8 and 9
have been filled in. For further details, see Laties, V. G. and H. L. Evans (in preparation).
unsteadiness on its feet and could no longer fly. In order to
tease out the effects of the ataxia upon schedule perform-
ance, we performed some control experiments to see what
ataxia itself could do to this performance.
One manner in which this was accomplished involved
binding the large toe of the pigeon back against the shank of
the foot, thereby producing a physically, rather than
chemically-induced ataxia. This led to the type of perform-
ance change shown in Fig. 1, bottom right. Note that the run
lengths became longer on the average, an effect opposite to
that seen with methylmercury.
More complete data for another bird are shown in Fig. 2,
which illustrates how run length was increased greatly during
the first foot-taping session but came down somewhat as the
pigeon learned how to get around even with one large toe
strapped against its leg. Variability was also increased. A
measure of the rate during runs is shown at the bottom of the
figure. Note that it was not changed by hobbling, even during
the first session. In contrast, this rate was invariably reduced
by methylmercury (Laties, V. G. and H. L. Evans, in prep-
aration).
We attempted to produce an unsteady gait in another
fashion by giving the pigeons ethyl alcohol. The results are
shown in Fig. 3 for the three birds studied. Again, the domi-
nant effect of the ethanol, which did indeed produce some
unsteadiness in the pigeons, was to increase the number of
responses the birds would make on the first key before
switching over to the second one and also to increase run
TABLE 1
SUMMARY OF THE RESULTS OF METHYLMERCURY, FOOT
TAPING AND ETHANOL ON THE FCN SCHEDULE
Methylmercury Foot Taping Ethanol
% Reinforced
Rate during runs
(R/sec)
Mean run length
(resp)
SD of run length
(resp)
J.
4.
t
= =
t t
t t
length variability. Running rates were unchanged. The re-
sults with these two control procedures are summarized in
Table 1.
We would tentatively conclude from these control proce-
dures that the ataxia-producing effects of methylmercury
cannot wholly account for the other behavioral effects that
we have seen.
-------
PROBLEMS IN INTERPRETING BEHAVIORAL EFFECTS
131
12-
10-
8-
^
^ "
-------
132
LATIES AND CORY-SLECHTA
BIRD 11
ETHfiNOL
BIRD 12 BIRD 24
>-
z
u
13
o
Ld
cr
SO-i
25-
0-J
SO-i
CONTROL
0-5 G/KG
25-i
Q-1
MM
10
i i r
20 28 1
10
I I I
20 28 1
1.0 G/KG
I
10
20 28
RUN LENGTH
FIG. 3. Effects of ethyl alcohol on performance on the fixed consecutive number schedule. Dosing via
intubation occurred 10 min. before the session. The effects of the prior treatment with methylmercury
(completed more than a year before) are still evident in the run length distributions of birds 11 and 24.
For further details, see Laties, V. G. and H. L. Evans (in preparation).
FIG. 4. Photographs depicting the phenomenon of lead-induced crop stasis in the pigeon. Gross necropsy revealed severely
distended crops (left) totally compacted with grit and food particles (middle). Although the remaining digestive tract appeared
normal (right), no food particles were found below the level of the crop.
-------
1 PROBLEMS IN INTERPRETING BEHAVIORAL EFFECTS
133
FIG. 5. Demonstration of the determination of crop passage time by the hall bearing test. Birds were force-fed a 2 mm stainless steel
ball bearing and radiographs were made 15 min later (0 hr test) to ensure its position in the crop (left). Subsequent radiographs were
made 24 hr later (24 hr test) to track passage out of the crop. In this case, as in all control conditions, the ball bearing had reliably
passed from the crop to the gizzard within 24 hr (right).
FIG. 6. Demonstration of increased crop passage time as determined by the ball bearing test. Radiographs revealed (he
stainless steel ball bearing to be located in the crop at 0 hr (left) and to have remained there 24 hr later (right). Subsequent
radiographs revealed it to remain immobile as long as 10 days afterwards, when the bird was sacrificed.
-------
134
LATIES AND CORY-SLECHTA
co
co
a
o
cc
o
u_
o
o
§
80 H
60 4
40 H
20 H
0J
1000
3L
LEAD ACETATE CONCENTRATION (ppm)
3000
FIG. 7. Time (in days) post-lead exposure to onset of crop stasis as a
function of lead acetate exposure concentration. Crop dysfunction
was determined by the ball bearing test. Two birds exposed to 0 ppm
lead acetate drinking solutions for over 100 days did not show crop
stasis and are represented by open circles. The two birds exposed to
1000 ppm showed crop stasis at 54 and 99 days, while both 3000 ppm
exposed animals displayed the effect in 10 days.
days. These findings extend the generality of lead-induced
crop dysfunction and suggest the phenomenon is not an ar-
tifact of crop or proventricular intubation. Additionally, in
each case, increases in crop passage time preceded signs of
overt toxicity by at least a week (but could have been more
since the x-rays were taken only twice weekly), making it
both a reliable and sensitive measure. Once the ball bearing
stayed in the crop, subsequent x-rays showed it to remain
immobilized as long as 10 days afterwards, suggesting that
the crop stasis may be an all-or-none phenomenon. To re-
peat, changes in crop passage time (and thus functional food
deprivation levels) were found to precede signs of overt tox-
icity. It would therefore be a mistake to conclude that be-
havioral changes seen in the birds prior to overt toxicity are
necessarily attributable to CNS effects.
The results of these crop intubations and the
proventricular intubation exposures of Barthalmus et al. [1]
are summarized in Fig. 8. We have not included the results of
Dietz et al. [4] here due to an apparent inconsistency be-
tween the dosing regimen and blood lead levels. These data
suggest that proventricular intubation produces overt toxic-
ity at lower concentrations than does crop intubation. Al-
though a dose-effect function is apparent, differences in in-
dividual susceptibility are also apparent.
Only one lead-exposed pigeon from this investigation
failed to demonstrate overt toxicity. Blood lead analysis
showed its levels to fall consistently below 200 /xg/dl. Most of
C
>501gl
to 200
to 55.90, 100
to 72
to 72
to 55, 57, 60, 70n
to 2001
fc
O 40 H
30 H
20 H
§ KH
0J
r
to 52
CORY-SLECHTA ET AL.
CROP INTUBATION
BARTHALMUS ET AL.
PROVENTRICULAR INTUBATION
CONTROL
5 10 25
DOSE OF LEAD ACETATE (mg/kg/day)
50
100
FIG. 8. Days post-lead exposure to onset of overt toxicity as a function of lead acetate exposure dose. Circles indicate birds
exposed via crop intubation by Cory-Slechta et al. (in press) while triangles represent birds exposed via proventricular intubation
by Barthalamus et al. [1] Arrows and numbers above symbols represent pigeons who never showed overt toxicity
through the number of exposure days indicated. These data suggest that proventricular intubation produces overt toxicity at lower
concentrations than does crop intubation. While a dose-effect relationship is suggested, large individual differences in susceptibil-
ity are also apparent.
-------
PROBLEMS IN INTERPRETING BEHAVIORAL EFFECTS
135
the birds exposed by Barthalmus et al., and all of those
exposed by Dietz et al. showed blood levels well above 200
jug/dl. However, because of differences in concentration and
duration of exposures, route of administration and method of
blood lead analysis, further research is needed to delineate
more clearly the parameters of lead exposure that result in
crop stasis. But, to reiterate, these observations, as well as
those of Barthalmus et al. and Dietz et al., certainly suggest
that the behavioral changes noted by them and by us may
have arisen not from lead-induced CNS changes, but from
starvation or, at best, some unknown combination of starva-
tion and lead.
No histopathological explanation for the crop stasis
emerged from examination of H & E stained tissue (Cory-
Slechta, D. A., R. H. Garman and D. S. Seidman, in
press). This raises the possibility of a subcellular mech-
anism, e.g., a decrease in the release of acetylcholine at the
neuromuscular junction of the crop sphincter. One severely
affected bird was serially treated with urecholine (a parasym-
pathomimetic known to stimulate mainly the GI tract).
(Urecholine (bethanechol chloride) from Merck, Sharp and
Dohme, West Point, Pa.) This treatment totally cleared the
bird's crop, such that within one week, ball bearing passage
time through the crop had returned to normal. No lead-
exposed bird who was not so treated ever showed any such
recovery during the time period we observed them. If one
mechanism of lead toxicity does involve disturbances in
cholinergic transmission, (e.g., [7,12]) the pigeon might pro-
vide a good model for exploring this possibility.
Effects of lead on the digestive system are certainly not
unique to avian species. Domestic animals with lead poison-
ing exhibit various degrees of derangement of the GI tract.
Young calves show anorexia and colic, while in sheep the
syndrome reportedly consists of anorexia, abdominal pain
and, often, diarrhea. At some time during the course of
poisoning, approximately 87% of dogs show GI signs consist-
ing of emesis, colic, diarrhea and anorexia [9]. Exposure of
lactating rats dams to relatively high concentrations of lead
decreased food consumption and necessitated the use of
pair-fed control rats [10]. In human adults, the syndrome of
acute abdominal colic due to lead poisoning consists of con-
stipation, followed by attacks of crampy diffuse abdominal
pain. When pain and constipation are severe, there may be
vomiting and anorexia with associated weight loss. In chil-
dren, the syndrome is similar: colic, anorexia, episodic
vomiting and constipation [9].
Whether the effects on the digestive system of these
different species involve similar mechanisms is as yet un-
clear. According to Casarett and Doull [2], these gastroin-
testinal signs of intoxication are probably related to
peripheral neuropathy rather than a direct effect on the in-
testinal mucosa. The crop impaction of the pigeon reported
here thus seems to be among a class of effects of lead on the
GI tract, rather than a species specific phenomenon. Such GI
changes may well be associated with changes in feeding be-
havior. The implications so engendered have been largely
ignored in studies assessing the effects of lead on food-
reinforced behavior.
SUMMARY
Both examples discussed here illustrate the need for great
care in interpreting data in behavioral toxicology. Suitable
control observations will always be required when the ac-
tions of a chemical are complex.
REFERENCES
1. Barthalmus, G. T., J. D. Leander, D. E. McMillan, P. Mushak
and M. R. Krigman. Chronic effects of lead on schedule-
controlled pigeon behavior. Toxic appl. Pharmac. 42: 271-284,
1977.
2. Casarett, L. J. and J. Doull. Toxicology: The Basic Science of
Poisons. New York: New York, McMillian, 1975, p. 479.
3. Cory-Slechta, D. A. and T. Thompson. Behavioral toxicity of
chronic postweaning lead exposure in the rat. Toxic, appl.
Pharmac. 47: 151-159, 1979.
4. Dietz, D. D., D. E. McMillan and P. Mushak. Effects of chronic
lead administration on acquisition and performance of serial
position sequences by pigeons. Toxic. Appl. Pharmac. 47: 377-
384, 1979.
5. Ferster, C. D. and B. F. Skinner. Schedules of Reinforcement.
New York: Appleton-Century Crofts, 1957, p. 415.
6. Gollub, L. R. and J. T. Urban. The accentuation of a rate differ-
ence during extinction. J. exp. analysis Behav. 1: 365-369, 1958.
7. Kostial, K. and B. Vouk. Lead ions and synaptic transmission
in the superior cervical ganglion of the cat. Br. J. Pharmac. 12:
219-222, 1957.
8. Laties, V. G. The modification of drug effects on behavior by
external discriminative stimuli. J. Pharmac. exp. Ther. 183:
1-13, 1972.
9. National Academy of Sciences. Lead: Airborne Lead in Per-
spective. Report of the committee on Biologic Effects of
Atomspheric Pollutants. Washington, D. C., 1972, p. 86 182-
184.
10. Sauerhoff, M. W. and I. A. Michaelson. Hyperactivity and
brain catecholamines in lead-exposed developing rats. Science
182: 1022-1024, 1973.
11. Slechta, D. A. The effects of chronic lead acetate administration
on fixed interval performance in the rat. Unpublished doctoral
dissertation. University of Minnesota, Minneapolis, Minnesota,
1977.
12. Silbergeld, E. K., T. J. Fales and A. M. Goldberg. The effect of
inorganic lead on the neuromuscular junction. Neurophar-
macology 13: 795-801, 1974.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 137-148. ANKHO International Inc., 1979.
Screening for Neurobehavioral Toxicity:
The Need for and Examples of Validation of
Testing Procedures
HUGH A. TILSON, CLIFFORD L. MITCHELL AND PATRICK A. CABE
Laboratory of Behavioral and Neurological Toxicology
National Institute of Environmental Health Sciences, Research Triangle Park,NC 27709
TILSON, H. A., C. L. MITCHELL AND P. A. CABE. Screening for neurobehavioral toxicity: The need for and examples
of validation of testing procedures. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 137-148, 1979.—The need for a sensitive
and reliable screen to assess environmental agents for potential behavioral and neurological toxicity is discussed.
Factors involving strategy, choice of animals and doses, route of administration, duration of study and requirements for the
selection of neurobehavioral tests are also evaluated. The primary emphasis concerns the need for standardization and
validation of neurobehavioral tests to be used in neurotoxicology. It is suggested that test validation be accomplished by
comparing the observed results of known neurotoxicants in animal models which are chosen to predict effects based on
reported human symptomatology. As a means of demonstrating how test validation is used in our laboratory, data from a
number of experiments concerning the effects of a variety of chemical agents on three measures of motor functioning were
discussed. The neurobehavioral effects of acrylamide, and agent known to produce "dying-back"axonopathies, were
assessed using separate techniques presumed to measure hindlimb and forelimb functioning and general motor activity. The
prediction that acrylamide will first decrease hindlimb functioning, while decreasing forelimb grip strength and motor
activity at higher doses, was confirmed. The validity of the hindlimb measurement was supported using a neurotoxicant,
carbon disulfide, known to affect motor functioning in a manner similar to acrylamide. The validity of the forelimb
technique was shown indirectly using normative data collected from rats of both sexes tested at various ages, i.e., males
were stronger than females and grip scores changed as a function of age. The relative sensitivities of the fore- and hindlimb
measurements were found to be approximately the same when used to assess the effects of known muscle relaxants, such
as phenobarbital and chlordiazepoxide. Finally, it was predicted and confirmed that an environmental agent believed to
affect behavior secondarily to effects on other organ systems would affect all measures of motor functioning at approx-
imately the same dose.
Neurobehavioral screening procedures Factors to be considered in screening
Examples of validation of neurobehavioral tests
THE Environmental Protection Agency (EPA) currently es-
timates that there are more than 60,000 chemicals in use
commercially, while there are approximately 1,000 new
chemicals introduced into the environment each year. Such a
massive number of chemical entities poses a problem in the
implementation of the Toxic Substances Control Act of 1976
(TSCA), which states, among other things, that chemicals
should be assessed for their behavioral and neurotoxic ef-
fects. In this regard, the lack of a sensitive and reliable
screening program for behavioral and neurological toxicity
testing is a clear impediment of the effectuation of TSCA
[5,21].
FACTORS TO BE CONSIDERED IN DEVELOPING A
NEUROBEHAVIORAL SCREEN
With any screening procedure, there are several aspects
which must be taken into account. Included in these are
strategy, choice of animals, doses, route of administration,
duration of study, and requirements for the neurobehavioral
tests. These have been discussed in detail elsewhere [5,21]
and will be discussed only briefly in this paper.
Screening Strategies
In Chapter 11, Effects on Behavior, in the National
Academy of Sciences Report of 1975 [13], a strategy for
screening was presented which adopts a sequential scheme
that proceeds through a series of increasingly specific, sen-
sitive determinations, guided by the results of preceding,
more general tests. For example, after determining lethal
doses, an elementary screen for biological effects begins
with relatively simple tests capable of reflecting a broad
range of impairments, a part of which includes those related
to nervous system function. Depending on the effects ob-
served, a substance progresses through a sequence of testing
procedures leading to a final decision to accept the substance
for marketing or reject it because the risks exceed the pre-
dicted benefits. The preliminary screen discussed in the Na-
tional Academy of Sciences Report consisted of procedures
typically used in the pharmaceutical industry to screen for
central nervous system activity, such as changes in motor
activity, gross observational ratings, alterations in reflex re-
sponses, and body weight fluctuations.
In our opinion, the preliminary screening tests with which
137
-------
138
TILSON, MITCHELL AND CABE
we are familiar and which are used in the pharmaceutical
industry are often not appropriate for use in neurotoxicol-
ogy. A major defect in these procedures is that, other than
for pain, tests for sensory and cognitive functioning are al-
most totally lacking. Yet, a major, early complaint with
many environmental toxicants concerns alterations in these
neurobehavioral functions. A more complete battery of
screening tests should be developed and this issue is dis-
cussed in more detail below.
Choice of Animals
The extrapolation of animal toxicological data to man is
always tenuous, but for obvious reasons, animals are neces-
sarily used. Unfortunately, there is no single animal model in
which effects correlate perfectly with toxicity in humans.
In the preliminary screening of a large number of known
or suspected environmental toxins, there are distinct eco-
nomic factors which must be taken into account. It is also
important that there be an adequate pharmacological and
toxicological data base for the species chosen for study, such
that meaningful interpretations of effects can be made and
appropriate hypotheses about mechanisms and loci of action
can be framed. For these reasons, the mouse or rat are fre-
quently preferred in the preliminary screen.
Other variables, besides species, must be considered. Ef-
forts should be made in a preliminary screen to minimize the
variance of the data. Since the estrous cycle of female ani-
mals might introduce additional variance into the data of
some behavioral tests [28], male animals are often preferred
in the preliminary screen. Another organismal variable is the
age of the subject. Variability is generally less in mature
adult organisms than in either developing organisms or older,
senescent animals.
We are well aware that primate or other species might
provide a more sensitive assay than rodents for threshold
estimates with many neurotoxicants [15]. Likewise, we are
aware of the importance of studying effects of environmental
toxins on the young, the old, the malnourished, the sick,
females as well as males, etc. in order to determine the popu-
lation at greatest risk. The purpose of a preliminary screen is
not, however, to determine the population at greatest risk.
Rather, it is to provide an initial, tentative evaluation of the
effect of agents on behavior and the nevous system. This
can best be accomplished by keeping the experimental vari-
ables to a minimim. In effect, we are applying the Principle
of the Blunt Ax proposed by Lincoln Moses, which states that
"if the ax chopped down the tree, it was sharp enough."
Moses [10] was pointing out that, if under some circum-
stances a simple statistical test might demonstrate the relia-
bility of a difference, a complicated analysis would than be a
waste of time. If the preliminary neurobehavioral screen re-
sults in either the rejection of a substance or in "flagging" it
for further (directed) study, it was sharp enough. It is only in
those cases where untoward effects are not observerd that
the screening ax must be sharpened before approving the
substance.
Choice of Doses
The detection of cumulative toxicity following repeated
exposure to subthreshold doses is a major goal in the devel-
opment of a screening program. Thus, a multiple-dosing re-
gimen is particularly appropriate since both quantitative and
qualitative changes in the response to environmental factors
can occur on repeated exposure, or even with time following
a single exposure [5]. A subacute, multiple-dosing regimen
used routinely in screening procedures is one which spans
about one-tenth of the expected life span of the tested animal
[3]. Neurobehavioral assessments should also be made for a
time following cessation of the dosing regimen, since it is of
interest to determine the permanency of any effects noted
during the dosing phase, or to note any delayed effects fol-
lowing cessation of dosing.
Choice of Tests
Two interrelated methodological problems which con-
front behavioral toxicology as it applies to environmental
agents are the insidious onset of effects and the subjective
nature of the complaints that are associated with earlier
stages of toxicosis. Because of these problems, there is lim-
ited agreement as to the sensitivity and utility of many
commonly used neurobehavioral tests and procedures. In-
deed, Evans and Weiss [5] have stated that "few rigorous
animal models are available to substantiate the kind of
human symptoms and functional changes that occur with
low-level exposure" (to a toxic agent). According to Dews
[4] there are no methods that have successfully predicted
when prolonged exposure to a low level of an agent will lead
to subtle and delayed behavioral effects in man.
The problem of selection of tests for use in
neurobehavioral toxicology is not based on the number of
tests available, but related to the apparent lack of rational
criteria for the choice of suitable methods. Evans and Wiess
[5] refer directly to this problem in their statement "too
many studies in behavioral toxicology have been purely de-
scriptive, with neither an attack on underlying mechanisms
nor a clear extrapolation to human health questions. . . "
This, in our opinion, is the major problem with selecting a
neurobehavioral test battery. Certainly, chemists do not
want to use an instrument which has not been calibrated
properly or which is not sensitive to the agent being studied.
The situation should be no different for neurobehavioral test-
ing. Thus, the present state of development of behavioral
toxicology suggests that methods need to be standardized
and validated before they can be maximally utilized in any
neurobehavioral testing program.
Test procedures of unknown utility cannot be validated
by testing them against substances producing unknown or
controversial effects. The most systematic approach to the
validation of sensitive and reliable methodologies for
neurobehavioral toxicology is to compare compounds
known to have specific neurotoxic effects in a battery of
tests chosen to detect a wide range of possible effects and to
overlap in terms of signs evaluated. The sensitivity and
selectivity of those methods assumed to measure the same
neurobehavioral function might be determined by generating
a profile of effects that will be characteristic for each com-
pound. Test validation will be achieved by demonstrating the
similarities between techniques presumed to measure similar
functions and by distinguishing between methodologies as-
sumed to assess different processes.
In the development of the research program at the Lab-
oratory of Behavioral and Neurological Toxicology of the
National Institute of Environmental Health Sciences, the
first step taken was to identify behavioral methods and
representative toxicants to be used in the validation and
standardization process [21]. It must be noted that those
tests chosen initially will not necessarily be those used ulti-
-------
VALIDATION OF BEHAVIORAL TESTS IN TOXICOLOGY
139
TABLE 1
SYMPTOMATOLOGY REPORTED BY HUMANS EXPOSED TO NEUROTOXINS
Function Affected
Symptomatology
Sensory
Motor
Arousal
Associative (cognitive)
Physiological and
consummately responses
Anosmia
Paresthesias in feet, fingers, toes
Visual deficits, photophobia, nystagmus
Auditory deficits, tinnitus
Perceptual dysfunctions, pseudohallucinations
Weakness in hands, arms, legs, paralysis
Incoordination, dizziness
Fatigue
Tremor, convulsions
Hyperactivity
Slurred speech
Nervousness, irritability, agitation, euphoria, psychosis
Apathy, lethargy, depression, compulsive behavior
Impaired short-term memory
Impaired long-term memory
Confusion, disorientation
Disrupted sleep-awake cycles
Hypothermia, hyperthermia, sweating
Loss or stimulated appetite
Loss or gain in body weight
TABLE 2
PRIMARY TEST BATTERY
Sensory Tests
Visual Orientation/Localization
Auditory Localization
Tactile Stimulation Orientation
Pain (Tail Flick to Hot Water)
Negative Geotaxis
Motor Tests
Spontaneous Motor Activity
Forelimb Grip Strength
Hindlimb Grip Strength
Tremor
Tests for Arousal Deficits
Emergence
Startle Response (Air Puff, Auditory Cue)
Tests for Associative (Cognitive) Functions
Rapid Escape/Avoidance Conditioning Test (RE/ACT)
Retention
Passive Avoidance
Tests of Physiological and Consummatory Responses
Body Weight
Rectal Temperature
Autonomic Signs
Respiration
mately. The final test battery will evolve slowly as knowl-
edge is gained through the validation process.
An adequate neurobehavioral screen should consist of
methods which have the potential to predict human toxicosis
symptomatology. Thus, those procedures thought to meas-
ure deficits analogous to neurological manifestations re-
ported by humans exposed to neurotoxins should be consid-
ered. Table 1 contains a list of such symptoms and, as can be
seen, they may be grouped into several categories including
areas of sensory function, motor strength and coordination,
emotionality, associative/cognitive factors and several other
effects (e.g., insomnia, anorexia, hyper- and hypothermia).
Table 2 lists the tests that are currently undergoing stan-
dardization and validation in our laboratory. These methods,
which have been termed the " primary test battery," are
grouped according to the category of human toxicosis symp-
tomatology for which they are presumed to measure. All of
the procedures are simple to perform and the entire battery
takes approximately eight hours for 2 to 3 people to complete
and examination of 40 animals. In practice, this is typically
done on two successive afternoons. The tests were originally
designed for rats, but have in most cases been adapted suc-
cessfully for mice. Animals receiving chemicals subacutely,
as well as chronically, have been tested routinely in this
battery and recently the tests have been used to assess ma-
ture offspring of mothers exposed to chemicals during gesta-
tion. Similar tests for the evaluation of neurobehavioral
functioning of animals prior to weaning are now under de-
velopment.
There is obvious overlap in many of the categories. For
-------
140
TILSON, MITCHELL AND CABE
TABLE 3
REPRESENTATIVE NEUROTOXICANTS CLASSIFIED ACCORDING
TO MECHANISMS OF ACTION AND PROJECTED FOR USE IN THE
VALIDATION PROCESS
General Mechanism of Neurotoxicity
Compound
Agent that produces demyelination Triethyl Tin
Agent that produces "dying back" neuropathies Acrylamide
Agents that produce mixed central and Methyl Mercury
peripheral neuropathies Inorganic Lead
Arsenic
Agents affecting specific CNS nuclear groups Salicylates
Agent whose mechanism of neurotoxicity in Arochlor 1254
humans is not yet well defined Kepone
Pharmacological tools d-Amphetamine
instance, a deficit noted in one of the sensory tests could be
associated with a motor impairment, whereas an effect ob-
served in one of the motor tests could be due to decreased
responsiveness to all stimulation or general malaise.
Nonetheless, an examination of the profile of changes pro-
duced by a given substance should provide considerable in-
formation about the neurobehavioral system(s) upon which
to focus subsequent work.
Neurotoxicants chosen for study were selected according
to the symptomatology produced and mechanism of action in
humans. As a means of providing a framework for future
work it was decided to study representative agents from a
standard classification of neurotoxicants, such as that described
by Norton [11]. This schema is relevant since it divides toxic
agents into general categories based upon symptomatology
and mechanism of action. For purposes of comparison be-
tween humans and animals, it was decided that adequate
information concerning the absorption, distribution,
metabolism, and dwell time in the body should be available
for both humans and rodents. This severely restricted the
number of possible agents (Table 3). In addition to represen-
tative neurotoxicants, various psychoactive drugs and/or other
experimental manipulations will be used in the test validation
process.
EXAMPLES OF THE TEST VALIDATION PROCESS
The remainder of this paper will focus on the validation
process and will attempt ot demonstrate how it works in our
laboratory. For purposes of illustration, data collected from
a variety of studies concerning the motor portion or our
neurobehavioral screen will be discussed.
Spontaneously occurring motor activity, forelimb grip
strength, and hindlimb extensor responses are used to gen-
erate a profile of motor related effects, which is then eval-
uated to determine the presence or absence of specific motor
dysfunction. For example, motor activity can be altered by
many factors, central and peripheral, and changes in the fre-
quency of this behavioral measure might be due to a variety
of reasons. Thus, it may be regarded as a general or rela-
tively nonspecific, but not necessarily insensitive, indicator
of toxicity. Alterations in forelimb grip strength or hindlimb
extensor thrust at doses lower than those required to pro-
duce changes in general motor activity can indicate the pres-
ence of a selective neurotoxic effect. A change in all three
measures of motor functioning at the same dose is suggestive
of a relatively nonspecific behavioral toxic reaction.
Description of Methods
Hindlimb extensor thrust was assessed using a device re-
ported elsewhere [1]. In this test, a rat is held by the tail and
its forelimbs placed on a Plexiglas ledge located approx-
imately 18 in. above the countertop. The hindlimbs of the rat
are placed on a T-bar l'/2 in. below and 3 in. laterally from
the ledge. The T-bar is attached to a strain gauge positioned
45° to horizontal. After sitting in the specified position for
approximately 5 sec., an air puff is delivered to the rump of
the rat. The deflection on the meter of the strain gauge is
taken as a measure of hindlimb thrust. An average of the
three highest readings of five trials is taken as the hindlimb
extensor response.
Forelimb grip strength was measured using a recording
grip meter described in detail elsewhere [2]. Briefly, animals
are held by the tail and permitted to grasp a wire ring 45mm
in diameter connected to a strain gauge. The force in grams
required to pull the subject away from the wire ring is meas-
ured in three trials, not counting those in which it holds the
ring with only one forepaw or jerks the ring. An average of
three readings was taken as the grip strength score.
Spontaneous motor activity was measured in commer-
cially available activity monitors (Automex, Columbus In-
struments, Columbus, OH) placed inside a sound-and
light-attenuated outer chamber. At the beginning of the test,
rats are placed individually into one of six plastic cages (11 x
7 x 5 in.) having perforated stainless steel lids and placed on
the activity measuring device. General motor activity was
measured in darkened conditions over a 9 min period.
All behavioral testing was done on a blind basis so as to
eliminate bias in the interpretation of responses. In addition,
testing was done between the hours of 10 a.m. and 2 p.m. to
control for diurnal variations.
In all of the following studies, the data were analyzed for
overall treatment effects using analysis of variance
(ANOVA) techniques [29]. If a significant treatment effect
was observed, independent group comparisons were made
using Fisher's Least Significant Difference (LSD) test [8].
The accepted level of significance in all cases wasp<0.05,
Effects of a Known Neurotoxin on Motor Functioning
One agent chosen for our validation studies was ac-
rylamide. Repeated exposure to this chemical results in a
progressive bilateral polyneuropathy and this agent has been
used extensively to study "dying-back" polyneuropathies
[18,19]. Acrylamide initially affects distal portions of larger
diameter nerve fibers, while with more prolonged exposure,
the fiber degeneration progresses, affecting more proximal
regions. Sensorimotor functioning of the lower extremities is
characteristically affected before that of the upper ex-
tremities, while more prolonged exposure affects function to
a greater extent. Regeneration of peripheral nerve fibers and
restoration of function usually occurs following cessation of
exposure [18].
The purpose of the following experiment was to deter-
mine the relative sensitivity and selectivity of the hindlimb
extensor measure to the effects of acrylamide on motor
-------
VALIDATION OF BEHAVIORAL TESTS IN TOXICOLOGY
141
120
+J100
1
I
|
3 80
60
40
Acrylamide
O Wmg/kg
A 20mg/kg
•p
-------
142
TILSON, MITCHELL AND CABE
g%j Controls
I I Acrylamide
•p tO.05
T
1
Weeks Postdosing
FIG. 2. Average hindlimb extensor (HLER) scores ± S.E. of rats
given 20 mg/kg of acrylamide (N=5)or control treatment (N=5) for
13 weeks and randomly selected for retesting at 1 and 5 weeks after
cessation of dosing. The asterisk indicates a significant difference
between groups (r-test,p<0.05). From Tilson et al. [26].
10 -
acrylamide treated animals (20 mg/kg group) were signifi-
cantly depressed (Fig. 6). There were no significant differ-
ences between groups 5 weeks after cessation of dosing.
These results are in accord with the predictions made
prior to the beginning of the experiment. Hindlimb function
was affected at a cumulative dose of acrylamide lower than
that required to affect forelimb grip strength or spontaneous
motor activity. More prolonged exposure to acrylamide af-
fected motor functioning to a greater extent. Furthermore,
recovery of function was observed following cessation of
dosing. Finally, these data are similar to those of a previous
report from our laboratory showing hindlimb functioning of
rats was affected before forelimb grip strength and motor
activity by acrylamide given 5 days per week for one month
[22]. Recovery of function was also observed following ces-
sation of dosing.
Effects of Carbon Bisulfide on Motor Functioning of Rats
In the preceding experiment, the prediction that ac-
rylamide will affect functioning in the hindlimbs prior to af-
fecting that in the forelimbs was confirmed. The purpose of
the next study was to determine the relative selectivity and
specificity of the behavioral methods using another agent
that produces a profile of neurotoxicity similar to that of
acrylamide. The agent chosen for investigation was carbon
disulfide (CS2), since this chemical has been shown in various
neuropathological and neurophysiological studies to affect
1
I
• Controls /N = 20)
• 10 mg/kg Acrylamide IN = 10)
A 20 mg/kg Acrylamide IN = 10)
# p<0.05 from Con trol
Weeks of Treatment
10
13
FIG. 3. Effects of 10 or 20 mg/ kg of acrylamide given orally, 3 times per week, on spontaneous motor
activity of rats tested at various times during the dosing regimen. Activity counts occurring during a 9
min period were converted to responses per min and square root transformed. Data are av-
erages ± S.E. of 10 animals per treatment group and 20 rats in control groups receiving either Iml/lOOg
of distilled water vehicle or no treatment. Analysis of variance for repeated measures was used to
assess overall effects of treatment and time. The asterisk indicates a significant difference between the
mean of the combined control groups and the treatment group (Fisher's Least Significant Difference
Test, p<0.05). From Tilson et al. [26].
-------
VALIDATION OF BEHAVIORAL TESTS IN TOXICOLOGY
143
I
(g^ Controls
I I Acrylamide
'p< 0.05
sssssssss'FX i
T
Weeks Postdosing
FIG. 4. Average activity counts (expressed as responses per min and
square root transformed) ± S.E. of rats given 20 mg/kg of ac-
rylamide (N=5) or control treatment (N=5) for 13 weeks and ran-
domly selected for retesting at 1 and 5 weeks after cessation of
dosing. The asterisk indicates a significant difference between
groups (/-test, p<0.05). From Tilson et al. [26].
108
104
§ 100
o
•5
t°96
92
Acrylamlde
o 10 mg/kg (N=IO)
A 20 mg/kg (N» 10)
0 I
4 7 10
Weeks of treatment
13
FIG. 5. Average grip scores ± S.E. of rats given 10 or 20 mg/kg of
acrylamide orally, 3 times weekly for 0 (predosing), 1, 4, 7, 10, and
13 weeks (N = 10 per group). Controls were dosed with distilled
water vehicle (N=10) or given no treatment (N = 10). Analysis of
variance for repeated measures was used to assess overall effects of
time and treatment. Asterisks indicate a significant difference be-
tween the means of the combined controls (N=20) and the treatment
group (Fisher's Least Significant Difference Test, p<0.05). From
Tilson el al. [26].
Controls
I'"™1] Acrylamide
*P <0,05
T
I
600
5
uj
§ 400
E
]E
ft
o
$ 200
Q
T (N • 20)
-
-
\
\
\
\
\
\
\
\
\
\
\
\
\
T
1 rV3 controls
1 — ics8
up (0.05
\
\
\
\
N
\
\
1 JL
\
\
\
\
\
\
\
i
predosing dosing dosing
wk.3 wk.6
-
" T
\
\
\
\
\
\
\
m
T
dosing
wk.6
(N-IO)
\
\
\
\
\
\
\
\
I
postdosir
wk.3
Weeks Postdosing
FIG. 6. Average forelimb grip scores ± S.E. of rats given 20 mg/kg
of acrylamide (N=5) or control treatment (N=5) for 13 weeks and
randomly selected for retesting at 1 and 5 weeks after cessation of
dosing. The asterisk indicates a significant difference between
groups (Mest, p<0.05). From Tilson et al. [26].
FIG. 7. The effects of CS2 exposure on the hindlimb extensor re-
sponse of rats. Data are average (+ S.E.M.) hindlimb scores of
animals (N=20 per group) prior to (PreDosing) and 3 or 6 weeks
after inhalation exposure to CS2 or vehicle (left graph). Data in the
right graph are average hindlimb scores of 10 animals in each group
after 6 weeks of dosing and 3 weeks after cessation of dosing. The
asterisk indicates that the treated group differs significantly from
vehicle exposed animals (Fisher's Least Significant Difference Test,
p<0.05). From Tilson el al. [24].
-------
144
TILSON, MITCHELL AND CABE
TABLE 4
EFFECTS OF CS, ON FORELIMB GRIP STRENGTH OF RATS
Time of Testing
Average Grip Score + SEM
Control CS2
Predosing
Dosing Week 3
Dosing Week 6
Dosing Week 6
Postdosing Week 3
20
20
20
10
10
466 ± 11
481 ± 15
520 ± 15
520 ± 22
552 ± 12
482 ± 13
519 ± 10
502 ± 10
499 + 9
542 ± 21
From Tilson et al. [24]
the functioning of lower extremities to a greater extent than
that in the upper extremities [16, 20, 27].
Male, albino rats of the Fisher strain were exposed to 2
mg of CS2/1 of air or air for 4 hours per day, 5 days per week
for 6 weeks (24). Twenty rats received CS2 while 20 served
as air controls. The neurobehavioral functioning of rats was
assessed in the week prior to dosing (Predosing) and after 3
and 6 weeks of dosing. Behavioral tests were administered at
least one our after removal of the rats from the inhalation
chambers. After cessation of exposures, ten rats from each
group were randomly chosen to be retested for recovery of
function 3 weeks later.
The hindlimb extensor response decreased significantly
after treatment with CS2 (Fig. 7). After 6 weeks of dosing,
CS2 treated rats showed significantly lower hindlimb scores
than controls. In addition, 3 weeks after cessation of dosing,
CS2 treated rats did not differ from controls, indicating a
recovery of function. On the other hand, forelimb grip
strength was not significantly affected at any time during
dosing with CS2, nor were there any effects noted following
cessation of dosing (Table 4). Motor activity occurring in a
novel environment was also not affected at any time by CS2
(Table 5).
As predicted, CS2 produced a significant decrease in the
hindlimb functioning of rats. While effects on forelimb grip
strength and motor activity were not observed during the
course of the repeated dosing regimen, these effects might be
expected to occur had dosing with CS2 continued for a longer
period of time. The profile of neurotoxicity observed with
CS2 in this study is similar to that observed with acrylamide
in the previous experiment and further demonstrates the
relative sensitivity of the hindlimb procedure.
Validation of the Forelimb Procedure
The experiments discussed thus far were designed to
demonstrate the validity of the hindlimb extensor technique.
However, it must be remembered that the three tests of
motor functioning are all part of a profile and before the
hindlimb technique can be shown to be a specific measure of
hindlimb dysfunction, the relative sensitivity and specificity
of the forelimb grip procedure must be demonstrated.
The most direct way to show the specificity of the forelimb
grip measurement would be to assess a chemical that affects
the functioning of the upper extremities prior to affecting
functioning in the lower extremities. Seppalainen [15], for
example, has reported that motor conduction velocities of
nerves in the upper extremities (ulnar and median nerves )
TABLE 5
EFFECTS OF CS2 EXPOSURE ON LOCOMOTOR ACTIVITY
OF RATS IN A NOVEL ENVIRONMENT
Average Counts per
9 Min ± SEM
Time of Testing
Control
CS2
Predosing
Dosing Week 3
Dosing Week 6
Dosing Week 6
Postdosing Week 3
20
20
20
10
10
709 ± 31
386 ± 30
283 ± 18
302 ± 31
451 ± 35
706 ± 24
433 ± 28
250 ± 23
253 ± 50
426 ± 42
From Tilson et al. [24]
were slowed in lead exposed workers, but the nerve veloci-
ties in the lower extremities were not significantly influ-
enced. These neurophysiological data correspond well to the
observation made by Goldstein et al. [6] that in adults, func-
tioning of the upper limbs is affected by lead to a greater
extent than in the lower limbs. The generality of this differ-
ential effect of lead on the functioning of upper and lower
limbs of rats has not yet been demonstrated, although this
work is scheduled to be done in our laboratory in the near
future.
In spite of the lack of direct evidence to demonstrate the
relative selectivity of the hindlimb and forelimb procedures,
the validity of the forelimb procedure can be inferred by data
from other types of experiments. For example, the forelimb
grip strength of rats should be correlated to some degree with
the size of the animal, i.e., older animals should have higher
grip scores than younger animals. Moreover, if the grip score
is dependent upon body weight, then similarly aged animals,
having different body sizes, should differ accordingly in their
grip scores, i.e., males should be stronger than similarly aged
females.
Data from a number of studies conducted in our labora-
tory on control male and female rats tested at different ages
were assimilated to generate Fig. 8. As can be seen, the grip
scores of male and female Fisher strain rats increase from
day 20 to day 230 of age. In addition, male rats score higher
than females beginning at about 120 days of age. It is our
experience that male Fisher rats weigh significantly more at
that time than the female rats (Squibb, unpublished observa-
tions).
Another way of demonstrating the validity of the forelimb
technique is to determine the relative sensitivities of the
hind- and forelimb procedures using psychoactive agents
having known muscle relaxant properties. If the two proce-
dures are, in fact, measuring similar components of muscle
exertion in the upper and lower extremities, then the
forelimb procedure should be at least as sensitive as the
hindlimb technique to the effects of these drugs.
Male rats of the F-344 strain weighting approximately 250
g at the beginning of the study were randomly assigned to
receive either 0, 3, 9, or 27 mg/kg of chlordiazepoxide or 0,
20, 40 or 80 mg/kg Na Phenobarbital intraperitoneally 30 min
prior to testing hindlimb and forelimb function. There were
ten rats per each treatment group.
Analysis of variance indicated that chlordiazepoxide and
-------
VALIDATION OF BEHAVIORAL TESTS IN TOXICOLOGY
145
800
600
3
s
J 400
_D.
S,
I
200
80
160
200
240
120
Days of Age
FIG. 8. Forelimb grip scores of male and female rats as a function of age. Data are average grip
scores (g) ± S.E. of 12 rats per point. Control male and female rats from 4 different experiments
conducted at 20, 60, 120 and 230 days of age were used to construct the curve. The grip scores of
the males are significantly more than the females at 120 and 230 days of age (/-test, p<0.05).
CHLORDIAZEPOXIDE
PHENOBARBITAL
100
80
60
5- 40
m
ID
< 20
100
80
60
40
20
OFORELIMB GRIP (N = 10)
DHINDLIMB EXTENSOR (N = 10)
*p<0.05
from control
27
20
40
80
DOSE (MG / KG)
FIG. 9. The effects of various doses of chlordiazepoxide and
phenobarbital on the fore- and hindlimb scores of rats. Each point is
the average percentage ± S.E. of control (0 mg/kg). There were 10
rats tested at each dose. The asterisks indicate that the mean of the
drugged group differed from that of the vehicle control group
(Fisher's Least Significant Difference Test, p<0.05).
phenobarbital significantly affected the motor functioning of
rats as measured by the fore- and hindlimb test (Fig. 9).
Those rats receiving 9 and 27 mg/kg of chlordiazepoxide had
forelimb grip scores that were significantly lower than con-
trols, while the effect at 3 mg/kg was not statistically signifi-
cant. The same animals receiving chlordiazepoxide and
tested for hindlimb dysfunction also showed decreased re-
sponses. However, only the effect at 27 mg/kg was statisti-
cally significant. Those animals receiving phenobarbital had
significantly decreased forelimb grip scores at 40 and 80
mg/kg, while the hindlimb measure was affected by the 80
mg/kg dose only. Phenobarbital had no effect on fore- and
hindlimb scores at any other doses.
The results of these experiments show that the forelimb
measure was as sensitive as, if not more sensitive, than the
hindlimb technique to chlordiazepoxide and phenobarbital.
That the two procedures were affected in a similar fashion by
psychoactive agents having muscle relaxant properties not
only demonstrates the relative sensitivity of the two tech-
niques, but supports the specificity of the hindlimb deficits
observed in the acrylamide and CS2 experiments.
Example of Putative Nonspecific Neurobehavioral Toxic
Reaction
The concept of a profile of effects generated by testing
known neurotoxins in a battery of tests having different, but
overlapping components, was generally supported in the ac-
-------
146
TILSON, MITCHELL AND CABE
0 3 10 0 3 10
Dose of FF-1 (mg/kg)
FIG. 10. The effects of FF-1 given to male and female rats for 6
months on hindlimb extensor response. Data are average hindlimb
scores (g) ± S.E. Asterisk indicates a significant difference from
vehicle treated control rats of the same sex (Fisher's Least Signifi-
cant Difference Test, p<0.05). From Tilson and Cabe [23].
60
6
40
20
Motor activity
Males Females
T
# p < 0.05 from control
*
-T
10
10
Dose of FF-1 (mg/kg)
FIG. 12. The effects of FF-1 given to male and female rats on spon-
tanous motor activity. Data are average responses occurring per
min ± S.E. in a 9 min test. The asterisks indicate that the mean of
the treated group differs statistically from that of the vehicle treated
animals (Fisher's Least Significant Difference Test, p<0.05). From
Tilson and Cabe [23].
rylamide and CS2 studies. In those experiments, it was pre-
dicited that hindlimb functioning would be affected prior to
spontaneous activity and forelimb grip strength. The purpose
of the present study was to examine the neurobehavioral
profile of an environmental agent that might affect behavior
secondary to effects on other organ systems.
800
600
8400
o.
'fe
01
en
S 200
Males
Females
* p<0.05 from control
0 3 10 0 3 10
Dose of FF-1 (mg/kg)
FIG. 11. The effects of FF-1 on the forelimb grip strength of male
and female rats treated with FF-1 for 6 months. Data are average
forelimb scores (g) ± S.E. The asterisk indicates a significant differ-
ence from vehicle control rats of the same sex (Fisher's Least Sig-
nificant Diffence Test, p<0.05). From Tilson and Cabe [23],
Polybrominated biphenyls (PBBs) have been shown to
induce hepatic microsomal enzymes and produce liver tox-
icity in laboratory animals [14,17] and alterations in the
functioning of these and other organ systems occurs in doses
lower than those required to produce measurable behavioral
changes [7, 9, 25]. These effects in laboratory animals and
the value symptomatology reported by humans exposed to
PBBs [30] suggest that they might influence behavior sec-
ondarily to effects on non-neural systems. If the PBBs in-
fluence behavior secondarily or affect motor functioning
nonspecifically, then it is predicted that they will depress all
three measures of motor functioning and these effects will be
observed at about the same dose.
Male and female rats of the F-344 strain were used as
described elsewhere [23]. A commercial mixture of PBBs,
Firemaster FF-1, was suspended in corn oil vehicle and
given by gavage, five days per week for a total of 130 doses.
There were 7-8 animals of each sex given 3 or 10 mg/kg per
dose, while 9 females and 15 males served as controls (corn
oil vehicle). The animals were tested for neurobehavioral
dysfunction after 1, 2, 4, and 6 months of dosing. Since no
effects on behavior were observed after 4 months of dosing,
only data collected at the 6 month test are reported.
Chronic exposure to PBBs had a significant effect on the
hindlimb extensor response of rats (Fig. 10). The males were
more consistently affected by PBBs than the females. Inde-
pendent group comparisons showed that males were affected
significantly by 3 and 10 mg/kg. PBB exposed female rats
had lower hindlimb scores than controls at 10 mg/kg but the
effect at 3 mg/kg was not significant.
As in the case of the hindlimb measure, the forelimb grip
scores of PBB treated rats were decreased after 6 months of
dosing with PBBs (Fig. 11). Independent group comparisons
indicated that male rats were significantly affected by 3 and
10 mg/kg of PBBs. Female rats receiving 3 mg/kg had lower
scores than controls but were not affected significantly
while those rats receiving 10 mg/kg had significantly reduced
grip scores.
The effects of PBBs on motor activity depended on the
baseline level of activity (Fig. 12). Female rats were much
more active than males and 3 and 10 mg/kg of PBBs signifi-
-------
VALIDATION OF BEHAVIORAL TESTS IN TOXICOLOGY
147
TABLE 6
PREDICTED NEUROBEHAVIORAL EFFECTS OF ACRYLAMIDE IN RATS AT VARIOUS
TIMES DURING DOSING*
Sensory
Visual
Auditory
Olfactory
Tactile
Pain
Vestibular
Motor
Spontaneous Motor Activity
Forelimb Function
Hindlimb Function
Tremor
Muscular Endurance
Arousal
Emergence
Startle Response
Learning
One-Way Active Avoidance
Retention
Physiological Functioning
Body Weight
Core Temperature
Autonomic Signs
Shorter Term/
Lower Dose
0
0
0
-
0
0
0
0
_
0
-
0
0
0
0
-
0
+
Longer Term/
Higher Dose
-
0
0
-
0
0
-
-
-
+
-
0
0
0
-
-
0
+
Recovery
-
0
0
0
0
0
0
0
0
0
-
0
0
0
-
0
0
0
*Response: + = measure is predicted to be elevated or rate increased; - = predicted to be
decreased in magnitude or rate decreased; 0 = no significant alteration in the measure is expected at
the time of testing.
cantly decreased the motor activity of the females. The male
rats were not affected significantly by either dose of PBBs.
In summary, the neurobehavioral effects of the PBBs de-
pend on the rate or magnitude of the behavioral dependent
variable. If the rate or magnitude of responding is high, then
the PBBs decrease that behavior. That the behavioral effect
observed with the PBBs was consistent in each of the 3 tests
and occurred at approximately the same dose is the type of
neurobehavioral profile indicative of a nonspecific toxic reac-
tion. Studies to confirm the generality of such a profile using
another agent that produces liver toxicity, such as carbon
tetrachloride [12], are planned.
SUMMARY AND CONCLUSIONS
In an attempt ot demonstrate how test methods in behav-
ioral toxicology can be standardized and validated, specific
examples concerning three motor tasks used in our labora-
tory have been discussed. Obviously, a great deal of work
will be required to standardize and validate all the tests used
in a neurobehavioral test battery such as that described in
Table 2. However, the requirement for behavioral and neuro-
logical assessment of a large number of chemical substances
only accentuates the need to conduct such research.
It is our contention that the success of the validation
scheme proposed in this report depends on the predictions
made prior to the beginning of the study. Such predictions
permit test validation by confirming the usefulness of any
given procedure for the type of neurobehavioral symptom
and representative chemical being assessed. With increasing
probability of success in making such predictions, the utility
of behavioral and neurological methods also increases.
It is important to indicate that projections which are not
confirmed are as valuable, if not more so, than correct pre-
dictions. Failure to confirm expectations for a given sub-
stance and procedure exposes the need for additional infor-
mation about the animal model in question. The data ob-
tained in validation studies of this type can only assist in the
development of other more refined techniques. Moreover,
such information can provide the basis for reevaluation or
reassessment of the mechanism of action of the agent under
investigation.
Another important feature of a test validation scheme as
proposed in this report is the profile of effects generated for
each substance. The pattern of correct and incorrect proj-
ections provides a basis for evaluating the competence of the
testing systems. Assessment of the profiles resulting from
the study of representative neurotoxins can lead eventually
to reevaluation of the principle factors underlying the animal
model.
-------
148
TILSON, MITCHELL AND CABE
For the purposes of illustration, Table 6 contains predi-
cted behavioral effects for acrylamide using the test battery
described in Table 2. Based upon a review of the human
neurotoxicology of acrylamide [12,19], a symptomatological
profile was generated. The pattern of expected effects for
acrylamide was then generated for the neurobehavioral test
battery currently used in our laboratory. Table 6 summarizes
these predictions in terms of the presence or absence of an
effect, the direction of the effect, and whether or not the
effect can be reversible or irreversible.
A similar profile of predicted effects could be generated
for the substances listed in Table 3 and, once the predictions
in terms of occurrence, onset, duration, and reversibility of
effect are correctly predicted, then the criteria for test
validation have been met. In addition, tests presumed to
measure similar functions can be compared for the dosage
level required to see an effect and for the capability to detect
an effect where one is expected or show no effect where
none is expected.
It is our belief that the acceptance of behavioral and
neurological procedures for use in toxicology in general and
neurotoxicology in particular will depend in large measure
upon their demonstrated validity, sensitivity, and specificity.
More importantly, validation of behavioral tests must occur
before environmental agents with unknown potential for
neurotoxicity can be assessed in any meaningful way.
ACKNOWLEDGMENTS
1. Special thanks go to Ms. Sue Bolton for her skillful assistance
in the preparation of this manuscript.
2. The technical assistance of Mr. Richard Rhoderick, Ms. Lyn
Reed, and Mr. Larry Judd in the performance of these experiments
is gratefully acknowledged.
REFERENCES
1. Cabe, P.A. and H.A. Tilson. The hindlimb extensor response: A
method for assessing motor dysfunction in rats. Pharmac.
Biochem. Behav. 9: 133-136, 1978.
2. Cabe,P. A., H. A. Tilson, C. L. Mitchell and R. Dennis. A
simple recording grip strength device. Pharmac. Biochem. Be-
hav. 8: 101-102, 1978.
3. Casarett, L.J. Toxicologic Evaluation. In: Toxicology: The
Basic Science of Poisons, edited by L. J. Casarett and J. Doull.
New York: MacMillan, 1975, pp. 11-25.
4. Dews, P. B. Epistemology of screening for behavioral toxicity.
Envir. Hlth. Perspec. 26: 37-42, 1978.
5. Evans, H. L. and B. Weiss. Behavioral Toxicology. In: Con-
temporary Research in Behavioral Pharmacology, edited by D.
E. Blackman and D. J. Sanger. New York and London: Plenum
Press, 1978, pp. 449-487.
6. Goldstein, N. P., J. T. McCall and P. J. Dyck. Metal
Neuropathy. In: Peripheral Neuropathy, edited by P. J. Dyck,
P. K. Thomas and E. H. Lambert. Philadelphia: W.B. Saunders
Co., 1975, pp. 1241-1262.
7. Luster, N., R. E. Faith and J. A. Moore. Effects of polybromi-
nated biphenyls (PBB) on immune response in rodents. Envir.
Hlth. Perspec. 23: 227-232, 1978.
8. Miller, R. G. Simultaneous Statistical Inference. New York:
McGraw-Hill, 1966.
9. Moore, J. A., M. . Luster, B. N. Gupta and E. E. McConnell.
Toxicological and immunological effects of a commercial poly-
brominated biphenyl mixture (Firemaster FF-1). Soc. of Tox-
icology, San Francisco, California, March 12-61, 1978, p. 153
(Abstract).
10. Moses, L. E. Quoted in: Goldstein, A., Biostatistics, an Intro-
ductory Text. New York: MacMillan, 1964, p. 62.
11. Norton, S. Toxicology of the Central Nervous System. In: Tox-
icology: The Basic Science of Poisons, edited by L. J. Casarett
and J. Doull. New York: MacMillan, 1975, pp. 151-169.
12. Pitchumon, C.S., R. J. Stenger, W. S. Rosenthal and E. A.
Johnson. Effects of 3,4-benzpyrene pretreatment on the
hepatotoxicity of carbon tetrachloride in rats. J. Pharmac. exp.
Ther. 181: 227-233, 1972.
13. Principles for Evaluating Chemicals in the Environment, Na-
tional Academy of Sciences, Washington, D. C., 1975.
14. Ringer, R. K. and D. Polin. The biological effects of polybromi-
nated biphenyls in avian species. Fedn. Proc. 36: 1894-1898,
1977.
15. Seppalainen, A. M. Peripheral nervous system in lead exposed
workers. In: Behavioral Toxicology, edited by C. Xintaras, B.
L. Johnson and I. deGroot. Cincinnati, Ohio: NIOSH, 1974, pp.
240-247.
16. Seppalainen, A. M. and M. Tolonen. Neurotoxicity of long-term
exposure to carbon disulfide in the viscose rayon industry: A
neurophysiological study. /. Work Environ. Hlth. 11: 145-153,
1974.
17. Slight, S. D. and V. L. Sanger, Pathologic feature of polybromi-
nated biphenyl toxicosis in the rat and guinea pig. J. Am. Vet.
Med.Assoc. 169: 1231-1235, 1976.
18. Spencer, P. S. and H. H. Schaumberg. A review of acrylamide
neurotoxicity. Part I. Properties, uses and human exposure.
Can. -J. Neural. Sci. 1: 143-150, 1974.
19. Spencer, P. S. and H. H. Schaumberg. A review of acrylamide
neurotoxicity. Part II. Experimental animal neurotoxicity and
pathologic mechanisms. Can. J. Neural. Sci. 1: 152-169, 1974.
20. Szenozikowski, S., J. Stetkiewicz, T. Wronska-Nofer, and I.
Zdrazkowska. Structural aspects of experimental carbon disul-
fide neuropathy. I. Development of neurohistological changes in
chronically intoxicated rats. Int. Arch. Arbeitsmed. 31: 135-149,
1973.
21. Tilson, H. A. and P. A. Cabe. Strategy for the assessment of
neurobehavioral consequences of environmental factors. Envir.
Hlth Perpec. 26: 287-299, 1978.
22. Tilson, H. A. and P. A. Cabe. The effects of acrylamide given
acutely or in repeated doses on fore- and hindlimb function of
rats. Toxicol. Appl. Pharmacol. 47: 355-362, 1979.
23. Tilson, H. A. and P. A. Cabe. Studies on the neurobehavioral
effects of polybrominated biphenyls in rats. Ann. N. Y. Acad.
Sci. in press.
24. Tilson, H. A., P. A. Cabe, E. H. Ellinwood, and L. P. Gon-
zalez. The effects of carbon disulfide on motor function and
responsiveness to rf-amphetamine in rats. Neurobehav. Tox-
icol. 1: 57-63, 1979.
25. Tilson, H. A., P. A. Cabe and C. L. Mitchell. Behavioral and
neurological toxicity of polybrominated biphenyls in rats and
mice. Env. Hlth. Perspec. 23: 257-263, 1978.
26. Tilson, H. A., P. S. Spencer and P. A. Cabe. Acrylamide
neurotoxicity in rats: A correlated neurobehavioral and
pathological study. Neurotoxicology. 1: 89-104, 1979.
27. Vasilescu, C. Motor nerve conduction velocity and elec-
tromyogram in carbon disulphide poisoning. Rev. Roum.
Neural. 9: 63-71, 1972.
28. Wang, G. H. Relation between spontaneous activity and oestr-
ous cycle in the white rat. Camp. Psychol. Monog. 2: 1-32,
1932.
29. Winer, B. J. Statistical Principles in Experimental Design. New
York: McGraw-Hill, 1962.
30. Valciukas, J. A., R. Lilis, M. S. Wolff and H. A. Anderson.
Comparative neurobehavioral study of a polybrominated
biphenyl-exposed population in Michigan and a nonexposed
group in Wisconsin. Envir. Hlth. Perspec. 23: 199-210, 1978.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 149-155. ANKHO International Inc., 1979.
Behavioral Epidemiology of Food Additives
BERNARD WEISS, CHRISTOPHER COX AND MARC YOUNG
Department of Radiation Biology and Biophysics
Division ofBiostatistics, and Environmental Health Sciences Center
University of Rochester School of Medicine and Dentistry
Rochester, NY 14642
AND
SHELDON MARGEN
Department of Nutritional Sciences, University of California, Berkeley, CA
AND
J. HICKS WILLIAMS
Department of Pediatrics, Kaiser-Permanente Medical Center, Santa Clara, CA
WEISS, B., C. COX, M. YOUNG, S. MARGEN AND J. H. WILLIAMS. Behavioral epidemiology of food additives.
NEUROBEHAV. TOXICOL. 1: Suppl. 1, 149-155, 1979.—Behavioral toxicology in the natural environment can be consid-
ered a special branch of epidemiology. Behavioral epidemiology, because it typically relies on complex functional criteria,
faces all of the problems of behavior measurement posed by uncontrollable variation, and amplified even further by
chemical exposure. Many such issues arose in a study of behavioral responses to artificial food colors in children.
Difficulties in employing Applied Behavioral Analysis in such a context run the gamut from selection of retrospective
criteria to appropriate statistical models.
Behavioral epidemiology
Randomization tests
Food colors Applied Behavior Analysis Time series Food additives
WHAT happens to people is more compelling than what
happens to rats. It also exerts more impact on public policy.
Would a saccharin ban have been lifted by Congress had a
few compulsive TAB drinkers testified to developing
cancer? Moreover, we may not always be able to find a
suitable animal model. High doses of amphetamine make
humans act schizophrenic; they make rats gnaw. The natural
environment, though, is such a thick web of frustration and
uncertainty that it drives us to the epistemologic imperfec-
tion but methodologic comfort of the laboratory. Here, we
will discuss the discomforts aroused by experiments in the
natural environment and their implications for what we've
called Behavioral Epidemiology.
Traditional epidemiology originated in infectious disease.
Like traditional toxicology, death or pathology was its pre-
ferred endpoint. Morbidity data, such as hospital admis-
sions, provide less satisfactory criteria because they are
more subject to blurring by diagnostic variability and varia-
tions in seeking medical care. Functional measures, like
those exemplified by behavior, magnify uncertainty. Aside
from the tentative significance intrinsic to functional criteria,
behavioral epidemiology faces, in exaggerated form, the
threats to interpretation posed by transient effects, variable
combinations of agents and exposure conditions, the stagger-
ing polymorphism of the human population, and wildly di-
vergent behavioral histories. A sobering illustration of how
such factors can combine into a puzzling montage is the
debate over what constitutes safe exposure levels of lead for
children. The population presumably most at risk has also
inherited exacerbated risks from many sources.
The discipline known as Applied Behavior Analysis
(ABA) offers a methodology for slipping some of these con-
straints, although at the cost of large samples. ABA grew out
of attempts to transfer the technology of operant condition-
ing to the modification of human behavior. The technology
could not be transferred wholesale. Elegant instrumentation
'The preparation of this report was supported in part by Grant MH-11752 from NIMH, Grant ES-01247 from NIEHS, and in part by a
contract with the U.S. Department of Energy at the University of Rochester Department of Radiation Biology and Biophysics and has been
assigned Report No. 3490-1605.
2The material in this paper is based in part on a study conducted under FDA contract 223-76-2040, with the Kaiser Foundation Research
Institute acting as primary contractor.
149
-------
150
WEISS ET AL.
had to be left behind and total environmental control was out
of the question. It retained the key element of precise spec-
ification of behavior, and developed recording schemes suit-
able for field use typically carried out by trained observers.
As in animal studies, single subject data and interventions
became the featured mode of investigation.
ABA shares common concerns with all behavioral ap-
proaches: stability, suitability, and sensitivity of baselines.
As a component of toxicology and epidemiology, it must also
incorporate an appreciation of dose or concentration as vari-
ables, the contribution of exposure duration and the impact
on particular target organs. Its unique problems and oppor-
tunities arise from its ability to cope with several behavioral
indices simultaneously, which implies the selection of sev-
eral criterion behaviors, which poses the issue of single sub-
ject designs and the associated analytical constraints.
The potential, and the difficulties of applied behavior
analysis as a tool of behavioral epidemiology can be illus-
trated by a recent experiment carried out in California. Its
aim was to determine whether synthetic food colors can
modify the behavior of susceptible children. It originated in
the claim [3] that some of the children clinically labelled as
hyperactive or hyperkinetic are actually exhibiting an ele-
vated sensitivity to certain properties of synthetic food
colors, flavors, and natural constituents chemically defined
as salicylates. The data bearing on this central question will
be published elsewhere. Here, we focus on the methodolog-
ical issues that emerged in the design, execution, and
analysis of the experiment.
Design Issues
Subjects and baseline selection. If acute effects are the
primary question of a study, as they were here, a
straightforward experimental design suggests itself. Select a
group of allegedly responsive children whose diets already
exclude the postulated offending substances and challenge
them with one or more of these substances to determine if a
reproducible reaction can be elicited. Such an approach ob-
viously exploits a biased sample of children. It is most rele-
vant when the validity of a phenomenon is at issue. It is a
more basic question than prevalence.
Choice of criteria. The psychological literature overflows
with behavior inventories, rating scales, and observational
schemes. All have virtues and disadvantages. Inventories
are useful in clinical assessment, as in helping a therapist
specify the problem behaviors of a client. Most inventories,
however, emphasize enduring behaviors or traits rather than
transient states, making them unsuitable for assessing acute
reactions. Rating scales, such as the Conners Parent-
Teacher Questionnaire [2], which is widely employed to help
diagnose hyperkinesis, provide a standard list of items, some
of which may be relevant, some irrelevant to a particular
subject. Standardization is an important virtue, however,
and so is the feature that rating scale items typically integrate
an observer's evaluation over a long time period (a day or
longer). This allows the rater some flexibility in weighing
specific incidents within their total environmental context.
Their disadvantages are parallel. They usually do not reflect
moment-to-moment variations in behavior. Often, the inte-
gration is secured at the cost of specificity in behavior, mak-
ing the items subject to varying interpretations by observers,
and, because of the long time spans often included, diluted
by the same distortions as all retrospective surveys.
Actual frequency counts of narrowly defined behaviors at
first seem an ideal solution. They are, certainly, for ques-
tions about the success of interventions. If, say, a child
squirms excessively in the classroom, and the teacher is in-
structed to pay attention the child only when it is sitting still,
the most direct criterion of effectiveness is the amount of
squirming. Toxicology and pharmacology, however, be-
cause their interventions occur along nonbehavioral dimen-
sions, may require a more comprehensive index of outcome.
Typically, to the disappointment of those who have seen the
productivity of observational methods in other settings, re-
cording the frequency of specific behaviors has generally
proven less sensitive to drugs than more global measures [6].
Narrow spectra of behaviors, combined with relatively brief
(and unrepresentative) sampling periods may be at fault, but
the empirical results are disappointing nevertheless.
Choosing observers. Highly-trained observers, thor-
oughly familiar with a specific coding scheme, represent an
ideal that cannot be practiced, in large-scale studies, except
at probably intolerable expense. The alternative is to select
observers in situ, so to speak: teachers, nurses, parents.
Because such observers are not trained specifically for such
a role, and, further, often must fulfill responsibilities of a
more demanding and immediate nature, they tend to produce
more variable information than trained observers. But they
have the advantage of unobtrusiveness, since they are part of
the subjects' customary environment, and, also, familiarity
with subtleties of the subjects' behavior not apparent to an
outside observer. They also may be more biased, but can be
provided with controls that minimize, or, at least, distribute
bias.
Data analysis. Operant experimenters are accustomed to
laboratory phenomena so robust that the data require little
additional treatment. Such transparency is not characteristic
of the natural environment. Important features may be
buried in variability, in serial dependencies, in shifting
baselines that cannot be extracted without some form of
statistical surgery. These considerations became practical di-
lemmas when we set out to determine the validity of Fein-
gold's assertions.
Specialists in Applied Behavior Analysis have been de-
bating for several years the suitability of various statistical
models and methods for data of this kind. Two sharp differ-
ences distinguish such data from the group designs typical of
psychology. First, if each subject is assessed by a unique set
of behaviors, and by a unique observer, grouping makes no
sense. Second, conventional repeated measures models not
only derive from group designs, but neglect the most critical
feature of the paradigm: serial dependency. That is, by
allowing several observations on the same experimental unit
(subject), repeated measures designs do allow within-subject
correlation to be modelled in the analysis. The experimental
units, however, must be sampled independently. (This is true
of any multivariate statistical procedure.) This requirement
is not generally met when the same subject is observed at
consecutive time points, unless some sort of washout occurs
between observation. Such washout is the basis for the tra-
ditional crossover designs. Because of the nature of the chal-
lenge in this study, serial dependency from day to day clearly
cannot be excluded. Therefore, any proposed statistical
analysis must take account of (model) this dependency. Two
methodologies are appropriate: time series analysis and ran-
domization tests.
A time series is a sequence (in time) of observations (of a
random process). The importance of this definition lies in its
-------
BEHAVIOR AND FOOD ADDITIVES
151
generality; statistical independence of the observations is not
assumed, nor are the observations assumed to have any
common probabilistic properties. Autocorrelations are im-
portant determinants of the behavior of a time series. The
concept of simple correlation is familiar as an index of asso-
ciation between two variables (such as I.Q. and shoe size)
measured on the same subject. If many observations (an
entire time series) are taken on the same subject, then the
first of them will be correlated with the second, and also with
the third, fourth, etc. In many situations, autocorrelations
decrease in magnitude as the lag or number of intervening
observations increase, so that the influence of the past les-
sens with time.
There are two approaches to modelling the structure of
the autocorrelations of a time series, corresponding to the
parametric and nonparametric statistical models employed in
group designs. The basic approach is to fit a parametric
model to the data. The three most common stationary mod-
els are autoregressive processes, moving average processes
and a combination of the two [5]. The first step in the
analysis involves selection of one of the above models. This
process involves the estimation and examination of the au-
tocorrelations (which behave differently for each model) and
some judgment, as well as any theoretical considerations
which apply.
The nonparametric alternative is known as spectral
analysis. This is basically an attempt to discover periodic
behavior in the series caused by patterns in the autocorrela-
tions. One basically tries to understand the behavior of the
series in terms of the autocorrelations themselves, rather
than specifying a model whose parameters determine the
autocorrelations.
Randomization tests (like much of modern statistical
methodology) originated with R. A. Fisher in the famous
experiment of the lady testing tea [4]. (Chapter 2 of Fisher's
classic reference still makes good reading.) The theory and
practice of such tests has experienced a revived interest with
the advent of third generation computers. The basic as-
sumption made by all such tests is that the randomness nec-
essary to generate the probability distribution of a test
statistic is provided, not by random sampling of any sort, but
by experimental randomization (the random assignment of
experimental units to treatments). In the simplest case, if we
have two groups, treatment and control, then we assign ex-
perimental units randomly to each of the two groups. If this
is done, then, under the null hypothesis of no treatment ef-
fect, the particular arrangement of the observed data into the
two groups may be considered simply a chance occurrence
produced by the random assignment. We may test the null
hypothesis by examining all possible rerandomizations of the
observed data into the two groupings (treatment and control)
of the same respective sizes (and category) and asking how
unusual the originally observed differences are among all
such permutations. As in any statistical test, unusualness is
measured by a statistic. Thep-value of the test is simply the
percentage of rerandomizations whose statistic exceeds the
value of the statistic for the observed data. The choice of the
statistic is dictated by the alternative hypothesis of interest,
i.e., a statistic is chosen which should be sensitive to the
particular alternative.
As the numbers in each of the treatment groups grow, the
number of rerandomizations quickly becomes enormous.
Instead of calculating an exact p-value in such situations by
generating all rerandomizations of the data, one can ran-
domly sample from the set of all possible rerandomizations
and estimate the true p-value. Monte Carlo studies have
indicated that 10,000 is a suitably large sample.
The California Study
Experimental design. We decided to select only children
successfully maintained (according to parents) on the diet,
then challenge them with specific agents. Such an approach
was feasible with synthetic colors, because only eight are
approved for food by FDA, and one, Orange B, is restricted
to sausage casings. There are 1500 synthetic flavors. We
chose to administer a monoblend of the seven colors (exclud-
ing Orange B) in proportions and amounts based on a diet
survey from which color intake was calculated by extrapola-
tion from published manufacturing practices [1].
Over a 77-day experimental period, an individual child
received eight color challenges. Every day, at a specified
time, the child consumed a specially-formulated soft drink.
On most days, it contained only a mix of caramel coloring
and cranberry powder. On challenge days, randomly in-
terspersed between Days 15-70, the drink contained the
monoblend. The assigned occasions, however, sometimes
had to be modified. Days on which the child consumed a
forbidden item were eliminated from analyses.
Important variables lay beyond our control. We chal-
lenged with a dose equivalent to our calculated mean daily
intake, although a more optimal design would have sought a
dose-effect function. We were limited by the anxiety of
human subjects review committees about the possibility of
severe reactions. We administered the color blend on the
average only once weekly because testimonial evidence
suggested reactions to a single infraction lasting several
days.
Specification of behaviors. Not only do standardized
rating scales strike some observers as vague, but the most
widely-used instruments for assessing hyperkinesis are
based on children older than the ones in our sample. Our
sample comprised 22 children, 15 boys and 7 girls, ranging in
age from 2/5 to 6/11. All allegedly had responded suc-
cessfully to the Feingold diet for at least three months before
entry into the study. Since no single standard scale or inven-
tory or coding scheme seemed suitable for all the children,
we decided to develop a set of target behaviors for each
child. The parent sorted a deck of 341 punched cards, each
labelled with an item from a standard inventory. On success-
ive sorts, the parent gradually narrowed the applicable items
to 7 aversive and 3 positive behaviors. The aversive behav-
iors were to be those associated with infractions. Positive
behaviors were included because it seemed possible that
some reactions might be expressed more as a reduction in
behaviors such as affection than as an eruption of behaviors
such as aggression. Table lisa list of target behaviors from a
representative subject.
We gauged the target behaviors in two ways. First, we
had the parents count frequency of occurrence during two
15-min periods within the 24 hr period following the drink.
One session took place within 3.5 hr of drink consumption,
the other at a later time. These times were determined on the
basis of parents' reports of when aversive behaviors seemed
most frequent. Time of drink consumption and of observa-
tions remained uniform throughout the study period.
At the end of the 24-hr period (the research day), the
parent recorded a global rating for each target behavior
based on the total day. He or she marked a point on a line
corresponding to "How much did it occur?" The position of
-------
152
WEISS ETAL.
9
8
7
g
cr
D
LJ -7
00 •->
SUBJECT 63
BEHflVIOR 3
o
»
*
PLPCEBO
INFRACTION
CHflLLENGE
6 ' 10 ' 20 ' 30 ' 40 ' SO ' 60 ' 70 ' 8'0
DRY
FIG. 1. Zero incidenceof a target behavior ("glazed eyes") selected as relevant by the
mother.
TABLE 1
TARGET BEHAVIOR: SUBJECT 71
1. Short attention span, distracted
2. Fidgets
3. Mood changes drastically
4. Acts as if driven by motor
5. Will not eat enough
6. Loose bowels
7. Temper outburst
8. Responds to social stimulation by talking, smiling, etc.
9. Helpful, cooperative
10. Affectionate
the point was converted into a score ranging from 1-9. The
parent also completed the Conners short form and supplied
additional information and comments, including a\ total count
of aversive behavior obtained from a wrist counter worn
during part of the day.
Parents were telephoned every weekday to prompt them
to complete and mail the forms, and to allow staff to
conduct a partly-structured interview about the previous
day's events. The interviewer and an independent listener
then also completed corresponding forms. A behavioral
specialist visited the home once weekly at a time during
which the parent was scheduled to conduct one of the 15-min
observations. She recorded the target behavior occurrences
independently and compared her results with those of the
parent. Such a procedure provides only a rough check of
observer reliability, but at least it was able to uncover gross
discrepancies in interpretation or a marked drift in criteria.
These visits, along with weekly visits by the research nutri-
tionist, served other functions as well. They helped maintain
a warm personal relationship between staff and parents, a
relationship crucial to a long, tedious, and demanding com-
mitment by the parents. We emphasize this point because it
is so often overlooked by experimenters who expect subjects
to share their own interest and dedication without the re-
wards accruing to the experimenter.
Appropriate choices of behavior. We are all aware of
dangers in the approach we adopted. One surely is reliance
on retrospective ratings, a tactic that some questions impose
because there are no alternatives, as in the studies of the
Michigan residents exposed to PBBs. Recall that we asked
parents to'specify behaviors associated with diet infractions
or that prevailed before adoption of the diet. All of the prob-
lems that accompany such approaches were amplified by the
often startling rapidity with which children change their be-
havioral repertoire. Figure 1 is an example of a behavior
selected by the parent as a characteristic aversive behavior.
It never occurred during the course of the study.
Rapid change was another phenomenon we saw. Figure 2
plots a target behavior almost never recorded by the parent
-------
BEHAVIOR AND FOOD ADDITIVES
153
9-
8-
7-
o 6-
t—*
i—
2 5-
(X
D
0—I
SUBJECT 68
BEHRVIOR 6
oo
ODtW ODDBEW
PLRCEBO
INFRRCTION
CHRLLENGE
o o
CD ttt tt O
Ot OCD
0 10 20 30 40 50 60 70 80
DRY
FIG. 2. Abrupt change in occurrence of a target behavior ("unable to stop a repetitive
activity"). Challenge days are connected by a line.
during the first part of the study. It then suddenly appeared,
and, in fact, demonstrated a statistically significant response
to challenge (p<0.01). Did the parent's criterion change?
Perhaps; but the other target behaviors did not reveal a simi-
lar pattern. Could the child have been "going through a
phase?" Might the earlier challenges have exerted a cumula-
tive impact?
One aspect of these data that surprised us, and that might
prove helpful for behavioral epidemiology research, is the
typically high correlations achieved among the global meas-
ures (total counts, day overall rating, and Conners score).
For most subjects, these coefficients were about 0.70. Such a
degree of concordance might suggest that, for some purpos-
es, global rather than detailed measures are feasible indices
of response or status. We must note, however, that without
specified target behaviors, our observers might have been far
less sensitive. Figure 3 (A and B) display daily ratings of two
behaviors from one subject. Both behaviors were highly re-
sponsive to color challenge. Yet, their patterning during the
study was markedly different.
Such peculiarities in patterning are partly what governed
our choice of randomization tests as the most appropriate
technique of data analysis. We simply could not meet the
assumptions required by parametric approaches. Further-
more, we were dealing with phenomena that, even if most of
the assumptions could be met, depart markedly from the
usual designs in many respects. Although the parametric
techniques employed in the first phase of the analysis did not
lead us too far astray, they did, however, prove insensitive
to subtle effects such as small elevations in the 15-min
counts.
Outcome. The detailed results of this study will be re-
ported elsewhere. For one child, the results were dramatic,
even astonishing in their magnitude and consistency. Two or
three other children showed significant reactions according
to one or more target behaviors, so that they would not have
been detected by global measures alone. The focus of this
paper, however, is our approach, the problems it entailed,
how they were handled, and the potential of this approach in
the arena of environmental toxicology, because even those
of us who strongly adhere to the operant tradition of inten-
sive study of single subjects seemed baffled about how to
transfer that tradition to toxicology. Perhaps we should stop
worrying about screening and standard setting and return to
the fundamental issue of what are the important variables.
-------
154
WEISS ET AL.
cc
oc
o
«
cr
9-
8-
7-
6-
5—
4-
3-
2-
1-
0-
o OOD
9-
8-
7-
6-
cr c
o: 5-
cc
D
cr
3-
2-
1-
0-
SUBJECT 55
BEHflVIOR 1
o =
PLflCEBO
INFRfiCTION
CHflLLENGE
CDOO ttO «OO
oo
O OB OX OD SO «O
CDOD O OO CJt OODOOD
0 10 20 30
40
DflY
50 60 70 80
SUBJECT 55
BEHOVIOR 2
o = PLflCEBO
« = INFRflCTIdN
* = CHflLLENGE
GOD ODOD
Qooott o o/oooo *
O OO O CD OtO » ODOO
too tto CD ¥o arm o o o oo o o o
0 10 20 30 40 50 60 70 80
DflY
FIG. 3. (A) Response to challenge of target behavior, "Bites, kicks, hits." (B) Response to
challenge of target behavior, "Throws things inappropriately." Lines connect challenge
days.
-------
BEHAVIOR AND FOOD ADDITIVES
155
REFERENCES
1. Certified Color Ir Committee. Guidelines for food man-
ufacturing practi. -. use of certified FD &C colors in food. Food
Technol. 22: 946-949, 1968.
2. Conners, C. K. Rating scales for use in drug studies with chil-
dren. Psychophartnac. Bull.: Pharmacotherapy of children.
Washington, DC: U.S. Government Printing Office, 1973, pp.
24-29.
3. Feingold, B. F. Why Your Child is Hyperactive. New York:
Random House, 1975.
6.
Fisher, R. A. The Design of Experiments, 5th edition. London:
Oliver and Boyd, 1949.
Glass, G. V., V. L. Willson and J. M. Gottman. Design and
Analysis of Time-Series Experiments. Boulder, Colo.: Colorado
Associated University Press, 1975.
Werry, J. S. Measures in pediatric psychopharmacology. In:
Pediatric Psychopharmacology: The Use of Behavior-Modifying
Drugs in Children, edited by J. S. Werry. New York: Brunner/
Mazel, 1978.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromolor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 157-161. ANKHO International Inc., 1979.
Psychological Test Methods: Sensitivity to Long
Term Chemical Exposure at Work
HELENA HANNINEN
Institute of Occupational Health, Haartmaninkatu 1, SF-00290 Helsinki 29, Finland
HANNINEN, H. Psychological test methods: Sensitivity to long term chemical exposure at work. NEUROBEHAV.
TOXICOL. 1: Suppl. 1, 157-161, 1979.—Five studies dealing with long term occupational exposure to carbon disulfide, a
mixture of organic solvents, toluene, styrene and lead are reviewed. All of the studies were cross-sectional, comprising
either a comparison between exposed and nonexposed groups or the determination of exposure-response relationships, or
both. The tests for the cognitive functions were known clinical intelligence and memory tests. The perceptual and
psychomotor tasks were the Santa Ana test, the Bourdon-Wiersma test for visual-motor speed and accuracy, the Symmetry
Drawing test and the Mira test. In four of the five studies the neurotoxic effect involved both cognitive and psychomotor
functions. In the carbon disulfide group, psychomotor retardation was the most pronounced effect: in the group exposed to
solvent mixtures the main effects were seen in the cognitive functions. The effects of styrene were limited to perceptual and
psychomotor disturbances. As the most sensitive methods have varied from study to study, the continued use of broad and
diverse psychological methodology in studies dealing with long term neurotoxic effects is proposed.
Environmental chemicals
Perceptual motor tests
Occupational exposure Behavioral effects Cognitive tests
THIS paper deals with the use of psychological test methods
for determining central nervous system (CNS) dysfunctions in
subjects with long term occupational exposure to toxic
agents. Five empirical studies will be reviewed, and eleven
of the test methods employed in these studies will be de-
scribed.
Primarily the use of intelligence and memory tests will be
dealt with, but some consideration will also be given to the
perceptual and psychomotor tests.
All of the studies were cross-sectional, comprising either
a comparison between exposed and nonexposed subjects or
the determination of exposure-response relationships within
the exposed group, or both.
METHOD
Table 1 gives a general view of the studies that provided
the data for this paper.
In the carbon disulfide (CS,) study the exposed group
consisted of 50 exposed workers in a rayon fiber factory and
50 workers with manifest CS2 poisoning. Fifty nonexposed
workers from the same factory served as controls and they
were matched with the exposed subjects in regard to age,
length of employment and the type of job [4].
The subjects with exposure to solvent mixture were car
painters from 27 repair shops in Helsinki. The mean ex-
posure level corresponded to 32% of the Finnish threshold
limit value for solvent mixtures. Locomotive assistants and
engineers matched to the subjects according to age were
used as controls [6J
In both studies the group differences in mean perform-
ances were evaluated with /-tests for the differences in sepa-
rate variables, and by discriminant function analyses. In the
car painter study the effect of possible differences in the
initial intelligence levels was controlled by separate tests for
group differences in a subsample where car painters were
matched with respect to the preoccupational intelligence
level; the test results obtained during military service were
used in the matching.
The subjects with pure toluene exposure were photogra-
vure printers. Age matched controls were picked out from
the control group of the car painters [10].
Workers with styrene exposure were chosen from the
reinforced polyester plastic industry and their controls were
construction workers [9,11].
In each of the two latter mentioned studies analyses of
exposure-response relationships were included in the data
analyses. The individual toluene doses were evaluated on the
basis of exposure histories obtained through interviews and
noting the yearly consumption of toluene in the printing
shops. Pertaining to the styrene investigation the mandelic
acid concentrations in urine samples were used as a measure
of the exposure level.
In order to investigate lead exposure, workers were cho-
sen from two storage battery factories and one railway
engineering workshop. The controls were nonexposed indus-
trial workers. No individual matches were made but the
group structure of the controls corresponded to that of the
exposed groups in respect to age, sex, education and type of
job [5]. The effects of low lead exposure were studied in a
157
-------
158
HANNINEN
TABLE 1
OVERVIEW OF THE STUDIES
Exposure Agent
Carbon disulphide
Solvent mixture
Toluene
Styrene
Lead (total group)
Low exposure
Number of
Subjects
(50+) 50
100
26
98
88
49
Age
Mean
38
36
40
30
36
33
(years)
Range
25-50
21-60
27-55
17-55
22-58
22-53
Duration
Mean
11
15
20
5
8
6
of Exposure
Range
5-20
1-36
10-38
0.1-10
2-28
2-20
subgroup containing only workers whose blood lead levels
had been monitored during their entire exposure time and
had never exceeded the level of 70 ju,g/100 ml. Emphasis in
the data analyses was placed on the exposure-response rela-
tionship within the exposed group [7],
The Test Batteiy
A rather extensive test battery was used in all of the studies
[8]. The rationale for using a large test battery was the fact
that behavioral effects of chemicals must be considered by
and large unpredictable as long as little is known about the
main target sites and modes of action of various toxins in the
CNS.
The structure of the test battery remained principally the
same in all studies though the individual tests varied. All
of the test batteries included cognitive tests as well as tests
for perceptual and psychomotor functions. Table 2 presents
the tests most frequently used.
THE COGNITIVE TESTS
The intelligence tests were subtasks of the Wechsler
Adult Intelligence Scale (WAIS) [15], chosen because they
are internationally used, based on general population norms
and suitable for individual testing both with clinical and nor-
mal materials. Four subtasks of the Wechsler Memory Scale
(WMS) [16] and Benton's well-known visual memory test [1]
constituted the memory tests employed. They were chosen
for the same reason as the intelligence tests.
Similarities
(Sim) is a test for verbal concept formation. It consists of
13 orally presented paired words and the subject has to find a
similarity between the presented concepts.
Picture Completion
(PC) is a visual intelligence test that requires the subject
to discover a missing part of an incompletely drawn picture.
Block Design
(BD) is a test for visual abstraction and has been widely
used for detecting visual (spatial) disturbances in brain injury
patients. The subject is given 4 or 9 two-colored blocks
and for each of ten test items reproduces a two-colored de-
sign placed before him.
TABLE 2
THE TESTS USED TO MEASURE DIFFERENT PSYCHO-
LOGICAL FUNCTIONS
Test Category
Test
Intelligence
Memory
Sensory and Motor Test
Similarities (Wais)
Picture Completion (Wais)
Block Design (Wais)
Digit Span (Wais and WMS)
Logical Memory (WMS)
Associative Learning (WMS)
Visual Reproduction (WMS)
Benton Test
Bourdon-Wiersma
Santa Ana
Symmetry Drawing
Mira Test
The Digit Span
The Digit Span (DSp) test is included in both the WAIS
and WMS. It is more a memory test than a test of intelligence
as the task is to recall digit series immediately after hearing.
The Logical Memory
The Logical Memory (LogM) test measures verbal mem-
ory to things logically associated together. A short story is
read to the subject and he has to recall it.
Associative Learning
Associative Learning (Ass.L) consists of ten pairs of
words to be learned during three trials. In each trial the word
pairs are first read to the subject and then upon hearing the
first word of each pair he must recall the second.
To determine visual memory either the Visual Reproduc-
tion (Vis.R) from the WMS or the Benton test were used.
The Vis.R task contains three items, two of which have one
figure to be remembered and one with two figures to re-
member. The Benton test contains ten items, eight of which
contain three different figures to remember. The figures of
the Benton test are simpler than those of the Vis.R task.
-------
PSYCHOLOGICAL TEST METHODS
159
TABLE 3
SENSITIVITY OF THE COGNITIVE TESTS IN DIFFERENT STUDIES
Test
CS
Solvent
Mixture
Exposure Agent
Toluene
Styrene
Lead
Similarities
Picture Completion
Block Design
Digit Span
Logical Memory
Associative Learning
Visual Reproduction or Benton
- = no significant effect (+) = effect is slightly indicated + = effect is strongly indicated
When evaluating the effects as either slightly or strongly indicated, both the statistical significances of group
differences and the support given by additional statistical treatment (discriminant function analysis, factor analysis,
etc.) are taken into account. Exposure-dependent results are considered as stronger indications than the simple group
comparisons. For more accurate information of the empirical results, see referents [4, 5, 6, 7, 9, 10, 11].
COGNITIVE TEST RESULTS
Table 3 gives an overview of the sensitivity of the intelli-
gence and memory tests in the various studies conducted.
The car painters had the greatest number of statistically
significant test results whereas the styrene workers had
none.
The BD test which is generally considered to be sensitive
to CNS dysfunctions was among the three best dis-
criminators in the car painter study [6], and among the tests
correlated with the blood lead levels in the lead study [7]. In
contrast to this, PC and Sim are generally considered to be
more resistant to CNS damages. Group differences in these
particular tests must therefore be taken as warnings about
possible preoccupational intelligence differences between
the groups.
Concerning the car painter study this possibility was ruled
out and deemed highly improbable in the toluene study. In
the lead study, however, intelligence was a possible con-
founding factor in the group comparison, but not in respect
to the relationship found between the uptake level and per-
formances in the Block Design, Visual Reproduction and
Digit Span tests.
The memory disturbances were pronounced in the car
painter group, where all memory tests showed significant
impairment and two of them—the Associative Learning and
Digit Span tests—were among the three best discriminators
[6]. In other groups the statistical significance of results ob-
tained with the Digit Span test were at the 5% level (or else
nothing at all). The significance of results received from the
Logical Memory test followed these same trends as well.
The Associative Learning test was superior to the Logical
Memory test in the 2 studies where both of these tests were
employed. The results received from the Visual Reproduc-
tion tests were in line with the oral tests.
THE PERCEPTUAL AND PSYCHOMOTOR TESTS
One cannot evaluate the utility of cognitive tests in the
detection of toxic impairments without comparing their sen-
sitivity to that of other kinds of tests. For that reason four
tests measuring perceptual (visual) and psychomotor func-
tions, or the coordination of them, will be briefly discussed.
The Santa Ana test is a visual-motor speed test. The
equipment consists of a base plate with square depressions.
Each depression contains an accurately fitting peg with a
circular top. The task is to turn each peg in succession 180°
as fast as possible. This test requires both eye-hand coordi-
nation and coordination of ths wrist and finger movements.
The Bourdon-Wiersma test measures visual-motor speed
and accuracy. The test sheet has 50 rows each containing 25
groups of either three, four or five dots. The task is to draw a
line through all of the groups of four dots as quickly and
accurately as possible.
The Symmetry Drawing test is a method for detecting
disturbances in visual perception and visual-motor coordi-
nation. The cognitive functions are more involved in this test
than in the other perceptual motor tasks. Half of a figure of a
tree leaf is printed on the test sheet and the task is to draw
the other symmetric half. The number of reversions in the
figure has been the most used variable in this task.
The Mira Test is used to measure psychomotor behavior.
The task is to draw simple straight and broken lines without
visual control, i.e., without seeing the paper and pencil. The
subject sees the model at the beginning of each trial, but after
the first lines are drawn a screen is put between him and the
paper. The scores refer to the size of the drawing, to the
deviations from the model line and to the disorganization of
the movement pattern.
PERCEPTUAL AND PSYCHOMOTOR TEST RESULTS
Table 4 presents the results obtained by the perceptual
and psychomotor tests. It can be seen that the psychomotor
retardation was an essential part of ths CS2 effects. They
were clearly more pronounced in this group than were the
cognitive impairments [4].
It can also be seen that the effects of styrene were limited
to this behavioral domain: test variables measuring visual-
motor accuracy and psychomotor performance were the
-------
160
HANNINEN
TABLE 4
SENSITIVITY OF THE PERCEPTUAL MOTOR TESTS IN DIFFERENT STUDIES
Test
CS2
Solvent
Mixture
Exposure Agent
Toluene Styrene
Lead
Santa Ana
Bourdon-Wiersma, speed
Bourdon-Wiersma, accuracy
Symmetry Drawing
Mira
- = no significant effect (+) = effect is slightly indicated + = effect is strongly indicated
sensitive ones in respect to styrene exposure [9], The effect
was not a reduction in speed by a decrease in accuracy.
On the other hand the car painters with exposure to a
mixture of organic solvents showed less pronounced im-
pairments in the perceptual and psychomotor tasks than in
the cognitive ones.
The Santa Ana test yielded significant results in four of
the five studies. In two of the studies it was among the three
best detectors [4,7]. The Mira test proved valuable in all
studies where it was employed, though the best discriminat-
ing variable was not always the same.
DISCUSSION
The studies of long term effects due to occupational expo-
sure to toxics always produce more or less soft data when
compared with experimental studies made under laboratory
conditions, because some compromises and approximations
in the study design of field studies must always be made. If
the group comparison method is used, then the critical factor
is the matching of the controls to the exposed subjects. In
principle the match should take into account all possible fac-
tors that can affect the measured behavioral variables, or at
least the ones that are considered most essential. But in em-
pirical studies one often has to be satisfied with an approx-
imately similarity between the groups.
In determining exposure-response relationships the avail-
ability of reliable exposure data is the critical factor. Often,
however, only the current exposures are known and the
previous exposure levels and the total long term uptake must
be an approximate evaluation.
It must be assumed that possible false approximations,
poorly controlled confounding factors and intervening vari-
ables can, in these kinds of studies, produce noise that can
hide or obscure the real information being searched for.
The confounding factors, such as age when correlated
with the total exposure, or work loads connected with the
exposure or its absence, can bias the results. The intervening
variables, such as the different abilities and motivations of
the subjects who may compensate for their functional im-
pairments with increased effort, can mask the effect. The
comparison of results that are obtained in different settings
of subjects and by different study designs may help us to
recognize the real information from the noise.
At the present there is some cumulative evidence from
studies carried out in different countries about the usefulness
of cognitive tests in the detection of the early effects of lead
[2,14], carbon disulfide [12,13] and solvent mixtures [3] al-
though the results differ in respect to the'sensitivity of some
of the individual tests. Special consideration of the sources
of even minor discrepancies may be helpful in avoiding study
design weaknesses which can produce either false negative
or false positive results.
When discrepant results concern the sensitivity of certain
behavioral domains to toxic effects, the discrepancy may be
due to the different sensitivities of the tests used in different
studies. Increased conformity between the methods used in
the determination of toxic effects certainly would increase
the cumulation of consistent information about these effects.
This objective should not, however, discourage the efforts to
enrich the methodology with new methods that have proven
valuable in other domains of psychology.
The differing results in the studies reviewed in this paper
cannot, except for a few minor points, be explained by
methodological differences alone. Some of the differences
seem to indicate real differences in the effects caused by
different neurotoxic agents: the effects of styrene especially
as well as those of CSL, seem to differ from those of the other
toxic agents. This finding points to different modes of action
for these agents. Future research may throw more light onto
this question.
Nevertheless, the comparison of results of several empir-
ical studies reported here point to a relative unpredictability
of the behavioral effects we are searching for. This also
means that due to the present state of the available knowl-
edge in this area it is not wise to reduce the test battery only
to those methods that proved sensitive in some single study
on some single exposure, but to proceed with a varied psy-
chological methodology.
REFERENCES
1. Benton, A. L. The visual retention test: Clinical and experi-
mental applications. New York, NY: The Psychological Cor-
poration, 1972, p. 92.
2. Grandj'ean, P., E. Arnvik and J. Beckman. Psychological dys-
functions in lead-exposed workers: Relation to biological pa-
rameters of exposure. Scand. J. work envir. Hlt/i 4: 275-283,
1978.
-------
PSYCHOLOGICAL JEST METHODS
161
3. Hane, M., O. Axelson, J. Blume, C. Hogstedt, C. Sundell and
B. Ydreborg. Psychological function changes among house-
painters. Scand. J. work envir. Hlth 3: 91-99, 1977.
4. Hanninen, H. Psychological picture of manifest and latent car-
bon disulfide poisoning. Br. J. ind. Med. 28: 374-381, 1971.
5. Hanninen, H. Changes in performance, personality and subjec-
tive wellbeing as indicators of long-term exposure to toxic en-
vironments. In: Proceedings of the 7th International Congress
of Pharmacology, 5, edited by C. Dumont. Paris: Pergamon
Press, 1979.
6. Hanninen, H., L. Eskelinen, K. Husman and M. Nurminen.
Behavioral effects of long-term exposure to a mixture of organic
solvents. Scand. J. work envir. Hlth 2: 240-255, 1976.
7. Hanninen, H., S. Hernberg, P. Mantere, R. Vesanto and M.
Jalkanen. Psychological performance of subjects with low expo-
sure to lead. J. occup. Med. 20: 683-689, 1978.
8. Hanninen, H. and K. Lindstrom. Behavioral Test Battery for
Toxicopsychological Studies. Used at the Institute of Occupa-
tional Health in Helsinki, 2nd revised edition. Helsinki: Institute
of Occupational Health, 1979.
9. Harkonen, H., K. Lindstrom, A-M. Seppalainen, S. Asp and S.
Hernberg. Exposure-response relationship between styrene ex-
posure and central nervous functions. Scand. J. work envir.
Hlth 4: 53-59, 1978.
10. Kalliokoski, P. Toluene Exposure in Finnish Publication Photo-
gravure Plants. A thesis submitted to the faculty of the
Graduate School of the University of Minnesota. Minnesota:
University of Minnesota, 1978.
11. Lindstrom, K., H. Harkonen and S. Hernberg. Disturbances in
psychological functions of workers occupationally exposed to
styrene. Scand. J. work envir. Hlth 2: 129-139, 1976.
12. Schneider, H. Moglichkeiten der Psychodiagnostik bei
neurotoxischen Expositionen. In: Adverse Effects of Environ-
mental Chemicals and Psvchotropic Drugs, Vol. 2, edited by M.
Horvath. Amsterdam: Elsevier, 1976, pp. 187-196.
13. Tuttle, T., G. D. Wood and C. Grether. Behavioral and Neuro-
logical Evaluation of Workers Exposed to Carbon Bisulphide.
Cincinnati, Ohio: U.S. Department of Health, Education and
Welfare, 1976, p. 156.
14. Valciukas, J. A., R. Lilis, J. Eisinger, W. E. Blumberg, A.
Fischbein and I. J. Selicoff. Behavioral indicators of lead
neurotoxicity: Results of a clinical field study. Int. arch, occup.
envir. Hlth 41: 217-236, 1978.
15. Wechsler, D. The Measurement and Appraisal of Adult Intelli-
gence. Baltimore: The Williams and Wilkins Co., 1958, p. 297.
16. Wechsler, D. A standardized memory scale for clinical use. J.
Psycho!. 19: 87-95, 1945.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 163-174. ANKHO International Inc., 1979.
Quantitative Analysis of Rat Behavior Patterns
in a Residential Maze1
JUERG ELSNER, ROLAND LOOSER AND GERHARD ZBINDEN
Institute of Toxicology, Swiss Federal Institute of Technology and University of Zurich, CH-8603
Schwerzenbach, Switzerland
ELSNER, J., R. LOOSER AND G. ZBINDEN. Quantitative analysis of rat behavior patterns in a residential maze.
NEUROBEHAV. TOXICOL. 1: Suppl. 1, 163-174, 1979.—A method for monitoring spontaneous locomotor patterns of
rats during one day is described. The animals' locomotion is registered in a residential maze by 18 optical gates connected
to a computer. Status changes of each optical gate are stored on a disk file and can be retrieved for complete session
reconstruction and data analysis. The general features of a rat's behavior in the maze are discussed. Quantitative analyses
and statistical comparisons between two sessions spaced two weeks apart and between a group of 4 control animals and 4
rats treated in utero with methylmercury chloride are performed. Following parameters are analysed as functions of time
and maze location: locomotor and local activity, occupational duration and time per visit in the maze compartments.
Angular dependences of path decisions and regional preferences of crossings at the alley bifurcations are observed. No
changes of the measured parameters can be observed between the first and second sessions. Methylmercury treatment
results in a consistently lower local activity during the night period and in differences of path preferences.
Residential maze
Activity
Locomotion patterns
Behavioral toxicology
Methylmercury chloride
THE most critical stage in assessing the behavioral toxicity
of new substance is the first one, when nothing or little is
known about the effects on the behavior of living organisms.
Behavioral toxicology traditionally uses psychophar-
macological methods, which were developed to detect drug
actions. There exists a great variety of different tests measur-
ing very accurately specific symptoms resulting from in-
teractions with receptors and transmitter mechanisms. Few
methods have been described which give a first general idea
about the spectrum of possible effects on the behavior.
Experiments in behavioral toxicology should fulfill two
crucial requirements: the first is the stipulation that all types
of behavioral effects can be recognized. The other is that
non-specific disturbances of the nervous system, manifesting
themselves in more generalized and diffuse subtle symp-
toms, should also be detected and measured. Since behavior
is a very complex phenomenon that by itself cannot be de-
fined and assessed, one may conclude that the best approach
to behavioral toxicology would be to apply a battery of tests
consisting of all known and established behavior assessment
methods. Each test might then contribute to the overall in-
formation about the alteration of behavior. Such a procedure
would be time-consuming and costly, and could thus only be
applied to selected substances. A simple test procedure is
therefore needed, which would give relevant clues about the
types of behavior, if any, which may be altered by a chemical
substance. Selected follow-up experiments may sub-
sequently circumscribe more precisely the nature of the ob-
served effects. This article explores the usefulness of a resi-
dential maze procedure as an initial and general screen for
behavior alteration induced by chemical substances.
A residential maze is mainly a structured environment
with a certain degree of complexity, in which the animal's
behavioral interaction with its environment can be studied.
No specific task is asked to be performed. But since the
animal resides in the maze for a prolonged period of time, it
is inevitable that certain behavioral patterns will develop,
which are not likely to be random but structured and orga-
nized. The main task of the behavior analysis, therefore,
does not simply consist in measuring activity, but must at-
tempt to define the degrees of randomness and organisation
of the behavior. In the process of defining the behavior, the
animal itself is allowed to give the answer, and no predefined
criteria will influence ideas about what "normal" behavior
should be, and by how much these behaviors may vary. The
only relevant criteria will be the stability within or across
animals or animal groups.
One of the important advantages of such an experimental
procedure is the ability to obtain a large amount of varied
behavioral information from the animals in a toxicological
experiment. At the same time the manipulations of the ani-
mals by the experimenter is minimal.
METHOD
The Residential Maze
The design of the residential maze is based on the work of
Norton [4], who measured the activity of a group of 4 rats during
'This work was supported in part by the Swiss National Science Foundation.
163
-------
164
ELSNER, LOOSER AND ZBINDEN
WATER
FOOD
FIG. 1. The residential maze. The optical gate locations are indi-
cated by dotted lines, numbered in groups of 6: the blind alleys (1-6),
the circular alley segments (11-16) and the spoke alleys (21-26).
Roman numbers refer to the alley crossings confined by optical
gates. Blind alley 3 contains a water spout and 5 a wide wire mesh
with access to food pellets. At the alley end of number 1, there is an
opening at the top, measuring 2x9 cm. Its end-wall may be re-
moved.
several days in an eight-shaped maze crossed by a straight
alley, and on the work of Baettig [1,5], who studied the ex-
ploratory behavior of a single rat during 6 minute sessions in a
rather complex maze structure. In our residential maze the
activity patterns of only one rat at a time are observed for
periods of 24 hours.
The maze configuration is a compromise between Nor-
ton's residential maze and Baettig's exploration maze (Fig.
1). It has a sixfold rotation axis of symmetry. Six concentric
alleys emerge from the center and are connected near the
periferal ends by a hexagonal alley. An infrared optical gate
is located in each alley segment. It is connected to a digital
computer through an interface system described elsewhere
[2]. The alleys are 9 cm wide and 12 cm high.
The maze may be logically subdivided into 3 radial com-
ponents containing 6 individual compartements each, num-
bered in clockwise order: (a) the blind alleys containing gates
1 through 6, (b) the circular alley segments with gates 11
through 16, and (c) the inner spoke alleys with gates 21
through 26. Likewise 6 circumferential compartements may
be defined. They are numbered I through VI. A seventh
compartement (VII) occupies the center of the maze and is
confined by gates 21 through 26.
Three blind alleys are empty (gates 2, 4, and 6), the other
contain a water spout (gate 3), a wide wire mesh letting ac-
cess to food pellets (gate 5) and one has an opening at the top
measuring 2x9 cm (gate 1). The front wall in alley number 1
can be removed for the introduction of the rat into the maze.
The maze is manufactured of 3 mm thick aluminum sheets
screwed at the top to a 6 mm thick Plexiglas plate. This
assembly is fixed on one end to a wooden plate by two
hinges, allowing it to be lifted for removal of the rat and for
cleaning at the end of the session. The floor is covered by a
thin, easily removable hardplastic sheet, which is exchanged
after each session for cleaning purposes. The maze can be
illuminated for daylight simulation by two computer-
switched lamps of 15 W each, located above gates 12 and 15.
The maze can be rolled on two slides into a wooden box,
providing about 20 dB of sound attenuation. The alluminum
sheets defining the maze boundaries contain round holes 5
mm wide for ventilation. In addition, the wooden enclosure
is ventilated by a fan.
Animals and Experimental Procedure
Two groups of 4 male ZUR:SIV-Z rats were used, each
group belonging to one litter. One group (animals Cl through
C4) was untreated. The dam of the other (aminals HI
through H4) was treated per os on days 6 through 9 of gesta-
tion (the day of conception being counted as day zero) with 2
mg/kg methylmercury chloride, a dose reported to have no
effect on general motility levels and motor coordination of
the pups as evaluated by swimming test, but impairing to
some extent DRH performance at the age of 90 days [3].
Both groups were previously used for an ultrasound vocali-
zation study, being isolated from their littermates for 1 min-
ute per day on days 1 through 20 of their life, and on day 21 in
a taste preference test (saccharine solution and tap water).
Thus, the animals may be considered as naive with regard to
the maze exposure, but they were handled daily in the first 3
weeks of their life.
The experiment started when the rats were 60 days old
and lasted for 4 weeks. Each rat visited the maze twice for
one day, the visits being spaced 2 weeks apart. A maze ses-
sion started at 4 p.m. in the afternoon and lasted until 3:50
p.m. of the following day. The following schedule was used
for both visits:
Mon Tue Wed Thur Fri
1st Week Cl C2 HI H2
2nd Week C3 C4 H3 H4
One week before and during the whole experimental proce-
dure the animals were housed in individual cages. Weight
gain and water and food consumption were monitored.
A rat was introduced at 4 p.m. into the illuminated maze
and left undisturbed for 23 hr 50 min. At 7 p.m. the light was
shut and lit again at 7 a.m. Every event as recorded by the
optical gates was recorded by a PDP11/34 computer and
stored on a disk file. Each record in the file contained the
time of day in a resolution of 1/50 of a second, the gate
number and its polartiy (break or release). The computer-
controlled on and off switch of the light was also contained in
the data stream.
The animal's weight before and after the maze session
and its water and food consumption in the maze were meas-
ured. The defecation pattern was monitored by sketching
onto a diagramm each defecation bolus as found on the floor
after a day's residence in the maze.
RESULTS
Baseline Behavior in the Maze
A typical computer output of one day of a rat's locomotor
activity in the residential maze is represented in Fig. 2. The
24 hours are divided into 12 strips of 2 hours each, labeled
with the time of day. The optical gates are grouped in the
-------
RAT BEHAVIOR PATTERNS IN A RESIDENTIAL MAZE
165
129
MINUTES
FIG. 2. Activity history of a one day session in the residential maze. The individual characters in the label Cl 1 refer to the
group (C), the animal number (1) and the session number (1). The time is represented in the horizontal axis in strips of 2
hours each. The optical gates are grouped in the vertical axis according to the maze regions I-VI defined in Fig. 1 (e.g.
region I contains gates 21, 1, and 11, represented upwards in this order). Above each strip, a heavy line indicates that the
light is on, and a thin line that it is off.
-------
166
ELSNER, LOOSER AND ZBINDEN
ordinate according to areas I through VI limited by the dot-
ted lines. Within each gate group, the lower point corre-
sponds to the blind alley, the higher to the spoke, and the one
on the dotted line to the gate in the circular alley. The repre-
sentation is circular in that the transition curves from region
I to VI and from VI to I are not shown. Thin lines represent
gate transitions and heavy lines visits in the gates. The
straight line above each strip indicates the illumination status:
a heavy line means the lights are on, and a thin line means
they are off.
This representation allows a qualitative inspection of a
rat's activity in detail. A rising curve showing visits in each
region represents a clockwise walk through the circular al-
ley, and an identical falling curve indicates a coun-
terclockwise path. The crossing of one or more areas without
any visits in them, indicates a crossing of the center square
through the spoke of alleys. Backturns can be spotted
through slope inversions.
A one day session can be somewhat artificially sub-
divided into four main periods: the exploration phase be-
tween 4 and 5 p.m., followed by the post-exploration phase
from 5 to 7 p.m., the nocturnal phase from 7 p.m. to 7 a.m.
and the diurnal phase from 7 a.m. to 3:50 p.m.
Besides showing a very high locomotor activity, the
exploration phase is characterized by a high randomness of
behavior. Every location in the maze is visited often during
this time (in the order of 10 to 30 times each) and almost
every path from one gate to another is walked through at
least once. Also the timing of the activity reflecting the rat's
walking speed is very irregular.
In the post-exploration period, the rat returns back to its
diurnal rest, only occasionally interrupted by bursts of
higher activity lasting up to 10 minutes. For resting the rat
generally retires into a blind alley (most often blind alley
number 4), the head turned towards the maze center.
The beginning of the nocturnal phase is almost indistin-
guishable from the post-exploration behavior. After 2 to 3
hours, a more or less sustained activity can be observed,
showing repeated visits of blind alley 3 presumably for drink-
ing and blind alley 5 for eating. A very high local activity (as
defined by successive crossings of the same gate) can be
observed in the food compartment, probably induced by eat-
ing activity. Also the blind alley 1 with the opening at the top
is visited very frequently during this time, showing a high
local activity as well. Before the switch-on of light, the ac-
tivity is lower again, rising steeply right at the end of the dark
period.
The begin of the diurnal phase is characterized by a
period of very high activity, comparable to the one during
exploration. Closer inspection of sequential features
suggests a qualitatively different kind of behavior as com-
pared to the exploration activity. It looks more orderly or
even stereotype, containing for example rapid successive full
circles around the circular alley.
The high activity early in the morning is followed by a
very rapid decline of activity, resulting in a calm diurnal
phase, which is only occasionally interrupted by short activ-
ity bursts. It is noteworthy that the choice of the sleeping
compartement seems to be more or less definitive, as after
short excursions the animal returns invariably back to the
blind alley it has left.
Quantitative Analysis
In order to get a quantitative evaluation, the behavioral
LOCOMOTION
Cll
FIG. 3. Three-dimensional histogram representing the locomotor
activity of Cll in function of time and gate number. The height of
each cell is proportional to the number of visits in a gate for one
hour. The gates are grouped into blind alleys (gates 1-6), circular
alley segments (gates 11-16) and spoke alleys (gates 21-26). The
vertical axis is scaled in proportion to the total visits in all gates and
during the whole day. The background "walls" represent the sum of
the values they are facing to. The lines are thick as long as the light is
on, and thin, when it is off.
data in the maze have been analysed by several statistical
procedures. The first analysis concerns locomotor activity
defined as the number of gate crossings excluding repetitions
of the same gate. The locomotor activity of the first session
of animal Cl is represented in Fig. 3 as a function of the time
of day and the gate number. The height of each cell in this
three-dimensional histogram is proportional to the number of
visits in a gate for one hour. The ordinate axis is scaled as
percent of the total of visits in all gates and during the whole
day. The background "walls" represent the sum of these
values, the left wall showing the sum over all gates, and the
back wall the sum over the whole day. The lines are drawn
heavily, as long as the lights are on, and lightly when they are
off.
The time evolution demonstrates quantitatively what has
been discussed in the preceding section. The distribution
over the gates shows that the rat spreads its activity rather
evenly over the whole maze. This even distribution is mod-
ulated in a way which locates most locomotor activity in the
circular alley (gates 11-16), and less in the spokes (gates
21-26) and blind alleys (gates 1-6). Focusing attention only
-------
RAT BEHAVIOR PATTERNS IN A RESIDENTIAL MAZE
167
LOCflL ACTIVITY 6'
Cll
FIG. 4. Three-dimensional histogram representing the local activity
(defined as the number of repetitive breaks and releases of the same
optical gate) of Cl 1 in function of time and gate number. For repre-
sentational details refer to Fig. 3.
6*-
LOG flCT / VISIT f
Cll
FIG. 5. Three-dimensional histogram representing the mean local
counts per visit of Cll in function of time and gate number. This
figure is calculated as the quotient of the values in Fig. 4 divided by
those represented in Fig. 3. The mean values in the back "walls" are
magnified six times in order to be seen. For representational details
refer to Fig. 3.
onto the blind alley visits, the gate 1 was visited markedly
more often than the other five. Within the other blind alleys,
food (gate 5) and water (gate 3) locations are next in fre-
quency of visits. Differences within the other areas of the
maze show also a concentration of the visiting rates around
area I (gates 11, 16, and 21). In the topography of the tem-
poral evolution of these visiting frequencies, one can ob-
serve that this preference becomes more pronounced in later
stages of the experiment. This suggests that occupational
habits are created during the maze session.
The next analysis focuses on the local activity, defined as
successive breaks and releases of the same gate. Figure 4 is
structured in the same way as Fig. 3. Compared to the
locomotion, the local activity shows much more marked
differences between the values in different location in the
time-gate space. However, the local activity roughly paral-
lels in time the locomotor activity. Additional high peaks of
local activity during otherwise relatively calm night periods
constitute the main difference between the two measures.
These peaks are mainly due to feeding activity in the blind
alley 5 and some "sniffing" activity in location 1.
The parallelism of the local and the locomotion activity
counts is demonstrated in Fig. 5, which represents the
number of local counts per visit as a function of time and
maze location. Besides markedly different values in the blind
alleys and random fluctuations over time, local activity
counts per visit are constant.
The next analysis pertains to the time spent in the indi-
vidual maze locations in proportion to the total time avail-
able (Fig. 6). The time spent outside a gate was evenly dis-
tributed between the areas defined by the last visited gate
and the next one. The figure shows that the time spent in the
areas of the circular (gates 11-16) and the spoke (gates 21-26)
alleys is relatively constant and low. Most time is spent in
the blind alleys (gates 1-6) during the whole session, exclud-
ing the exploration phase, when the occupational time is
more evenly distributed among all maze areas. Thus, in con-
trast to the fact that the blind alleys are visited at about same
or at even lower frequencies than the rest of the maze,
these visits last much longer. This fact is represented in Fig.
7, where the mean time per visit is plotted, showing a large
difference between the blind alley values (gates 1-6) and
others.
The behavior pattern of single animals can be observed
-------
168
ELSNER, LOOSER AND ZBINDEN
I
s;
I
DURflTION
Cll
TIME PER VISIT )6'
Cll
FIG. 6. Three-dimensional histogram representing the occupational
duration in the maze compartments surrounding the gates, in pro-
portion to the total time available. For representational details refer
to Fig. 3.
FIG 7. Three-dimensional histogram representing the mean time
spent per visit in the compartments surrounding the optical gates.
This figure is calculated as the quotient of the values in Fig. 6 di-
vided by those represented in Fig. 3. For representational details
refer to Fig 3.
also in pooled data of a group of animals. These values,
besides having been smoothed to a certain extent by the
pooling, show the same characteristics as those of single
animals. Moreover, the behavior of rats in the maze exhibit a
remarkable stability. This is illustrated in Figs. 8 and 9,
where the time course of the mean locomotor and local ac-
tivities of the same 4 animals are plotted, comparing the data
of the first session with the second session two weeks later.
The heavily drawn mean values are accompanied by lightly
drawn confidence limits weighed in such a way, as to show
f-test comparison results: no overlap of the two areas means
a statistically significant difference at the 0.05 level. These
plots demonstrate that even on a hourly basis and in consid-
ering separately the three main radial maze areas, the behav-
ior in the maze is reproducible. This fact is confirmed also by
paired z-tests made for all discussed parameters, in classify-
ing the data into the three radial compartments and into the
four main periods of the sessions. These tests result in fairly
high probabilities of null hypothesis rejection error prob-
abilities (allp>20% and mostp>50%).
The next analysis studies the frequency distributions of
the time differences between two successive breaks and re-
leases of alternative gates (locomotion) and of the same gate
(local activity). Figure 10 shows that the frequency of time
differences decreases with increasing intercount time. The
shape of the curves demonstrate that they can not be ex-
plained by simple Poisson statistics, since the logarithmic
transformation of the relative frequency distributions did not
result in a straight line. However, the superposition of two or
more exponential distributions would result in the observed
shape. This fact indicates that the animal may be in two or
more different distinct states of activity, each following a
Poisson distribution with its own characteristic mean transi-
tion and repetition time. More detailed analysis is needed to
study these aspects and to separate out distinct states of
behavior.
The last study of the maze data deals with the relative
frequencies of transition between the gates. Due to the maze
geometry, not every transition is possible. Table 1 shows the
matrix of transition counts in the first session of rat Cl. In
paranthesis are those observed during the first hour of explo-
ration. Impossible and null transitions in the main diagonal
are left empty.
The transition frequencies can be ordered in several man-
-------
RAT BEHAVIOR PATTERNS IN A RESIDENTIAL MAZE
169
LOCOMOTION
CONTROLl
COUNTS/MIN)
CONTROL2
LOCRL RCTIVITY (COUNTS/MIN)
CONTROLl CONTROL2
TIME OF DRY
FIG. 8. Mean locomotion counts per minute (heavy lines) and confi-
dence limits (thin lines) for two sessions spaced two weeks apart for
the control group. The values for one hour, calculated as the mean of
the 4 control rats is represented as plateaus in the curves. A heavy
line on the time axis indicates that the light is on. The top figure
shows the values of all gates, and the three lower ones the values of
the three radial compartments (blind, circular and spoke alleys from
top to bottom). Note the difference in scales between the top and
other graphs.
ners. Two ways of ordering are discussed here. The first,
illustrated in Fig. 11, analyses the decision behavior at
different types of bifurcations. The diagrams at the bottom of
Fig. 11 have to be viewed as representing the decisions about
which way to follow, if the animal is located in the bottom
compartment and is proceeding upwards. The bar graphs
above each diagram represent the corresponding decision
frequencies, pooled over the six symmetrical situations in the
maze. The top four hashed bar graphs are from the first
sessions of the individual rats Cl through C4 and the black
bar graphs at the bottom represent the pooled transition fre-
quencies. Note the even distribution in the first case, the
mirror symmetry between the second and third case and the
remarkably symmetric distributions in the fourth and fifth
case of the pooled data. These bar graphs indicate clearly an
angular dependency of the bifurcation decisions, i.e. the ob-
tuse angles are preferred. The only deviation of an angular
24100 04100
TIME
DRY
FIG. 9. Mean local activity counts per minute for two sessions
spaced two weeks apart. For representational details refer to Fig. 8.
function is seen in the first case. It may be explained by the
fact that this situation results from a turn-around in the blind
alley.
A second way of ordering the transition frequencies is
according to the crossings of the seven squares I through VII
and is represented in Fig. 12. In accordance with the
locomotor activity, squares I, III, and V are preferentially
crossed. The central square VII is crossed about as often or
even less often than all the others, although there are more
alleys leading to it. This finding may be explained by the
observed angular dependency of bifurcation decisions.
Both orderings of the frequency distributions are rela-
tively stable in comparing the first and second sessions of the
control group. A chi square analysis of contingency tables
results in high probabilities of rejection errors of the null
hypothesis (all but one p>10% and half of them >50%).
Two measures remain to be presented briefly: the defeca-
tion pattern in the maze as compiled at the session end is
represented in Fig. 13. It demonstrates the defecations per
area plotted against radial distances from the center. The
accuracy of these data may be confounded by coprophagia.
Nevertheless, a very consistent steep and monotonously fall-
ing curve as a function of the distance from the center re-
sults. Thus, although the center of the maze is less often
-------
TABLE 1
Cll TRANSITION FREQUENCIES
21 22 23 24
58 (7) tt
17 (3) 21 (5)
20 (3) 27 (10)
8 (4) 12 (3)
23 (4) 43 (2)
19(4) 26(1)
11 63(3) 9(0) 43(14) 22(6) 13(4) 15(3)
12 26(8) 21(6) 42(5) 2916) 10(3) 11(4)
13 25(11) 17(3) 27(4) 24(6) 16(4) 6(2)
14 20(7) 29(4) 24(7) 17(5) 7(3) 12(1)
15 44(4) 17(3) 29(10) 28(6) 11(3) 7(1)
16 46(6) 27(1) 28(13) 35(8) 12(2) 13(0)
2* 61(9) 12(4) 18(5) 2(0) 9(0) 20(4) 13(2) 2(0)
22 22 (4) 17 (2) 16 (4) 1 (0) 1 (0) 10 (4) 15 (1) 5 (1)
23 31(3) 12(2) 4(0) 10(2) 2(0) 3(1) 10(2) 20(9)
24 22(7) 13(3) !0(4) 22(5) 8(1) 6(3) 0(0) 5(1)
25 37(3) 11(0) 16(5) 14(2) 21(4) 7(0) 1(0) 1(0) [—'
. _ (/j
26 IS (2) 8 (3) 9 (5) I (0) 10 (1) 12 (1) 10 (1) I (0) Z
W
Frequency matrix of transitions between gates. The rows order the data according to the originating gates, and the columns according to the
final gates. Impossible and null transitions (the main diagonal) are left blank. In parentheses are the values of the first hour only. Q
O
Cfl
W
X3
>
Z
O
N
2
3
o
w
z
-------
RAT BEHAVIOR PATTERNS IN A RESIDENTIAL MAZE
171
Cl 1
LOCOMOTION
LOCRL RCTIVITY
2 1 6 8 18 12
INTER COUNT TIME (SEC)
FIG. 10. Frequency distribution and logarithmically transformed
relative frequency distributions of inter-count times for locomotor
(top) and local activity counts (bottom), calculated from the whole
one day session Cll. The values are recorded at a resolution of 20
ms but represented at a resolution of 0.1 sec.
visited and the time spent in it is much lower than in other
locations, it elicits the most defecations.
During the maze sessions the rats lose up to 20 g of
weight, although food and water consumption is of the same
order of magnitude as compared to home-cage data.
C*l
FIG. 11. Relative bifurcation frequencies of the four individual con-
trol rats (upper bar graphs) and for the pooled data (bottom) calcu-
lated from the whole one day sessions. The diagrams below the bar
graphs indicate the represented decisions about which way to fol-
low, if the animal is located at the bottom compartment and is pro-
ceeding upwards. The bar graphs are ordered according to the result-
ing path directions in the diagrams. The values are pooled over the
six symmetrical situations in the maze.
The Behavior of Methylmercury-Treated Rats in the Maze
Group comparisons of locomotor and local activities be-
tween control and treated groups of both maze sessions are
shown in Figs. 14 and 15. The locomotor activity was not
affected, while a consistent drop in the nocturnal local activ-
ity in the blind alleys and in the spokes can be observed. A
two-factor analysis of variance was performed with the
locomotor and local activities, with the occupational dura-
tions and with the time per visit values, one factor being the
sessions and the other the treatment. No comparison re-
sulted in a difference above chance level in either factor.
The transition frequencies of methylmercury treated ani-
mals are represented in Figs. 16 and 17. They differ signifi-
cantly from the controls at both experimental sessions, as
analysed by a chi square contingency table procedure: all
but one of the 12 comparisons resulted in ap<5%, and half of
them in ap<\%. This finding has to be confirmed by a repe-
tition of the experiment.
DISCUSSION
The described experimental procedure for the assessment
of psychotoxic effects on the behavior of rats was chosen
mainly for the potentially high information content of the
resulting data. To monitor all actions of an animal in the
structured environment of the residential maze during a pro-
longed period of time, is an interesting method for behavioral
research in itself. In order to be useful for toxicology, the
measured parameters have to be stable and reproducible and
at the same time highly sensitive to effects caused by chemi-
cal agents. It is also desirable that the resulting data are well
interpretable with respect to underlying basic mechanisms of
behavior, in order to lead the way to subsequent closer in-
vestigation of observed effects with more specific methods.
As far as stability and reproducibility are concerned, the
study demonstrates that successive sessions of the same
animal in the residential maze result in similar behavior,
measured by locomotor and local activity, occupational du-
rations in different maze compartments (absolute and per
visit), and by several measures of compartment transition
frequencies. This finding shows that the residential maze
procedure permits longitudinal behavioral toxicology studies
in which the animals can be used as their own controls.
-------
172
ELSNER, LOOSER AND ZBINDEN
C*l
II III IV V VI VII
C*l
12345
R CflRBITRflRY UNITS)
FIG. 12. Relative crossing frequencies of areas I through VII as
defined in Fig. 1. The values are calculated over the whole one day
sessions of the four control rats (upper bar graphs) and of the data
pooled over the group (bottom).
FIG. 13. Normalized frequency distribution of the number of defe-
cation boli per area found on the floor at the end of the session for
the four control rats. The radial distance from the center R takes
following values: 1 for the center area, 2 for the inner spoke alley
segments, 3 for the outer spoke alley segments, 4 for the circular
alley and 5 for the blind alleys. Upper bar graphs represent the
individual values and the lower graph the values pooled over the
group.
-------
RAT BEHAVIOR PATTERNS IN A RESIDENTIAL MAZE
173
LOCOMOTION
CONTROL
COUNTS/MIN)
MMC
LOCflL RCTIVITY (COUNTS/
CONTROL MMC
20.60
2MI00 0Mi00 08i00 12i98
TIME OF DRY
16.00
TIME OF DRY
FIG. 14. Comparison between the mean locomotion counts per
minute of the two sessions of the four control rats, and those of the
methylmercury treated rats. For representional details refer to Fig. 8.
FIG. 15. Comparison between the mean local activity counts per
minute of the two sessions of the 4 control rats, and those of the
methylmercury treated rats. For representational details refer to
Fig. 8.
-------
174
ELSNER, LOOSER AND ZBINDEN
H*l
FIG. 16. Relative
treated rats.
bifurcation frequencies of the four methylmercury
For representational details refer to Fig. 11.
H*l
I II III IV V VI VII
FIG. 17. Relative crossing frequencies of the four methylmercury
treated rats. For representational details refer to Fig. 12.
REFERENCES
1. Baettig, K. and H. U. Wanner. Die spontane Lokomotion von
Ratten in einer Kombination von Labyrinthgaengen und Offen-
feld. Helv. Physiol. Ada 25: 249-261, 1967.
2. Eisner, J. and R. Wehrli. Interface systems in behavioral re-
search. Behav. Res. Meth. Instrum. 10: 259-263, 1978.
3. Muesch, H. R., M. Bornhausen, H. Kriegel and H. Greim.
Methylmercury chloride induces learning deficits in prenatally
treated rats. Archs Toxicol. 40: 103-108, 1978.
4. Norton, S., B. Culver and P. Mullenix. Development of noctur-
nal behavior in albino rats. Behav. Biol. 15: 317-331, 1975.
5. Uster, H. J., K. Baettig and H. H. Naegeli. Effects of maze
geometry and experience on exploratory behavior in the rat.
Anim. Learn. Behav. 4: 84-88, 1976.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 175-178. ANKHO International Inc., 1979.
Comparison of Neurobehavioral Effects
Induced by Various Experimental Models of
Ataxia in the Rat1'2
F. B. JOLICOEUR, D. B. RONDEAU, A. BARBEAU
Department of Neurobiology, Clinical Research Institute of Montreal, 110 Pine Avenue West, Montreal, Quebec
AND
M. J. WAYNER
Brain Research Laboratory, Syracuse University, 601 University Avenue, Syracuse, NY 13210
JOLICOEUR, F. B., D. B. RONDEAU, A. BARBEAU AND AND M. J. WAYNER. Comparison ofneurobehavioral effects
induced by various experimental models of ataxia in the rat. NEUROBEHAV. TOXICOL 1: Suppl. 1, 175-178, 1979.—The
purpose of the present study was to design a standard battery of tests capable of quantitatively characterizing ataxia and
concomitant neurological signs in the rat. In addition to a systematic analysis of the walking gait of animals, tests for
activity, catalepsy, rigidity and various reflexive responses were included in the battery. The standardization of the test
system was performed by determining and comparing neurobehavioral effects produced by 3-acetyl pyridine, acrylamide,
pyrithiamine and thiamine deficiency, four experimental treatments reported to induce ataxia in animals. Results indicate
that profiles ofneurobehavioral disturbances accompanying ataxia in animals varied distinctively with each experimental
treatment.
Ataxia 3-Acetyl pyridine
Neurobehavioral toxicology
Thiamine deficiency Pyrithiamine Neurobehavioral effects
ATAXIA frequently results from administration of
neurotoxicants to experimental animals. However a stan-
dard method and procedure for detecting and measuring
ataxia in experimental animals has not been developed.
Usually, reports of ataxic symptoms in animals are based on
qualitative observations. Possible concomitant neurological
damage is rarely assessed. The purpose of the present study
was to devise a standard battery of tests capable of quan-
titatively characterizing ataxia in the laboratory rat. The
selection of the various tests was based on their proven abil-
ity to measure neurobehavioral changes induced by drugs or
other manipulations in animals The standardization of the
test battery was carried out by determining and comparing
profiles of the neurobehavioral effects produced by 3-acetyl
pyridine, acrylamide, thiamine deficiency and pyrithiamine,
four treatments known to induce ataxia in animals.
A single injection of 3-acetyl pyridine in rats (75mg/kg)
produces within 24 hours signs of cerebellar ataxia and dam-
age to the medulla oblongata and climbing fibers of the cere-
bellum of rats [2]. Microscopic examination of the CNS re-
veals lesions as early as 7 hours after injection [3]. The ataxia
resembles, both histologically and biochemically the
olivocerebellar atrophy originally described by Holmes [7].
Recent studies have demonstrated that the lesions are as-
sociated with significant changes in the levels of certain
animo acids in specific regions of the CNS [1].
Chronic administration of acrylamide in doses of 10-50
mg/kg in animals results in peripheral neuropathy char-
acterized behaviorally by proprioception impairments,
hindlimb paralysis, and progressive ataxia [6,10]. His-
topathological examinations reveal distal axonal degenera-
tion of peripheral motor and sensory nerve cells [16]. The
neuropathy is first seen with a cumulative dose of 400 mg/kg
and is most prominent with a cumulative dose of 500 mg/kg
[4].
Chronic deficiency of vitamin B, results in pervasiva
metabolic and biochemical alterations in the nervous system.
Thiamine deficiency induces a peripheral neuropathy of the
"dying back" type which involves both sensory and motor
nerve fibers [13]. When rats are chronically fed a thiamine-
'The authors gratefully acknowledge the contribution of Dr. R. F. Butterworth, E. Hamel, F. Belanger and S. Gariepy. F.B.J. and D. B. R.
were supported by the conseil de la Recherche en Sante du Quebec and the Medical Research Council of Canada respectively.
2Requests for reprints should be addressed to: Dr. Andre Barbeau, M.D., Department of Neurobiology, Clinical Research Institute of
Montreal, 110 Pine Avenue West, Montreal, Quebec, Canada H2W 1R7.
175
-------
176
JOLICOEURETAL.
free diet, a variety of neurobehavioral disturbances such as
anorexia, piloerection, tremors, hypokinesia and ataxia de-
velop at about 30-40 days from the start of the diet [12,19].
Chronic administration of pyrithiamine, an antimetabolite
of thiamine, results in central histopathological changes
mostly localized in the pons and medulla oblongata [19].
Pyrithiamine also causes axonal degeneration of peripheral
nerves [19]. Manifestations of ataxic symptoms have been
observed in animals injected daily for 18 days with 0.5 mg/kg
pyrithiamine [5].
METHOD
Animals
Fifty four male Sprague Dawley rats, 275-350 g in weight,
were used. They were divided into nine groups of six animals
each. Food, which consisted of standard Purina Rat Chow,
and water were available ad lib except when specified in the
procedure.
Procedure
For 3-acetly pyridine, two groups of animals were used.
One group received an acute intraperitoneal injection of 75
mg/kg 3-acetyl pyridine dissolved in physiological saline.
Volumes of injection were 1 ml/kg. Animals in the control
group were injected with an equal volume of saline. Animals
were tested for ataxia and other neurological symptoms at 6,
12, 24, 48, and 72 hours following injection.
Two groups of animals were included in the acrylamide
model. Animals in the experimental group received ten suc-
cessive daily injections of 50 mg/kg acrylamide. Control
animals were injected with saline. Acrylamide was dissolved
in physiological saline and administered intraperitoneally in
a volume of 1 ml/kg. Starting on the second day of ac-
rylamide administration, neurobehavioral tests were carried
out 30 minutes following the daily injection procedure.
For thiamine deficiency, one group of rats was given a
thiamine-free diet (ICN Life Sciences, Nutritional Biochem-
ical) throughout the experiment. Since this regime results in
hypophagia with ensuing body weight losses, animals in the
control group received standard rat chow in daily rations
equivalent to the amounts consumed by the thiamine deficient
animals. Neurobehavioral tests were performed on days 7,
14, 21, 27, 30 and 33: starting on day 35, tests were carried
out daily until day 44.
To study the effects of pyrithiamine, three groups of ani-
mals were given the thiamine free diet and assigned to one of
the following experimental treatments as described by Gubler
[5]. Pyrithiamine treated rats were administered 100 /ag/kg of
thiamine and 0.5 mg/kg of pyrithiamine. Thiamine deficient
animals received saline and control animals were injected
with 100 /u-g/kg of thiamine. These treatments were given
daily for 18 consecutive days. All substances were dissolved
in saline and administered subcutaneously in volumes of 2
ml/kg. Animals were tested for neurological symptoms daily
throughout the experiment.
The neurobehavioral effects induced by the various
treatments were assessed by means of the following tests.
The tests were performed in the order they are listed here. A
more detailed description of testing procedures can be found
in the original report of this study [9].
Locomotor activity. Spontaneous locomotor activity was
measured for two minutes by means of a photocell activity
apparatus (Lehigh Valley Electronics).
Catalepsy. Intensity of catalepsy was determined by plac-
ing the animal's front paws on a horizontal bar (1 cm in
width) suspended 10 cm above the table. Time spent in that
position, up to a maximum of 60 seconds, was recorded.
Rigidity. The rat was suspended by its front paws grasp-
ing a metal rod (0.5 cm diameter) which was held by the
experimenter about 50 cm above the table. The time the
animal remained on the bar (maximum 60 sec) was recorded.
A prolonged grasping response has been correlated with di-
rect measures of muscle rigidity [17].
Landing foot spread. After staining the hindpaws with
ink, the animal was held horizontally 30 cm above a table
covered with absorbent paper. The rat was dropped and the
distance between the prints of each hindlimb was measured.
This procedure has proved to be useful in detecting
peripheral neuropathy in rats [11].
Gait analysis. After staining the hindfeet with ink, the
animal was walked through an enclosed 90 cm long corridor
with a paper covered floor. When two consecutive strides
were obtained, the stride width, length and angle between
consecutive steps on contralateral sides were calculated ac-
cording to the procedure of Lee and Peters [11].
Reflexive responses. The presence of the righting and
corneal reflexes as well as of a normal reaction to tail pinch
was verified according to the procedure of Irwin [8]. The
animal's ability to shift its weight during gravitational pull
and the position of its hindlimbs when held vertically were
then checked [15]. Subsequently, the presence or absence of
a normal extension of the forelimbs when the animal, held by
the tail, was lowered briskly toward a table top was also
recorded [9]. Then, the animal, placed on a table, was lifted
by the tail and the presence of a normal extension of the
hindlimbs was noted [9]. Finally in the traction test, the ani-
mal was held by the tail and pulled horizontally on a table:
the presence of a spontaneous hunched posture was record-
ed [9].
RESULTS
On each model, data obtained on activity, catalepsy,
rigidity, landing foot spread and the three gait components
were analysed by individual ANOVA's and the appropriate
post hoc tests [18]. Results obtained on the various reflex
tests were analysed by means of Fisher Exact Probability
test [14]. In all cases, a difference between groups was con-
sidered significant if it had a probability of random occur-
rence of less than 5 percent.
For 3-acetyl pyridine, results chained at 6, 12, 24, 48 and
72 hours were included in the statistical analyses. It was
found that 3-acetyl pyridine treated animals displayed more
catalepsy and muscle rigidity, and had larger landing foot
spreads than control animals at each of the five test periods.
The treated animals were also found to be significantly less
active than controls 6 hours after the injection but not at the
other test periods. For the three components of gait analysis,
no significant difference between the groups was found 6
hours after injection. However starting at 12 hours and for
the remainder of the post injection test periods, treated ani-
mals were consistently ataxic as revealed by significantly
smaller angles and stride lengths as well as by larger widths
between steps. Analyses of the results obtained with the
various reflex tests indicated the following significant ef-
fects. At 24, 48 and 72 hours, treated animals had lost the
righting reflex and the ability to maintain a normal hunched
posture during the traction test. Starting at 12 hours and
enduring for the remainder of the experiment, treated ani-
-------
NEUROBEHAVIORAL TESTS AND ATAXIA
177
mals displayed an abnormal hindlimb position characterized
by the feet being retracted and held closely to the body.
Finally a disturbance in the weight shift response of treated
animals was found at the 72 hours post injection test period.
For acrylamide, data obtained in each of the nine daily
test periods were included in the statistical analyses. In
comparison to controls, acrylamide treated animals man-
ifested significantly higher scores of catalepsy in all nine test
periods. They also displayed significantly larger landing foot
spreads than control animals in the fourth, sixth, seventh,
eighth and ninth test period. Significant gait disturbances in
acrylamide treated animals were found in all test periods
including the first test period following acrylamide adminis-
tration. The ataxia was characterized by significantly smaller
stride angel and length and by larger widths between steps.
The activity and rigidity scores of acrylamide treated animals
'did not differ significantly from those of control animals.
Results on the various reflex tests revealed that in ac-
rylamide treated animals the righting reflex was absent in the
eighth and ninth test period and that the hindlimb extension
response was impaired in the ninth test period. An abnor-
mality in hindlimb position manifested by foot dropping and
an inability to maintain a hunched position during the trac-
tion test were also found on the last test period. All other
reflexes were unaffected by acrylamide.
For thiamine deficiency, results were analysed in two
parts. First, a statistical analysis was performed on the data
obtained on days 7, 14, 21, 27, and 33 of the experiment.
This analysis indicated that only transient and sporadic ef-
fects were produced during this initial phase of thiamine defi-
ciency. On day 21, thiamine deficient animals displayed sig-
nificantly less locomotor activity than pair fed controls. A
significant decrease in locomotion was also found on day 33
but not on day 27. The gait angle and width of deficient
animals were respectively decreased and increased on day 27
while length of stride was unchanged. On day 33 stride length
was significantly decreased in thiamine deficient rats but the
other two gait parameters remained unaffected. The second
part of the analysis dealt with the final phase of the experi-
ment, i.e. days 35 to 44. During that phase, all thiamine
deficient animals lost the righting reflex, displayed impaired
weight shift responses and eventually died. These effects
were not seen in pair fed controls. The time of occurrence of
the neurological symptoms and of death in the thiamine de-
ficient group varied from animal to animal. Because of this,
and in order to uniformly compare groups, the results ob-
tained on days when individual thiamine deficient animals
lost their righting reflex were retained for analysis. Results
collected in yoked pair fed controls on these days were also
included in the analysis. No significant group differences
were found for activity, catalepsy, rigidity and landing foot
spread. Also, aside from the righting reflex and weight shift
response, no other reflexes were affected in thiamine defi-
cient animals. Finally, gait analysis revealed that in thiamine
deficient rats the angle and length of strides were signifi-
cantly smaller than those of pair fed controls.
Data obtained on days 1, 3, 6, 9, 12, 15, 16 and 17 of
pyrithiamine administration were included in the statistical
analysis. No significant group differences were detected for
activity and rigidity. The landing foot spread of pyrithiamine
treated rats was significantly larger than thiamine deficient
animals throughout the experiment but did not differ signif-
icantly from controls. For catalepsy, the scores of
pyrithiamine animals were significantly higher than controls
on day 9. The catalepsy endured until the end of the experi-
ment except for day 16 where differences between the two
groups failed to reach statistical significance. Pyrithiamine
had a minimal effect on the gait of treated animals. The width
of steps in pyrithiamine treated animals was significantly
larger than controls and this only difference did not occur
before day 17 of pyrithiamine administration. Similarly, re-
flexes were not affected until day 17 when pyrithiamine ani-
mals lost their righting reflex and their ability to maintain a
normal body posture during the traction test. Following the
seventeenth injection the toxic effects of pyrithiamine pre-
cipitated. By day 18 two animals had died and the remaining
rats were completely debilitated. In addition to the abnor-
malities found on day 17, these animals were incapable of
sustaining locomotion, lost the forelimb extension reflex and
could not emit a normal weight shift response.
An overall summary of the results obtained with all four
models is presented in Table 1 where significant differences
between treated animals and their respective controls are
given for each treatment.
DISCUSSION
As expected, the four experimental treatments of this
study induced ataxia as revealed by the gait analyses. How-
ever the overall pattern of neurological signs accompanying
the uncoordinated gait varied distinctively from treatment to
treatment.
Of all treatments, 3-acetyl pyridine produced the most
diversified profile of neurobehavioral effects. This is not
surprising in view of the known pervasive neurotoxic actions
of this substance. A decrease in locomotor activity, an in-
crease in landing foot spread, the presence of catalepsy as
well as the appearance of a distinctive muscle rigidity were
all apparent 6 hours after the administration of 3-acetyl
pyridine which closely parallels the known time course of
neuropathological changes induced by this substance [3].
However the first evidence of ataxia in this study was not
found until the 12 hour test period following injection of
3-acetyl pyridine.
The neurobehavioral effects of acrylamide were different
in several aspects. Ataxia and catalepsy were seen after the
first injection of acrylamide while an increase in landing foot
spread was not found until the fourth injection. The induc-
tion of a strong and persistent catalepsy by acrylamide was
unexpected and it indicates that this substance might have
widespread pharmacological effects in the CNS, aside from
its documented neuropathological actions in the periphery.
The rapid onset of ataxia after a single injection of 50 mg/kg
is surprising in view of the known dose related neuro-
pathological effects of acrylamide. As mentioned earlier,
evidence of peripheral neuropathy is first detected after
cumulative doses of 400 mg/kg [16]. This suggests that the
early ataxic gait of acrylamide treated animals might appear
before the neuropathological changes can be observed. The
abnormalities observed in the righting, traction and forelimb
extension reflexes as well as in the hindlimb position might
be more directly related to neuropathy since they did not
occur until the eighth injection, which corresponds to a
cumulative dose of 400 mg/kg of acrylamide. In the hindlimb
position test, acrylamide treated animals displayed foot
dropping, an effect not seen with the other treatments of this
study.
Contrary to expectations, thiamine deficiency and
pyrithiamine did not yield similar profiles of neurobehavioral
-------
178
JOLICOEUR ET AL.
TABLE 1
SUMMARY TABLE: OCCURRENCE OF NEUROLOGICAL SYMPTOMS IN ALL TREATMENTS
Treatments
Thiamine 3-Acetyl
Deficient Pyrithiamine Acrylamide Pyridine
Gait Analysis
Length of steps ••!•
Width of steps »t «t
Angle of steps ••!•
General Signs
Motor activity ••!•
Catalepsy •
Rigidity
Landing foot spread
Reflexes (Loss of . .)
Righting reflex • •
Corneal reflex
Traction •
Forelimb extension •
Hindlimb extension NT
Hindlimb position
Weight shift • •
Tail pinch
• 4. ••!•
• t «t
• i «4-
•;
• •
•
•t «t
• •
• •
• NT
• •
•
•Significant difference from control group (p<0.05) UDirection of change NT = not tested
effects. Only the righting reflex and weight shift response
were affected similarly by both treatments. However all
three components of gait were disturbed in thiamine deficient
animals while only the width of stride was altered in
pyrithiamine treated animals. Thiamine deficiency, unlike
pyrithiamine, decreased locomotor activity. Catalepsy as
well as abnormal traction and forelimb extension reflexes
were observed with pyrithiamine but not with thiamine defi-
ciency. These striking discrepancies in the effects indicate
that thiamine deficiency and administration of pyrithiamine,
an antimetabolite of thiamine, affect the nervous system by
distinct biochemical and/or neuropathological mechanisms.
Taken together, the results indicate that the battery of
neurobehavioral tests used in this study constitutes a sensi-
tive and reliable technique for detecting, quantifying and
differentiating various ataxic syndromes in rats. The use of
such standardized tests should prove to be valuable in
studies of animal models of ataxia and in investigations of
neurobehavioral effects induced by various neurotoxic sub-
stances.
REFERENCES
1. Butter-worth, R. F., E. Hamel, F. Landreville and A. Barbeau.
Cerebellar ataxia produced by 3-acetyl pyridine in rat. Can. J.
Neural. Sci. 5: 131-133, 1978.
2. Desclin, J. C. Histological evidence supporting the inferior olive
as the major source of cerebellar climbing fibers. Brain Res. 77:
365-384, 1971.
3. Desclin, J. C. and J. Escubi. Effects of 3-acetyl pyridine on the
central nervous system of the rat, as demonstrated by silver
methods. Brain Res. 77: 341-364, 1974.
4. Gipon, L., P. Schotman, F. G. I. Jennekins and W. H. Gispen.
Description of the acrylamide syndrome in rats. Neuropath.
appl. Neurobiol. 3: 115-123,1977.
5. Gubler, C. J., B. L. Adams, B. Hammond, E. C. Yvan, S. M.
Guo and M. Bennion. Effect of thiamine deprivation and
thiamine antagonists on the level of gamma aminobutyric acid
and on 2-oxoglutarate metabolism in the rat brain. J.
Neurochem. 22: 831-836, 1974.
6. Hamblin, D. O. The toxicity of acrylamide. A preliminary re-
port. S.E.D.E.S. Paris, 195-199, 1956.
7. Holmes, G. A form of familial degeneration of cerebellum.
Brain. 30: 466-471, 1907.
8. Irwin, S. Comprehensive observational assessment.
Psychopharmacologia 13: 222-257, 1968.
9. Jolicoeur, F. B., D. B. Rondeau, E. Hamel, R. F. Butterworth
and A. Barbeau. Measurement of ataxia and related neurolog-
ical signs in the laboratory rat. Can. J. Neural. Sci. 6: 209-215,
1979.
10. Kuperman, A. S. Effects of acrylamide on the central nervous
system of the cat. J. Pharmac. exp. Ther. 123: 180-192, 1958.
11. Lee, C. C. and P. J. Peters. Neurotoxiciy and behavioral effects
of thiamine in rats. Envir. Hlth. Persp. 17: 34-43, 1976.
12. McCandless, D. S., S. Schenkerand M. Cook. Encephalopathy
of thiamine deficiency. J. din. Invest. 47:2268-2289, 1968.
13. Schoental, R. and J. B. Cavanagh. Mechanisms involved in the
'Dying Back' process. Neuropath, appl. Neurobiol. 3: 145-157,
1977.
14. Siegel, S. Non-Parametric Statistics for the Behavioral Sci-
ences.New York: McGraw-Hill, 1956.
15. Snyder, D. R. and J. J. Braun. Dissociation between behavioral
and physiological indices of organo-mercurial ingestion. Toxic.
appl. Pharmac. 41: 277-284, 1977.
16. Spencer, P. S. and H. H. Schaumberg. A review of acrylamide
neurotoxicity Part II. Can. J. Neural. Sci. 1: 152-169, 1974.
17. Steg, G. Efferent muscle innervation and rigidity. Acta Physiol.
Scand. 61:Suppl. 225, 5-53, 1964.
18. Winer, B. J. Statistical Principles in Experimental Design. New
York: McGraw-Hill, 1971.
19. Yoshimura, K., Y. Nishibe, Y. Inoue, S. Hirono, K. Toyoshima
and T. Minesita. Animal experiments on thiamine avitaminosis
and cerebral function. J. Nutr. Sci. Vitaminol. 22: 429-437,
1976.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 179-186. ANKHO International Inc., 1979.
Methodological Problems in the Analysis of
Behavioral Tolerance in Toxicology
GIORGIO BIGNAMI
Section ofPsychopharmacology, Laboratorio dl Farmacologia, Istituto Superiore di Sanitd, 1-00161 Roma, Italy
BIGNAMI, G. Methodological problems in the analysis of behavioral tolerance in toxicology. NEUROBEHAV. TOXI-
COL. 1: Suppl. 1, 179-186, 1979.—The analysis of selected data on differential behavioral tolerance to drugs and other
chemicals leads to a series of tentative methodological proposals with potential interest for the purposes of toxicology.
These data show a wide range of different relations between tolerance induced by continued exposure to treatment per se
and tolerance dependent on specific treatment-behavior interactions, such as behavioral testing in the treatment state and
unfavorable consequences on reinforcement density of response changes induced by treatment. Consequently, when
tolerance phenomena occurring with a particular type of treatment deserve an in-depth analysis, a sequential strategy
should assess (i) critical factors in short-term compensation for behavioral deficits (acute behaviorally augmented
tolerance), (ii) relations between sensitization and tolerance phenomena (particularly in the case of agents with long-lasting
and/or cumulative physiological-biochemical effects), with special regard to tolerance development in the absence of
measurable changes in the lower dosage ranges, and (iii) factors responsible for behaviorally augmented tolerance in
medium- and long-term experiments. The latter analysis may require the investigation of different relations between time of
treatment and time of testing, and different treatment-induced changes in reinforcement density. Specific and non-specific
transfer of coping responses across situations must also be considered, as well as changes in response topographies,
interindividual differences in rate of tolerance development as a function of size and direction of the original treatment
changes, and several other cues which can facilitate the understanding of the phenomena observed. Several lines of work
indicate that the search for separate types of mechanisms underlying different components of tolerance may have greater
heuristic value than approaches based on continuum models of differential tolerance when attempts are made to single out
critical physiological-biochemical mechanisms underlying various types of tolerance phenomena.
Behavioral tolerance Differential tolerance Behaviorally augmented tolerance Methods in behavioral toxicol-
ogy Scopolamine Antimuscarinics Organophosphate anticholinesterases Amphetamine
Stimulants Narcotic analgesics Cannabis derivatives CNS depressants (miscellaneous)
IN the year 63 B.C. the King of Pontus Mithridates (or, more especially [23,32]). The emphasis here will be on data and
correctly, Mithradates) the Sixth, repeatedly defeated by working models with special interest from a methodological
Roman forces led first by Lucullus and then by Pompey, and viewpoint. In terms of the present state of knowledge it must
menaced by a mutiny of his own troops led by his son Phar- be understood that a substantial portion of the relevant
naces, attempted suicide by taking poison. A fatal intoxica- examples will be drawn from behavioral pharmacology,
tion did not follow; therefore, the king ordered a Gallic rather than from behavioral toxicology.
mercenary to slay him. This is how official history describes
the death of Rome's greatest enemy in Asia Minor. The IMPLICATIONS OF SOME SHORT.TERM TOLERANCE PHENOMENA
chronicle or legend, however, credits Mithridates with a
life-long self-experimentation with increasing doses of var- It is well known [32] that behavioral changes during acute
ious toxicants, carried out to create a protection against intoxication are often modified too quickly to allow a re-
homicidal attempts. This story has been told here since it course to the more obvious explanations, such as those based
shows that an interest in tolerance has existed for thousands on the chemical's disposition. In fact, many data on CNS
of years, and that experiments on desensitization by re- depressants such as ethanol, hypnotic-sedatives, tranquiliz-
peated exposure to noxious substances were started well ers, narcotic analgesics, and cannabis derivatives (at least at
before the beginning of the scientific age. In more recent relatively high doses) indicate that acute changes in sensitiv-
times the attempts to analyze not only tolerance per se, but ity are the result of a treatment-behavior interaction trigger-
also the numerous irregularities observed in its development, ing compensatory mechanisms which tend to reverse the be-
have produced a great amount of data which have relevance havioral imbalances produced by the treatments themselves
at the descriptive, the explanatory, and the methodological (for example, impairment of postural regulation and motor
levels. coordination; reduction of consummatory responses; de-
The area covered by studies on tolerance is obviously too crease in food or water reinforcement or increase in shock
broad to be analyzed here in a systematic fashion. Several density as consequences of response changes in operant
excellent reviews on behavioral tolerance are available (see tasks).
179
-------
180
BIGNAMI
Fairly simple tests can be carried out in order to assess
whether or not a short-term modification of treatment effects
can be ascribed to compensation due to treatment-behavior
interactions. Consider, for example, the reduced response-
reinforcement ratio often described in studies on cannabis
derivatives using DRL tasks. During a test started 3 hours
after treating monkeys with 0.75 mg/kg A9-THC the animals
can show considerable impairment in the first hour and re-
markable (although incomplete) recovery in the second hour.
Replication of the experiment with an interval of 4 hours
between treatment and testing yielded identical results. This
shows that short-term recovery is due to a factor other than
passage of time from drug administration, and is caused by a
mechanism other than waning of treatment consequences by
drug metabolization (or by any other change taking place
independently of the drug-behavior interaction) [39].
Another example can be drawn from studies on shuttle-
box avoidance to assess dose-response curves in the case of
neuroleptics. In fact, if a balanced cross-over-design is used,
and tests are given to rats after increasing doses of chlor-
promazine either with or without an interval of one hour
between injection and testing, the response reduction is less
if the animals are allowed to run continuously [31].
The methodological implications of these simple exam-
ples are self-evident, considering the large number of toxi-
cants of different types which have depressant effects on the
CNS. The suggestion is that dose-response curves over time
should be assessed both within and between animals in order
to separate attenuation and disappearance of effects due re-
spectively to waning of the intoxication per se and to
treatment-behavior interactions.
THE RELATIONS BETWEEN SENSITIZATION AND TOLERANCE
Empirical data have shown in several instances the exist-
ence of a continuum which includes (i) progressive increase
in the size of treatment effects with repeated exposure, (ii)
little or no change in sensitivity, and (iii) progressive
tolerance. This has special relevance in the case of treat-
ments with long-lasting effects on physiological-biochemical
systems, as is the case with organophosphate compounds
which inhibit cholinesterase. For example, an experiment
comparing the effects on water intake of 0.2-1.0 mg/kg of
DFP given intramuscularly on alternate days showed little
deviation from the control level at the lowest dose, rapid
deterioration and death at the highest dose, and a distur-
bance lasting several days, but ending in tolerance, at inter-
mediate (0.4 and 0.6 mg/kg) dosage levels [18]. The 0.2 and
0.4 mg/kg groups differed in enzyme inhibition only so far as
rate of reduction of enzyme activity was concerned, but not
with respect to final levels. Furthermore, challenges by other
treatments known to exert modified effects in tolerant ani-
mals (for example, increased sensitivity to the effects of
scopolamine on drinking) indicated that tolerance was ob-
tained both in the 0.4 mg/kg group and in the 0.2 mg/kg
group, i.e., both in the presence and in the absence of a
measurable disturbance during tolerance development. In
other words, if a bias created by a treatment with cumulative
effects is introduced very gradually, a profound change in
the organism's physiological condition and behavioral reac-
tivity can occur without any obvious sign of intoxication.
The relevance of this phenomenon for toxicology is obvious.
The methodological implications of the aforementioned
data and several others on organophosphates (for review and
discussion see [7, 46, 48, 49]) is that compounds with long-
lasting and cumulative effects on physiological-biochemical
systems reflecting themselves in important behavioral
changes should be tested by a battery of relatively simple
methods (e.g., one or two activity tests, food and water in-
take, measurements of simple food- or water-reinforced op-
erants, measurements of an avoidance response) in the
course of continued exposure with different dosages. Treat-
ments should be adjusted so as to determine the relations
between (i) schedules causing progressive deterioration, (ii)
schedules inducing deterioration, but leaving room for the
development of tolerance, (iii) schedules possibly leading to
tolerance in the absence of measurable changes, and (iv) a
genuine no effect level. Obviously, in order to separate (iii)
and (iv) appropriate challenges must be found for each type of
treatment. Within a given category, however, there should
be no difficulties of standardization, which makes it surpris-
ing that experiments such as that mentioned above on DFP
and water intake [18] have not been replicated more widely
in recent years.
Depending on the interest in a particular type of com-
pound, and on the features of the agent to be tested, more
and more specialized questions may have to be asked. For
example, anti-anxiety agents do not induce a maximal disin-
hibition of responses suppressed by punishment during ini-
tial exposures. If an appropriate Geller-type schedule is
used, with a high response rate in a component with intermit-
tent positive reinforcement, and a very low response rate in
another component with both positive reinforcement and
punishment, the initial reduction of high rates often goes
hand in hand with a modest increase of punished behavior.
With repeated treatment the former change is attenuated,
while the latter becomes more and more marked [11, 22, 41,
52]. Although some experiments show that the more obvious
explanation (selective tolerance to the depressant compo-
nent of the treatment action, leading to a progressive
emergence of a disinhibiting component) may be too
simplistic [22], one cannot overemphasize the relevance of
these and related findings for behavioral toxicology. In fact,
there are many situations in which shifts in the balance be-
tween response-enhancing and response-disinhibiting ef-
fects of a given treatment, interacting with those created by
the contingencies which control behavior, may lead to un-
predictable changes in the type of risk to which human sub-
jects may be exposed.
Selective attenuation of response reduction, accompanied
by maintenance (or even progressive increase) of response
enhancements has often been described in the case of tests in
which the former type of change has favorable, and the latter
unfavorable, consequences at the reinforcement level (see
later). The peculiarity of the shift obtained with anti-anxiety
agents is that it takes place in the direction of a higher and
higher punishment frequency via a more and more marked
response disinhibition. The meaning of this phenomenon for
the purposes of preclinical psychopharmacology and for the
analysis of the mechanisms by which an anti-anxiety effect is
obtained is outside the scope of the present discussion.
There remains the fact that from a toxicological viewpoint a
more accurate assessment of acute, subacute, and chronic
effects of the drugs should be carried out, particularly with
regard to weakening of response control by signals related to
potentially harmful events.
THE ANALYSIS OF DIFFERENTIAL TOLERANCE
The representative data so far discussed suggest that in
-------
METHODOLOGY IN THE STUDY OF BEHAVIORAL TOLERANCE
181
order to gain ample information on the behavioral effects of
various treatments it is often necessary to analyze longitudi-
nal changes in the profile of such effects as a function of
different treatment-behavior interactions. This analysis can
now be extended by making recourse to a somewhat unusual
case, that of antimuscarinics (for more complete data and
discussion see [2, 5, 6]). In fact, the illustration of data show-
ing a complete separation of different types of tolerance can
provide an appropriate contrast for a better understanding of
ordinary cases which are characterized by a continuum of
various types of tolerance phenomena.
The Unusual Case of Central Muscarinic Blockade: Com-
plete Separation Between Different Types of Tolerance
If scopolamine is administered to rats prior to each train-
ing session on a FI schedule the characteristic response pat-
tern induced by such schedules develops with considerable
delay. Furthermore, this phenomenon is strictly related to
slowly developing tolerance, since animals drugged after
each of an extended series of training sessions exhibit a re-
duced treatment effect when eventually treated before ses-
sions, relative to animals not drugged at all during training
and then given a pre-test scopolamine challenge [16,17]. In
other words, these analyses suggest a slowly developing
tolerance on a metabolic basis or some other functional basis
independent of any specified treatment-behavior interactions
(e.g., receptor changes).
At the opposite extreme scopolamine effects on step-
down avoidance are reduced in rats already at the second
exposure, and post-test treatment does not allow such
tolerance to develop [43,44]. Independent checks in the same
laboratory on the effects of repeated treatment in a different
situation (measurement of locomotor activity) did not show
any tolerance [45]. In other words, the fast tolerance ob-
served in the passive avoidance test appears to be due
entirely to the treatment-behavior interaction. Prima facie
evidence favors the hypothesis that animals stepping down
from the platform in the drug state more frequently than in
the control state and receiving many extra shocks can ac-
quire an adequate coping response via substitute
physiological-biochemical mechanisms not affected by the
treatment.
In between the two extremes intermediate results are ap-
parently obtained in some situations, for example, active-
passive avoidance tasks which require crossing from one
side to the other of a shuttle box during certain signals and
response withholding during other signals. If appropriate
stimulus combinations are used the initial effect of
scopolamine (or any other centrally acting antimuscarinic)
consists mainly of a passive avoidance disruption, and
tolerance develops over a series of test sessions in the treat-
ment state [1, 5, 15]. Also in this instance a comparison of
pre- and postsession treatments showed that tolerance could
not develop in the absence of a treatment-behavior interac-
tion [15]. Additional checks on the role of reinforcement
density were made with another go-no go task with similar
features, except for the important fact that response sup-
pression during the no-go stimulus complex was induced by
extinction, instead of an avoidance contingency. The results
indicated that in the absence of punishment for hyperre-
sponding the maximal disinhibition produced by initial expo-
sure is fully maintained in a series of sessions which would
suffice to obtain tolerance in the face of a passive avoidance
contingency [29].
Still other experiments contributed to the demonstration
of the unusual features of differential tolerance developing in
the presence of a scopolamine-induced alteration of rein-
forcement density. For example, in the active-passive
avoidance task such tolerance does not consist of a shift to
the right of the dose-response curve, since animals having
regained response control in the course of a series of 1 mg/kg
challenges were shown to be insensitive to doses up to 100
mg/kg [5].
Preliminary evidence in favor of a substitute-system ex-
planation of behaviorally augmented tolerance to scopolamine
was obtained by studying the interactions between drug
treatment and frontal lesions which can induce a passive
avoidance disruption similar to that induced by the drug.
Animals made tolerant to scopolamine prior to lesioning
showed a full fledged lesion effect, which simply confirmed
that cortical areas are not critical for the induction of re-
sponse changes by central muscarinic blockade. After the
lesion deficit had also been compensated by retraining, how-
ever, sensitivity to scopolamine was again at a high level,
while sham-operated animals showed maintenance of the
drug tolerance acquired previously [4]. (This type of result,
incidentally, provides little information on the nature of the
mechanisms involved, but constitutes strong evidence for a
role of systems not primarily involved in the drug-induced
change in the development of behaviorally augmented
tolerance.) Last, but not least, animals made fully tolerant to
scopolamine in the go-no go task showed an EEG desyn-
chronization indistinguishable from that observed after first
exposure to the drug [28].
Similar indications have been obtained in still other situa-
tions. For example, if repeated scopolamine challenges are
given to pigeons trained in a multiple FR-FI schedule one
can observe two types of changes. The first, consisting of
prolonged pauses which reduce reinforcement density, dis-
appears within a few treatment sessions, while the second,
consisting of an alteration of the FI response pattern, is
maintained for an extended series of sessions [5]. Therefore,
all the evidence suggests that in the case of central mus-
carinic blockade tolerance developing very slowly as a func-
tion of repeated treatment per se [16,17] and fast tolerance
developing only in the face of unfavorable consequences at
the reinforcement level are separated by a wide gap, and
presumably require entirely separate neuropsychological
and physiological-biochemical explanations.
Observational data have indicated that scopolamine-
treated animals which have recovered response control,
thereby meeting again reinforcement requirements, can be-
have differently from untreated animals. (Such differences in
response topographies, obviously, would not be expected, if
behaviorally augmented tolerance was served by the same
mechanisms responsible for tolerance developing as a func-
tion of continued exposure per se (see [23]). Specifically, the
hyperactivity and hyperreactivity induced by the drug was
not modified by development of tolerance in the go-no go
avoidance task. Furthermore, untreated rats performing at
asymptotic level tended to remain quiet, or turned their
heads around, or reared during presentation of passive
avoidance signals. The same animals during scopolamine
sessions, but after development of tolerance, often avoided
self-inflicted punishment by taking brief runs towards the
opposite side of the shuttle-box and then stopping just short
of the midline [15]. A similar weeding of response compo-
nents leading to noxious consequences, paralleled by a main-
tenance of other components of the drug-induced change,
-------
182
BIGNAMI
has been observed during active avoidance training with or
without intertrial response punishment. In fact, active
avoidance responses to the discrete signal were markedly
enhanced by scopolamine in both conditions, while the tend-
ency towards a higher intertrial activity was quickly sup-
pressed only in the face of intertrial punishment [3].
The case of scopolamine has been illustrated in some de-
tail not to suggest that experiments of a comparable com-
plexity should be carried out systematically for toxicological
purposes, but to facilitate the identification of methodolog-
ical suggestions which do not necessarily require complex
testing and evaluation procedures. Several important aspects
of this process need to be discussed at this point, although it
must be admitted that most inferences are quite tentative at
the present state of knowledge.
Methodological Considerations
1. Initial steps in the search for differential (behaviorally
augmented) tolerance. It is now widely accepted that the
study of behavioral tolerance must assess the relative roles
of treatment per se and of representative treatment-behavior
interactions in bringing about a change in sensitivity. If the
substance has a short metabolic cycle and does not induce a
long-lasting change in physiological-biochemical systems
some important preliminary information can be obtained
simply by comparing the consequences of treatments ad-
ministered respectively before and after behavioral tests. If,
on the contrary, the effects of each exposure are prolonged,
one can make recourse to tests yielding stable scores in spite
of extended intervals between successive sessions (e.g.,
avoidance), and use a 2x2 design with treatment vs placebo
as one factor and testing vs no testing as the other factor. In
both instances animals treated after testing (or animals
treated and not tested) must be treated before a subsequent
testing session at a time when substantial tolerance has de-
veloped in animals tested in the treatment state.
Ad hoc variations of this strategy have allowed a separa-
tion of components of tolerance with quite different char-
acteristics. For example, A9-THC treatment prior to
avoidance testing of rats leads quickly to a tolerance which is
subsequently maintained for an extended period of time
(even in the absence of renewed treatment and testing),
while repeated dosing in the absence of testing leads to a
tolerance which is subsequently lost in about two weeks [36].
Results of this type obviously suggest that it is necessary to
search for separate (although often interacting) mechanisms,
namely (i) changes in a chemical's disposition and/or
changes in a chemical's effects on specific targets (e.g.,
drug-receptor interactions) taking place as a function of
treatment per se, and (ii) shifts of response control from (a)
target system(s) whose functioning is modified by treatment
to (an)other system(s), as a consequence of treatment-
behavior interactions triggered by changes in reinforcement
density.
2. Further analysis of treatment-behavior interactions in
differential tolerance. If the treatment studied is deemed to
be worth an in-depth analysis one must proceed in an at-
tempt to diversify the tests used, so as to obtain different
combinations of treatment consequences at the response and
the reinforcement levels. For example, efforts should be
made not only to compare situations in which the treatment
modifies, or conversely does not alter, reinforcement den-
sity, but also to compare situations in which a given change
in reinforcement density depends on modifications of re-
sponse rate in opposite directions (e.g., reduction of a high
VI rate and increase of a low DRL rate; for the rationale, see
later). Furthermore, an effort should be made to identify
situations in which treatment-induced changes are in a
favorable direction, as in the case of enhancements of re-
sponse rate and reduction of shock rate in avoidance tasks
(particularly if one uses a population of animals with a wide
range of baseline avoidance performances). In fact, such
changes are often maintained or even magnified with re-
peated exposure, providing a dramatic contrast with fast
tolerance developing in the face of adverse treatment conse-
quences at the reinforcement level (the discussion in [23]
shows that this phenomenon is not limited to stimulants).
3. Checks on dose-response relations and carry-over of
tolerance across tests. Independent confirmation of the rel-
ative roles of different components in tolerance can be ob-
tained by analyzing dose-response relations before, during,
and after tolerance development. In fact, if tolerance con-
sists mainly of a shift of the dose-response curve to the right,
then behaviorally augmented components—if any—should
provisionally be viewed as if they were on a continuum with
components obtained by repeated exposure to treatment per
se. It would be incorrect in the absence of additional evi-
dence to start systematic biochemical or physiological
searches for separate tolerance mechanisms. Conversely, if
behaviorally augmented tolerance generates an overall in-
sensitivity over a wide range of doses—and this, of course,
can be thoroughly verified only in the case of substances
showing a wide ratio between toxic level and effective
level—then one is authorized to think in terms of substitute
mechanisms intervening in response control in the face of
adverse consequences of a bias in the primarily affected
system (see the scopolamine example).
Results obtained by dose-response studies can be con-
fronted with analyses of carry-over of tolerance from one
situation to the other. Quite obviously, carry-over of
tolerance has a different meaning depending on whether or
not the behaviors measured have a common learning ele-
ment, which can be analyzed by conventional experiments
on transfer of learning in the absence of treatment [38]. (For
nonspecific carry-over of essential coping responses in the
face of profound disturbances induced by treatment see a
later section.)
4. Interpretation of tolerance in relation to type of initial
change. All too often one forgets that an accurate analysis of
the type of behavioral change produced by a given agent can
help to understand the nature of treatment-behavior interac-
tions in tolerance development. For example, if a substance
has a general depressant and other incoordination effects,
reflected in a drastic reduction of activity, impairment of
balance, suppression of food and water intake, and impair-
ment in a wide variety of operant and other learned behav-
iors, then homeostatic mechanisms are likely to be put to
work to compensate for these deficits, even independently of
any formal testing carried out by the experimenter. Alterna-
tively, one can state that coping responses acquired by ani-
mals treated in their home environments and severely dis-
turbed by the intoxication may transfer nonspecifically to a
wide range of test situations [23].
Conversely, if the treatment consequences are more sub-
tle, nonspecific transfer phenomena will play a lesser role,
particularly in test situations in wnich reinforcement re-
quirements are quite different from those of the home en-
vironments. In fact, as shown by the examples above and by
the amphetamine data to be discussed later, in studies on
-------
METHODOLOGY IN THE STUDY OF BEHAVIORAL TOLERANCE
183
tolerance to treatments which cause enhanced response
rates and unfavorable changes in reinforcement density in
DRL or passive avoidance paradigms conclusions have been
in favor of a predominant role of the treatment-behavior in-
teraction.
Although the real world is much more complex than
suggested by any crude dichotomy, it is a fact that inves-
tigators working mainly with treatment effects of the former
type have tended to support the view that tolerance induced
by continued exposure per se and behaviorally augmented
tolerance are strictly related to each other, and must be
placed on the same continuum [7, 23, 32, 37, 38]. Con-
versely, investigators working mainly with effects of the lat-
ter type have rather emphasized qualitative differences be-
tween tolerance developing as a function of repeated treat-
ment per se and tolerance due to specific treatment-behavior
interactions [5, 6, 7, 23, 24, 25, 26, 27, 39, 40, 50, 51]. This
controversy is more than a matter of semantics, since at
some point it influences the choice between quite different
strategies in the analysis of underlying mechanisms (see la-
ter).
Even setting aside those cases in which new coping re-
sponses have been observed during development of
tolerance (see the scopolamine example illustrated in a prev-
ious section), an increasing number of tolerance phenomena
are now being interpreted as depending on situation-specific
interactions. For example, an apparently simple phenom-
enon such as behaviorally augmented tolerance to the
analgesic action of morphine and related agents must be
subdivided into different components, since at a behavioral
level classically conditioned changes have been clearly sepa-
rated from instrumental learning to cope with specified
treatment consequences [30, 53, 54, 55, 56, 57]. Furthermore,
it has recently been shown that development of tolerance to
narcotic treatment can be prevented by the application of
nociceptive stimulation concurrently with each drug experi-
ence [21]. This phenomenon, to be viewed in conjunction
with absence of tolerance (or attenuated tolerance) to the
discriminative stimulus properties of drugs [9, 19, 20, 47] and
with the marked tolerance to the unconditioned stimulus
properties of drugs [12, 13, 61], favors the hypothesis of an
active regulation of the physiological consequences of treat-
ment, with the biological significance of such consequences
acting as the major factor in the regulation.
Any further discussion of these and other comparable
phenomena would lead far away from the goals of the pre-
sent analysis. Nevertheless, the warning must be that pres-
ently available criteria for the study and classification of
tolerance phenomena may have to change quickly when
adequate explanations are provided for those results which
are still controversial.
5. Caveats on methods for the analysis of differential
tolerance. The methodological criteria so far identified are
all too easily overextended. In fact, a line of analysis which
allows one to identify a major source of variance in the case
of a particular treatment may well account for a much
smaller portion of the variance in the case of other agents, or
even prove to be completely useless under certain circum-
stances.
Presence vs absence of changes in reinforcement density
have been repeatedly emphasized, therefore some data from
the literature on amphetamine and related agents will be dis-
cussed to show the limits in the use of this discriminant. In fact,
one finds both instances in which a reinforcement density
model can account for most of the observed variation, and
instances in which other factors must be responsible for
differences in rate of tolerance development.
The original studies by Schuster and coworkers [50,51]
and those carried out in recent years (see e.g. [10, 14, 26, 27,
58, 59]) have shown that consequences of response changes
on reinforcement density have considerable importance both
in the case of consummatory responses with a high survival
value, and in that of responses trained with specified rein-
forcement contingencies (food- and water-reinforced oper-
ants, avoidance responses). For example, tolerance has been
shown to develop to the disruptive drug effects on DRL
responding, on response withholding in passive avoidance,
and on response sequences, but not to modification of the FI
response pattern, to facilitation of avoidance, nor to the ef-
fects observed in time-out periods.
As one turns to responses which do not belong to con-
summatory repertoires such as feeding and drinking, and
have not been trained with specified reinforcement con-
tingencies imposed by the experimenter, differential
tolerance to amphetamine presents considerable difficulties
of interpretation. For example, enhancements of locomotor
activity and stereotypes tend to be maintained with repeated
treatment, or even to become more and more pronounced
(see, e.g. [8, 33, 35, 51, 59, 60, 62]. One should also note that
measurements of activity have sometimes indicated an ap-
parent reduction of the effect with repeated treatment; this
phenomenon, however, is due to sensitization rather than to
tolerance, since it goes hand in hand with a progressive in-
crease of behavioral responses usually triggered by higher
doses, such as some stereotypes [62]).
Other responses, on the contrary, show clear-cut
tolerance, as exemplified by abatement of perseveration in a
Y maze without a parallel reduction of overall activity
[33,34]. Furthermore, it has been shown that such selective
tolerance does not depend on testing in the treatment state,
since it takes place also in animals treated after testing ses-
sions [34]. In other words, either such abatement of the
drug-induced perseveration is not behaviorally augmented,
or a behaviorally augmented component must be ascribed to
the development of coping responses opposing perseveration
in the home environment, and to a subsequent transfer of
such responses to the testing situation. Observational data fit
to clarify this point are apparently not available.
It is obvious that by mentioning the difficulties in identify-
ing sources of variance with certain types of treatments one
scratches only the surface of a difficult problem. For exam-
ple, in the case of amphetamine and related agents one
should analyze at each step the treatment factors which
might provide cues about biochemical substrates of different-
ial tolerance, by considering the complex changes over time
of different components of the neurochemical action of the
drugs. The problem, however, is that a correlational
(biochemistry-behavior) study might proceed at random if it
is not supported by working hypotheses at a behavioral
(neuropsychological) level which appear to account for a
considerable portion of the observed variance.
6. Behavioral analyses versus studies on mechanisms under-
lying tolerance. Important advances in the analysis of
tolerance (particularly differential tolerance) have been
made in those areas in which working models at a behavioral
(or neuropsychological) level have allowed identification of
specific targets for physiological-biochemical analyses of
underlying mechanisms. (A bolder statement paying homage
to the late Robert McCleary might be that the further one
moves away from considering behavioral responses simply
-------
184
BIGNAMI
as convenient dependent variables, the greater the probabil-
ity that one will encounter substantial explanations [42].)
A satisfactory illustration of this point would require an
extensive discussion of the studies on a specific class of
compounds, for example, those on organophosphate agents
which have permitted the establishment of a multifactorial
model of tolerance development. In fact, a considerable por-
tion of the observed behavioral variance can be ascribed to
well-documented changes in muscarinic receptor sensitivity
[48,49], and another portion to biochemical mechanisms
which allow the return of acetylcholine towards normal
levels in spite of a continuing enzyme depression, possibly
by a selective fast recovery of some of the molecular forms
of cholinesterase [7,46]. Furthermore, a third portion must
tentatively be ascribed to a shift of response control from
functionally impaired cholinergic neuronal systems to other
systems in the face of the severe disturbances in the or-
ganism's homeostasis, since there is evidence for a behav-
iorally augmented component of tolerance which cannot be
ascribed to the aforementioned mechanisms [7,46].
These and other advances with a similar meaning indicate
that there can be at least heuristic value in postulating sepa-
rate mechanisms for different components of behavioral
tolerance; although—it should go without saying—what ul-
timately counts in the real world is the end product of higher
order interactions between various mechanisms. Con-
versely, working models which tend to consider all different-
ial tolerance phenomena simply as a series of different levels
on the same continuum appear to have reduced heuristic
value. In fact, either they lead to the search for a single type
of mechanism, or they create difficulties in the search for
different types of mechanisms by de-emphasizing those ex-
treme situations in which one or the other of several interact-
ing factors can be singled out.
This discussion should be terminated here, not only for
reasons of space, but also to avoid confusion between a
selective analysis of observed phenomena with direct
methodological relevance and an extensive review of behav-
ioral tolerance. For the sake of coherence, however, one
more point must be made.
7. The attitude of the experimenter. If one looks back at the
history of studies on tolerance it appears obvious that the
identification of critical sources of variance and the propo-
sals with high heuristic value for the analysis of underlying
mechanisms have descended most of the time from a careful
exploitation of apparent irregularities which could easily
have been dismissed as disturbing variability. This applies,
for example (i) to the finding of remarkable differences in
tolerance development between animals showing behavioral
changes in opposite directions upon initial exposure, and (ii)
to the observation of different response topographies in in-
stances in which automated recording yielded apparently
identical performances.
This is not to affirm that standardized (run-of-the-mill)
approaches in behavioral toxicology should be discarded al-
together. In fact, once a particular methodological question
has been solved for a given type of treatment, a considerable
amount of routine work may be necessary in order to extend
the analysis to a wider range of treatment schedules, to addi-
tional compounds of the same class, to different animal
strains or species, and so forth. However, the investigator
who runs a fully automated laboratory and never stops to
think about differences between two individual records, or
never takes the time to check whether or not similar re-
sponse topographies correspond to a particular pattern in the
records, might be overtaken by others more capable—as
Rabelais might have said—to break the bone in order to suck
the nourishing marrow.
CONCLUDING REMARKS
This discussion has to leave out two important extremes
on a very wide continuum of problems. On the one hand, no
attempt has been made to analyze results and explanatory
models which are universally known, or methods which are
sufficiently well established (this applies, for example, to the
wide field of cross-tolerance analysis, although some asym-
metrical results can create remarkable methodological prob-
lems [23]). On the other hand, only hints have been made to
data and tentative models which are still too controversial to
provide short- and medium-term methodological proposals
for the purposes of toxicology.
The universe of behavioral tolerance phenomena with
potential relevance for toxicology appears to be expanding
rapidly. In fact, one sees today a fast development of new
information such as that on classical conditioning
phenomena in behavioral tolerance, on differential tolerance
depending on the type of stimulus function exerted by a
treatment in a particular context (e.g., no tolerance to a CS
function vs tolerance to a US function at the same dose of
the same agent), and more generally, on the many ways in
which organisms can actively regulate the physiological con-
sequences of treatment as a function of the changes in
adaptiveness triggered by the treatment itself. A simple
methodological solution to cope with these and other un-
avoidable complications is, unfortunately, not available.
REFERENCES
1. Bignami, G. Anticholinergic agents as tools in the investigation
of behavioral phenomena. In: Proceedings of the Vth Interna-
tional Congress of the Collegium Internationale Neuro-
psychopharinacologicum, Washington, D.C., 1966, edited by
H. Brill. Amsterdam: Excerpta Medica, I.C.S. 129, 1967, pp.
819-830.
2. Bignami, G. Nonassociative explanations of behavioral changes
induced by central cholinergic drugs. Acta Neurobiol. exp. 36:
5-90, 1976.
3. Bignami, G., L. Amorico, M. Frontali and N. Rosic. Central
cholinergic blockade and two-way avoidance acquisition: The
role of response disinhibition. Physiol. Behav. 1: 461-470, 1971.
4. Bignami, G., G. Carro-Ciampi and M. Albert. Effects of frontal
lesions on "go-no go" avoidance behaviour in normal and
scopolamine-treated rats. Physiol. Behav. 3: 487-493, 1968.
5. Bignami, G. and G. L. Gatti. Repeated administration of central
anticholinergics: Classical tolerance phenomena versus be-
havioural adjustments to compensate for drug-induced deficits.
In: Sensitization to Drugs, Proceedings of the European Society
for the Study of Drug Toxicity, Xth Meeting, Oxford, 1968,
edited by S. B. de C. Baker and J. Tripod. Amsterdam: Ex-
cerpta Medica, I.C.S. 181, 1969, pp.40-46.
6. Bignami, G. and H. Michalek. Cholinergic mechanisms and av-
ersively motivated behaviors. In: Psychopharmacology of Av-
ersively Motivated Behavior, edited by H. Anisman and G. Big-
nami. New York: Plenum Press, 1978, pp. 173-255.
-------
METHODOLOGY IN THE STUDY OF BEHAVIORAL TOLERANCE
185
7. Bignami, G., N. Rosic, H. Michalek, M. Milosevic and G. L.
Gatti. Behavioral toxicity of anticholinesterase agents:
Methodological, neurochemical, and neuropsychological as-
pects. In: Behavioral Toxicology, edited by B. Weiss and V. G.
Laties. New York: Plenum Press, 1975, pp. 155-215.
8. Browne, R. G. and D. S. Segal. Metabolic and experiential fac-
tors in the behavioral response to repeated amphetamine.
Pharmac. Biochem. Behav. 6: 545-552, 1977.
9. Bueno, O. F. A. and E. A. Carlini. Dissociation of learning in
marihuana tolerant rats. Psychopharmacologia 25: 49-56, 1972.
10. Campbell, J. C. and L. S. Seiden. Performance influence on the
development of tolerance to amphetamine. Pharmac. Biochem.
Behav. 1: 703-708, 1973.
11. Cannizzaro, G., S. Nigito, P. M. Provenzano and T. Vitikova.
Modification of depressant and disinhibitory action of
flurazepam during short term treatment in the rat. Psychophar-
macologia 26: 173-184, 1972.
12. Cappell, H. and A. E. LeBlanc. Gustatory avoidance condition-
ing by drugs of abuse. In: Food Aversion Learning, edited by N.
W. Milgram, L. Krames and T. M. Alloway. New York: Plenum
Press, 1977, pp. 133-167.
13. Cappell, H., A. E. LeBlanc and S. Herling. Modification of the
punishing effects of psychoactive drugs in rats by previous drug
experience. J. comp. physiol. Psychol. 89: 347-356, 1975.
14. Carlton, P. L. and D. L. Wolgin. Contingent tolerance to the
anorexigenic effects of amphetamine. Physiol. Behav. 1: 221-
223, 1971.
15. Carro-Ciampi, G. and G. Bignami. Effects of scopolamine on
shuttle-box avoidance and go-no go discrimination: Response-
stimulus relationships, pretreatment baselines, and repeated
exposure to drug. Psychopharmacologia 13: 89-105, 1968.
16. Chamey, N. H. and G. S. Reynolds. Development of behavioral
compensation to the effects of scopolamine during fixed-interval
reinforcement. J. exp. Analysis Behav. 8: 183-186, 1965.
17. Charney, N. H. and G. S. Reynolds. Tolerance to the behav-
ioral effects of scopolamine in rats. Psychopharmacologia 11:
379-387, 1967.
18. Chippendale, T. J., G. A. Zawolkow, R. W. Russell and D. H.
Overstreet. Tolerance to low acetylcholinesterase levels: Mod-
ification of behavior without acute behavioral change.
Psychopharmacologia 26: 127-139, 1972.
19. Colpaert, F. C., J. J. M. D. Kuyps, C. J. E. Niemegeers and P.
A. J. Janssen. Discriminative stimulus properties of fentanyl
and morphine: Tolerance and dependence. Pharmac. Biochem.
Behav. 5; 401-408, 1976.
20. Colpaert, F. C., C. J. E. Niemegeers and P. A. J. Janssen.
Studies on the regulation of sensitivity to the narcotic cue.
Neuropharmacology 17: 705-713, 1978.
21. Colpaert, F. C., C. J. E. Niemegeers and P. A. J. Janssen.
Nociceptive stimulation prevents development of tolerance to
narcotic analgesia. Eur. J. Pharmac. 49: 335-336, 1978.
22. Cook, L. and J. Sepinwall. Reinforcement schedules and ex-
trapolation to humans from animals in behavioral pharmacol-
ogy. Fedn. Proc. 34: 1889-1897, 1975.
23. Corfield-Sumner, P. K. and I. P. Stolerman. Behavioral
tolerance. In: Contemporary Research in Behavioral Phar-
macology, edited by D. E. Blackman and D. J. Sanger. New
York: Plenum Press, 1978, pp. 391-448.
24. Elsmore, T. F. The role of reinforcement loss in tolerance to
chronic A9-tetrahydrocannabinol effects on operant behavior of
rhesus monkeys. Pharmac. Biochem. Behav. 5: 123-128, 1976.
25. Ferraro, D. P. Effects of A9-trans-tetrahydrocannabinol on
simple and complex learned behavior in animals. In: Current
Research in Marijuana, edited by M. F. Lewis. New York:
Academic Press, 1972, pp. 49-95.
26. Fischman, M. W. and C. R. Schuster. Tolerance development
to chronic methamphetamine intoxication in the rhesus mon-
key. Pharmac. Biochem. Behav. 2: 503-508, 1974.
27. Fischman, M. W. and C. R. Schuster. Long-term behavioral
changes in the rhesus monkey after multiple daily injections of
rf-methyl-amphetamine. J. Pharmac. exp. Ther. 201: 593-605,
1977.
28. Florio, V., G. Bignami and V. G. Longo. EEG patterns during
the behavioural desensitization to scopolamine in rats. Int. J.
Neuropharmac. 8: 405-411, 1969.
29. Frontali, M., L. Amorico, L. De Acetis and G. Bignami. A
pharmacological analysis of processes underlying differential
responding: A review and further experiments with
scopolamine, amphetamine, lysergic acid diethylamide (LSD-
25), chlordiazepoxide, physostigmine, and chlorpromazine. Be-
hav. Bio/. 18: 1-74, 1976.
30. Hayes, R. L. and D. J. Mayer. Morphine tolerance: Is there
evidence for a conditioning model? Science 200: 343-344, 1978.
31. Janku, I. The influence of delayed and immediate exposure to
trials upon the effect of chlorpromazine on conditioned
avoidance behaviour. Psychopharmacologia 6: 280-285, 1964.
32. Kalant, H., A. E. LeBlanc and R. J. Gibbins. Tolerance to, and
dependence on, some non-opiate psychotropic drugs. Pharmac.
Rev. 23: 135-191, 1971.
33. Kokkinidis, L. and H. Anisman. Behavioural specific tolerance
following chronic d- or /-amphetamine treatment: Lack of in-
volvement of p-hydroxynorephedrine. Neuropharmacology 17:
95-102, 1978.
34. Kokkinidis, L. and H. Anisman. Abatement of stimulus persev-
eration following repeated d-amphetamine treatment: Absence
of behaviorally augmented tolerance. Pharmac. Biochem. Be-
hav. 8: 557-563, 1978.
35. Kokkinidis, L., J. Irwin and H. Anisman. Shock-induced
locomotor excitation following acute and chronic amphetamine
treatment. Neuropharmacology 18: 13-22, 1979.
36. Larsen, F. F. and G. T. Pryor. Factors influencing tolerance to
the effects of A9-THC on a conditioned avoidance response.
Pharmac. Biochem. Behav. 7: 323-329, 1977.
37. LeBlanc, A. E., R. J. Gibbins and H. Kalant. Behavioral aug-
mentation of tolerance to ethanol in the rat. Psychophar-
macologia 30: 117-122, 1973.
38. LeBlanc, A. E., R. J. Gibbins and H. Kalant. Generalization of
behaviorally augmented tolerance to ethanol, and its relation to
physical dependence. Psychopharmacologia 44: 241-246, 1975.
39. Manning, F. J. Acute tolerance to the effects of delta-9-tetra-
hydrocannabinol on spaced responding by monkeys. Pharmac.
Biochem. Behav. 1: 665-671, 1973.
40. Manning, F. J. Role of experience in acquisition and loss of
tolerance to the effect of A9-THC on spaced responding. Phar-
mac. Biochem. Behav. 5: 269-273, 1976.
41. Margules, D. L. and L. Stein. Increase of "antianxiety" activity
and tolerance of behavioral depression during chronic adminis-
tration of oxazepam. Psychopharmacologia 13: 74-80, 1968.
42. McCleary, R. A. Response specificity in the behavioral effects
of limbic system lesions in the cat. J. comp. physiol. Psychol.
54: 605-613, 1961.
43. Meyers, B. Some effects of scopolamine on a passive avoidance
response in rats. Psychopharmacologia 8: 111-119, 1965.
44. Meyers, B. and M. A. Lazarus. Diminished responsivity on a
passive avoidance task to a second administration of
scopolamine. Psychol. Rep. 20: 175-178, 1967.
45. Meyers, B. and R. C. Wilchin. Some effects of scopolamine on
locomotor activity in rats. Psychon. Sci. 17: 174-175, 1969.
46. Michalek, H., A. Meneguz, G. M. Bisso, G. Carro-Ciampi, G.
L. Gatti and G. Bignami. Neurochemical changes associated
with the behavioural toxicity of organophosphate compounds.
In: Advances in Pharmacology and Therapeutics, Vol. 9, Tox-
icology, edited by Y. Cohen. Oxford: Pergamon Press, 1979, pp.
187-201.
47. Miksic, S. and H. Lai. Tolerance to morphine-produced dis-
criminative stimuli and analgesia. Psychopharmacology 54:
217-221, 1977.
48. Russell, R. W. Cholinergic substrates of behavior. In:
Cholinergic Mechanisms and Psychopharmacology, edited by
D. J. Jenden. New York: Plenum Press, 1977, pp. 709-731.
-------
186
BIGNAMI
49. Russell, R. W., D. H. Overstreet, C. W. Cotman, V. G. Carson,
L. Churchill, F. W. Dalglish and B. J. Vasquez. Experimental
tests of hypotheses about neurochemical mechanisms underly-
ing behavioral tolerance to the anticholinesterase, diisopropyl
fluorophosphate. J. Pharmac. exp. Ther. 192: 73-85, 1975.
50. Schuster, C. R., W. S. Dockens and J. H. Woods. Behavioral
variables affecting the development of amphetamine tolerance.
Psychopharmacologia 9: 170-182, 1966.
51. Schuster, C. R. and J. Zimmerman. Timing behavior during
prolonged treatment with ^/-amphetamine. J. exp. Analysis Be-
hav. 4: 327-330, 1961.
52. Sepinwall, J., F. S. Grodsky and L. Cook. Conflict behavior in
the squirrel monkey: Effects of chlordiazepoxide, diazepam,
and N-desmethyldiazepam. J. Pharmac. exp. Ther. 204: 88-102,
1978.
53. Siegel, S. Evidence from rats that morphine tolerance is a
learned response. J. comp. physiol. Psychol. 89:498-506, 1975.
54. Siegel, S. Morphine analgesic tolerance: Its situation specificity
supports aPavlovian conditioning model. Science 193: 323-325,
1976.
55. Siegel, S. Morphine tolerance acquisition as an associative
process. J. exp. Psychol.: Anim. Behav. Proc. 3: 1-13, 1977.
56. Siegel, S. Morphine tolerance: Is there evidence for a condition-
ing model? (Reply to Hayes and Mayer) Science 200: 344-345,
1978.
57. Siegel, S., R. E. Hinson and M. D. Krank. The role of predrug
signals in morphine analgesic tolerance: Support for a Pavlovian
conditioning model of tolerance. J. exp. Psychol.: Anim. Behav.
Proc. 4: 188-196, 1978.
58. Thompson, D. M. Repeated acquisition of behavioral chains
under chronic drug conditions. J. Pharmac. exp. Ther. 188:
700-713, 1974.
59. Thornhill, J. A., M. Hirst and C. W. Gowdey. Variability in
development of tolerance to repeated injections of low doses of
^/-amphetamine in rats. Can. J. Physiol. Pharmac. 55: 1170-
1178, 1977.
60. Tormey, J. and L. Lasagna. Relation of thyroid function to
acute and chronic effects of amphetamine in the rat. J. Phar-
mac. exp. Ther. 128: 201-209, 1960.
61. Vogel, J. and B. A. Nathan. Reduction of learned taste aver-
sions by pre-exposure to drugs. Psychopharmacology 49: 167-
172, 1976.
62. Weston, P. F. and D. H. Overstreet. Does tolerance develop to
low doses of d- and 1-amphetamine on locomotor activity in
rats? Pharmac. Biochem. Behav. 5: 645-649, 1976.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 187-188. ANK.HO International Inc., 1979.
Morphological Studies of Toxic Distal
Axonopathy
HERBERT H. SCHAUMBURG
Rose F. Kennedy Center, Albert Einstein College of Medicine, Bronx, NY 10461
SCHAUMBURG, H. H. Morphological studies of toxic distal anoxopathy. NEUROBEHAV. TOXICOL. 1: Suppl. 1,
187-188, 1979.—Distal axonopathy is the most common form of toxic injury to the peripheral nervous system.
Morphological studies of the experimental distal axonapathies produced by acrylamide monomer and hexacarbons have
lead to a reapparaisal of the dying-back hypothesis. These studies have also provided a rationale for many of the clinical
findings in humans with distal axonopathies, and have been especially helpful in elucidating the effects of axonal
neurotoxins on the central nervous system.
Toxic distal axonopathy Morphology
A RECENT classification of toxic disorders of the peripheral
nervous system (PNS) has described three types of patholog-
ical reaction: nerve cell death (neuronopathy), segmental
demyelination (myelinopathy) and distal axonal degenera-
tion (distal axonopathy) [3]: Distal axonopathy is the princi-
pal focus of this report since it is the most common form of
toxic disease of the nervous system.
In 1972, Dr. Spencer and I began our experimental animal
studies of dying-back peripheral nerve degeneration pro-
duced by acrylamide. At that time, the dying-back hypoth-
esis was a generally accepted concept in neuropathology. In
brief, this idea suggested that a toxic derangement of
neuronal metabolism would cause the nerve cell body to
withdraw metabolic support from its axonal process so that
degeneration would commence in the nerve terminal and, as
neuronal metabolism became progressively impaired, axonal
degeneration would move, in a seriate, retrograde (dying-
back) fashion towards the cell body. If the intoxication was
stopped it was assumed that axonal regeneration would
commence and the animal (or human) would eventually
recover. Neurons with the longest and largest diameter
axons were thought to be most vulnerable to dying-back
degeneration since they supported the greatest metabolic
load.
Our studies have indicated that dying-back was an in-
accurate and misleading term to describe the pathological
process, and suggested that the designation central-
peripheral distal axonopathy was more appropriate [5]. The
need for this somewhat cumbersome term was emphasized
by the fact that for many clinicians and neuropathologists,
the appellation dying-back had become synonymous with
peripheral neuropathy, despite the previous demonstration
that varying degrees of CNS degeneration sometimes ac-
companied the peripheral changes. The need to give equal
emphasis to the central and peripheral nervous system
became self-evident when we demonstrated widespread dis-
tal axonal degeneration in the CNS occurring concurrently
with similar changes in the PNS.
Experimental Studies of the PNS
Acrylamide. Our initial studies of acrylamide-induced
PNS axonal degeneration were based on the commonly-held
assumption that dying-back nerve fiber degeneration would
begin in the nerve terminals and proceed in a seriate fashion
up the fibers of the longest nerve (sciatic). Accordingly, we
focused on the hindpaw of the cat where we examined the
spatio-temporal pattern of axon degeneration in sensory
terminals supplying pacinian corpuscles, the equatorial zone
of nearby muscle spindles, and adjacent motor nerve termi-
nals supplying extrafusal muscle fibers. Cats were daily
injected with 1 mg/kg of an aqueous solution of acrylamide
and tissue was obtained at stages of intoxication. His-
topathological examination revealed that the most sensitive
structure was the 7-11 /u.m axons supplying pacinian cor-
puscles which began to degenerate before the larger 11 /am
axons supplying the annulospiral muscle spindle terminals.
Both sensory endings were more vulnerable than the jux-
taposed motor nerve terminals. Pacinian corpuscle terminals
in the forepaw began to degenerate at the same time as those
in the hindpaw. In addition, by exploring axonal regions
proximal to the nerve terminals of pacinian corpuscles and
muscle spindles, it was apparent that degeneration here
sometimes preceded the onset of terminal degeneration.
Thus, these studies with acrylamide directly challenged two
of the basic tenets of the dying-back hypothesis, namely that
axonal length and diameter dictated vulnerability and that
degeneration began at the axon terminal before proceeding
proximally [4].
Hexacarbons. The hexacarbon studies were undertaken
after completion of the initial acrylamide investigation in
order to evaluate further the dying-back process in the PNS.
Our studies with the hexacarbons (MBK, n-hexane and
2,5-hexanedione) were conducted in cats and rats, and
employed various routes of administration, depending on the
physical properties of the compound. Intoxicated cats and
rats lost weight, then developed an unsteady gait and
187
-------
188
SCHAUMBURG
hindlimb foot drop. With continued intoxication forelimb
distal weakness appeared and, after very prolonged intoxi-
cation, animals became quadriparetic and were unable to sit
or stand. At no stage did they display truncal ataxia or head
tremor, clinical features which had been prominent in cats
intoxicated with acrylamide. Histopathological studies re-
vealed that the large myelinated fibers of the tibial nerve
branches to the calf muscles displayed paranodal giant
axonal degeneration before similar changes were evident in
the more distal plantar nerve fibers in the hindpaw. The
nerves to the calf muscles contain the largest diameter
myelinated axons in the hindlimb, and their special sensitiv-
ity to hexacarbons suggested that axon diameter might play a
major role in determining vulnerability. The importance of
axonal length was also apparent because the hindlimbs,
containing long nerves, were consistently involved before
forelimbs, supplied with shorter nerves. Surprisingly, re-
generating fibers were also seen scattered among the de-
generating axons during the course of intoxication. Another
important observation in the hexacarbon studies was that
degeneration did not begin in the terminal and travel steadily
up the nerve fiber, but rather began in a multifocal pattern on
the proximal sides of multiple nodes of Ranvier in distal
nerve fibers. These findings confirmed our earlier observa-
tions in acrylamide animals, namely, that the axonal degen-
eration in dying-back neuropathies did not begin at the
terminal and progress toqards the cell body in a seriate
fashion, as the term dying-back had imDlied [61
Experimental Studies of the CNS
In contrast to the PNS changes, the distribution of CNS
axonal degeneration produced by acrylamide and the
hexacarbons compounds appeared identical. A striking and
consistent finding in the hexacarbon neuropathies was the
development of axonal degeneration in certain areas of the
central nervous system at the same time changes were first
noted in the most vulnerable areas of the PNS. In general,
these changes were found in the distal regions of long spinal
nerve tracts and later, in shorter CNS pathways. As in the
PNS, tractal degeneration appeared to spread toward the
neuron cell body with continued intoxication. The distribu-
tion of the changes in asymptomatic, hexacarbon-intoxicated
animals showed that the rostral regions of long, ascending
sensory tracts (gracile fasiculus, dorsal spinocerebellar) and
caudal regions of long, descending motor pathways (cor-
ticospinal) were first affected. The sensory pathways were
more sensitive than the descending fibers, and the gracile
nucleus and mossy fibers and white matter of the anterior
cerebellar vermis were the first CNS structures to display
axonal swellings. In animals with more prolonged intoxica-
tion and moderate clinical impairment these same areas dis-
played an advanced stage of axonal degeneration accom-
panied by breakdown of the myelin sheath and a mild as-
trocytic proliferation [7]. In cats with very prolonged intoxi-
cation, and severe weakness in all extremities, axonal swel-
lings were present in the distal optic tract, superior col-
liculus, lateral geniculate body and mammillary bodies [2],
These studies confirmed many ideas about the dying-back
process that had been drawn from earlier studies of the
spatio-temporal pattern in the PNS. The distal regions of
long and large CNS fibers were susceptible to degeneration,
and axon length appeared an important factor in determining
vulnerability.
Recovery from Hexacarbon Intoxication
Several of the cats with very prolonged intoxication were
allowed to recover for periods of up to two years. These
animals gradually regained weight, became mobile and, after
a year's recovery, appeared to have regained strength in all
extremities. However, on attempting to walk, all animals
manifested stiff-legged, unsteady hindlimb movements that
resulted in a swaying motion of the pelvis. Attempts at run-
ning resulted in a reeling movement and the animals fell fre-
quently to either side. Histopathological examination re-
vealed that the gracile nuclei consistently displayed the most
severe changes in these recovered animals with a striking
axonal loss, diffuse gliosis and loss of neurons. In the lateral
columns of the lumbar spinal cord there was moderate
axonal loss accompanied by mild gliosis, and gliosis and fiber
loss were prominent in the white matter of the anterior ver-
mis of the cerebellum. Based on these data we concluded
that the degeneration in the corticospinal tracts and cerebel-
ler vermis probably accounted for the residual spastic-ataxic
gait in these animals.
Human Correlation
In brief, the clinical findings in humans with acrylamide
and n-hexane intoxication correlate well with the data ob-
tained from animal studies. Humans with acrylamide intoxi-
cation initially develop a severe loss of peripheral vibratory
sensation, numbness and an unsteady, ataxic gait; while
those with chronic n-hexane exposure display distal sym-
metrical weakness and sensory loss. Individuals recovering
from severe n-hexane peripheral neuropathy have developed
residual spasticity, and visual impairment has been demon-
strated following prolonged exposure to n-hexacarbons [1,8].
REFERENCES
1. Koroblin, R., A. K. Asbury, A. J. Sumner and S. L. Niesen.
Glue-sniffing neuropathy. Archives of Neural. 32: 158-162, 1975.
2. Schaumburg, H. H. and P. S. Spencer. Environmental hydro-
carbons produce degeneration in cat hypothalamus and optic
tract. Science 199: 199-200, 1978.
3. Schaumburg, H. H. and P. S. Spencer. Toxic neuropathies.
Neurology 29: 429-431, 1979.
4. Schaumburg, H. H., H. M. Wisniewski and P. S. Spencer. Ul-
trastructural studies of the dying-back process. I. Peripheral
nerve terminal and axonal degeneration in systemic acrylamide
intoxication. Journal of Neuropath. Exp. Neurology 33: 260-295,
1974.
5. Spencer, P. S. and H. H. Schaumburg. Central-peripheral distal
axonopathy—the pathology of dying-back polyneuropathies. In:
Progress in Neuropathology, Vol. 3, edited by H. Zimmerman.
New York: Grune and Stratton, 1977, pp. 253-295.
6. Spencer, P. S. and H. H. Schaumburg. Ultrastructural studies of
the dying-back process. III. The evolution of experimental
peripheral giant axonal degeneration. Journal of Neuropath.
Exp. Neural. 36: 276-299, 1977.
7. Spencer, P. S. and H. H. Schaumburg. Ultrastructural studies of
the dying-back process. IV. Differential vulnerability of PNS and
CNS fibers in experimental central-peripheral distal
axonopathies. Journal of Neuropath. Exp. Neural. 36: 300-320,
1977.
8. Yamamura, Y. n-Hexane polyneuropathy. Folia psychiat.
Neural. Jap. 23: 45-57, 1969.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 189-191. ANKHO International Inc., 1979.
Cellular Responses to Neurotoxic Compounds
of Environmental Significance
PETER S. SPENCER
Neurotoxicology Unit, Albert Einstein College of Medicine, 1410 Pelham Parkway, Bronx, NY 10461
SPENCER, P. S. Cellular responses to neurotoxic compounds of environmental significance. NEUROBEHAV.
TOXICOL. 1: Suppl. 1, 189-191, 1979.—Many neurotoxic chemicals of environmental significance can be conveniently
classified according to their cellular site of action in the nervous system. This paper considers neurotoxins which damage
the nerve cell body (neuronopathy), its axonal process (axonopathy), or the myelin sheath which segmentally enwraps the
myelinated axon (myelinopathy). Each of these three conditions can be reproduced in experimental animals for study of the
mechanisms and consequences of neurotoxic damage. Detailed morphological examination of toxic distal axonopathies
have stimulated biochemical studies which promise to yield a precise explanation of the molecular basis for this common
type of neurotoxic disease. It seems possible that a precise description of the molecular basis for toxic distal axonal
degeneration is within sight.
Neurotoxins Neuronopathy Axonopathy Myelinopathy
MANY neurotoxic chemicals of environmental significance
can be conveniently classified according to their cellular site
of action in the nervous system [14]. This paper considers
neurotoxins which damage the nerve cell body
(neuronopathy), its axonal process (axonopathy), or the
myelin sheath which segmentally enwraps the myelinated
axon (myelinopathy). Each of these three conditions can be
reproduced in experimental animals for study of the mech-
anisms and consequences of neurotoxic damage. For exam-
ple, the rat develops a neuronopathy restricted to neurons in
peripheral ganglia following administration of Adriamycin
[2], a central and peripheral myelinopathy from acetyl ethyl
tetramethyl tetralin [12], and a central-peripheral distal
axonopathy from hexacarbon solvents, acrylamide or carbon
disulfide [9].
TOXIC NEURONOPATHY
The anthracycline antibiotic, Adriamycin, occupies an in-
creasingly important position in the treatment of malignant
tumors. Administered intravenously to cancer patients, the
drug binds to nucleic acid species and inhibits the growth of
rapidly dividing malignant cells. Although neurological dis-
ease has not been reported in patients treated with Ad-
riamycin, this is a reproducible response in the rat. After a
single, intravenous injection of 10 mg/kg, animals develop an
abnormal limb posture and later die. The morphological
substrate of this neurological illness has been determined by
light and electron microscope examination of the peripheral
nervous system. Fluorescence studies [6] have demonstrated
that Adriamycin binds to the cell nuclei of sensory and au-
tonomic ganglia, sites where the blood-nerve barrier is nor-
mally porous [4]. Within a few hours, histological changes
appear in neuronal nuclei of sensory ganglia as a clumping
and focal clearing of chromatin. After several days, many of
the more severely affected, larger neurons undergo nuclear
and cytoplasmic degeneration, presumably because the
anabolic machinery of the cell has been shut down. The sur-
rounding satellite cells, which also bind Adriamycin, but do
not undergo breakdown, cluster to occupy the degenerative
site of the neuron. This selective loss of the nerve cell body
is accompanied by a rapid, secondary breakdown of the
axon—both the centrally-directed process to the spinal cord
and the peripheral axon supplying sensory terminals such as
stretch receptors in the muscle spindle. Deficient spindle
sensory input probably accounts in significant part for the
bizarre limb posturing seen in sensory neuronopathies. The
limbs are not weak since Adriamycin fails to damage the
motor nerve cell or axon, presumably because the former is
located in the spinal cord and protected from the drug by the
blood-brain barrier.
In summary, Adriamycin produces a primary degenera-
tion of neurons in sensory and autonomic ganglia, secondary
breakdown of the axon, and permanent denervation of cer-
tain sensory receptors. Although other examples of toxic
sensory neuronopathy can be cited, e.g., methyl mercury
[4], such diseases are surprisingly rare in view of the ease
with which circulating toxins can gain direct access to the
neurons of peripheral ganglia.
TOXIC MYELINOPATHY
Primary breakdown of myelin is a more common re-
sponse to systemic intoxication. Alkyl tins, cuprizone,
hexachlorophene, tellurium and acetyl ethyl tetramethyl tet-
ralin (AETT) are examples of myelinotoxins [1]. The fra-
gance comound AETT produces a progressive neurological
disease in chronically intoxicated animals. When AETT is
daily administered to the rat, either by an oral or percutane-
ous route, the animal gradually develops a bizarre, arched-
189
-------
190
SPENCER
back posture and subsequently develops limb weakness.
Both of these signs are proceeded by hyperirritability and an
extraordinary blue discoloration of internal organs including
the brain.
The morphological substrate of AETT intoxication is
complex, comprising an early, slow progressive and wide-
spread ceroid neuronopathy, and a later, primary demyeli-
nation. Demyelination begins in the large diameter fibers of
central and peripheral nervous tissue. In the latter, affected
myelin segments undergo ballooning to form large
intramyelinic vacuoles in which phagocytic cells eventually
reside. These cells selectively removed the damaged myelin
sheath, leaving lengths of axons denuded but intact.
Schwann cells then envelop the bare portions of axon and
produce short, remyelinated segments. This combination of
demyelination and remyelination, structural events as-
sociated with blockade and restoration of nerve impulse
transmission can account for the progressive limb weakness
which eventually reaches a plateau.
The cellular mechanisms of AETT myelinopathy have yet
to be resolved. However, several morphological observa-
tions point to the myelin sheath rather than the myelin-
producing cell (Schwann cell or oligodendrocyte) as the
likely locus of toxic damage.
TOXIC DISTAL AXONOPATHY
Damage to the axon of the nerve cell, especially the distal
ends of long and large axons (distal axnonopathy), probably
constitutes the single most common neuropathological re-
sponse to neurotoxic agents. Prolonged, low-level intoxica-
tion is usually required to produce clinical disease
(peripheral neuropathy) which is typically expressed in the
form of distal, symmetrical hindlimb weakness and sensory
loss. (Schaumburg, this volume.)
The distal axonopathy produced by neurotoxic hexacar-
bons [11] has been studied most thoroughly. The first identi-
fiable pathological event is a focal axonal swelling which
appears on the proximal side of one or more nodes of Ran-
vier in the sub-terminal parts of the nerve fibers. These
axonal swellings contain large numbers of neurofilaments
and other organelles which normally move along the axon.
Similar changes may be found in animals intoxicated with
acrylamide or carbon disulfide. Axoplasmic organelles seem
to accumulate at nodes of Ranvier because the axonal trans-
port systems are locally defective. The portion of axon
below the position of transport blockade then undergoes de-
generation, thereby its sensory or motor innervation.
The neurotoxic property of the hexacarbon solvents is
attributable, in large part, to the water-soluble neuro-
toxic metabolite 2,5-hexanedione [3]. This compound is
more potent than its related metabolites 2,5-hexanediol,
5-hydroxy-2-hexanone, 2-hexanol, 2-hexanone (M«BK), or
«-hexane, and consistently appears as a metabolite of all the
other compounds both in vivo and in tissue culture models of
hexacarbon neuropathy [7,15]. The neurotoxic property of
2,5-hexanedione (2,5-HD) is associated with the gamma
spacing of the carbonyl groups; other compounds of longer
chain length and non-symmetrical gamma carbonyl group-
ings are also neurotoxic. By contrast, compounds which lack
the gamma spacing of the carbonyl groups (e.g., 2,4-
hexanedione) fail to exhibit neurotoxicity [7,9].
Knowledge of the mechanism of cellular damage in distal
axonopathies has advanced considerably in recent years.
The concept that toxic chemicals damage the nerve cell
body, causing long and large axons to die-back from their
distal ends, has been replaced by a theory of direct damage
to the nerve fiber. Organophosphates, isoniazid, acrylamide,
disulfiram and hexacarbons are all believed to damage the
nerve fiber directly. Direct evident for this thesis exists in
the case of diisofluorophosphate [5] and 2,5-hexanedione [8].
When 2,5-HD is applied repetitively to a peripheral nerve,
giant axonal swellings of the type seen in systemic intoxica-
tion develop beneath the site of application. No such dwell-
ings appear when non-neurotoxic 2,4-hexanedione is used.
The biochemical lesion(s) underlying distal axonopathy
remain to be elucidated. Especially puzzling is how a number
of chemically distinct compounds can induce closely similar
patterns of cellular breakdown. For example, the hexacar-
bons, acrylamide and carbon disulfide (CS2) are all capable
of producing giant axonal swelling as a prelude to degenera-
tion. One explanation may be that such compounds act at
nearby or related sites in metabolic pathways, the net
biochemical defect being common for all three compounds
[13]. Intense interest is currently focussed on glycolysis and
associated energy tranformation pathways because these are
implicated in toxic and vitamin deficiency states associated
with distal axonopathy. 2,5-HD, CS2 and acrylamide each
inhibit in vitro the major glycolytic enzymes glyceraldehyde
phosphate dehydrogenase and phosphofructokinase, the de-
gree of inhibition being dependent both on the concentration
of toxin and the duration of preincubation of toxin and
enzyme [9]. A comparable inhibition of enzymes in the nerve
fiber axon presumably would raise the demand for their re-
placement by the nerve cell body. Failure to resupply the
entire axon would leave distal regions with an inadequate of
glycolytic enzymes resulting in a depletion of energy
supplies in this region. This may produce a blockade of
energy-dependent axonal transport at energy-intensive
nodes of Ranvier, and initiate the sequence of pathological
changes which leads to distal nerve fiber degeneration [13].
CONCLUSIONS
The application of advanced investigative techniques has
provided neuropathological assays to detect very early struc-
tural changes in the nervous system [10], allowing one to
identify and distinguish between three major types of
neurotoxic disease: neuronopathy, myelinopathy and
axonopathy. Detailed morphological examination of toxic
distal axonopathies have stimulated biochemical studies
which promise to yield a precise explanation of the molecu-
lar basis for this common type of neurotoxic disease. It
seems possible that a precise description of the molecular
basis for toxic distal axonal degeneration is within sight.
REFERENCES
1. Cammer, W. Toxic demyelination: biochemical studies and
hypothetical mechanisms. In: Experimental and Clinical
Neurotoxicology, edited by P. S. Spencer and H. H. Schaum-
burg. Baltimore, MD: Williams and Wilkins, 1980.
2. Cho, E-S., P. S. Spencer and B. S. Jortner. Doxorubicin. In:
Experimental and Clinical Neurotoxicology, edited by P. S.
Spencer and H. H. Schaumburg. Baltimore, MD: Williams and
Wilkins, 1980.
-------
CELLULAR RESPONSES TO NEUROTOXIC COMPOUNDS
191
3. DiVincenzo, G. D., M. L. Hamilton, C. J. Kaplan and J. Dedi-
nas. Characterization of the metabolites of methyl n-butyl
ketone. In: Experimental and Clinical Neurotoxicology, edited
by P. S. Spencer and H. H. Schaumburg. Baltimore, MD:
Williams and Wilkins, 1980.
4. Jacobs, J. M. Vascular permeability and neural injury. In: Ex-
perimental and Clinical Neurotoxicology, edited by P. S.
Spencer and H. H. Schaumburg. Baltimore, MD: Williams and
Wilkins, 1980.
5. Lowndes, H. E. and T. Baker. Toxic site of action in distal
axonopathies. In: Experimental and Clinical Neurotoxicology,
edited by P. S. Spencer and H. H. Schaumburg. Baltimore, MD:
Williams and Wilkins, 1980.
6. Mendell, J. R. and Z. Sahenk. Interference of neuronal process-
ing and axoplasmic transport by toxic chemicals. In: Experi-
mental and Clinical Neurotoxicology, edited by P. S. Spencer
and H. H. Schaumburg. Baltimore, MD: Williams and Wilkins,
1980.
7. O'Donoghue, J. L. and W. J. Krasavage. Identification and
characterization of M«BK neurotoxicity in laboratory animals.
In: Experimental and Clinical Neurotoxicology, edited by P. S.
Spencer and H. H. Schaumburg. Baltimore, MD: Williams and
Wilkins, 1980.
8. Politis, M. J., R. G. Pellegrino and P. S. Spencer. Ultrastruc-
tural studies of the dying-back process. V. Axonal neurofila-
ments accumulate at sites of 2,5-hexanedione application: evi-
dence for nerve fiber dysfunction in experimental hexacarbon
neuropathy. J. Neurocytol., submitted for publication.
9. Sabri, M. I. and P. S. Spencer. Toxic distal axonopathy:
biochemical studies and hypothetical mechanisms. In: Experi-
mental and Clinical Neurotoxicology, edited by P. S. Spencer
and H. H. Schaumburg. Baltimore, MD: Williams and Wilkins,
1980.
10. Spencer, P. S., M. C. Bischoff and H. H. Schaumburg.
Neuropathological methods for the detection of Neurotoxic
Disease. In: Experimental and Clinical Neurotoxicology, edited
by P. S. Spencer and H. H. Schaumburg. Baltimore, MD:
Williams and Wilkins, 1980.
11. Spencer, P. S., D. Couri and H. H. Schaumburg. n-Hexane and
methyl n-butyl ketone. In: Experimental and Clinical
Neurotoxicology, edited by P. S. Spencer and H. H. Schaum-
burg. Baltimore, MD: Williams and Wilkins, 1980.
12. Spencer, P. S., G. Foster, A. B. Sterman and D. Horoupian.
Acetyl ethyl tetramethyl tetralin. In: Experimental and Clinical
Neurotoxicology, edited by P. S. Spencer and H. H. Schaum-
burg. Baltimore, MD: Williams and Wilkins, 1980.
13. Spencer, P. S., M. I. Sabri, H. H. Schaumburg and C. Moore.
Does a defect of energy metabolism in the nerve fiber underlie
axon degeneration in polyneuropathies? Annls Neural. 5: 501,
1979.
14. Spencer, P. S. and H. H. Schaumburg. Classification of
neurotoxic disease: A morphological approach. In: Experi-
mental and Clinical Neurotoxicology, edited by P. S. Spencer
and H. H. Schaumburg. Baltimore, MD: Williams and Wilkins,
1980.
15. Veronesi, B., E. R. Peterson and P. S. Spencer. Reproduction
and analysis of MrtBK neuropathy in organotypic tissue culture.
In: Experimental and Clinical Neurotoxicology, edited by P. S.
Spencer and H. H. Schaumburg. Baltimore, MD: Williams and
Wilkins, 1980.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 193-206. ANKHO International Inc., 1979.
Physiological and Neurobehavioral Alterations
During Development in Lead Exposed Rats
DONALD A. FOX
Department of Animal Physiology, University of California, Davis, CA 95616
FOX, D. A. Physiological and neurobehavioral alterations during development in lead exposed rats. NEUROBEHAV.
TOXICOL. 1: Suppl. 1, 193-206, 1970.—Neonatal rats were exposed to lead (Pb) from parturition to weaning via the milk
of dams'which consumed 0 (tap water), 0.02% or 0.2% PbAc, solutions. To determine if this regimen altered physiological
and neurobehavioral development, responses to a battery of sensory-motor tests were evaluated during maturation and as
adults. The tests were: visual evoked responses (VER), temperature regulation, maximal electroshock seizure patterns,
reflex patterns, and neuromuscular performance. Overall results revealed that the Pb-exposed group compared to controls
exhibited delayed maturation, altered developmental patterns and long-term CNS disturbances. Additionally, low-level
strychnine administration during development caused additive interactions with both Pb groups, uncovering subtle effects
of toxicant exposure. These sensitive and quantifiable techniques proved useful for assessing CNS functioning following
perinatal insult, and except for the VER, are simple to conduct and cost efficient because they require a minimal amount of
personnel training, equipment cost and time invested per animal. These screening tests also suggest further areas of study
and may indicate the mechanism(s) responsible for the deficit.
Leadtoxicity Visual evoked response development Seizure development Behavioral development Tem-
perature regulation development Behavioral toxicology Drug-toxicant interactions Screening tests
DESCRIPTIONS of the neurologic sequelae of lead (Pb)
encephalopathy in children frequently include reports of
visual-motor deficits, recurrent grand mal seizures, retarded
behavioral development and electroencephalographic ab-
normalities [12, 57, 73, 80]. However, few investigators have
examined the extent of functional impairment in the
presumed asymptomatic pediatric population despite reports
of neurobehavioral deficits in this population [6,88] and .in
animals following low-level Pb exposure during development
[9, 13, 26, 27, 32, 54, 55, 74].
Sensory, motor and integrative systems function differ-
ently in the neonate and in the adult. During the early
postnatal or preweaning period in the rat, a time when the
most dynamic changes are occurring in the biochemical and
morphological components of the nervous system, the CNS
appears to be particularly sensitive to toxic insult [17,38].
During this developmental period different neural substrates,
responsible for the physiological and behavioral activity of
the organism, possess different spatio-temporal rates of ma-
turation [37,77]. The rates at which these dynamic changes
occur are subject, within certain genetic limits, to modifica-
tion by both internal and external environmental influences.
Recent evidence [16, 26, 27, 55, 64, 75] demonstrates that
subtle functional impairment occurs in animals subjected to
toxic insult during the prenatal and/or early postnatal period
at exposure levels below those producing observable abnor-
malities.
This report focuses on the physiological and neuro-
behavioral alterations during development resulting
from low-level neonatal Pb exposure. The primary purpose
of these studies was to determine the effects of this exposure
protocol on the ontogeny of (1) visual evoked responses
(VER), (2) temperature regulation, (3) maximal electroshock
seizure (MES) patterns, (4) reflex patterns and neuromuscu-
lar performance, and (5) somatic indices of development.
This battery of tests examining the sensory, motor and
integrative systems of the neonate was chosen to examine
the specificity of effect and to determine at what age the Pb
treatment effects first appeared. In order to study the inter-
active effects of Pb exposure and additional stressors, to
examine possible mechanisms of action of Pb, and to
validate our assessment techniques, some pups were in-
jected with a subconvulsive dose of strychnine prior to test-
ing. Strychnine, a neuropharmacological agent with known
mechanisms of action [21,28], has been used to assess the
maturation of inhibitory synapses [61,92].
Pb AND VER
The neurophysiological development of the visual system
following neonatal Pb-exposure was examined because
visual-motor and visual-perceptual deficits have been fre-
quently reported in children following Pb poisoning [8, 12,
57] and because long-term deficits in visual discrimination
learning have been reported in animals after early Pb-
exposure [9,13]. The first appearance of the VER establishes
the time at which this sensory system begins to function.
During CNS development two components of the VER
emerge in an orderly sequence. According to one group of
investigators [65] the ontogenetically early appearing, long-
latency, positive-negative complex (P2N2) is a manifestation
of activity in the nonspecific corticipetal radiation (superior
colliculus and pretectal region of the midbrain), whereas the
193
-------
194
FOX
Age (days)
Control
Lead
12
20
90
100
100 msec
FIG. 1. Patterns of visual evoked response in developing and adult rats following neonatal Pb exposure. Dams' were consuming 0.2% PbAc
or tap water. Average form of 20 VERs. Arrows indicate stimulus onset. Positive polarity downwards. a=P2; b=N2; c=Pl; d=Nl.
ontogenetically later-appearing, positive-negative complex
(P1N1) reflects activity in the specific sensory system (lat-
eral geniculate nucleus). Thus this method may provide a
means for differentiating lesion sites in the visual system
following perinatal insult [26, 49, 66, 89]. By measuring the
change in peak latency of each component of the VER with
age, the rate at which the neural generating systems respon-
sible for that component matures can be compared between
groups following exposure to a toxic agent [59, 65, 67].
In the first study, conducted in collaboration with Drs. J.
P. Lewkowski and G. P. Cooper, the effects of exposing
Long-Evans neonatal rats to Pb via dams' milk (dams'
-------
NEUROTOXIC EFFECTS OF LEAD DURING DEVELOPMENT
195
u
< 50
i I
Nl 0.2% Pb
PI 0.2% Pb
Nl Control
Control
°'—#-
o
15 16 17 18
AGE (DAYS POSTNATAL)
N2 0.2% Pb
N2 Control
P2 0.2 %Pb
P2 Control
z
5 200
AGE (DAYS POSTNATAL)
FIG. 2. Mean peak latencies of visual evoked response in developing and adult rats following neonatal
Pb exposure. Points presented indicate the means ± SE for six animals per treatment per age. The
data were analyzed by an ANOVA and the Student f-test and are presented in the text of paper. (Top)
Mean ± SE for PI and Nl. (Bottom) Mean ± SE for P2 and N2.
consuming 0.2% PbAc2) from parturition to weaning on brain
electrical activity were determined in acute experiments per-
formed on 84 male rats on Days 9-13, 16 and 20 and on 12
adult male rats (90-100 days old) [26]. Rats were anes-
thetized by pentobarbital sodium (30 mg/kg; IP) and the in-
stantaneous power spectrum was monitored on-line (Nicolet
Spectrum Analyzer, Model UA-500A), in order to maintain
the animals in a state of moderate to light anesthesia. VERs
were recorded on the visual cortex following a brief (10 ju-sec)
light flash produced by a Grass photostimulator. To evaluate
the responses, 20 VERs were averaged (Mnemotron CAT,
Model 400B) to produce one averaged VER. Five such
summations were made for each animal. An analysis of vari-
ance (ANOVA) with repeated measures showed no signifi-
cant time effect with respect to the five summations, so the
average of the five summations was used. The latencies of
each component (PI, Nl, P2, N2) were measured at the point
of greatest amplitude and calculated from the time of
-------
196
F0>
TABLE 1
LATENCIES (MSEC) OF VER COMPONENTS IN DEVELOPING AND ADULT RATS FOLLOWING NEONATAL LEAD
EXPOSURE3
Age
(Days)
9
10
11
12
13
16
20
90-100
Treatment
C
0.2% Pb
C
0.2% Pb
C
0.2% Pb
C
0.2% Pb
C
0.2% Pb
C
0.2% Pb
C
0.2% Pb
C
0.2% Pb
I
82.4
107.8
61
78,
53,
61
46.
60,
36.
49.
.5
,9
.1
.0
,2
.4
.0
,6
>
-
-
+
+
±
+
+
±
+
±
±
±
3.0
2.1cf
&
&
&
»
i
92.8
121.7
74.2
95.1
64.7
86.8
55.8
71.8
51.6
61.8
-
-_
± 3.8
± 2.5cf
±3.2
± 2.8f
±1.8
± 1.61
± 1.3
±2.5f
± 2.9
±1.7d
P,
224.6 ±
311.7 ±
153.7 ±
203.9 ±
138.1 +
174.2 ±
122.7 ±
134.6 ±
102.5 ±
126.2 ±
86.3 ±
110.5 ±
78.7 ±
99.6 ±
68.7 t
81.9 ±
9.6
26.3bd
3.2
12. 2e
1.8
3.3f
6.4
6.2
8.0
4.6d
3.7
3.8e
3.8
6.4e
4'5H
0.9d
N,
296.2 ±
384.9 ±
246.8 ±
294.2 ±
184.8 ±
275.9 ±
206.6 ±
277.4 ±
201.1 ±
203.4 ±
177.6 ±
193.2 ±
172.7 ±
211.3 ±
106.5 ±
129.7 ±
9.7
26.2bd
5.1
8.1e
7.1
7.6f
17.1
13.5
5.8
8.2
4'3H
5.8d
8.2
19.1
10.6
11.0
Values presented are means ± SE for 6 animals per treatment per age. The data were analyzed by four separate two-way
ANOVAs. Significant effects from these ANOVAs are presented in the text. Comparisons among treatment cell means are
presented here using the Student r-test. Significant differences between means are indicated by superscripts d, e and f.
Three of six animals exhibited no response.
?Four of six animals exhibited no response.
Significantly different from control value at p<0.05.
^Significantly different from control value at p<0-01.
Significantly different from control value at p<0.001.
stimulus onset. For this experiment and all others reported
herein, the experimenter was unaware of the animals' exper-
imental condition.
Rats exposed to Pb compared to controls showed no
changes in growth rate, brain wet weight, or brain dry weight
at any of the ages examined or in age of eye opening. At 21
days of age, the Pb concentration in the blood and brain of
controls were 6 ± 1 and 7 ± 1 yu,g%, respectively, while Pb-
exposed rats had 65 ± 9 and 53 ± 4 /xg%, respectively. By
Day 90, the Pb concentrations in blood and brain of both
groups were approximately 6 and 12 /u,g%, respectively [26].
The effects of Pb on the VER were: (1) to delay the ap-
pearance of each component, (2) to alter the waveform, and
(3) to increase the latencies of each component (Figs. 1 and
2). For example, on Day 9 when 100% of the control pups
developed the initial long-latency P2N2 complex, only 50%
of the Pb-exposed pups exhibited this response. Similarly on
Day 12, when 100% of the controls developed the short-
latency P1N1 complex, only 33% of the Pb-exposed pups
exhibited this response. Figure 1 (Lead, Days 20 and 90) is a
typical example which illustrates the altered waveform pres-
ent in approximately 30% of the Pb-exposed animals be-
tween Days 16 and 90. A similar alteration, attributed to a
demyelination of the rapidly conducting rate nerve fibers,
has been reported to occur in adult rats dosed subacutely
with an organic mercurial compound [41]. The latencies of
the VER components were all significantly longer (about
30%) in the Pb group compared to controls as analyzed by an
overall ANOVA.
Post-hoc comparisons of the means for the PI and Nl
components in the Pb group compared to controls revealed
highly significant differences at all ages, while similar com-
parisons for the P2 and N2 components revealed differences
at all ages except Day 12 for the P2 wave and Days 12, 13, 20
and 90 for the N2 wave (Table 1). The latter was due to the
wide variability found in this component.
The data from this study demonstrates that low-level de-
velopmental Pb exposure in age- and weight-matched
neonatal rats results in immediate and long-term
neurophysiological disturbances in a primary sensory sys-
tem. The effect upon the specific projection system probably
represents a decreased conduction velocity within the CNS.
Similar findings in peripheral nerves have been reported in
industrial lead workers with blood Pb concentrations below
70 Aig/100 ml [69]. An effect upon the nonspecific projection
system may have more serious implications for the long term
health of the animal, since it is well established that this
nonspecific system plays a fundamental role in the elabora-
tion of arousal, learning, attention, perception, emotional
behavior, as well as in the control of neocortical electrical
rhythms [42].
The long-term (90-100 days of age) effects observed in
-------
NEUROTOXIC EFFECTS OF LEAD DURING DEVELOPMENT
197
TABLE 2
PREWEANING BODY WEIGHTS (G) OF LEAD-EXPOSED AND SEIZURE-TESTED RATS
10
14
18
21
NoPb
No MES, No Strychnine
MES
MES + Strychnine
All Pups
0.02% Pb
No MES, No Strychnine
MES
MES + Strychnine
All Pups
0.2% Pb
No MES, No Strychnine
MES
MES + Strychnine
All Pups
9.0 ± 0.2
8.7 ± 0.2
8.5 ± 0.2
8.7 ± 0.1
9.6 ± 0.2
9.7 ± 0.2
9.6 ± 0.2
9.6 ± 0.1
9.2 ± 0.2
9.3 ± 0.2
9.1 ± 0.2
9.2 ± 0.1
14.3 ± 0.5
14.6 ± 0.4
14.2 ± 0.5
14.4 ± 0.3
15.1 ± 0.4
14.6 ± 0.4
14.7 ± 0.3
14.8 ± 0.2
14.4 ± 0.3
14.5 ± 0.4
14.2 ± 0.4
14.4 ± 0.2
19.7 ± 0.6
20.0 ± 0.5
19.9 ± 0.6
19.8 ± 0.3
19.6 ± 0.5
20.1 ± 0.4
20.0 ± 0.4
19.9 ± 0.1
19.5 ± 0.5
18.9 ± 0.4
18.6 ± 0.5
19.0 ± 0.3
27.6 ± 0.7
27.3 ± 0.6
26.9 ± 0.7
27.2 ± 0.4b
27.2 ± 0.7
27.2 ± 0.6
27.1 ± 0.5
27.2 ± 0.3C
25.8 ± 0.5
25.4 ± 0.4
25.2 ± 0.5
25.4 ± 0.3bc
36.7 ± 0.9df
34.6 ± 0.7
34.7 ± 0.8
35.0 ± 0.5b
35.0 ± 0.8
34.5 ± 0.7
34.3 ± 0.7
34.6 ± 0.5°
33.2 ±0.7
32.7 ± 0.5f
31.8 ± 0.6d
32.6 ± 0.4bc
43.8 ± 1.0d
41.6 ±0.8
40.6 ± 1.0
42.2 ± 0.6b
43.6 ± 1.0e<
41.5 ± 0.8 ,
42.3 ± 0.8f
42.6 ± 0.5C
41.8 ± 1.0
38.6 ± 0.7def
38.6 ± 0.9def
39.8 ± 0.5bc
aValues presented are means ± SE of body weights with 26-41 pups per lead exposure level by seizure condition (treatment cells).
The data were analyzed by (1) an overall unweighted means ANOVA with repeated measures on all the data, (2) an unweighted means
ANOVA at each age, and (3) multiple pairwise comparisons among treatment cell means at each age by Tukey's HSD test. Significant
effects from the overall ANOVA are presented in the text of paper. Significant Pb main effects from ANOVA at each age are indicated
on the rows labelled "All Pups" and "All Pups" groups which share the same superscript differed significantly as indicated below.
Significant differences between means at each age are indicated on the treatment cell means and groups with a prime superscript notation
differed from those groups having the same superscript. Thus f identifies a treatment cell mean which differed from all other treatment
cell groups of the same age labelled with an f. bcp<0.001; dep<0.01; fp<0.05.
animals whose blood and brain Pb concentrations do not
differ from those of controls, suggests a permanently altered
biochemical or cytoarchitectural substrate. This was re-
flected in increased latencies in the VER and a decreased
ability to follow repetitive flashes [26]. The latter effect indi-
cates that early Pb exposure caused a persistent delay in
CNS maturation since it has been shown that the ability of
the cerebral cortex to follow repetitive stimulation at increas-
ingly higher frequencies improves with age [22,59].
Pb AND TEMPERATURE REGULATION
Altered environmental temperatures have been shown to
influence the susceptibility of animals to Pb poisoning [5, 7,
62, 93]. However, the ontogeny of temperature regulation,
an exemplary of integrative functioning, has not been exam-
ined following exposure to this toxin. Rats, being
poikilothermic at birth, slowly develop the ability to main-
tain a constant body temperature over a wide range of am-
bient environmental temperatures [2, 25, 30]. This ability
develops during the second postnatal week [2, 25, 30],
thereby providing an early developmental period in which to
examine the effects of perinatal insult.
This second study, done in collaboration with Dr. R. L.
Bornschein, was designed to examine the effects of exposing
neonatal rats to Pb using the same exposure level and exper-
imental design as in the first study [26]. From Day 3 to 16
surface body temperature of 24 male control and 18 Pb-
exposed rats (one per litter) was determined and recorded.
Each pup was removed from the litter, immediately moni-
tored (YSI Telethermometer Model 43, Probe No. 708) on
the ventral surface (above heart), and isolated in a 1000 ml
polypropylene beaker for 60 min (ambient temperature,
22 ± 1°C). At the completion of this time the surface body
temperature was recorded, animals weighed and returned to
their home cages.
Initial body temperature (35.7 ± 0.3°C) was the same for
both groups of rats. However, following 60 min of isolation
there was a significant decrease in the final body temperature
of the Pb-exposed neonates compared to controls on Days
3-14 as indicated by an overall ANOVA (Fig. 3). Post-hoc
comparison of daily means revealed highly significant differ-
ences at all ages examined (Fig. 3).
The ontogenetic pattern of body temperature regulation in
controls is comparable to that observed by others [25]. A
similar pattern is seen in the Pb-exposed group, however,
there is a two to three day maturation lag in the development
of the thermoregulatory mechanisms. This is more clearly
seen in Fig. 4, which illustrates the degrees lost by each
group of pups on each experimental day. In addition to the
maturational delay, the slightly altered slope of the devel-
opmental curve in the Pb group (Fig. 3) suggests a functional
disturbance in this integrative process.
Pb AND MES
Recurrent grand mal seizures occur in approximately 50%
of the pediatric population exhibiting the effects of Pb
encephalopathy [57, 73, 80]. However, alterations in seizure
susceptibility in the presumed symptomatic pediatric popu-
lation have not been reported. Two reports recently ap-
peared which examined seizure thresholds in adult rodents
-------
198
FOX
TABLE 3
BLOOD AND BRAIN Pb CONCENTRATIONS (Mg%>) IN Pb-EXPOSED DAMS AND THEIR
OFFSPRING
Age
(Days)
No Pb
0.02%
0.2%
Blood
Brain
Blood
Brain
Blood
Brain
10
2 +
0.4 ±
21.7 +
6.3 ±
49.6 ±
18.8 ±
Pb-Exposed Offspring
21 60
0
0
1.1
1.0
4.2
1.3
2.2 ±
0.6 ±
25.2 ±
12.5 ±
89.4 ±
81.9 ±
0.1
0.1
3.3
1.6
15.4
23.3
3
3.2
2.5
6.9
5.0
16.3
± 0
± 0.6
± 0.3
± 1.8
± 0.5
± 2.4
Dams
21
2.4 ± 0.3
-
29.0 ± 3.5
-
71.9 ± 11.0
-
aValues presented are means ± SE with 5-13 animals per Pb treatment at each age. PbAc2
drinking solutions were provided ad lib to dams throughout lactation (Days 0-21), while controls
received tap water.
following either acute [1] or chronic developmental [71] Pb
exposure. However, neither study examined the effects of
Pb exposure on seizure responsiveness during development
and in addition, both studies were confounded by the failure
to use weight-matched experimental and control animals.
This third study, done in collaboration with Drs. S. R.
Overmann and D. E. Woolley, was designed to evaluate
changes in seizure development in neonatally Pb-exposed
rats by using the MES test [27]. The MES test is an experi-
mental model of grand mal convulsions and an overall indi-
cator of the level of CNS excitation [78,81]. Like the VER,
specific components of the MES response emerge in an or-
derly sequence during development [48,85]. In this study a
similar protocol as in the first two studies was used, how-
ever, an additional exposure level of Pb was included (0.02%
Pb acetate solution to the dam). At ten days of age pups
within each litter (3 males, 3 females) were assigned to one of
three seizure regimen conditions (No MES, No Strychnine;
MES; MES+Strychnine) as shown in Fig. 5. Repeated sei-
zure testing on the same pup with the same treatment was
carried out at 10, 12, 14, 16, 18, 20 and 60 (adults) days of
age. The total number of neonates studied was 235 from 69
dams while the total number of adults studied was 100. MES
seizures were induced by a constant current, 100 /xA, sine
wave (60 Hz) stimulus delivered for 0.2 sec to the animal's
scalp through silver disk electrodes with a stimulus of 100
mA for neonates and 1.0 mA/g body weight for adults [27, 85,
90, 91]. Four electronic timers were activated simulta-
neously with the shock to the animals and were turned off
manually at the end of each phase of the seizure: forelimb
flexion, forelimb extension, hindlimb flexion, and hindlimb
extension. A stopwatch was used to time postseizure de-
pression, determined as the time to recover righting reflex
(RRR). Total tonus was defined as the total duration of the
tonic seizure, from its onset to its completion.
Each rat was injected sc with either saline or one of two
doses of strychnine sulfate (125 or 500 /xg/kg; Merck) 30 min
prior to seizure testing. Strychnine sulfate was diluted in
physiological saline and given in a volume of 20 jul. Adult
rats were not injected prior to seizure testing. The 500 (Jig/kg
strychnine injection dramatically altered the developmental
MES pattern described above: all animals, regardless of age
or Pb exposure, responded with grade 5 seizures during MES
testing. Therefore, except where noted, discussion of results
using strychnine administration only refer to the 125 /ag/kg
dose.
Low-level Pb exposure had no effect on body growth,
whereas high-level Pb exposure resulted in a small depres-
sion (6-7%) of body growth at 14, 18 and 21 days of age
(Table 2). The overall AN OVA on body weight indicated
that there was a decrease due to Pb, an increase due to age,
and no effect due to MES (or MES+Strychnine) testing.
Blood and brain Pb values at weaning in the 0.2% Pb group
are similar to those reported in the first study, while values
for the 0.02% Pb group are 25.2 ± 3.3 and 12.5 ± 1.6 /ag%,
respectively (Table 3). Note that at 60 days of age the brain Pb
concentrations are still elevated in both Pb groups compared
to controls. No sex differences were found at any age with
regard to the seizure parameters examined, so all data pre-
sented represent the pooled values for males and females.
MES seizure ontogeny in rats is characterized by the se-
quential appearance of graded seizures of increasing sever-
ity: grade 1—clonus; grade 2—tonic florelimb flexion (FF);
grade 3—FF followed by tonic forelimb extension (FE);
grade 4—FF plus tonic hindlimb flexion (HF) and FE; and
grade 5—FF, HF, FE and tonic hindlimb extension (HE).
Neonatally Pb-exposed rates exhibited: (1) earlier on-
togenetic appearance of tonic seizures, (2) increased seizure
severity during development and (3) long-term disturbances
in seizure responsiveness. Low-level strychnine administra-
tion during development caused additive interactions with
both Pb groups, uncovering subtle effects of toxicant expo-
sure. Alterations in seizure responsiveness were observed in
both high and low Pb-exposed rats.
The most sensitive and reliable measure for detecting
early maturational differences between No Pb and Pb-
treated rats was the percent (Chi-square tests) of animals
exhibiting clonus (grade 1) versus the percent exhibiting
tonus (grade 2 or greater). In both No Pb and 0.02% Pb-
exposed rats approximately 90% responded with clonus on
Day 10 while approximately 70% responded with clonus on
Day 12. On Day 12 the remainder exhibited tonic seizure
grades 2 and 3 (Table 4). In contrast, the 0.2% Pb-exposed
rats responded with approximately 25% clonus on Days 10
and 12, while the remainder exhibited tonic seizure grades 2,
3 and 4 (Table 4).
Figures 6 and 7 illustrate the cumulative percentage of
neonates exhibiting each seizure phase. For example, a pup
-------
DISTRIBUTION OF NEONATAL RATS
Pb
Age
(Days)
10
12
14
16
18
20
Grades
Treatment
Drug
Treatment
Saline
Strychnine
Saline
Strychnine
Saline
Strychnine
Saline
Strychnine
Saline
Strychnine
Saline
Strychnine
0
95a
100s
79a
87g
21C
40kg
0
0
0
0
0
0
1
0.02%
90b
82h
67b
68h
15e
14k
0
0
0
0
0
0
0.2%
26ab
llgh
23ab*
ogh*
4ce
Og
0
0
0
0
0
0
0
oa
og
2a
Og
16
20
0
0
0
0
0
0
2
0.02%
5b
14h
7b
5h
20
19
0
0
0
0
0
0
TABLE 4
EXHIBITING VARIOUS SEIZURE GRADES DURING MES3
0.2%
47ab
56gh
40ab
39gh
17
28
0
0
0
0
0
0
0
2b
Og
19
7
*
49
20k*
60a
53gh
20
7
2
0
3
0.02%
5C
oh
24
18
54
38g
49t
10gt
21*
0*
2
0
0.2%
nbc
33gh
27
28
+
44 T
Okgt
Cl *
29a
6h*
13
0
4
0
0
0
0
0
0
12
Og
18cd
20*
10
7
5
0
4
0.02%
0
5
2
0
12
10
43C
52ij
18
20
13
0
0.2%
9
0
8
0
27
44s
44d
22j
20
11
4
0
0
2
0
0.
71
*
2
20*
22e
27g
70
87
93
100
5
0.02%
0
0
0
9k
*
0
19*
geff
38St
61
80
85
100
0.2%
2
0
2t
33ikt
*
28*
27ft
72git
67
89
91
100
EUROTOXIC EFFECTS O
W
O
o
C|
5
3
o
o
<
td
r
^^
hj
W
2
aValues presented are the percent of neonates exhibiting each seizure grade (see Results section for definition of each grade) during MES with 15-48 animals per
Pb treatment by seizure group. For each Pb treatment at each age, bold-faced numbers indicate the MES grade shown by the major proportion(s) of saline or
strychnine injected pups. Strychnine (125 Mg/kS s-c-) was injected 30 min prior to MES testing. The effects of lead exposure are evaluated statistically by x2 analyses
of values for pups within the same seizure grade, age, and drug treatment (saline or strychnine); values sharing the same superscript letter differed at the following
levels of significance: abshp<0.01; cdlip<0.05; efk0.05
-------
200
FOX
36 n
8 9 10 II \Z 13 14 15
AGE (days)
FIG. 3. Final body temperature of control and neonatally fto-
exposed rats following 60 min isolation from dam. Points indicate
means ± SE for 18-24 male rats. Initial temperature was the same
for both groups, 35.7 ± 0.3°C. Ambient temperature was
22.0 ± 1.0°C. The data were analyzed by an ANOVA and are pre-
sented in the text of paper. Post hoc comparison of means within
days are indicated by superscript: p<0.001 for a; p<0.02 for b.
EXPERIMENTAL DESIGN PER LITTER
-CT
Pb EXPOSURE
TO LACTATING
DAM
(0,.02,.2%PbAc2
Drinking Solution)
-No MES, No STRYCHNINE
-MES
-MES+STRYCHNINE
(125 or 500/ig/kg)
L9
-No MES, No STRYCHNINE
-MES
-MES+STRYCHNINE
(125 or 500/ig/kg)
FIG. 5. Experimental design per litter for assignment of pups to
seizure treatment conditions at ten days of age. At parturition litters
were reduced to three males and three females and dams were pro-
vided one of three drinking fluids—tap water, 0.02% or 0.2% lead
acetate (PbAc), in aqueous solution. One pup of each sex was not
seizure tested (No MES, No Strychnine), was tested for maximal
electroshock seizure (MES) pattern or was dosed sc with strychnine
prior to MES testing (MES+Strychnine).
demonstrating HE, will also exhibit FF, FE and HF, and
therefore, was included in the number demonstrating each of
these phases. Presented in this way the developmental data
show even more clearly that a greater proportion of pups in
the 0.2% Pb-exposed group exhibited a higher seizure grade
than in the other two groups at 10, 12 and 14 days of age.
On Day 12 an additive effect, indicated by a shift in sei-
zure distribution, was observed between the high Pb group
and the low-level strychnine injection (Table 4). On Day 16
strychnine administration increased the severity of the sei-
zures by about one seizure grade in both Pb groups. The low
Pb group responded almost equally with grade 4 (52%) and 5
(38%) seizures, while the high Pb group exhibited predomi-
B 3
Q CONTROL
0
Ji
rttrV
4567
IZ 13 14 15 16
AGE (days)
FIG. 4. Degrees lost by control and neonatally Pb-exposed rats fol-
lowing 60 min isolation from dam. Bars indicate mean ± SE for
18-24 male rats. See Fig. 3 for details.
FORELIMB FLEXION .
100 r ^GRADE 2 e,n,p ,8 o',r
75
50
V)
§25
Q.
UJ
CE
FORELIMB EXTENSION.
->GRADE3
Yb,h,i i
O 10 12 14 16 10 12 14 16
C HINDLIMB FLEXION h,i,j HINDLIMB EXTENSION
" >GRADE4 Jj'k||,mrGRADE5
o,t.
75
50
25
o,p,q
k,l,m,n
10 12 14 16 18 20 10 12 14 16 18 20
AGE (Days)
FIG. 6. Development of MES in Pb-exposed saline-injected neonatal
rats: Tonic Phases. Points presented indicate cumulative percentage
of neonates exhibiting a response with 42-48 animals per Pb treat-
ment condition. Statistical significance levels reported are based on
X2 analysis. Note that comparisons illustrate both Pb treatment and
developmental (within Pb treatment) differences. Pb effects are
compared with superscripts a through g. Pb groups sharing the same
superscript differed at the following levels of significance: p<0.001
for a,b,c,d; /?<0.02 for e; p<0.05 for f,g. Developmental effects are
compared with superscripts h through t. Developmental groups shar-
ing the same superscript differed at the 0.02 level of significance.
-------
NEUROTOXIC EFFECTS OF LEAD DURING DEVELOPMENT
201
FLEXION
100 r
ONo Pb
A0.02%Pb
0.2%Pb
FORELIMB EXTENSION
>GRADE 3
14
16
10
12
14
16
100
ffl
I
X
75
50
25
HINDLIMB FLEXION
->GRADE4
a*l
HINDLIMB EXTENSION
rGRADE 5
e,g
10 12 14 16 18 20 10 12 14 16 18 20
AGE (days)
FIG. 7. Development of MES in Pb-exposed strychnine-injected
(125 /ig/kg) neonatal rats: Tonic Phases. Points presented indicate
cumulative percentage of neonates exhibiting a response with 15-22
animals per Pb treatment condition. Statistical significance levels
reported are based on x2 analysis. Note that comparisons illustrate
both Pb treatment and saline versus strychnine injected treatment
differences. Pb groups sharing the same superscript differed at the
following levels of significance: p<0.001 for a,b,c,d;p<0.02 for e,f:
p<0.05 for g. Asterisks indicate difference from similarly aged and
Pb-treated saline injected neonatal rats (see Fig. 6) at the following
levels of significance: p<0.005 for ***; p<0.02 for **; p<0.05 for *.
nantly grade 5 seizures (72%). By contrast, strychnine-
injected No Pb pups exhibited primarily (53%) grade 3 sei-
zures (Table 4).
In the high Pb group, strychnine injection increased the
cumulative incidence of FF, FE, HF, and HE at 12 days of
age, when compared with saline-injected pups in the same
group (Fig. 7). At 14 days of age strychnine treatment com-
pared to saline treatment increased the incidence of HF and
HE, but only in the 0.2% Pb group. At 16 days of age
strychnine increased the incidence of HF and HE in both Pb
groups, but not in the No Pb group (Fig. 7). Thus, at this age,
the combination of MES-(-Strychnine uncovered a subtle ef-
fect of 0.02% Pb exposure not detected by the MES test
alone.
The additive interactions between Pb exposure and
strychnine during maturation may be attributable to Pb-
induced alterations in postsynaptic inhibition since
strychnine, a drug which removes background inhibition in
the spinal cord by blocking postsynaptic inhibition [21,28], at
both doses modified the seizure pattern in both Pb groups to
a greater degree than in No Pb pups. In addition, the more
25
20
15
_FORELIMB EXTENSION
QNoPb
A0.02%Pb
0.2%Pb
10
o
-------
202
FOX
TABLE 5
SUMMARY OF ANALYSIS OF VARIANCE OF DURATIONS OF MES SEIZURE PHASES IN Pb AND STRYCHNINE TREATED
NEONATAL RATS
Seizure
Phase
Examined
Forelimb
Extension
Hindlimb
Extension
Total
Tonus
Recovery
of Righting
. Reflex
Age
Studied
20
18
16
20
18
20
18
16
20
18
16
Pb
(0, 0.02, 0.2%)
ttt
ttt
ttt
ttt
ttt
ttt
ttt
ttt
ttt
ttt
ttt
Treatment Groups Included in
Strychnine
(0, 125, 500 Mg/kg)
ttt
ttt
ttt
ttt
ttt
ttt
ttt
. ttt
ttt
NS
ttt
ANOVA Analyses3
Pb
(0, 0.02%)
tt
NS
t
'tt
NS
tt
NS
t
NS
NS
NS
Strychnine
(0, 125 Mg/kg)
NS
t
NS
NS
tt
NS
NS
NS
NS
NS
NS
Arrows presented indicate results of analyses using an overall unweighted means analysis of variance (ANOVA) with 15-48 animals
per Pb treatment by seizure condition. The data analyzed by the ANOVA was partitioned into a 3x3 and then a 2x2 factorial design
in order to examine the effects of all Pb and strychnine treatments and the low Pb and low dose strychnine treatment, respectively.
Levels of significance as determined by the ANOVA are represented by arrows and are indicated as follows: p<0.01 for ttt;p<0.02
for tt;p<0.05 for t.
100
2 "
-------
NEUROTOXIC EFFECTS OF LEAD DURING DEVELOPMENT
203
TABLE 6
AGE AT APPEARANCE OF AUDITORY STARTLE, EYE OPENING AND AIR RIGHTING3
No Pb
0.02% Pb
0.2% Pb
All Pups
Auditory Startle
No MES, No Strychnine
MES
MES + Strychnine
All Pups
11.
12.
.9 ±
.2 ±
12.3 ±
12.
1 ±
Eye Opening
Air
No MES, No Strychnine
MES
MES + Strychnine
All Pups
Righting
No MES, No Strychnine
MES
MES + Strychnine
All Pups
13.
13.
13.
13.
16.
17.
17.
17.
,5 ±
7 ±
,7 ±
6 ±
8 ±
2 ±
3 ±
1 ±
O.lg
0.1
O.lg
O.ld
O.lb
0.1
0.2
O.lbc
0.2g
0.1
0.2g
0.1
11.8
12.0
12.2
12.0
13.8
13.8
14.1
13.9
17.0
17.3
17.3
17.2
±
±
±
±
±
±
±
+
±
+
±
±
0.2
0.2
0.2
O.ld
0.2b
0.1
0.1
O.lb
0.1
0.1
0.2
O.lc
11.7
12.0
12.3
12.0
13.7
13.9
13.7
13.8
17.3
17.8
17.6
17.6
±0.1h
± 0.2
±0.2h
± 0.1
± 0.2
± 0.2
+ 0.2
±0.1C
± 0.3
± 0.2
± 0.2
±0.1C
11.8
12.1
12.2
13.7
13.8
13.8
17.0
17.4
17.4
±0.1C
± 0.1
±0.1e
± 0.1
± 0.1
± 0.1
± O.lef
± O.le
±0.1f
aValues presented are means t SE with 25-38 animals in each Pb exposure by seizure regimen condition.
Statistical significance levels reported are based on x2 analyses of the number of pups in which landmark first
appeared on each day's teating. Presentation of mean ± SE values was chosen to facilitate data reduction. Groups
sharing the same superscript differed at the following levels of significance: Pb exposure effects - bp<0.01,
cp<0.05, d0.10>p>0.05; Seizure regimen effects - ep<0.01, fp<0.05, gh0.10>p>0.05.
hibited the least significant effect while the
MES+Strychnine group exhibited the most significant effect
(Fig. 9B).
Analysis of the HF duration data for those rats exhibiting
grade 4 seizures revealed no differences between any Pb or
preweaning treatment condition. However, the mean HF du-
ration for these flexor rats was 10-11 sec compared to only
3-5 sec mean duration for those rats exhibiting both HF and
HE. This suggests that the altered seizure pattern in adults
was due to greater inhibition on extensor than on flexor
motoneurons. Woolley [91], using DDT or pentobarbital,
demonstrated that when the duration of flexion exceeded 5
sec, extension was not likely to occur.
The results of this investigation demonstrated that
neonatal Pb exposure in rats produced two temporally-
distinct alterations in MES responses: one during prewean-
ing development and the other at 60 days of age. Neonatal
rats exposed to Pb exhibited earlier ontogenetic appearance
of tonic seizures, increased duration of recovery of righting
reflex and increased seizure severity during development.
Although small (6-7%) differences in body weight were
found between the No Pb and the 0.2% Pb-exposed group
during development, these effects were not seen prior to Day
14 and therefore cannot account for the observed alterations
in seizure responses in the 0.2% group on Day 10. At 60 days
of age, regardless of neonatal seizure treatment, most Pb-
exposed rats did not undergo hindlimb extension during the
MES, whereas the controls did.
Pb AND REFLEX PATTERNS AND NEUROMUSCULAR PERFORM-
ANCE
Measures of reflex ontogeny and neuromuscular devel-
opment such as, auditory startle, air righting, eye opening,
and performance on negative geoxtaxis and forelimb sus-
pension, provide simple, yet useful, data on the overall
sensory-motor competence of the developing neonate. Al-
terations in these measures have been induced by a variety
of perinatal insults [11, 20, 52, 55, 63, 72, 76, 86, 92].
This study, done in collaboration with Drs. Overmann
and Woolley, uses the same animals as in the MES study
[55]. Age at first appearance of full eye opening (both eyes),
auditory startle and air righting as well as testing on the
negative geotaxis and forelimb suspension were determined
as described previously [31,55,92]. From nine to 15 days of age
each pup was observed for eye opening and auditory startle
while air righting was examined from 15 days of age until the
appearance of a criterion response. From four to eight
days of age each pup was evaluated on the negative geotaxis,
a measure of neuromuscular performance and coordination.
On Days 10 to 14 pups were tested on forelimb suspension, a
measure of forelimb muscular strength and endurance.
Age of full eye opening was delayed approximately one-
half day by both low and high Pb exposure compared to
controls, but was unaffected by MES testing. The develop-
ment of the air-righting reflex was delayed one-half day by
the high Pb treatment while MES testing delayed the appear-
ance of this reflex an additional half-day in all seizure-tested
rats compared to non-seizure tested rats. Pb exposure had no
effect on the first appearance of the auditory startle response
whereas MES+Strychnine treatment delayed the onset of
this response one-half day in all pups (Table 6).
These results agree with others [63] who found no effect
of maternal Pb exposure on age at appearance of the auditory
startle response and approximately one-half day delay in full
eye opening and appearance of the air-righting reflex in rat
offspring of normal body weight. Delayed eye opening was
also reported in Pb-exposed mice with markedly impaired
body growth [70]. The effect on eye opening in exposed ro-
-------
204
FOX
z 50
rr
h-
10 40
m
0
<5 30
UJ OU
CO
o
S 20
fj
I0.
tf
IV ONoPb
.. A.02%Pb
}• .2% Pb
*>'
fa.b
j[
Pi
r
50 b T
4*
5
i i i i i i
45678
DAYS OF AGE
FIG. 10. Negative geotaxis performance of control and postnatally
Pb-exposed rats. Points indicate means ± SE with 28-30 perform-
ance scores (two scores per litter—one for females and one for
males) per data point. Pups were placed head downward on an in-
clined plane and the latency to a criterion turn up the plane was
measured. ANOVA on the groups' scores showed the following
relations: No Pb vs 0.02% Pb -0.10>p>0.05: No Pb vs 0.2% Pb—
p>0.10: 0.02% Pb vs 0.2% Pb—p<0.05. For pairwise comparisons
within days: ap<0.05 "0.10>p>0.05.
dents probably reflects Pb-induced delay in normal somatic
ontogeny.
Negative geotaxis, a test using gravity as an eliciting
stimulus, was conducted by placing rat pups head down-
ward longitudinally aligned on an inclined plane (20° from
horizontal). Pups were scored on the direction and latency to
turn 135° from the starting line [55]. All pups regardless of
treatment showed a developmental decrease in the latency
for negative geotaxis-induced turning (Fig. 10). The low Pb-
exposed rats exhibited a borderline significant increase in
latency (were slower) compared to the other two groups. No
differences were found with regard to the proportion of right
or left turns between groups. The results from this devel-
opmental test are internally consistent with the air-righting
reflex data in this study. These data suggest that Pb acts to
impair muscular coordination.
In contrast to the above findings, no Pb group differences
were found with respect to the forelimb suspension test. A
slight effect due to MES testing was found in all Pb treatment
groups [55]. These data suggest that Pb exposure does not
alter forelimb muscular strength and endurance.
Pb AND SOMATIC INDICES OF DEVELOPMENT
In collaboration with Drs. Overmann and Woolley,
hematocrit, brain and organ weight data were obtained from
litter mates of rats used in the MES study, providing addi-
tional indices of the effects of Pb exposure on somatic devel-
opment. Hematocrit, organ weight data, water intake and
growth measures, were also collected for dams as physiolog-
ical indicators of the effects of maternal Pb exposure [55].
Dams' drinking water with 0.2% PbAc2 added had de-
creased fluid consumption by approximately 15% compared
to the No Pb and 0.02% groups. Despite this reduced fluid
intake, no differences in dam body weights due to Pb expo-
sure were found during the lactation period. Dams' drinking
either 0.02% or'0.2% Pb solutions throughout lactation had
hypertrophied kidneys, reduced hematocrits, and normal
weight adrenals.
Pups, like dams, at weaning in both Pb groups exhibited
enlarged kidneys and reduced hematocrits. In addition,
pups in 0.2% group had decreased adrenal weights compared
to the other two groups. In agreement with Fox et al. [26], no
differences in whole brain (minus cerebellum) or cerebellum
weight were found between groups.
PHYSIOLOGICAL AND BIOCHEMICAL BASIS FOR Pb EFFECTS
Although the physiological and biochemical basis for the
developmental and long-term effects of low level Pb expo-
sure are not clear, delays in synaptogenesis [46] and myelin
formation [40], alterations in neurotransmitter systems [19],
brain electrolytes [29] and amino acid transport [44] have all
been reported following neonatal Pb exposure. It is also
possible that the interference of Pb in the normal processes
of transmitter release, demonstrated in peripheral synapses
[15, 39, 45] may permanently alter synaptic function if this
disturbance occurs during the early developmental period.
Factors which influence the development of the VER, tem-
perature regulation, MES, reflex patterns and neuromuscu-
lar performance include synaptogenesis, myelination, ma-
turation of neurotransmitter systems, and metabolic changes
in the rat brain during development [10, 34, 36,43, 49, 50, 51,
59, 60, 65, 67, 82, 85]. Thus, any or all of the above
neurotoxic effects of Pb may account for the observed data.
CONCLUSIONS
This battery of developmental sensory-motor tests pro-
vides sensitive and quantifiable techniques for assessing de-
velopmental and long-term alterations in CNS functioning (in
the same animal, except for the VER as conducted herein)
following perinatal insult. All of these tests, except the VER,
are simple and cost efficient because they can be performed
with a minimal amount of: (1) personnel training, (2) equip-
ment cost, and (3) time invested per animal, thereby making
them ideal screening tests. In addition, these tests suggest
further areas of study, hopefully leading to the underlying
mechanism(s) responsible for the deficit.
ACKNOWLEDGEMENTS
Figures and tables have been reprinted with permission as fol-
lows: Figures 1 and 2 and Table 1 from D. A. Fox, J. P. Lewkowski
and G. P. Cooper [26]: Figures 5 and 10 and Tables 2 and 6 from S.
R. Overmann, D. A. Fox and D. E. Woolley [55]: Figures 6, 7 and 9
and Tables 3, 4 and 5 from D. A. Fox, S. R. Overmann and D. E.
Woolley [27]. Figure 8 was drawn from the data in the latter refer-
ence while Figs. 3 and 4 represented unpublished data by D. A. Fox
and R. L. Bornschein. Original research was supported by N1H
Grants ES-00127 (D. A. Fox), ES-05094 (D. A. Fox), ES-05057 (S.
R. Overmann) and ES-01503 (D. E. Woolley).
I thank Dr. D. E. Woolley for her thoughtful criticism of this
manuscript. The expert assistance of Ms. Mary Lou Rodriguez in
typing this manuscript is gratefully acknowledged.
-------
NEUROTOXIC EFFECTS OF LEAD DURING DEVELOPMENT
205
REFERENCES
1. Adler, M. W. and C. H. Adler. Toxicity to heavy metals and
relationship to seizure thresholds. Clin. Pharmac. Ther. 22:
774-779, 1977.
2. Adolf, E. F. Ontogeny of physiological regulations in the rat. Q.
Rev. Biol. 32: 89-137, 1957.
3. Alvares, A. P., S. Leigh, J. Cohn and A. Kappas. Lead and
methyl mercury: effects of acute exposure on cytochrome P-450
and the mixed function oxidase system in the liver. J. e\p. Meet
135: 1406-1409, 1972.
4. Alvares, A. P., A. Fischbein, S. Sassa, K. E. Anderson and A.
Kappas. Lead intoxication: Effects on cytochrome P-450
mediated hepatic oxidations. Clin. Pharmac. Ther. 19: 183-190
1976.
5. Baetjer, A. M., S. N. D. Joardar and W. A. McQuary. Effect of
environmental temperature and humidity on lead poisoning in
animals. Archs envir. Hlth. 1: 463-477, 1960.
6. Baloh, R. W. The effects of chronic increased lead absorption
on the nervous system—A review article. Bull. L. A. Neural.
Soc. 38: 91-99, 1973.
7. Blackman, S. S. The lesions of lead encephalopathy in children.
Bull. Johns Hopkins Hasp. 61: 1-62, 1937.
8. Bradley, J. E. and R. J. Baumgartner. Subsequent mental de-
velopment of children with lead encephalopathy as related to
type of treatment. J. Pediatr. 53: 311-315, 1958.
9. Brown, D. Neonatal lead exposure in the rat: Decreased learn-
ing as a function of age and blood lead concentration. Toxic.
appl. Pharmac. 32: 628-637, 1975.
10. Buchanan, A. R. and R. M. Hill. Temperature regulation in
albino rats correlated with determinations of myelin density in
the hypothalamus. Proc. Soc. exp. Biol. Med. 66: 602-608,
1947.
11. Butcher, R., K. Hawver, K. Kazmaier and W. Scott. Postnatal
behavioral effects from prenatal exposure to teratogens. In:
Basic and Therapeutic Aspects of Perinatal Pharmacology,
edited by P. Morselli, S. Garattini and F. Sereni. New York:
Raven Press, 1974, pp. 171-176.
12. Byers, R. and E. Lord. Late effects of lead poisoning on mental
development. Am. J. Dis. Child. 66: 471-494, 1943.
13. Carson, T., G. Van Gelder, G. Karas and W. Buck. Slowed
learning in lambs prenatally exposed to lead. Archs envir. Hlth.
29: 154-156, 1974.
14. Chow, C. P. and H. H. Cornish. Effects of lead on the induction
of hepatic microsomal enzymes by phenobarbital and 3,4-
benzpyrene. Toxic, appl. Pharmac. 43: 219-228, 1978.
15. Cooper, G. P. and D. Steinberg. Effects of cadmium and lead on
adrenergic neuromuscular transmission in the rabbit. Am. J.
Physiol. 232: C128-C131, 1977.
16. Coyle, I., M. J. Wayner and G. Singer. Behavioral
teratogenesis: A critical evaluation. Pharmac. Biochem. Behav.
4: 191-200, 1976.
17. Dobbing, J. Undernutrition and the developing brain. In: Devel-
opmental Neurobiology, edited by W. A. Himwich. Springfield,
Illinois: C. C. Thomas, 1970, pp. 241-261.
18. Domer, F. R. and J. C. Llera. Blood-brain barrier permeability
changes caused by lead exposure and amphetamine in mice.
Res. Communs Psychol. Psychiatr. Behav. 3: 101-108, 1978.
19. Dubas, T. C., A. Stevenson, R. L. Singhal and P. D. Hrdina.
Regional alterations of brain biogenic amines in young rats fol-
lowing chronic lead exposure. Toxicology 9: 185-190, 1978.
20. Eayrs, J. and W. Lishman. The maturation of behaviour in
hypothyroidism and starvation. Br. J. Anim. Behav. 3: 17-24,
1955.
21. Eccles, J. C. Pharmacology of central inhibitory synapses. Br.
Med. Bull. 21: 19-25, 1965.
22. Ellingson, R. J. and R. C. Wilcott. Development of evoked
responses in visual and auditory cortices of kittens. J.
Neurophysiol. 23: 363-375, 1960.
23. Esplin, D. W. and. J. W. Freston. Physiological and phar-
macological analysis of spinal cord convulsions. J. Pharmac.
exp. Ther. 130: 68-80, 1960.
24. Fischbein, A., A. P. Alvares, K. E. Anderson, S. Sassa and A.
Kappas. Lead intoxication among demolition workers: The ef-
fect of lead on the hepatic cytochrome P-450 system in humans.
J. Toxic. Envir. Hlth. 3: 431-437, 1977.
25. Fowler, S. J. and C. Kellogg. Ontogeny of thermoregulatory
mechanisms in the rat. J. comp. physiol. Psychol. 89: 738-
746, 1975.
26. Fox, D. A., J. P. Lewkowski and G. P. Cooper. Acute and
chronic effects of neonatal lead exposure on development of the
visual evoked response in rats. Toxic, appl. Pharmac. 40: 449-
461, 1977.
27. Fox, D. A., S. R. Overmann and D. E. Woolley.
Neurobehavioral ontogeny of neonatally lead exposed rats. II.
Maximal electroshock seizures in developing and adult rats.
Neurotoxicology 1: 148-169, 1979.
28. Franz, D. N. Central nervous system stimulants. In: The Phar-
macological Basis of Therapeutics, 5th edition, edited by L. S.
Goodman and A. Oilman. New York: Macmillan Publishing
Company, 1975, pp. 359-361.
29. Goldstein, G. W. and I. Diamond. Metabolic basis of lead
encephalopathy. Res. Publs Ass. Res. nerv. ment. Dis. 53:
293-303, 1974.
30. Hahn, P., O. Koldovsky, J. Krecek, J. Martinek and Z. Valek.
Endocrine and metabolic aspects of the development of
homeothermy in the rat. In: Somatic Stability in the Newly
Born. Boston: Little, Brown and Company, 1961, pp. 131-155.
31. Hard, E. and K. Larsson. Development of air righting in rats.
Brain Behav. Evolut. 11: 53-59, 1975.
32. Hastings, L., G. Cooper, R. Bornschein and I. Michaelson.
Behavioral effects of low level neonatal lead exposure. Phar-
mac. Biochem. Behav. 1: 37-42, 1977.
33. Heim, L. M. and P. S. Timiras. Gonad-brain relationship: Pre-
cocious brain maturation after estradiol in rats. Endocrinology
72: 598-606, 1963.
34. Hensel, H. Neural processes in thermoregulation. Physiol. Rev.
53: 948-1017, 1973.
35. Hurley, L. S., D. E. Woolley, F. Rosenthal and P. S. Timiras.
Influence of manganese on susceptibility of rats to convulsions.
Am. J. Physiol. 204: 493-496, 1963.
36. Jacobsen, S. Sequence of myelinization in the brain of the al-
bino rat. A. Cerebral cortex, thalamus, and related structures.
J. comp. Neural. 121: 5-29, 1963.
37. Jacobson, M. Developmental Neurobiology. New York:
Plenum Press, 1978.
38. Kostial, K., T. Malijkovic and S. Jugo. Lead acetate toxicity in
rats in relation to age and sex. Archs Toxikol. 31: 265-269, 1974.
39. Kober, T. and G. P. Cooper. Lead competitively inhibits
calcium-dependent synaptic transmission in the bullfrog sym-
pathetic ganglion. Nature, London 262: 704-705, 1976.
40. Krigman, M. R., M. J. Druse, T. D. Traylor, M. H. Wilson, L.
R. Newell and E. L. Hogan. Lead encephalopathy in the devel-
oping rat: Effect upon myelination. J. Neuropath, exp. Neural.
33: 58-73, 1974.
41. Lehotzky, K. and I. Meszaros. Alteration of electroencephalog-
ram and evoked potential in rats induced by organic mercury.
Ada pharmac. tax. 35: 180-184, 1974.
42. Lindsley, D. B. The role of nonspecific reticulo-thalamocortical
systems in eotion. In: Physiological Correlates of Emotion,
edited by P. Black. New York: Academic Press, 1970, pp. 147-
188.
43. London, E. D., F. J. Vocci and G. Butterbaugh. Age-dependent
modification of the maximal electroshock convulsive threshold
by central monoamine reduction in rats. Pharmacologist 19:
214, 1977.
44. Lorenzo, A. V. and M. Gewirtz. Inhibition of [I4C] tryptophan
transport into brain of lead exposed neonatal rabbits. Brain Res.
132: 386-392, 1977.
45. Manalis, R. S. and G. P. Cooper. Presynaptic and postsynaptic
effects of lead at the frog neuro-muscular junction. Nature,
London 243: 354-356, 1973.
-------
206
FOX
46. McCauley, P. T. and R. J. Bull. Lead-induced delays in synap-
togenesis in the rat cerebral cortex. Fedn Proc. 37: 740, 1978.
47. Miller, E., P. Sapienza, T. C. Michel, V. L. Olivito, E. L. Earl
and E. J. Van Loon. Some aspects of neurotoxic effects of lead
in neonate beagle dogs. Fedn Proc. 34: 287, 1975.
48. Millichap, J. G. Development of seizure patterns in newborn
animals. Significance of brain carbonic anhydrase. Proc. Soc.
exp. Biol. Med. 96: 125-129, 1957.
49. Mourek, J., W. A. Himwich, J. Myslivecek and D. A. Callison.
The role of nutrition in the development of evoked cortical re-
sponses in rat. Brain Res. 6: 241-251, 1967.
50. Myers, R. D. Temperature regulation: Neurochemical systems
in the hypothalamus. In: The Hypothalamus, edited by W.
Haymaker, E. Anderson and W. Nauta. Springfield, Illinois: C.
C. Thomas, 1969, pp. 506-523.
51. Noback, G. R. and D. P. Purpura. Postnatal ontogenesis of
neurons in cat neocortex.7. comp. Neural. 117: 291-307, 1961.
52. Olson, K. and G. Boush. Decreased learning capacity in rats
exposed prenatally and postnatally to low doses of mercury.
Bull. Environ. Contam. Toxic. 13: 73-79, 1975.
53. O'Tuama, L. A., C. S. Kim, J. T. Gatzy, M. R. Krigman and P.
Mushak. The distribution of inorganic lead in guinea pig brain
and neural barrier tissues in control and lead-poisoned animals.
Toxic, appl. Pharmac. 36: 1-9, 1976.
54. Overmann, S. Behavioral effects of asymptomatic lead expo-
sure during neonatal development in rats. Toxic, appl. Phar-
mac. 41: 459-471, 1977.
55. Overmann, S. R., D. A. Fox and D. E. Woolley.
Neurobehavioral ontogeny of neonatally lead-exposed rats. I.
Reflex development and somatic indices. Neurotoxicology 1:
125-147, 1979.
56. Pentschew, A. and F. Garro. Lead-encephalo-myelopathy of
the suckling rat and its implications on the porphyrinopathic
nervous diseases. Acta Neuropath. 6: 266-278, 1966.
57. Perlstein, M. A. and R. Attala. Neurologic sequelae of plum-
bism children, din. Pediatr. 5: 292-298, 1966.
58. Petito, C. K., J. A. Schaefer and F. Plum. Ultrastructural char-
acteristics of the brain and blood-brain barrier in experimental
seizures. Brain Res. 127: 251-267, 1977.
59. Purpura, D. P. Morphophysiological basis of elementary evoked
response patterns in the neocortex of the newxorn cat. Ann.
N.Y. Acad. Sci. 92: 840-859, 1961.
60. Purpura, D. P. Ontogenetic models in studies of cortical seizure
activities. In: Experimental Models of Epilepsy, edited by D. P.
Purpura, J. K. Penry, D. M. Woodbury, D. B. Tower and R. D.
Walter. New York: Raven Press, 1972, pp. 531-556.
61. Pylkko, O. O. and D. M. Woodbury. The effect of maturation
on chemically-induced seizures in rats. J. Pharmac. exp. Ther.
131: 185-190, 1961.
62. Rapoport, M. and M. Rubin. Lead poisoning: A clinical and
experimental study of factors influencing seasonal incidence in
children. Am. J. Dis. Child. 61: 245-257, 1941.
63. Reiter, L., G. Anderson, J. Laskey and D. Cahill. Devel-
opmental and behavioral changes in the rat during chronic expo-
sure to lead. Envir. Hllh. Perspect. 12: 119-123, 1975.
64. Rodier, P. Behavioral teratology. In: Handbook of Teratology,
Vol. 4, edited by J. Wilson and F. Fraser. New York: Plenum
Publishing Corp., 1978, pp. 397-428.
65. Rose, G. H. and D. B. Lindsley. Visually evoked electrocortical
responses in kittens: Development of specific and nonspecific
systems. Science 148: 1244-1246, 1965.
66. Salas, S. and S. Schapiro. Hormonal influences upon the ma-
turation of the rat brain's responsiveness to sensory stimuli.
Physiol. Behav. 5: 7-11, 1970.
67. Scheibel, M. E. and A. B. Scheibel. Selected structural-
functional correlations in postnatal brain. In: Brain Develop-
ment and Behavior, edited by M. D. Sterman, D. J. McGinty
and A. M. Adinolfi. New York: Academic Press, 1971, pp. 1-21.
68. Scoppa, P., M. Roumengous and W. Penning. Hepatic drug
metabolizing activity in lead-poisoned rats. Experientia 29:
970-972, 1973.
69. Seppalainen, A. M., S. Tola, S. Hernberg and B. Kock. Subclin-
ical neuropathy at "safe" levels of lead exposure. Archs Envir.
Hlth. 30: 180-183, 1975.
70. Silbergeld, E. and A. Goldberg. Hyperactivity: A lead-induced
behavior disorder. Envir. Hlth Perspect. 7: 227-232, 1974.
71. Silbergeld, E. K., L. P. Miller, S. Kennedy and N. Eng. Lead
and seizures: Role of-y-aminobutyric acid. Soc. Neurosci. Abst.
3: 323, 1977.
72. Smart, J. and J. Dobbing. Vulnerability of developing brain. II.
Effects of early nutritional deprivation on reflex ontogeny and
development of behaviour in the rat. Brain Res. 28: 85-95, 1971.
73. Smith, H. D., R. L. Baehner, T. Carney and W. J. Majors. The
sequelae of pica with and without lead poisoning. Am. J. Dis.
Child. 105: 609-616, 1963.
74. Sobotka, T., R. Brodie and M. Cook. Psychophysiological ef-
fects of early lead exposure. Toxicology 5: 175-191, 1975.
75. Spyker, J. Behavioral teratology and toxicology. In: Behavioral
Toxicology, edited by B. Weiss and V. G. Laties. New York:
Plenum Press, 1975, pp. 311-349.
76. Spyker, J. and D. Avery. Neurobehavioral effects of prenatal
exposure to the organophosphate diazinon in mice. J. Toxic.
Envir. Hlth. 3: 989-1002, 1977.
77. Sterman, M. B., D. J. McGinty and A. N. Adinoki. Brain De-
velopment and Behavior. New York: Academic Press, 1971.
78. Swinyard, E. A. Electrically induced convulsions. In: Experi-
mental Model of Epilepsy, edited by D. P. Purpura, J. K. Penry,
D. M. Woodbury, D. B. Tower and R. D. Walter. New York:
Raven Press, 1972, pp. 433-458.
79. Taylor, P. M. Oxygen consumption in new-born rats. J. Physiol.
154: 153-168, 1960.
80. Thurston, D. L., J. N. Middelkamp and E. Mason. The late
effects of lead poisoning. J. Pediatr. 47: 413-423, 1955.
81. Toman, J. E. P., E. A. Swinyard and L. S. Goodman. Proper-
ties of maximal seizures, and their alteration by anticonvulsant
drugs and other agents. J. Neurophysiol. 9: 231-239, 1946.
82. Vernadakis, A. and D. Woodbury. Electrolyte and amino acid
changes in rat brain during maturation. Am. J. Physiol. 203:
748-752, 1962.
83. Vernadakis, A. and P. S. Timiras. Effects of whole-body
x-irradiation on electroshock seizure responses in developing
rats. Am. J. Physiol. 205: 177-180, 1963.
84. Vernadakis, A., J. J. Curry, G. J. Maletta, G. Irvine and P. S.
Timiras. Convulsive responses in prenatally irradiated rats.
Expl Neural. 16: 57-64, 1966.
85. Vernadakis, A. and D. M. Woodbury. The developing animal as
a model. Epilepsia 10: 163-178, 1969.
86. Wasterlain, C. Effects of neonatal seizures on ontogeny of re-
flexes and behavior. Eur. Neural. 15: 9-19, 1977.
87. Werboff, J., I. Goodman, J. Havlena and M. Sikov. Effects of
prenatal x-irradiation on motor performance in the rat. Am. J.
Physiol. 201: 703-706, 1961.
88. Wiener, G. Varying psychological sequelae of lead ingestion in
children. U.S. Publ. Hlth Rep. 85: 19-24, 1970.
89. Wilson, E., J. Bogacz and E. Garcia-Austt. Development of
cortical visual evoked responses in the albino rat. Acta neural.
latinoam. 12: 91-105, 1966.
90. Woodbury, L. A. and V. D. Davenport. Design and use of a new
electroshock seizure apparatus, and analysis of factors altering
seizure threshold and pattern. Archs int. Pharmacodyn. 62:
97-107, 1952.
91. Woolley, D. E. Effects of DDT and of drug-DDT interactions on
electroshock seizures in the rat. Toxic, appl. Pharmac. 16:
521-532, 1970.
92. Woolley, D. E. Effects of DDT on the nervous system of the rat.
In: The Biological Impact of Pesticides in the Environment,
edited by J. Gillett. Corvallis, Oregon: Oregon State University
Press, 1970, pp. 114-124.
93. Yanase, M. Effects of environmental temperature on chronic
lead poisoning. J. Nagoya City Univ. Med. Ass. 15: 55-81, 1964.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 207-212. ANKHO International Inc., 1979.
A Preliminary Test Battery for the Investigation
of the Behavioral Teratology of Selected
Psychotropic Drugs1
RICHARD E. BUTCHER AND CHARLES V. VORHEES
Behavioral Sciences Unit, Division of Inborn Errors of Metabolism
Children's Hospital Research Foundation, Cincinnati, OH 45229, USA
BUTCHER, R. E. AND C. V. VORHEES. A preliminary test battery for the investigation of the behavioral teratology of
selected psychotropic drugs. NEUROBEHAV. TOXICOL. 1: Suppl. 1, 207-212, 1979.—Pregnant Sprague-Dawley rats
received 25 mg/kg of prochlorperazine, 20 mg/kg of fenfluramine, 75 mg/kg of propoxyphene or 200 mg/kg of diazepam daily
between the 7th and 20th days of gestation. Vehicle control groups and a positive control group (vitamin A 40,000
lU/kg/day) were similarly prepared. Observations of reproductive performance were made and the offspring examined in a
battery of neurobehavioral tests. Fenfluramine and prochlorperazine produced abnormalities in both the reproductive
measures and neurobehavioral testing. Propoxyphene produced developmental delays and other signs of "pure" behav-
ioral teratogenesis in that these effects were not anticipated in any of the observations of reproductive performance.
Diazepam appeared to have the mildest effect on all the measurements taken. The test methods used in this study appear to
be a reasonable initial approach to the development of neurobehavioral screening procedures which are comprehensive,
sensitive, and usable.
Behavioral teratology Prochlorperazine Fenfluramine Diazepam Propoxyphene
INCREASINGLY, a mandate for the inclusion of
psychotoxicity studies are included in guidelines for the es-
tablishment of the safety of drugs and other chemicals, and
instructions for such investigations are part of the drug re-
productive guidelines in Britain, France and Japan. Although
no comparable requirements exist in the United States the
FDA is currently engaged in revising the teratology and re-
productive guidelines for new drugs and food additives [13]
and this meeting is evidence of the EPA interest in this area.
Our laboratory has had a particular interest in devel-
opmental psychotoxicology and for about the past 8 years
has examined the post-natal functional consequences of pre-
natal chemical exposure (behavioral teratology) [4, 6, 7, 8, 9,
10, 12, 16, 23]. Under contract to the FDA, we have also
undertaken large scale studies of the behavioral effects from
developmental exposure to food additives and psychotropic
drugs. As part of this latter effort four psychotropic drugs
were examined using a test battery that represents a first
attempt to devise a test series that would be usable, sensi-
tive, and comprehensive.
Usability may be defined as the battery's applicability to
large scale testing within reasonable cost. Sensitivity must
include the conventional indices of reliability and validity
and may be operationally defined as the ability of the battery
to disclose differences between negative controls and treat-
ments known to produce neurological and behavioral im-
pairments. Comprehensiveness is perhaps the most difficult
inasmuch as the range of possible behavioral tests is almost
unlimited. At present a limited variety of behavioral func-
tions must be covered under the guidelines of other countries
which dictate a minimal level of comprehensiveness. Test
protocols of the magnitude suggested by these criteria have
only rarely been undertaken heretofore and have not been
designed with current guidelines in mind [1,15].
Our experience in investigating the behavioral teratologic
potential of selected psychotropic drugs affords an opportu-
nity to evaluate the tentative battery in terms of these criteria
and to compare the behavioral results with more usual meas-
ures of fertility.
METHOD
Adult Sprague-Dawley rats (Laboratory Supply, In-
dianapolis, IN) were used for breeding. Females weighed
about 260 g at conception, males about 400 g. Date of con-
ception (expelled vaginal plug) was considered Day 0 of ges-
tation (GO) and all females were primiparous. Daily on Days
G7-20 females were gavaged with one of the following: 25
mg/kg of prochlorperazine edisylate (Pz) (courtesy of Smith,
Kline & French), 20 mg/kg of fenfluramine HC1 (Ffl) (cour-
tesy of A. H. Robins), 75 mg/kg of propoxyphene HC1 (Pp)
(courtesy of Eli Lilly & Co.), 40,000 lU/kg vitamin A palmi-
'Support by FDA project 223-76-3026 and by NIH grant HD-05221. The authors thank Linda Jonas for her assistance with data analysis.
207
-------
208
BUTCHER AND VORHEES
tate (12 mg/kg) (USV Pharmaceuticals) or saline to the Con-
trols. All drugs were given in solution in saline in a volume of
5 ml/kg except vitamin A (the positive control group), which
was solubilized with 12% sorethytan oleate and given in a
volume of 1 ml/kg. Doses were based on preliminary dose
ranging experiments demonstrating that these doses were
not embryotoxic.
A fourth test drug, diazepam (courtesy of Hoffman-
LaRoche), was not water soluble and was, therefore, sus-
pended at a concentration of 10 mg/ml in a solution of 5%
acacia. Preliminary studies indicated that diazepam could be
administered at 200 mg/kg if the offspring were fostered at
birth. This quantity of diazepam was used throughout the
entire treatment period and the offspring were, on Days 1-4
of postnatal life, fostered to dams which had received
equivalent volumes of acacia during gestation. The diazepam
dams were used to rear the acacia offspring in the expecta-
tion, based on the pilot investigation, that drug withdrawal
would produce disruptions of nurturing behavior. This ex-
pectation was not confirmed in subsequent measurements of
survival and growth. The cross fostering procedure and the
necessity of a separate vehicle (acacia) control group, how-
ever, prevent direct comparison to the data from experi-
mental groups using the saline vehicle. The data are, there-
fore, presented separately in the Results section.
All dams were weighed daily during treatment and all
were allowed to litter normally and nurture their offspring.
Parturition was considered Postnatal Day 0 (PNO). On Day
PN1 all litters were examined externally for malformations,
sexed, weighed and any dead fetuses removed. Litters with
more than 12 were reduced to 12, balancing for sex. Behav-
ioral testing was conducted on not more than 8 offspring per
litter (4 males and 4 females) selected at random on Day PN1
and marked for testing. Testing began on Day PN3 and ex-
tended to Day PN70. Offspring were weaned on Day PN21.
Due to the amount of time required to conduct some adult
tests, only males were examined in spontaneous alternation,
Biel maze, active avoidance and passive avoidance tests.
The following number of dams delivered litters in each
group: Pz 13 litters, Ffl 14 litters, Pp 10 litters, vitamin A 9
litters, and saline 15 litters. In the diazepam experiment the
numbers were diazepam 9 and acacia 8. Five Pz litters were
eliminated because they did not meet the requirement of 6
live progeny: one additional litter was eliminated two weeks
after birth due to the dams developing an inner ear infection:
no Ffl, Pp or vitamin A litters were discarded and only one
saline litter was discarded because of too few offspring. In
the diazepam experiment one acacia litter was lost due to
small litter size, though the dam was still used in the cross
fostering procedure and no diazepam litters were discarded.
The 2 litter disparity was, for purposes of cross fostering,
handled as follows. Although there were only 7 usable acacia
litters there were 8 acacia dams, which accommodated 8 of
the 9 diazepam litters, the final diazepam litter was fostered
to an extra untreated lactating dam that had also just deliv-
ered. The 7 acacia litters were cross fostered to their
matched diazepam dams and the 2 remaining diazepam dams
were eliminated (diazepam and acacia dams were matched
for weight and day of conception prior to group assignment).
Behavioral Tests
Surface righting. Observed daily from Day 3. Rats were
tested until all test pups righted in =£2 sec on 2 out of 2 trials
on a given day [5,24].
Cliff avoidance. Observed daily from Day 6 until each test
pup, when placed on an edge with forepaws and nose just
over the edge, showed retraction and/or sideward movement
away from the edge [5,24].
Swimming development. Observed on Days PN6, 8 and
10 with experienced raters [5, 20, 24]. Test pups were placed
in a tank of water (26.7°C) for a period of 5-15 sec and
observed for three aspects of swimming: direction, angle in
the water (or head position) and use of limbs.
Negative geotaxis. Observed daily on Days PN6-12. Test
pups were timed for completing a 180° turn when placed in a
head down position on a 25° inclined plywood surface [2].
Preweaning open field. On Days PN15-17 half of the
males and half of the females were observed for 3 min/day in
a circular open field (dia. 45.7 cm) for number of section
entries (the floor was marked into 20 sectors), number of
rearings and latency to exit the central start circle [5].
Postweaning open field. Observations were made on 3
consecutive days (PN41-43) for 3 min/day and scored for
number of sections entered, pattern of sections entered, la-
tency to begin exploration, number of rearing instances and
number of fecal pellets deposited. The open field was circu-
lar and 91.4 cm in dia. and twice the scale of the preweaning
open field. Animals were those not tested in the preweaning
open field [5,24].
Spontaneous alternation. On Day PN45 all males were
given 4 alternation trials in a non-reinforced T-maze (stem
45.7 cm, arms 50.8 cm each) and scored for the number of
alternations in each pair of trials.
Biel water maze. This task has been described in detail
elsewhere [3,11]. Basically the test consists of 2 phases, de-
termination of straight channel swimming speed (5 trials),
followed by maze solution (6 trials), recording time and er-
rors. All males were tested on days PN50-53.
Active avoidance. On days PN65-70 half the males were
given trials in a wheel turn active avoidance apparatus de-
scribed in detail elswhere [17]. On each trial a warning white
noise came on for 9 sec during which a one-half turn of the
wheel terminated the trial and an avoidance was scored: if no
turn was made by the end of the 9 sec warning interval,
scrambled footshock was delivered to the grid floor and
metal walls (0.75 ma) until a wheel turn response was made.
Passive avoidance. Half the male subjects were tested for
3 consecutive days, one day of training and 2 days of reten-
tion testing, beginning between days PN55-57, in a 2 cham-
bered passive avoidance apparatus. On each day the animal
was placed in the lighted side, the guillotine door raised and
the rat allowed to enter the dark side. Entry into the dark
side was timed and on the first day resulted in the divider
door closing and the onset of a 1 sec, 1.0 mA shock being
delivered through a scrambler to the grid floor. On the next
two days (retention) no shock was administered when the
animal entered the darkened chamber [5,24].
Rotorod. The rod was 11.4 cm in dia. with its surface
roughened by a mixture of paint and sand. All rats were
given 2 trials/day on two consecutive days between days
PN60-65. On each trial the rat was placed on the wheel and
the wheel was gradually accelerated until it reached 30 RPM
at which point the trial was timed until the rat fell or up to a
limit of 3 min [5,24]. Rats were scored for the amount of time
on the rod as a percentage of controls and for the percentage
of animals in each group reaching the 3 min criterion.
-------
TEST BATTERY FOR BEHAVIORAL TERATOLOGY
209
Statistical Analyses
Body weight and most behavioral data were analyzed
using unweighted means analyses of variance procedures.
Individual group comparisons were made using Newman-
Kuels tests [18,19]. Data in proportions (mortality) was
analyzed using Fisher's test for uncorrelated proportions
[14].
RESULTS AND DISCUSSION
A summary of the effects resulting from prenatal exposure
to the water soluble drugs is presented in Table 1. The details
of these results will be presented elsewhere. An inspection
of these data suggests that with the exception of Pz all the
test compounds (including the vitamin A positive control
group) had some effect on measures of fertility and offspring
weight. Particularly strong effects were observed in the Pz
group in which fewer liters with 6 or more pups were pro-
duced, the length of gestation and offspring mortality was
increased, and the weight of the surviving offspring was re-
duced at the PN7 weighing.
TABLE 1
SUMMARY OF EFFECTS FROM PRENATAL EXPOSURE TO
PSYCHOTROPIC DRUGS*
TABLE 2
SUMMARY OF EFFECTS FROM PRENATAL EXPOSURE TO
PSYCHOTROPIC DRUGS POSTWEANING TESTS*
Measure
No. of litters delivered
No. of litters with 6+ progeny
Length of gestation
Offspring mortality
Offspring weight
PN7
PN14
PN21
Surface righting
Cliff avoidance
Swimming
Direction
Angle
Negative geotaxis
Pivoting locomotion
Preweaning open field
Ambulation
Rearing
Start latency
Pz
0
+
+
-
0
0
0
0
0
0
+
0
0
0
0
Ffl
0
0
0
+
-
0
0
0
0
-
-
-
(- +)
0
0
+
Pos. Cont.
Pp Vit. A
0 0
0 0
0 0
0 0
0
0 —
0
Q
0
0
—
-t-
+ -
+ 0
+ +
0
*Symbol code: minus sign = a significant decrease in the
dependent measure; plus sign = a significant increase in the
dependent measure; zero = no significant differences. One sign =
p<0.05, 2 signs = p<0.01, except for pivoting where the (- +) set
of symbols means a significant decrease followed by a significant
increase in pivoting time across days of testing.
Further inspection of Tables 1 and 2 with emphasis placed
on the behavioral data permits some interesting comparisons
with the physical measures. The results of the behavioral
testing of the Ffl animals are generally consistent with the
increased offspring mortality and reduced offspring weights
observed in this group. In this case behavioral examination
Measure
Postweaning open field
Ambulation
Rearing
Start latency
Defecation
Spontaneous alternation
Straight channel
swimming speed
Biel water maze errors
Wheel turn avoidance
Passive avoidance retention
Rotorod performance
Body weight (PN70)
Pz
0
-
0
0
0
0
0
0
0
-
0
Ffl
+
+
0
+
0
0
0
0
0
0
0
Pp
+
+
0
0
0
0
0
+
0
-
0
Pos. Cont.
Vit. A
+
+
0
+
0
—
0
+
0
-
0
*Symbol use: same as for Table 1
provides confirmation of the functional relevance of the
growth retardation and a description of the nature of the
developmental delay. The results of testing the Pz offspring,
in which fewer differences from controls were observed,
suggest that the effects of this compound at this dosage are
less remarkable than the observed reproductive abnor-
malities. A strikingly different conclusion may be drawn
from the Pp test results. Here, in contrast to the absence of
effects on reproduction, a large increase in behavioral ab-
normalities was observed. The pattern of departures from
saline control values in the Pp animals can be characterized
by a slight developmental delay and increased activ-
ity/reactivity. At the dose level used in this study these ef-
fects were not presaged by the fertility or growth measures
and constitute an occurrence of "pure" behavioral
teratogenesis.
The females receiving diazepam during gestation were
uniformly observed to have a slight to moderate hypotonia
during the treatment period which was followed by slight
agitation for the 1-3 day period immediately following termi-
nation of treatment. The data from the diazepam treated sub-
jects and acacia controls are displayed in Tables 3-5. A com-
parison by /-test of the results from the acacia and saline
vehicle controls revealed a significant (p<0.05) difference
only in preweaning open field rearing frequency. The di-
azepam animals were found to weigh significantly less than
acacia females at the end of gestation and a marginally sig-
nificant (p<0.06) increase in offspring mortality was ob-
served.
Behaviorally, prenatal diazepam exposure produced a
significant delay in swimming development (angle). A sig-
nificant effect was found in passive avoidance in which the
experimental offspring actually withheld the formally
punished response longer than acacia treated controls. Simi-
larly, an increased percentage of diazepam offspring were
able to attain criterion performance on the rotorod test. Both
these latter effects, in the absence of abnormalities in other
measures of response inhibition and locomotor coordination,
are difficult to interpret until additional data are available.
-------
210
BUTCHER AND VORHEES
TABLE 3
DIAZEPAM SUMMARY: REPRODUCTIVE PERFORMANCE AND TESTS OF BEHAVIORAL DEVELOPMENT*
Treatment
Measure
Acacia
Diazepam
Sig.
Length of gestation
Litter size
Preweaning mortality
Gestation weight (G19)
Lactation weight (L21)
Offspring weight (PN21) M
F
Cliff avoidance
Negative geotaxis (PN6)
Surface righting reflex
Startle reflex
Swimming development (PN6, 8, 10)
Direction (PN8)
Angle (PN8)
Angle (PN10)
Limb usage (PN8)
Pivoting (PN7, 9, 11)
Time
No. of 90° turns
22.4 ±
10.4 ±
0.2
0.7
(8)
(8)
5% (4/74)
361.8 +
302.0 ±
36.7 ±
35.4 ±
9.3 ±
38.3 ±
9.3 ±
13.4 ±
2.0 ±
1.5 ±
2.0 ±
1.0 ±
17.2 ±
5.4 H-
7.9
10.0
4.0
3.9
0.4
5.7
0.6
0.3
0.1
0.3
0.2
0.1
2.7
0.7
(8)
(7)
(7)
(7)
(7)
(7)
(6)t
(7)
(5)f
(5)t
(7)
(5)f
(7)
(7)
22.4 ±
10.7 ±
13%
330.1 ±
326.3 ±
37.4 ±
36.3 ±
9.7 ±
38.9 ±
9.8 ±
12.8 ±
1.9 ±
0.9 ±
1.4 ±
0.9 ±
15.7 ±
5.4 ±
0.2
1.0
(9)
(9)
(12/96)
10.1
8.1
2.8
2.3
0.5
3.7
0.5
0.4
0.1
0.2
0.1
0.04
2.2
0.7
(9)
(9)
(9)
(9)
(9)
(9)
(9)
(8)f
(8)f
(8)t
(9)
(8)t
(9)
(9)
NS
NS
p<0.06
p<0.05
NS
NS
NS
NS
NS
NS
NS
NS
NS
p<0.05
NS
NS
NS
*Values represent the mean ± SEM on a per litter basis with the number of litters tested shown in parentheses. Unless shown
separately, male and female data have been combined.
fN's are slightly reduced in these cases due to technician's failure to complete test or equipment malfunction.
TABLE 4
DIAZEPAM SUMMARY: TESTS OF ACTIVITY*
Measure
Treatment
Acacia
Diazepam
Sig.
Preweaning open field (PN15-17)
Ambulation
Rearing
Start latency
Postweaning open field (PN40-45)
Ambulation
Rearing
Start latency
Defecation
M
F
M
F
M
F
M
F
49.5
4.0
-f
+
7.6 ±
45.5
63.7
9.
.1
±
±
+
12.5 ±
1.2 ±
1.0
2.
.1
1.4
±
±
±
2,
0
1
4
6
1
1
0,
0
0
.6
.9
.6
.5
.5
.4
.9
.3
.3
.7
0.6
(26)
(26)
(26)
(12)
(10)
(12)
(10)
(12)
(10)
(12)
(10)
52.9 ± 3.7 (35)
3.9 ± 0.4 (35)
11.2 ± 2.6 (35)
46.0 ± 2.9
57.1 ± 3.1
9.6 ± 1.3
9.9 ± 1.1
1.8 ± 0.4
1.9 ± 0.6
1.9 ± 0.3
2.1 ± 0.4
(14)
(13)
(14)
(13)
(14)
(13)
(14)
(13)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
*Values represent the mean score across all 3 days of testing ± SEM with the number of animals tested shown in parentheses.
-------
TEST BATTERY FOR BEHAVIORAL TERATOLOGY
211
TABLE 5
DIAZEPAM SUMMARY: TESTS OF POSTWEANING BEHAVIORAL PERFORMANCE*
Treatment
Measure
Acacia
Diazepam
Sig.
Spontaneous alternation (PN45)
Biel maze-straight alley (PN50-53)
Biel maze-errors (PN50-53)
Wheel turn active avoidance (PN65-70)
24h passive avoidance (PN55— 60)
Rotorod (PN60-65) M
F
0,
113,
31
102
71%
.17 ± 0.01
.4
.4
.7
± 12.
± 1.
± 23.
22%
82%
,2
,1
6
(52)
(27)
(27)
(13)
(11)
(27)
(22)
0
115,
32,
156,
69'
.15 ±
.4 ±
.5 ±
.8 ±
%
0.01
10,
1.
12,
.7
.1
.8
53%
79%
(56)
(32)
(32)
(17)
(15)
(32)
(29)
NS
NS
NS
NS
p<0.05
p<0.05
NS
*Values for Biel maze and avoidance tasks represent means ± SEM. For spontaneous alternation values represent the percentage of
trials on which the animals alternated. For rotorod the values represent the percentage of animals reaching the 3 min performance
criterion. N's shown in parentheses.
Diazepam could be judged among the psychotropic com-
pounds tested to have the lowest potential as a behavioral
teratogen. Tervo et al. [21], however, have suggested that
prenatal exposure to diazepam at much lower doses than
those used here produces developmental abnormalities.
The test battery used in these studies reveals a wide spec-
trum of behavioral outcomes some of which contrast sharply
with physical measures of reproduction. We believe these
behavioral results, although limited to a single dose, add
substantially to the ability to accurately appraise risk in the
use of these compounds. Although the battery is not as com-
prehensive as is required (sensory tests are needed addi-
tions), the sensitivity of the battery appears completely
adequate to detect the effects of prenatal vitamin A expo-
sure, and in our laboratory, proved to be a usable test system
which was applied without undue difficulty.
REFERENCES
1. Almli, C. R. and R. S. Fisher. Infant rats: Sensorimotor on-
togeny and effects of substantia nigra distraction. Brain Res.
Bull. 2: 425-459, 1977.
2. Altaian, J. and K. Sudarshan. Postnatal development of
locomotion in the laboratory rat. Anim. Behav. 23: 896-920,
1974.
3. Biel, W. C. Early age differences in the maze performance of
the albino rat. J. genet. Psychol. 56: 439-453, 1940.
4. Brunner, R. L., M. S. McLean, C. V. Vorhees and R. E.
Butcher. A comparison of behavioral and anatomical measures
of hydroxyurea induced abnormalities. Teratology 18: 379-384,
1978.
5. Brunner, R. L., C. V. Vorhees, L. Kinney and R. E. Butcher.
Aspartame: Assessment of developmental psychotoxicity in a
new artificial sweetner. Neurobehav. Toxicol. 1: 79-86, 1979.
6. Butcher, R. E. Behavioral testing as a method for assessing risk.
Envir. Health Pers. 18: 75-78, 1976.
7. Butcher, R. E., R. L. Brunner, T. Roth and C. A. Kimmel. A
learning impairment associated with maternal hypervitamin-
osis-A in rats. Life Sci. 11: 141-145, 1972.
8. Butcher, R. E., K. Hawver, T. Burbacher and W. Scott. Behav-
ioral effects from antenatal exposure to teratogens. In: Aberrant
Development in Infancy, edited by N. Ellis. Potomac, Mary-
land: Erlbaum Assoc., 1975, pp. 161-167.
9. Butcher, R. E., K. Hawver, K. Kazmaier and W. Scott.
Postnatal behavioral effects form prenatal exposure to terato-
gens. In: Basic and Therapeutic Aspects of Perinatal Phar-
macology, edited by P. L. Morselli, S. Gerattini and F. Serini.
New York: Raven Press, 1975, pp. 171-176.
10. Butcher, R. E., W. J. Scott, K. Kazmaier and E. J. Ritter.
Post-natal effects in rats of prenatal exposure to hydroxyurea.
Teratology 7: 161-165, 1973.
11. Butcher, R., C. Vorhees and H. Berry. A learning impairment
associated with induced phenylketonuria. Life Sci. 9: 1261-
1268, 1970.
12. Butcher, R. E., C. V. Vorhees and C. A. Kimmel. A learning
impairment from maternal salicylate treatment in rats. Nature
NewBiol. 236:211-212, 1972.
13. Collins, T. F. X. Reproduction and teratology guidelines: Re-
view of deliberations by the national toxicology advisory com-
mittee's reproduction panel. J. envir. pathol. Toxicol. 2: 141-
147, 1978.
14. Guilford, J. P. Fundamental Statistics in Psychology and Edu-
cation. New York: McGraw-Hill, 1965.
15. Irwin, S. Comprehensive observational assessment: la. A sys-
tematic, quantitative procedure for assessing the behavioral and
physiologic state of the mouse. Psychopharmacologia 13: 222-
257, 1968.
16. Kimmel, C. A., R. E. Butcher, C. V. Vorhees and H. J.
Schumacher. Metal-salt potentiation of salicylate-induced tera-
togenesis and behavioral changes in rats. Teratology 10: 293-
300, 1974.
17. Kinney, L. and C. V. Vorhees. A comparison of methylpheni-
date induced active avoidance and water maze performance
facilitation. Pharmac. Biochem. Behav. 10: 437-439, 1979.
18. Kirk, R. E. Experimental Design: Procedures for the Behav-
ioral Sciences. Belmont, Calif.: Brooks/Cole Publishing, 1968.
19. Kramer, C. Y. Extension of multiple range tests to group means
with unequal numbers of replications. Biometrics 12: 307-310,
1956.
20. Schapiro, S., M. Salas and K. Vukovich. Hormonal effects on
ontogeny of swimming ability in the rat: Assessment of central
nervous system development. Science 168: 147-151, 1979.
-------
212 BUTCHER AND VORHEES
21. Tervo,D.,C. Kellogg, A. Miller and J.lson. Prenatal diazepam 23. Vorhees, C. V., R. L. Brunner, C. R. McDaniel and R. E.
exposure in rats: Effects on growth and behavioral development Butcher. The relationship of gestational age to vitamin A in-
of the offspring. Soc. Neurosci. 4: 128, 1978 (Abstract). duced post-natal dysfunction. Teratology 17: 271-276, 1978.
22. Vorhees, C. V. Some behavioral effects of maternal hyper- 24. Vorhees, C. V., R. E. Butcher, R. L. Brunner and T. J.
vitaminosis A in rats. Teratology 10: 269-273, 1974. Sobotka. A developmental test battery for neurobehavioral tox-
icity in rats: A preliminary analysis using MSG, calcium car-
rageenan and hydroxyurea. Toxic, appl. Pharmac. in press,
1979.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 213-215. ANKHO International Inc., 1979.
Assays for Behavioral Toxicity: A Strategy for
the Environmental Protection Agency
1,2
BERNARD WEISS AND VICTOR G. LATIES
Department of Radiation Biology and Biophysics
and
Environmental Health Sciences Center, School of Medicine and Dentistry
The University of Rochester, Rochester, NY 14642
WEISS, B. AND V. G. LATIES. Assays for behavioral toxicity: A strategy for the environmental protection agency.
NEUROBEHAV. TOXICOL. 1: Suppl. 1, 213-215, 1979.—Broad agreement on specific approaches or standardized test
batteries for assessing behavioral toxicity is unlikely to emerge in the foreseeable future. EPA should reject test stan-
dardization in any case, however; standardization stifles progress and, in addition, may bypass unique properties of new
types of substances. The optimal strategy is to prescribe a set of functions, such as sensory, motor, and complex perform-
ance processes, leaving it to the manufacturer to select adequate tasks. Adequacy would be judged by EPA staff, in
consultation with advisory panels, and resolved, in most cases, by a dialogue with the manufacturer.
Behavioral toxicity Standardization of tests Screening strategies, behavior
THIS collection of papers provides the most emphatic
statement so far of how essential it is for the Environmental
Protection Agency to shun test standardization. No obvious
approach or battery of behavioral tests emerges as univer-
sally suitable for toxicity screening. Too many choices are
available for any subset to elicit wide agreement, a reflection
of behavioral toxicology's youth and vigor. Furthermore,
although it is being called upon to provide simple screening
methods, we really need a strategy aimed at the total
assessment of risks, one that starts in the laboratory but
concludes with monitoring of human populations once a
chemical is marketed. The broad choice of assessment
methods offered by behavioral toxicology, however, is not a
situation to bemoan. It reflects not chaos, but flexibility.
Such flexibility commands a price, however. It cannot be
purchased without a perceptive review of the strengths and
suitability of the tools selected for any particular evaluation.
How these are to be weighed was the salient theme of this
conference.
A behavioral analog of the Ames test, the bacterial
mutagenic assay, is an impossible dream. Mutagenesis is
triggered by a limited, if not a single, collection of mech-
anisms. Behavior, in contrast, is the most diverse of all
biological functions, and is subserved by an extraordinary
range of mechanisms. The best we can expect from any test-
ing scheme is a restricted sampling of mechanisms, perhaps
amplified selectively by a sequential narrowing of specific
questions emerging during assessment.
Several current proposals for screening envisage test bat-
teries is made up of relatively simple elements. The absence
of complex measures and detailed analysis is assumed to be
compensated for by large numbers of animals and the broad
scope of the tests. Such batteries can satisfy only afirst stage
assessment. First stage must be emphasized. Behavioral sci-
entists have turned to complex procedures and instrumenta-
tion for reasons surpassing a fascination with slick gadgetry
and theoretical minutiae. They have done so because most
easy questions already have been answered. Further, behav-
ioral toxicology cannot emphasize only simple questions. Its
responsibility is to demonstrate that particular behavioral
deficits do not occur under particular circumstances; to do
so it must meet standards of experimental rigor and test re-
producibility that do not come cheap.
An intermediate strategy would limit the number of
different situations to be evaluated and analyze one, or a
few, in great detail. For example, one can build on the mas-
sive empirical foundation of operant technology and study
different behaviors in the same situation. An animal trained
to make a particular response in the presence of one set of
environmental stimuli, for example, might be trained in the
same situation to emit another kind of response to another
set of stimuli. Such an approach might achieve operational
'Based on the proceedings of a workshop on Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor
Function organized by the Southwest Foundation for Research and Education and sponsored by the United States Environmental Protection
Agency.
2The preparation of this paper was supported in part by Grant ES-01247 from NIEHS and in part by a contract with the U. S. Department of
Energy at the University of Rochester Department of Radiation Biology and Biophysics and has been assigned Report No. 3490-1615.
213
-------
214
WEISS AND LATIES
simplicity without sacrificing functional and conceptual
complexity. It has worked successfully in behavioral phar-
macology, and its reliance on automation reduces many of
the problems of reliability afflicting the more conventional
batteries.
Psychophysical approaches designed to achieve a finely-
tuned assessment of sensory capacity are characterized by
both complexity and specificity, but may require extended
training and expensive instrumentation. Although they might
not be feasible for early screening, such techniques play a
significant role in behavioral toxicology. First, they repre-
sent the culmination of a sequential screening strategy whose
final steps would consist of specifying the precise impact of
an agent on various sensory systems. Second, they can tell
us which functions and at which parameter values the toxic-
ity of an agent is first expressed, providing cues for human
monitoring and guides for ancillary criteria such as morphol-
ogy and biochemistry. Third, they provide the basis for
validating simpler and less comprehensive assessment tech-
niques.
There is no inherent conflict among the various screening
strategies. They are all essential. They simply play different
roles, all of which are required to fulfill EPA's responsibility.
They usually are called upon at different stages of evalua-
tion.
No technique, no combination of techniques, no single
subdiscipline will achieve widespread adoption, however,
unless it is accompanied by demonstrations of sensitivity and
selectivity. Any approach must show responsiveness to rel-
atively small amounts of a chemical and respond differently
to different chemicals. One way to develop such char-
acteristics is by designing methods directed at specific
classes of agents. If a test battery is aimed at characterizing
the sequelae of a well-studied agent such a acrylamide,
measures of impairment can be refined to provide a well-
focused image of acrylamide toxicity. Such reference sub-
stances help us to hone and maintain our tools. Although we
have almost no alternatives to this approach, we still remain
somewhat uncertain, however, of our tools' predictive
power. It is as though we had devised a new test of intelli-
gence, then determined its validity by correlating it with the
Stanford-Binet instead of an independent criterion.
One source of independent criteria is coordinate data
from morphology, biochemistry, physiology, and phar-
macokinetics. If, for example, a portion of the cerebral cor-
tex known to subserve vision is damaged by an agent such as
methylmercury, and vision testing indicates correlated defi-
cits in function, we feel confident in our choice of visual
tests. But we rarely enjoy such a luxury. We already know
that the central nervous system possesses such a huge func-
tional reserve that major damage may be inflicted before any
overt impairment appears; "silent damage" is the term used
in the methylmercury literature. Such discrepancies are
exaggerated by behavioral and neural compensatory mech-
anisms. Conversely, what independent verification can we
extract about agents that leave no easily verifiable traces?
Not all behaviorally active agents produce nervous system
pathology or enduring neurochemical changes. The mech-
anisms of mercury vapor toxicity are unknown, with no
morphological or chemical guides. Yet, its advanced stages
are marked by neurological impairment such as tremor, and
its earliest stages by characteristic psychological complaints.
Furthermore, the central nervous system may not be the
target organ at all in some intoxications. Early, non-specific
symptoms may arise from damage to other organ systems.
Kidney and liver disease may lead to nervous system dys-
function by indirect routes. The inhibition of peristalsis in
the pigeon crop by lead demonstrates how totally unex-
pected mechanisms may underlie a change in behavior. Last,
an agent may well exert a nearly simultaneous affect on mor-
phology and behavior that are not at all related. No one can
assume that biological questions always stimulate
straightforward answers.
The predictive power of a test, or system of tests,
emerges also as a statistical issue. It is most vividly illus-
trated by the problems of behavioral teratology. Even potent
teratogens may damage selectively only small proportions of
offspring, which is why postnatal test batteries typically in-
clude many different indices. Equivalent problems, perhaps
even magnified because dose-related manipulations are dif-
ficult to perform in that context, prevail in human studies. So
far, most attempts to validate test systems have confined
themselves to individual tests and group statistics, noting,
whether test "X", say, differentiates between groups of
treated and untreated animals. When many different tests
and measures are available, however, multivariate statistical
techniques enable investigators to express an entire array of
findings in a compact format of combined indices that may
stimulate the design of new techniques. Nor should behav-
ioral toxicology rely solely on parametric statistics. If only a
few animals in a large group respond adversely to a toxic
agent, such measures may inadequately reflect the impact of
the agent. The distribution of responses in a group represents
important data and should not be buried in group calcula-
tions. Adequate statistical techniques are available for such
analyses as well.
This discussion has emphasized problems and inconsis-
tencies. Can the Environmental Protection Agency, faced
with such quandaries, carry out its responsibilities under the
Toxic Substances Control Act and other legislation? This
conference confirms that it can, and also indicates the strat-
egies to adopt in doing so.
EPA should begin by rejecting any pressure to proclaim
standardized tests. Freezing a test battery into regulatory
practice is helpful only to underemployed lawyers, because
it fosters debate about the minor details of regulatory lan-
guage, rather than about scientific content. But without ex-
plicit descriptions of particular procedures, how is it possible
to achieve enough uniformity for regulatory consistency?
One way is to follow the examples of Japan and Britain, both
of which now require behavioral toxicology for new drugs.
Neither nation specifices guidelines or precise tests. Instead,
the adequacy of behavioral testing is determined in a di-
alogue between manufacturer and regulator. EPA's Office of
Toxic Substances, however, because of U.S. regulatory
traditions and practices, is forced to specify more concretely
what it considers to be adequate data. It should do so by
adopting a functional approach, that is, by specifying those
aspects of behavior it considers critical in evaluating toxic-
ity.
Although they may express divergent opinions on many
issues, most practitioners of behavioral toxicology, like
those at this conference, agree on certain principles. They
agree that both sensory and motor function should be eval-
uated. They agree that more complex behaviors, such as
discriminative and learning processes need to be included in
most assessments. They agree that postnatal evaluation can-
not stop with early reflex measures, but must include per-
formance after maturation. They agree that reproductive be-
havior processes are crucial indices of toxicity. EPA
-------
ASSAY STRATEGIES
215
strategy, then, would be to list a set of functions accom-
panied by possible assessment methods for each function. It
would be made clear that such methods do not exclude addi-
tional, perhaps newer and more sensitive methods, espe-
cially since academic, industrial, and government scientists
now are engaged in such active pursuit of better methods
that it would be surprising if many did not soon emerge.
We envisage a scenario, then, in which a manufacturer
brings to EPA a portfolio of results on behavioral tests with a
new agent. The choice of tests reflects the sponsor's opinion
about the most likely effects of concern. EPA staff then re-
views the portfolio, exercising professional judgment about
the adequacy of the techniques and the data. They take into
account the background of the manufacturer's scientists, the
scientific history of the chosen methods, data treatment, re-
producibility, comparative data from reference substances,
and perhaps other criteria upon which, say, a referee for a
journal might base a recommendation for acceptance, re-
vision, or rejection of a manuscript. EPA staff would also
exploit the advice of review panels and individual scientists
with special qualifications. Should EPA's conclusions con-
flict with those of the manufacturer, we foresee that most
such cases will be resolved in a dialogue rather than in court,
provided that the science itself is not blatantly inadequate.
Both the public, represented by EPA, and the manufac-
turer would gain from such an arrangement. Rather than
being locked into obsolete, often inappropriate standards,
testing methods could evolve in parallel with the scientific
development of the discipline, and the accretion of other
toxicologic knowledge. Rather than squandering resources
and talent on standardized tests unsuitable for a particular
question, behavioral toxicology could then liberate its extra-
ordinary potential to ask more adequate, exacting, and spe-
cific questions to everyone's benefit.
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Functio
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, p. 217. ANKHO International Inc., 1979.
Final Comments
When I helped spawn this Workshop, I realized that the
reward was problematical. When Peter Spencer stated that
ours is the first interdisciplinary meeting of a neurotoxicity
group, a reward is in sight if this Workshop encourages pur-
suit of an activity analogous to psychopharmacology—
psychotoxicology.
Obviously, it would be presumptuous of me to attempt
literally to summarize this symposium. I can recall the ques-
tions and issues raised by Norbert Page in his introductory
remarks and note how close we came to answering and
clarifying them. The convening of this Workshop is a result
of a Toxic Substances Control Act requirement to develop
test data for behavioral disorders. Since neither behavior nor
behavioral is defined in the law, its administrators will do so.
It was my intention that an authoritative group expert in
experimental investigation of behavior have a cohesive set of
recommendations before the regulatory groups adopt less
scientific guides in defining behavior. I encountered an
analogous situation shortly after I joined the Environmental
Protection Agency. I quickly discovered that the regulators
had redefined hydrocarbons to include compounds that no
organic chemists would consider as hydrocarbons. The re-
sult is that the lesion described by pathologists as prolifera-
tive hydrocarbon glomerulonephritis is frequently caused by
ketones, alcohols and other nonhydrocarbon compounds. I
assumed that something similar would happen to' 'behavioral''
when regulators defined it without adequate input from some
group such as this one.
A second incentive was my experience before a Congres-
sional oversight committee on the Michigan incident with
polybrominated biphenyls and at a National Institutes of
Health meeting on the behavioral effects of these com-
pounds. On both occasions the neurological assessment of
behavioral effects was based considerably on anecdotes. The
signs and symptoms could not be clearly differentiated from
possible effects on endocrine organs.
Page's first question was directed to strengthening routine
general toxicity tests to provide more sensitive indicators for
carrying out tests for behavioral and neural toxicity. Such
indicators are derived from simple observation of the dosed
animals and scoring the effects. The outcome determines
whether behavioral or neurological tests are necessary.
Papers presented at this meeting allow for such an approach.
However, it is not foolproof and will miss neurological or
behavioral effects in some instances. For instance, prom-
azine was screened in the antimalarial program. The neuro-
logical effects, which are obvious, were missed by the ob-
server. The useful effects of the promazines were discovered
13 years later.
Should we require much neuropathology, neurochemistry
and neurophysiology measurement during routine toxicity
testing? This question will be resolved only after much de-
bate. Two extreme views are current. The pathology
oriented investigator of behavior is not convinced of a true
effect unless he can demonstrate a true lesion in the integrity
of neuronal tissue. The other extreme of this concept is rep-
resented by those behaviorists who contend that elec-
trophysiology and lesion placement cannot contribute sig-
nificantly to an understanding of behavior. The resolution of
this question for our purposes will probably be based on an
intermediate position.
What existing tests for sensory, motor or cognitive effects
are well enough developed and validated so that they can be
proposed as standards? Several elegant methods for outwit-
ting animals or for preventing the animals from outwitting
the operator were presented at this Workshop. They were
presented on the assumption that these methods could be
used as tests for predicting behavioral toxicity in humans.
For which of these methods is the assumption valid? Much
of the discussion of these methods was centered on test de-
tails as though this assumption is justified.
What areas require priority for further research on test
methods? It became obvious during the discussions that
many such areas exist. One such area is further investigation
of correlates of behavior. An example of incomplete under-
standing correlates to behavior was furnished by return of
behavior to "normal" in the presence of persistent
neuropathology. Failure of correlates of a phenomenon to
predict an event is not limited to behavioral science. I am not
discouraged by such failures; they have been resolved in
other areas.
How well have we met the point raised by Dr. Page? I feel
that some have been met, others not. What is needed are
tests that will reliably detect neurological effects and behav-
ioral effects of chemicals.
JOSEPH SEIFTER
Office of Toxic Substances
U. S. Environmental Protection Agency
217
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, p. 219. ANKHO International Inc., 1979.
AUTHOR INDEX
ANDO, K 45
BARBEAU, A 175
BAUTISTA, S 113
BIGNAMI, G 179
BRADFORD, L. D 73
BRADY, J. V 73
BUTCHER, R. E 207
CABE, P. A 137
COLOTLA, V. A 113
CORY-SLECHTA, D. A 129
COX, C 149
DEWS, P. B 119
ELSNER, J 163
FLYNN, E. R 93
FLYNN, J. C 93
FOX, D. A 193
MACPHAIL, R. C 53
MARGEN, S 149
MAURISSEN, J. P. J 23
MCMILLAN, D. E 105
MERIGAN, W. H 15
MITCHELL, C. L 137
MOODY, D. B 33
PAGE, N. P 3
PATTON, J. H 93
REITER, L. W 53
RICE, D. C 85
RODRIGUEZ, R 113
RONDEAU, D. B 175
SCHAUMBURG, H. H 187
SEIFTER, J 9, 217
SPENCER, P. S 189
STEBBINS, W. C 7, 33
CAUSE, E 9 TAKADA, K 45
GELLER, 1 7, 9 TILSON, H. A 137
GOULD, D. H '
VORHEES, C. V.
207
HANNINEN, H 157
HARTMANN, R. J 9
HIENZ, R. D 73
WAYNER, M. J 175
WEISS, B 149, 213
WENGER, G. R 119
JOLICOEUR, F. B 175 WILLIAMS, J. H 149
LATIES, V. G 129, 213
LOOSER, R 163
LORENZANA-JIMENEZ, M 113 ZBINDEN, G
219
WOOD, R. W 67
YOUNG, M 149
163
-------
Test Methods for Definition of Effects of Toxic Substances on Behavior and Neuromotor Function
Neurobehavioral Toxicology, Vol. 1, Suppl. 1, pp. 221-225. ANKHO International Inc., 1979.
SUBJECT INDEX
3-Acetyl pyridine
comparison of neurobehavioral effects induced by various
experimental models of ataxia, 175
Acquisition
behavioral toxicity, 105
performance, 105
serial position sequences, 105
Activity
operant conditioning of infant monkeys for toxicity testing, 85
quantitative analysis of rat behavior patterns in a residential
maze, 163
Affective behavior
behavioral toxicology, 93
effects of pre- and post-natal lead exposure, 93
learning, 93
social interaction, 93
Ammonia
inhaled substances, 67
nitrous oxide, 67
reinforcing properties, 67
toluene, 67
Amobarbital
behavioral assessment of risk-taking and psychophysical
function, 73
behavioral toxicology, 73
Amphetamine
methodological problems in the analysis of behavioral
tolerance, 179
methods in behavioral toxicology, 179
Antimuscarines
methodological problems in the analysis of behavioral
tolerance, 179
methods in behavioral toxicology, 179
Applied behavior analysis
behavioral epidemiology, 149
food additives, 149
food colors, 149
randomization tests, 149
time series, 149
Ataxia
3-acetyl pyridine, 175
comparison of neurobehavioral effects induced by various
experimental models, 175
pyrithiamine, 175
thiamine deficiency, 175
Atropine
behavioral toxicity, 119
caffeine, 119
schedule-controlled responding, 119
spontaneous motor activity, 119
testing for behavioral effects of agents, 119
Auditory discrimination
carbon monoxide, 9
discrimination behavior, 9
environmental contaminants, 9
ketones, 9
polybromide biphenyl, 9
Auditory threshold
behavioral assessment of risk-taking and psychophysical
function, 73
drug-affected sensory threshold changes, 45
kanamycin, 45
LSD-25, 45
pilocarpine, 45
quinidine, 45
trailwise tracking method, 45
visual threshold, 45
Axonopathy
cellular responses to neurotoxic compounds of environmental
significance, 189
Behavioral development
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
Behavioral effects
psychological test methods: sensitivity to long term chemical
exposure at work, 157
Behavioral epidemiology
applied behavior analysis, 149
food additives, 149
food colors, 149
randomization tests, 149
time series, 149
Behavioral teratology
diazepam, 207
fenfluramine, 207
preliminary test battery for the investigation of selected
psychotropic drugs, 207
prochlorperazine, 207
propoxyphene, 207
Behavioral tolerance
amphetamine, 179
antimuscarinics, 179
cannabis derivatives, 179
CNS depressants, 179
methodological problems in the analysis of behavioral tolerance,
179
methods in behavioral toxicology, 179
Behavioral toxicity
acquisition, 105
atropine, 119
caffeine, 119
performance, 105
schedule-controlled responding, 119
screening strategies for the Environmental Protection Agency,
213
serial position sequences, 105
spontaneous motor activity, 119
standardization of tests, 213
testing for behavioral effects of agents, 119
Behavioral toxicology
affective behavior, 93
amobarbital, 73
amphetamine, 179
antimuscarinics, 179
behavioral assessment of risk-taking and psychophysical
function, 73
cannabis derivatives, 179
chlordiazepoxide, 73
CNS depressants, 179
crop dysfunction, 129
diazepam, 73
d-methylamphetamine, 73
effects of pre- and post-natal lead exposure, 93
221
-------
222
ethanol, 129
lead, 129
learning, 93
methodological problems in the analysis of behavioral tolerance,
179
methods, 179
methylmercury, 129
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
quantitative analysis of rat behavior patterns in a residential
maze, 163
social interaction, 93
Caffeine
atropine, 119
behavioral toxicity, 119
schedule-controlled responding, 119
spontaneous motor activity, 119
testing for behavioral effects of agents, 119
Cannabis derivatives
methodological problems in the analysis of behavioral tolerance,
179
methods in behavioral toxicology, 179
Carbon disulfide
psychological test methods: sensitivity to long term chemical
exposure at work, 157
Carbon monoxide
auditory discrimination, 9
discrimination behavior, 9
environmental contaminants, 9
Chlordiazepoxide
behavioral assessment of risk-taking and psychophysical
function, 73
behavioral toxicology, 73
Chronic inhalation
effects of solvents on schedule-controlled behavior, 113
CNS depressants
methodological problems in the analysis of behavioral tolerance,
179
methods in behavioral toxicology, 179
Cognitive tests
psychological test methods: sensitivity to long term chemical
exposure at work, 157
Comparative behavioral toxicology
hearing loss, 33
toxic effects on sensory systems in experimental animal models,
33
Crop dysfunction
behavioral toxicology, 129
ethanol, 129
lead, 129
methylmercury, 129
Diazepam
behavioral assessment of risk-taking and psychophysical
function, 73
behavioral toxicology, 73
preliminary test battery for the investigation of the behavioral
teratology of selected psychotropic drugs, 207
Discrimination behavior
auditory behavior, 9
carbon monoxide, 9
environmental contaminants, 9
ketones, 9
polybromide biphenyl, 9
Discrimination reversal
operant conditioning of infant monkeys for toxicity testing, 85
Drug
3-acetyl pyridine, 175
ammonia, 67
amobarbital, 73
amphetamine, 179
antimuscarinics, 179
atropine, 119
caffeine, 119
cannabis derivatives, 179
carbon disulfide, 157
carbon monoxide, 9
Chlordiazepoxide, 73
diazepam, 73, 207
ethanol, 129
fenfluramine, 207
kanamycin, 45
lead, 85, 93, 129, 193
LSD-25, 45
d-methylamphetamine, 73
methylmercury, 15, 129
methylmercury chloride, 163
narcotic analgesics, 179
nitrous oxide, 67
organophosphate anticholinesterases, 179
pentobarbital, 73
pilocarpine, 45
polybrominated biphenyl, 9
prochlorperazine, 207
propoxyphene, 207
pyrithiamine, 175
quinidine, 45
scopolamine, 179
thiamine, 175
toluene, 67, 113
Drug-affected sensory threshold
auditory threshold, 45
kanamycin, 45
LSD-25, 45
pilocarpine, 45
quinidine, 45
trailwise tracking method, 45
visual threshold, 45
Drug-toxicant interactions
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
Environmental chemicals
psychological test methods: sensitivity to long term chemical
exposure at work, 157
Environmental contaminants
auditory discrimination, 9
carbon monoxide, 9
discriminative behavior, 9
ketones, 9
polybrominated biphenyl, 9
Environmental Protection Agency
behavioral toxicity, 213
screening strategies, 213
standardization of tests, 213
Environmentally significant neurotoxins
axonopathy, 189
cellular response, 189
myelinopathy, 189
neuronopathy, 189
Ethanol
behavioral toxicology, 129
crop dysfunction, 129
lead, 129
methylmercury, 129
-------
223
Experimental animal models
comparative behavioral toxicology, 33
hearing loss, 33
toxic effects on sensory systems, 33
Fenfluramine
a preliminary test battery for the investigation of the behavioral
teratology of selected psychotropic drugs, 207
Fixed ratio
operant conditioning of infant monkeys for toxicity testing, 85
Food additives
applied behavior analysis, 149
behavioral epidemiology, 149
food colors, 149
randomization tests, 149
time series, 149
Food colors
applied behavior analysis, 149
behavioral epidemiology, 149
food additives, 149
randomization tests, 149
time series, 149
drug-toxicant interactions, 193
screening tests, 193
seizure development, 193
temperature regulation development, 193
visual evoked response development, 193
Learning
effects of pre- and post-natal lead on affective behavior, 93
Locomotion patterns
quantitative analysis of rat behavior patterns in a residential
maze, 163
Long term chemical exposure
behavioral effects, 157
carbon disulfide, 157
cognitive tests, 157
environmental chemicals, 157
perceptual motor tests, 157
sensitivity, 157
LSD-25
auditory threshold, 45
drug-affected sensory threshold changes, 45
trailwise tracking method, 45
visual threshold, 45
Hearing loss
comparative behavioral toxicology, 33
toxic effects on sensory systems in experimental animal models,
33
Inhaled substances
ammonia, 67
nitrous oxide, 67
reinforcing properties, 67
toluene, 67
Kanamycin
auditory threshold, 45
drug affected sensory threshold changes, 45
trailwise tracking method, 45
visual threshold, 45
Ketones
auditory discrimination, 9
discrimination behavior, 9
environmental contaminants, 9
Lead
behavioral toxicology, 129
crop dysfunction, 129
ethanol, 129
methylmercury, 129
operant conditioning of infant monkeys for toxicity testing, 85
Lead exposure
affective behavior, 93
behavioral toxicology, 93
learning, 93
pre- and post-natal exposure, 93
social interaction, 93
Lead toxicity
behavioral development, 193
behavioral toxicology, 193
Measurement technique
motor activity, 53
toxicity testing, 53
d-Methylamphetamine
behavioral assessment of risk-taking and psychophysical
function, 73
behavioral toxicology, 73
Methylmercury
behavioral toxicology, 129
crop dysfunction, 129
ethanol, 129
lead, 129
neurotoxicity, 15
visual fields, 15
visual thresholds, 15
Methylmercury chloride
quantitative analysis of rat behavior patterns in a residential
maze, 163
Method
motor activity: a survey of methods with potential use in toxicity
testing
operant conditioning of infant monkeys (Macaco fascicularis) for
toxicity testing, 85
Methodological problems
behavioral tolerance, 179
behavioral toxicology, 179
Morphology
toxic distal axonopathy, 187
Motor activity
measurement technique, 53
toxicity testing, 53
Myelinopathy
cellular responses to neurotoxic compounds of environmental
significance, 189
Neonatal exposure
operant conditioning of infant monkeys for toxicity testing, 85
Neurobehavioral effects
3-acetyl pyridine, 175
experimental models of ataxia, 175
pyrithiamine, 175
thiamine deficiency, 175
-------
224
Neurobehavioral screening procedures
examples of validation of neurobehavioral tests, 137
factors to be considered in screening, 137
neurobehavioral toxicity, 137
Neurobehavioral tests
examples of validation, 137
factors to be considered in screening, 137
neurobehavioral screening procedures, 137
neurobehavioral toxicity, 137
Neurobehavioral toxicity
examples of validation of neurobehavioral tests, 137
factors to be considered in screening, 137
neurobehavioral screening procedures, 137
Neuronopathy
cellular responses to neurotoxic compounds of environmental
significance, 189
Neurotoxicants
sensory disturbances, 23
somatosensory system, 23
vibration sensitivity, 23
Neurotoxicity
methylmercury, 15
visual fields, 15
visual thresholds, 15
Neurotoxins
axonopathy, 189
cellular response, 189
environmental significance, 189
myelinopathy, 189
neuronopathy, 189
Nitrous oxide
ammonia, 67
reinforcing properties of inhaled substances, 67
toluene, 67
Operant behavior
effects of solvents on schedule-controlled behavior, 113
Operant conditioning
activity, 85
discrimination reversal, 85
fixed ratio, 85
lead, 85
neonatal exposure, 85
toxicity testing, 85
visual discrimination, 85
Pentobarbital
behavioral assessment of risk-taking and psychophysical
function, 73
Perceptual motor tests
psychological test methods: sensitivity to long term chemical
exposure at work, 157
Performance
acquisition, 105
behavioral toxicity, 105
serial position sequences, 105
Pharmacological toxicants
behavioral assessment of risk-taking and psychophysical
function, 73
Pilocarpine
auditory threshold, 45
drug-affected sensory threshold changes, 45
trailwise tracking method, 45
visual threshold, 45
Polybrominated biphenyl
auditory discrimination, 9
discrimination behavior, 9
environmental contaminants, 9
Prochlorperazine
a preliminary test battery for the investigation of the behavioral
teratology of selected psychotropic drugs, 207
Propoxyphene
a preliminary test battery for the investigation of the behavioral
teratology of selected psychotropic drugs, 207
Psychological test methods
behavioral effects, 157
carbon disulfide, 157
cognitive tests, 157
environmental chemicals, 157
perceptual motor tests, 157
sensitivity to long term chemical exposure at work, 157
Psychophysical function
amobarbital, 73
behavioral toxicology, 73
chlordiazepoxide, 73
diazepam, 73
d-methylamphetamine, 73
pentobarbital, 73
Pyrithiamine
comparison of neurobehavioral effects induced by various
experimental models of ataxia, 175
Quantitative analysis
activity, 163
behavioral toxicology, 163
locomotion patterns, 163
methylmercury chloride, 163
rat behavior patterns in a residential maze, 163
Quinidine
auditory threshold, 45
drug-affected sensory threshold changes, 45
trailwise tracking method, 45
visual threshold, 45
Randomization tests
applied behavior analysis, 149
behavioral epidemiology, 149
food additives, 149
food colors, 149
time series, 149
Rat behavior patterns
activity, 163
behavioral toxicology, 163
locomotion patterns, 163
methylmercury chloride, 163
quantitative analysis, 163
residential maze, 163
Rate-dependent effects
effects of solvents on schedule-controlled behavior, 113
Reinforcing properties
ammonia, 67
inhaled substances, 67
nitrous oxide, 67
toluene, 67
Residential maze
activity, 163
behavioral toxicology, 163
locomotion patterns, 163
methylmercury chloride, 163
quantitative analysis of rat behavior patterns, 163
-------
225
Risk-taking
amobarbital, 73
auditory thresholds, 73
behavioral toxicology, 73
chlordiazepoxide, 73
diazepam, 73
d-raethylamphetamine, 73
pentobarbital, 73
pharmacological toxicants, 73
psychophysical functions, 73
Safety-testing
atropine, 119
behavioral toxicity, 119
caffeine, 119
schedule-controlled responding, 119
spontaneous motor activity, 119
Schedule-controlled responding
atropine, 119
behavioral toxicity, 119
caffeine, 119
spontaneous motor activity, 119
testing for behavioral effects of agents, 119
Screening factors
examples of validation of neurobehavioral tests, 137
neurobehavioral screening procedures, 137
neurobehavioral toxicity, 137
Screening strategies
behavioral toxicity, 213
Environmental Protection Agency, 213
standardization of tests, 213
Screening tests
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
Seizure development
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
Sensory disturbances
neurotoxicants, 23
somatosensory system, 23
vibration sensitivity, 23
Sensory systems
comparative behavioral toxicology, 33
experimental animal models, 33
hearing loss, 33
toxic effects, 33
Serial position sequences
acqusition, 105
behavioral toxicity, 105
performance, 105
Social interaction
effects of pre- and post-natal lead on affective behavior, 93
Solvents
chronic inhalation, 113
operant behavior, 113
rate-dependent effects, 113
temporal discrimination, 113
toluene, 113
Somatosensory system
neurotoxicants, 23
sensory disturbances, 23
vibration sensitivity, 23
Spontaneous motor activity
atropine, 119
behavioral toxicity, 119
caffeine, 119
schedule-controlled responding, 119
testing for behavioral effects of agents, 119
Standardization of tests
behavioral toxicity, 213
screening strategies for the Environmental Protection Agency,
213
Temperature regulation development
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
Temporal discrimination
effects of solvents on schedule-controlled behavior, 113
Thiamine deficiency
comparison of neurobehavioral effects induced by various
experimental models of ataxia, 175
Time series
applied behavior analysis, 149
behavioral epidemiology, 149
food additives, 149
food colors, 149
randomization tests, 149
Toluene
ammonia, 67
effects of solvents on schedule-controlled behavior, 113
nitrous oxide, 67
reinforcing properties of inhaled substances, 67
Toxic distal axonopathy
morphology, 187
Toxic effects
comparative behavioral toxicology, 33
hearing loss, 33
sensory systems in experimental animal models, 33
Toxicity testing
activity, 85
discrimination reversal, 85
fixed ratio, 85
lead, 85
measurement technique, 53
motor activity, 53
neonatal exposure, 85
operant conditioning, 85
visual discrimination, 85
Trailwise tracking method
auditory threshold, 45
drug-affected sensory threshold changes, 45
kanamycin, 45
LSD-25, 45
pilocarpine, 45
quinidine, 45
visual threshold, 45
Visual discrimination
operant conditioning of infant monkeys for toxicity testing, 85
Visual evoked response development
physiological and neurobehavioral alterations during
development in lead exposed rats, 193
Visual fields
methylmercury, 15
neurotoxicity, 15
visual thresholds, 15
Vibration sensitivity
neurotoxicants, 23
sensory disturbances, 23
somatosensory system, 23
Visual threshold
auditory threshold, 45
drug-affected sensory threshold changes, 45
kanamycin, 45
LSD-25, 45
methylmercury, 15
neurotoxicity, 15
pilocarpine, 45
quinidine, 45
trailwise tracking method, 45
visual fields, 15
------- |