U

Chapter
Choice of Animal Subjects
in Behavioral Analysis
William J. Jackson
Contents
I. Introduction
II. Origin of the Albino Laboratory Rat
III. The Laboratory Rat in Behavioral Research
IV. Advantages of Rat Models
V. Disadvantages of Rat Models
VI. Strain Selection
A. The Wistar Rat Colony
B. The Long-Evans Strain
C. Strains from Columbia University
D. Sprague-Dawley Rats
E. Holtzman Rats
F. N/Nih Rats
G. Wild Norway Rats
VII. Inbred Rat Strains Selected for Various Behavioral Traits
A. Rat Strains Selected for Preference of — and
Sensitivity to —Alcohol
B. ACI Strain
C. Strains Bred for Various Serotonin Receptors
D. Roman Strain
E. Maudsley Strains
F. Tryon’s Maze-Bright and Maze-Dull Rats
G. Spontaneous Hypertensive Rats
H. Flinders Sensitive Line and Flinders Resistant Line
I. Dahl Salt Sensitive Rats
2
Methods of Behavior Analysis in Neuroscience
VIII. Comparison of Various Rat Strains for Behavioral Characteristics
IX. Mice in Behavioral Research
X. Pigeons and Other Species Performing Traditional Non-human
Primate Tasks
XI. Non-Human Primates
A. Advantages
B. Disadvantages
C. Commonly Used Non-human Primates in Biomedical
Research
D. Basic Behavioral Differences Among Monkey Species
E. Primate Cognitive Skills
F. Transfer of Training
XII. Discussion
References
I. Introduction
Many researchers using behavioral techniques are not primarily interested in animal
behavior, as such. Typically, behavioral animal research in physiology and pharmacology
is designed to provide a model for human processes, and great effort is given
toward the development of animal models that reflect behavioral processes shared
by animals and humans.
1
Whenever using animals as research subjects, behavioral

physiologists, pharmacologists, and geneticists commonly describe their work as
animal models of human characteristics, or justify their work on the basis of relevance
to human pathology. The search for treatment and cure of illness with behavioral
implications will continue to lean heavily upon animal models. Animal subjects
will assist in evaluation of the effectiveness of putative treatments, and in providing
further insight into underlying physiologic mechanisms of human pathology. Subsequent
chapters of this book focus on many of these models. Anxiety, addiction,
taste-aversion, attention deficit, and disorders of learning and memory are examples
of behavioral disorders that are often studied
via
animal models.
Behavioral researchers usually plan most aspects of their research projects to
fine detail, but may fail to give the same level of planning toward selection of the
species that is to be used as a model. The goal of this chapter is to provide a glance
at some of the most popular species used for behavioral neuroscientific research.
Most of the discussion is given to rats and non-human primates, but there is a section
on mice and notes about other species as well.
II. Origin of the Albino Laboratory Rat
Contemporary strains of albino laboratory rats were bred from captured wild Norway
rats. The wild Norway rat is believed to have originated from temperate regions of
© 2001 by CRC Press LLC
Choice of Animal Subjects in Behavioral Analysis
3
Asia and southern Russia. As civilization developed, these animals found a suitable
ecological niche in the castoffs and trash heaps of man and, as economic pests, spread
rapidly over the world. Norway rats were common throughout Europe and the British
Isles by the early 18th century. By 1775, Norway rats were common in the northeastern
U.S.
2
Because the large number of rats represented an economic hazard, special breeds
of rat-catching terrier dogs were bred. In a roundabout way, the breeding of these dogs
was responsible for the beginnings of the albino laboratory rat.
The recorded speculation of early breeders of albino strains was that they were
products of the “sport” of rat baiting, which was outlawed by decree.
3
Rat baiting
involved the release of 100 or 200 newly trapped rats into a fighting pit. A trained
terrier dog was then put into the pit, and a measurement of the time until the last
rat was killed was taken. Wagers were placed on the speed of the various dogs. Rat
baiting required that many rats be trapped and retained in pounds. Historical records
relate that albinos were removed from these pounds and kept for show purposes and
breeding. Many albino show rats were tamed and offspring were selected for docility
and color. Because the captured wild Norway rats were fierce and difficult to handle,
the more docile albinos were the stock from which early European laboratory
researchers selected their first animals.
III. The Laboratory Rat in Behavioral Research
After serving for a hundred years as a subject in behavioral and physiological
research experiments, the albino rat is a generally accepted model. However, early
researchers had to justify their selection of animals, as opposed to humans, and it
would be valuable for us to remember their rationale. Memories of the initial battles
to gain acceptance for animal models of human cognition have dimmed over time;
today, studies using animals are commonly accepted.
Behavioral investigation using rats as subjects in the U.S. arose from work at
Clark University Biological Laboratory during the 1890s. According to Miles,
4
Stewart was working with wild gray rats to determine the effects of alcohol, diet,
and barometric change on the activity of the animals as early as 1894. The feralcaptured

gray rats were fierce and difficult to handle, and Stewart was forced to
switch to the more docile white rats by 1895. The rats were trained to run through
the maze to earn food reward. Kline
5
invented several problem boxes that served as
prototypes for contemporary devices still used to measure cognition and learning/
memory in rodents. Small used a maze patterned after the garden maze found
at Hampton Court Palace in England to measure observable behavior that would
indicate learning by the rats. Use of the white rat as a research subject in behavioral
research was given a great impetus by the investigations of Watson
6
at Chicago
University. Watson’s ideas firmly cemented the white rat as a fixture of experimental
studies in behavior.
7
At the same time of Stewart’s behavioral work in the 1890s, Donaldson was
using the white rat at the University of Chicago for anatomical and physiological
research. Donaldson’s rat colony became the parent stock at the Wistar Institute of
4
Anatomy and Biology in Philadelphia. In 1924, a book entitled
The Rat
14
by H. H.
Donaldson gave impetus to the white rat as an acceptable research subject and
provided wisdom about the use of animal subjects in biomedical research. In one
comment from
The Rat
, Donaldson stated that “in enumerating the qualifications of
the rat as a laboratory animal, and in pointing out some of its similarities to man,
it is not intended to convey the notion that the rat is a bewitched prince or that man
is an overgrown rat, but merely to emphasize the accepted view that the similarities
between mammals having the same food habits tend to be close, and that in some
instances, at least, by the use of equivalent ages, the results obtained with one form
can be very precisely transferred to the other.”
IV. Advantages of Rat Models
Rats are commonly employed as animal subjects in contemporary medical research,
and general acceptance of the white rat as an animal research subject has increased
in synchrony with an increased appreciation of the value of behavioral research.
Since rats are small, clean, relatively inexpensive, easily handled and maintained,
widely available, have short twenty-one day gestational periods, and a short two to
three year lifespan, their use as research subjects offers many advantages. These
advantages are amplified in application to research problems that require large
numbers of animals. Likewise, the relatively short twenty-one day gestation and
approximate three year lifespan of the rat provide a practical opportunity to study
the stages of development and aging.
V. Disadvantages of Rat Models
Despite the numerous advantages offered by the laboratory rat model, there are
difficulties that must be considered. First, it is more difficult to draw parallels
between rodent and human behavior and physiology, than to compare non-human
primates with humans. The behavioral characteristics of the rodent subject are more
primitive, making behavioral comparison more complicated. Second, it is more
difficult to establish stimulus control of the rodent’s behavior in training paradigms.
Often, it is necessary to use aversive electrical shock or drastic food deprivation to
motivate the rodent subjects. These severe control procedures further complicate the
comparison with humans, since such control measures are unacceptable for human

research.
VI. Strain Selection
Once it has been decided to use rats as subjects in a given research endeavor, the
question of strain selection becomes important. There are many outbred and inbred
5
strains of rats available on the commercial market, and the effects of many common
treatments differ according to the strain selected. Early in my scientific training I
received a lesson about this important fact by finding that hippocampectomy in Sprague-
Dawley rats increased several forms of activity, and increased one type of error
in a Lashley III maze. However, when the same hippocampectomy was effected in
Fischer strain rats, the animals consistently became less active, and did not make the
same Lashley III error. Normally, Fisher rats are much less active than normal Sprague-
Dawley rats, and the hippocampectomy may have removed cortical inhibition of
behavioral tendencies that were already present in the normal unoperated controls.
There are many published papers that show differences among strains of laboratory
rats regarding the effects of various treatments, shown in the following sections.
A. The Wistar Rat Colony
The first rats brought from Chicago to the Wistar Institute in Philadelphia by Donaldson
in 1906 became the parent stock of a rat colony, whose offspring were sold to
research facilities throughout the U.S. and many other countries until 1960. The
Wistar Institute was a leader in determining laboratory animal husbandry practices
necessary to support a large rat colony. By 1922, the colony had a total population
of about 6000 rats. The commercial rights to the sale of the Wistar Rat were sold
in 1960.
From the beginning, the Wistar Institute maintained a random bred, heterogeneous
colony. Therefore, there was considerable variability in the commercial colony
maintained by the Institute. It is unknown whether albino lines other than those
provided by Donaldson were introduced into the Wistar Colony. According to Lindsey,
2
it is documented that outside breeders were brought into the colony to boost
breeding production.
Most of the albino laboratory rats used in the U.S. are linked to the colony of
the Wistar Institute. In previous discussion, the role of Donaldson in the development
of the albino laboratory rat model was mentioned. Donaldson’s rat colony became
the parent stock at the Wistar Institute of Anatomy and Biology. Even Donaldson
himself did not know whether stock from the European labs found its way to the
U.S., or whether the first albinos in the U.S. were derived from wild rats captured
in the U.S.
8,9

B. The Long-Evans Strain
Prior to 1920, two members of the faculty at the University of California at Berkley,
J. A. Long and H. M. Evans, were interested in the estrous cycle of the rat. To
support their research interests, Long and Evans established one of the leading strains
of rats that continues to bear their names. The origins of the Long-Evans rat colony
were described in a monograph entitled
The Oestrous Cycle in the Rat and Its
Associated Phenomena
. In that monograph, it was stated that the colony descended
from a cross made around 1915 “between several white females and a wild gray
6
male” that was trapped along the bank of Strawberry Creek, which ran through the
university campus. Dr. Leslie Bennett, a colleague of Dr. Long, is quoted as saying
that the white females were supplied by the Wistar Institute.
2

The Long-Evans rats exhibited varied fur color. The coats represented in the
colony are black, gray, and hooded. The hooded animals are characterized by pigmented
fur on their heads and often along the spine. These animals have a pigmented
iris, and their visual acuity far exceeds that of the albino strains.
C. Strains from Columbia University
The Crocker Institute of Cancer Research at Columbia University began in 1913 to
inbreed six major bloodlines of rats to be used in cancer research. Researchers at
the institute had noticed that certain rats were more susceptible to sarcomas induced
by feeding tapeworm eggs. The inbreeding program was designed to determine if
the susceptibility to sarcoma was genetic.
2
Dr. Maynie R. Curtis was the chief
developer of the inbred rat colony. Dr. Curtis purchased a few breeding pairs of rats
from each of four local breeders whose names were August, Fischer, Marshall, and
Zimmerman. The obtained rats had different coloring, and these external characteristics
were used as markers to help identify the various strains. The Marshall rats
were albinos. The Fischer and Zimmerman rats were non-agouti piebalds, but did
carry the albino gene. The August rats were the most varied, and included some
with pink eyes. In 1941, a group of rats with red eyes were obtained from a breed
in Connecticut. These animals were seed stock for brother-sister mating and are
progenitors of several popular inbred strains. The first litter of pedigreed rats at
Columbia were from mating number 344, and were the first representatives of the
Fischer 344 strain.
D. Sprague-Dawley Rats
There appears to be little record of the origin of the Sprague-Dawley strain. The
primary stock is believed to have been established by Robert W. Dawley, who was
a physical chemist at the University of Wisconsin. Mr. Dawley included his wife’s
maiden name, which was Sprague, to name the rats. Mr. Dawley later established
Sprague-Dawley, Inc. to advance the commercial sale of his rats. Lindsey
2
cites a
letter from Mr. Dawley to the National Institutes of Health (NIH), dated July of
1946, in which he states that the original parents were a hybrid hooded male rat of
exceptional size, and vigor, that was genetically half albino. He was mated to a
white female, and subsequently to his white female offspring for seven generations.
The origin of the hooded male is unknown, but the first white female is believed to
be from the Wistar colony. Selection was on the basis of many factors, including
high lactation, rapid growth, vigor, good temperament, and high resistance to arsenic
trioxide. The original company continues today under the name Harlan Sprague-
Dawley.
7
There are many sub-lines of the Sprague-Dawley animals, and they are often
used in behavioral research. They are randomly bred strains, but differ from one
another. The Sprague-Dawley strain is also sold by the Charles River Co. There is
variability in the stock marketed by the different suppliers. For example, Pollock
and Rekito 10
found that the Sprague-Dawley rats marketed by Harlan differed from
Sprague-Dawley rats marketed by the Charles River Co. in regard to hypertensive
response to chronic L-NAME-induced nitric oxide synthase inhibition.
E. Holtzman Rats
A major sub-line of the Sprague-Dawley line is the Holtzman rat. These rats are
provided by a company established by E. C. Holtzman, who was a former employee
of the Sprague-Dawley Co. Sprague-Dawley animals were the original seed stock
of the Holtzman line.
F . N/Nih Rats

The National Institutes of Health (NIH), through systematic interbreeding of eight
inbred rat strains, created a heterogeneous stock of rats.
11 This strain, called N/Nih,
has been maintained by a strict breeding policy to ensure that mating pairs of
subsequent generations are distantly related. Predictably, the large genetic variability
among the N/Nih rats results in varied phenotypes. Several selective breeding programs
that originated with N/Nih rats have developed strains of inbred rats with
characteristics of great interest to behavioral research. These inbred strains include
rats chosen for high and low alcohol consumption and sensitivity. The genealogy
of the animals maintained by the NIH can be found on their web site
(http://www.nih.gov/od/ors/dirs/vrp/s&slst.htm#ratinbred), or by correspondence
with them.
G. Wild Norway Rats
Not many researchers are hardy enough to explore the use of wild rats as animal
subjects, although some have.
12,13
Discussion of the laboratory rat would be incomplete
without some mention of how the typical albino laboratory rat differs from the original
wild stock. Albino laboratory rats were originally selected for docility, i.e., a reduced
tendency to flee from humans or to struggle and bite when handled. There are implications
of this selection for docility. For example, novel objects that induce avoidance
or fear in wild rats often elicit approach or apparent curiosity in the albino laboratory
strains. Albino males do not attack other male rats with the intensity typical of wild
male rats. The food preferences of the wild rats also differ from the albino strains.
12,14
Differences in behavior are paralleled by changes in growth and in the relative weight
of the adrenal glands. Wild rats are genetically heterogeneous, and thus there is
considerable behavioral variation among wild rats. Barnett
15
used the techniques of
8
ethology to observe the innate behavioral characteristics of wild Norway rats in groups.
Under these conditions, it is obvious that rats have a complex social structure that is
not generally measured in typical laboratory testing. Barnett also describes rat behavior
that most of us have casually noticed, but generally have not understood. Barnett
explained and illustrated the role of many body postures and gestures in rat society.
Barnett has also reviewed techniques for testing wild Norway rats.
15

VII. Inbred Rat Strains Selected for Various
Behavioral Traits
With the explosion of interest in the genetic basis of physiology and behavior, there
are many papers in which various inbred strains of rats are compared to outbred
strains according to the criteria of interest.
Populations of outbred rats such as the
Wistar, and hetereogenous strains such as the N/Nih rats, manifest considerable
variability along almost any behavioral or physiological attribute. Selective breeding
programs for high or low manifestations of various phenotypes have resulted in
inbred strains that are useful for many areas of behavioral research.
The names of these strains do not follow a systematic nomenclature. One attempt
to standardize the rat strains
16
recommended that the rat nomenclature system follow
that of inbred mice. The following quote from Festing and Staats presents the problem:
“In many cases a strain name has been changed whenever a strain has been transferred
to a new laboratory, and in other cases strains which have only a distant relationship
have been given the same name. This is particularly the case with strains descended

from Wistar outbred stock, which tend to be named ‘WIS’ or some other name
beginning with W.” The nomenclature system recommended by Festing and Staats has
not been universally followed. Many strains are named after the university or some
other prominent, but not obvious, feature of their development. The following sections
describe basic aspects of some common inbred strains used in behavioral research.
16

A. Rat Strains Selected for Preference of — and Sensitivity
to — Alcohol
Normal out-bred (e.g., Wistar) and heterogeneous rats (e.g., N/Nih) do not typically
drink much alcohol, but there are large individual differences in alcohol drinking
over a large population of these rats. By selectively breeding for high or low alcohol
intake and high and low neurosensitivity, lines of rats that exhibit these characteristics
regarding alcohol have been derived. The inbred alcohol-preferring rats typically
consume up to ten times the amount of alcohol taken by normal out-bred or heterogeneous
rats. The selectively bred rat lines include the alcohol-preferring (P),
alcohol-accepting (Alko Alcohol — AA), Sardinian alcohol-preferring, and high
alcohol drinking (HAD) rats.
17
These lines do not share all other behavioral traits.
9
Higher than normal alcohol drinking has also been observed in several groups
of rats that were selectively bred for other specific behavioral traits. These include
the Tryon Maze-Bright and Tryon Maze-Dull rats,
18
the Roman High- and Low-
Avoidance rats,
19
and the Fawn-Hooded rats that have serotonin receptor abnormality.
20,21
Typically, the ACI strain of rats will not voluntarily drink alcohol.
B. ACI Strain
The ACI line was originated by Curtis and Dunning at the Columbia University
Institute for Cancer Research. Initially, the primary phenotype of interest was susceptibility
to estrogen-induced tumors; however, there are also a number of behavioral
differences associated with this strain. For example, the Brh sub-line shows
low defecation response and high activity response in the open field test.
C. Strains Bred for Various Serotonin Receptors
Selective inbreeding of N/Nih rats has resulted in strains that vary in sensitivity to
5-HT
1A
receptor stimulation. Overstreet
21
has established a selective breeding program
for high (HDS) and low (LDS) sensitivity to the hypothermic response of the
5-HT
1A
agonist 8-OH-DPAT. These two rat lines are believed to differ in behavioral
tests of depression, but not of anxiety. The lines also differ in post-synaptic 5-HT
1A
receptors. Pre-synaptic mechanisms are not affected.
D. Roman Strain
The Roman strain of rats was selectively bred from Wistar stock for high and low
performance in two-way active avoidance learning.
22
RHA/Verh rats acquire active
avoidance (shuttle box) performance quickly, because they are less emotionally
reactive, but more active in regard to locomotion. RLA/Verh rats cope with the active
avoidance problem more passively, and become immobile when faced with the

avoidance task. They show increased defecation in the open field and increased
activity in the hypothalamic-pituitary-adrenal axis. The two lines of the Roman strain
differ in many respects at the behavioral and neurochemical level (for review).
19

E. Maudsley Strains
The Maudsley Reactive (MR) rats were selectively bred for high defecation in the
open field test.
23
Maudsley Non-Reactive rats (MNRA) were selectively bred for
low rates of defecation in the open field test. The strains were genetically selected
from outbred Wistar progenitors.
10
F. Tryon’s Maze-Bright and Maze-Dull Rats
The genetics of maze-learning ability was investigated by Tryon and his group of
researchers from the late 1920s through the 1940s. Tryon
24,25
was successful in
segregating animals that were adept at learning a variety of mazes (Tryon mazebright)
from those rats that learned less well (Tryon maze-dull). Tryon conducted
many experiments to identify the factors that were involved in maze learning. Some
of the differences between the two groups can be attributed to emotionality and
others seem to be a function of the ability of rats to perceive helpful guiding cues
outside of the maze.
G. Spontaneously Hypertensive Rats
The adult spontaneously hypertensive (SHR) rat includes behavioral changes in
its phenotype, in addition to changes in blood pressure. SHR rats are also impaired
in ability to perform learning and memory-related tasks and exhibit a decrease in
the expression and nicotine-stimulated function of brain nicotinic acetylcholine
receptors. These cholinergic factors are known to be important to learning performance.
26,27

H. Flinders Sensitive Line and Flinders Resistant Line
The Flinders Sensitive Line (FSL) and the Flinders Resistant Line (FRL) were
selectively bred at Flinders University, in Australia, by selective breeding for
differences in effects of the anticholinesterase, di-isopropylfluorophosphate (DFP)
on temperature, drinking, and body weight. The FSL rats are more sensitive to
DFP, as well as cholinergic agonists, and have more brain muscarinic receptors
than the FRL rats (for review).
28
Because of the known relationship between
cholinergic activity and many functions in humans and animals, these strains are
useful for a wide variety of research topics, including learning, memory, and
depression.
I. Dahl Salt Sensitive Rats
Another hypertensive rat strain is the Dahl salt sensitive rat. Correspondingly, there
is a Dahl salt resistant strain. The salt sensitive rats exhibit reduced learning ability,
and also reduction in nAChR in brain regions known to be important for learning
and memory. These include the hippocampus and the amygdala.
11
VIII. Comparison of Various Rat Strains for
Behavioral Characteristics
From the preceding sections, it is clear that the most common outbred, and certainly
the inbred rat strains, differ in regard to important baseline behavioral characteristics.
Some years ago, Harringon
29,30-37

and others presented behavioral standardization
data for several outbred and inbred rat strains. The behavioral comparisons included
activity in both stabilimeters,
30,31
open-field
38
and rotating wheels,
39
free operant
lever press levels,
37
cued lever press levels,
32,36
passive avoidance conditioning,
33
shuttle-box avoidance conditioning,
35
home cage behavior,
31
and runway learning.
34
Other comparisons measured anatomic and physiologic variables, including basal
metabolism Harrington and Hellwig,
40
organ weights,
41
and cholinesterase.
The value of these baseline studies is that they compare the building blocks of
behavior and some important physiologic substrates for a number of popular rat
strains. Each of these studies of elementary behavioral characteristics presents data
obtained from the behavior of approximately 500 rats. Though the studies were
conducted some years ago, they remain valuable as a catalog of standards by which
strains of rats may be specifically selected for particular behavioral traits.
IX. Mice in Behavioral Research
The major reason for the use of the mouse in behavioral research resides in the abundance
of knowledge about the mouse genome and in the large number of genetically
defined strains that display characteristics valuable for research. There is considerable
excitement about the possibilities of measuring data from genetically engineered strains.
The opportunities afforded by “knockout” of specific genes have garnered considerable
interest. Likewise, transgenic mice have been genetically engineered and altered by
injection of one or more genes, such as the human gene for apolipoprotein E (ApoE),
which has been linked with the pathogenesis of Alzheimer’s disease.
The small size of the mouse poses technical problems; however, there is commercially
available, specialized behavioral equipment designed for their use. It is
possible to obtain behavioral data from mice through many of the traditional paradigms
designed for use with the rat. Mice are often tested by mazes and simple
behavioral measures, such as drinking behavior.
One of the first things that will be noticed when reading research literature
pertaining to mice is the complicated system of nomenclature for the many strains.
A review of this system has been written by Lyon,
42
and is necessary for examination
of various strains. The vast number of mouse strains developed for biomedical
research exceed any other species. The Jackson Lab in Bar Harbor, Maine is a leader
in the development of the laboratory mouse.
12
One of the first steps toward learning more about the relationship of behavior to
the mouse genome will be behavioral comparisons of various strains. The NIH has
funded a multicenter effort toward that goal, and the Department of Psychology at the

State University of New York at Albany hosts a website that updates the results of the
multi-center effort. At the time of this writing, the website is available through
www.albany.edu.
X. Pigeons and Other Species Performing
Traditional Non-human Primate Tasks
Pigeons are able to perform visually complex discriminated operant problems, such as
non-spatial delayed matching-to-sample; however, there are important differences
between the strategies that the pigeons use in comparison with primate species. For
example, Cumming and Berryman 43 trained pigeons to match red, blue, and green disks
to a high level of accuracy. To test for transfer of the matching concept, a novel yellow
disk was then substituted for the blue disk. The pigeons regressed to chance levels of
accuracy on trials in which the sample was the yellow stimulus. The birds resorted to
a more concrete position strategy when the novel stimulus appeared. Jackson
44 tested
a series of similar transfers in rhesus monkeys and found that the monkeys continued
to match at 100% accuracy when presented with a novel yellow stimulus. However, if
the novel stimulus was not a disk (e.g., cross, triangle), then the monkeys did not transfer
the matching behavior either. These species differences relate very much to the Transfer
Index (TI) of Rumbaugh, 45 a useful tool for comparing cognitive skills between species.
Pigeons, in fact, do not form concepts of matching-to-sameness, but instead form
individual stimulus-response links that are not mediated by a matching concept.
There are numerous papers in the recent literature in which the titles indicate
results of experiments using rodents and other sub-primate species while performing
tasks traditionally associated with non-human primates. Delayed response tasks,
such as delayed matching-to-sample are examples of this. In most instances, these
behavioral tasks are actually quite different from the typical non-spatial delayed
response tasks traditionally administered to non-human primates. In the experience
of this author, rodents are surprisingly proficient at delayed tasks that involve a
spatial dimension, such as position or location; however, rodents are very much less
proficient in non-spatial tasks. Pigeons and other birds have highly developed visual
skills and can perform non-spatial visual tasks with high levels of accuracy.
XI. Non-human Primates
A. Advantages
As research subjects, non-human Primates offer several potential advantages over
other species. The behavioral capabilities of non-human primate species resemble
13
human capability along many criteria, and the near phylogenetic position of nonhuman
primate species to humans offers opportunity for modeling many aspects of
human anatomy and both normal and abnormal function.
46 Some behavioral tests
can be adapted to allow testing of both humans and non-human primates by the
same paradigm.
47,48 This offers a distinct advantage for the modeling of human
physiology and behavior.
B. Disadvantages
Despite the attractive advantages of using non-human primates as research subjects,
there are numerous disadvantages to their use. Non-human primates are expensive
to obtain and to maintain. Non-human primates are large, powerful, and potentially
dangerous animals. The similarity to man that makes them attractive as research
subjects also carries increased risk of animal-to-human disease transmission. Available
animals, whether feral captured or colony-reared, are from diverse populations;
thus there is great inter-subject variability among animals. The problems of such
variability are magnified by generally small subject populations, which are a consequence

of cost factors that often limit the number of subjects to be obtained and
maintained for a given project.
C. Commonly Used Non-human Primates in
Biomedical Research
Table 1.1, adapted from Whitney,
49
lists some of the commonly employed primate
species, and serves as a guide for the different family groups of non-human
primates.
D. Basic Behavioral Differences Among Monkey Species
Despite the large number of research studies that have been conducted upon various
species of non-human primates, it is difficult to evaluate which species to select for
a given project. Often, economic factors or availability determine the final choice;
however, there is considerable difference in basic behavior among the non-human
primates, and these differences have implication for the results of many treatments.
These special species differences should be considered, if the results of the selected
treatment are to be fully appreciated.
One comparison of basic behaviors among several common non-human primate
species was conducted by Davis.
50
The authors observed and measured lemurs, Old
World monkeys, and New World monkeys in a runway that measured 9 ft long by
18 in. wide by 24 in. high. Located on both ends of the runway was an “observation”
box (18 by 18 by 24 in.) into which an animal of the same species was introduced.
The monkey to be observed was introduced into the long central runway area. The
14
categories of behavior that were measured were: visual surveillance, cage manipulation,
social behavior, rapid energy expenditure, self-involved behavior, and vocalization.
The monkey species surveyed were: rhesus macaques, pig-tailed macaques,
stump-tailed macaques, Cebus (Apella) monkeys, Squirrel monkeys, Wooly monkeys,
and Lemurs. The following summarizes some of the characterizations of basic
behavioral tendencies for each species.
TABLE 1.1
Some Common Species of Non-human Primates
Prosimians
Tree Shrews
Tupaia glis
Galagos
Galago spp.
New World Monkeys
Capuchins Cebus spp.
Marmosets Callithris spp.
Owl Monkeys Aotus t rivirgatus
Spider Monkeys Ateles spp.
Squirrel monkeys Saimiri sciureus
Tamarins Saguinus spp.
Wooly monkeys Lagothrix spp.
African Monkeys
Guenons Ceropithecus spp.
Green monkeys Ceropithecus aethiops
Mangabeys Cerocebus spp.
Patas monkeys Erythrocebus patas
Talapoins Miopitecus talapoin
Baboons
Savanna baboons Papio cynocephalus
Hamadryad baboons Papio hamadryas
Asian Macaques
Celebes black ape Macaca nigra
Cynomolgus (Crab eating macaque) Macaca fasicularis
Japanese macaque Macaca fuscata

Pig-tailed macaque Macaca nemistrina
Rhesus Macaca mulatta
Stump-tailed macaque Macaca speciosa (a.k.a. arctoides)
Apes
Chimpanzees Pan troglotydes
Gibbons Hylobates spp.
Choice of Animal Subjects in Behavioral Analysis 15
The rhesus monkeys (Macaca mulatta) were described as curious and manipulative.
Rhesus spent more time in visual survey than either the stump-tailed or the
pig-tailed macaques. Like the other macaques, rhesus monkeys also spent more time
in presenting and grooming than was observed in the New World monkeys. Rhesus
monkeys are also highly manipulative. They spend considerable time licking and
examining objects with their hands.
Aside from the findings of Davis et al., rhesus monkeys are popular research
subjects due to their relatively low cost and high cognitive skill; however, these are
large, powerful, and potentially dangerous animals, and require special caution in
their handling. In the wild, the defensive strategy of the rhesus troupe is to form a
line of aggressive male defenders, who menace and threaten intruders. This defensive
line allows the females and young to evacuate to safer territory. Everyone who visits
a zoo or non-human primate housing area is confronted with the strong defensive
and attack behavior that is elicited by eye contact with rhesus monkeys. This
aggressive cage-rattling behavior in captive rhesus monkeys is a manifestation of
the wild defensive posture.
In the Davis et al. studies, the stump-tailed macaques (Macaca speciosa) spent
most of their time presenting and grooming. This preoccupation with presenting and
grooming by the stump-tailed monkeys resulted in lower scores in the remaining
behavioral categories.
The reports of Kling and Orbach 51 and Orbach and Kling52 regarding the docility
of stump-tailed macaques initiated great interest in this species for laboratory use.
Juvenile stump-tails are considerably more compliant than are juvenile rhesus monkeys
(Macaca mulatta). Larger stump-tailed macaques are noticeably more aggressive
than the stump-tailed juveniles, but still considerably less difficult to handle
than are rhesus monkeys. The defensive strategy of wild stump-tailed monkey
troupes is in contrast to that of rhesus monkeys. When threatened by intruders, wild
stump-tailed monkeys remain quiet and still and thus blend into the heavy foliage
cover that is characteristic of their normal habitat. This stillness in the face of
potential threat is the basis of the once prevalent idea that these monkeys are
extremely docile; however, quietude and docility are not the same. Adult stump-tails
become quite large and are known to become gradually more aggressive toward
human handlers.53 They are particularly dangerous if perched at eye level or above.
One characteristic of stump-tailed macaques that frequently elicits comment is their
musty body odor, but it is not so strong as to eliminate their use. An ethological
study of stump-tailed macaque gestures and facial expression has been presented by
Jones and Trollope.54
Pigtailed macaques (Macaca nemestrina) were intermediate to rhesus monkeys
and stump-tailed macaques in the Davis et al. tests. The only exception to this was
vocalization, which was the same as the other macaque species. Pig-tails become
very large and, as with stump-tailed macaques, may gradually become quite aggressive
toward their handlers. Our aged non-human primate project at the Medical
College of Georgia has been forced to abandon several well-trained pig-tails because
of extreme aggressiveness toward human handlers.
16 Methods of Behavior Analysis in Neuroscience
Cynomolgus or Crab-eating macaques (Macaca fasicularis) have become more
popular as biomedical research subjects, as imported rhesus monkeys are difficult
to obtain. These animals have very long tails. Their behavior and general cognitive
skill resembles that of rhesus monkeys.

Apella monkeys (Cebus Apella) are the mechanics of the monkey world. They
spend a large amount of time manipulating inanimate objects with their hands and
tongues. Like stump-tailed and pig-tailed macaques, the Davis et al. studies found
that they spent little time in visual survey, and, as other New World monkeys, they
spent large amounts of time pacing, swinging, and bouncing on all four legs. They
spent little time grooming and presenting.
Squirrel monkeys (Saimiri sciurea) were characterized by high levels of selfinvolved
behavior (licking, biting, manipulation of their own body) in the Davis et
al. tests. Also, many of the squirrel monkeys exhibited high levels of bouncing and
pacing, but this was an individual characteristic that was not exhibited by all squirrel
monkeys. Squirrel monkeys have a unique form of vocalization, characterized by
frequent high-pitched shrieks and less frequent cooing.
Wooly monkeys (Lagothrix humbolti) are known for frequent vocalizations, but
in the Davis et al. tests, this characterization varied considerably among individuals.
Wooly monkeys will frequently threaten and scold, but are not known for their
aggression toward handlers. Only rarely do they scratch or bite.
Chimpanzees (Pan troglotydes) are rare and special animals. The endangerment
of these animals as a species, in conjunction with the great cost of housing and caring
for them, results in few opportunities to work with these great apes. Most research
personnel that work with these animals are profoundly impressed by their similarity
to man, and many become ardent opponents of invasive research techniques to be
applied to them. One source of information about the anatomical, physiological, and
behavioral study of chimpanzees is provided by a volume edited by H. H. Reynolds. 55
Chimpanzees and other great apes, such as gorillas and orangutans, are generally
superior to both Old and New World monkeys in learning skills, if the learning task
is sufficiently difficult to distinguish the species. 56 The fact that chimpanzees can master
rudimentary language skills has aroused great interest. 57
E. Primate Cognitive Skills
An important assumption that underlies most non-human primate research is that it
is possible to align primate species in a graduated series of cognitive skill, so as to
increasingly approximate humans.58,45 Without entering into argument about the
theory of linear evolution, most would agree that the behaviors of great apes, such
as gorillas, chimpanzees, and orangutans, resemble human behavior more than do
the behaviors of other primates, such as lemurs or marmosets. One factor that
correlates highly with human similarity is development of the cerebral cortex. Among
the primates, cortical development is greatest in humans, followed in order by the
great apes, the lesser apes (e.g., gibbons) Old World monkeys, New World monkeys,
and finally the pro-simians.59,45
F. Transfer of Training
There is considerable difference among various non-human primate species in cognitive
capability. One cognitive skill that has been extensively studied across species
is transfer of training. Humans and animals have the ability to generalize aspects of
what has been learned in one situation to a different situation. Likewise, animals
and humans improve in generalization skills as a result of experience. One relevant
body of literature began with the classic series of experiments by Harlow. 60 This
work was continued and expanded by his many colleagues and students. Harlow
demonstrated that, after repeated experience with discrimination problems, animals
learned strategies for solving problems that allowed them to become more proficient
at solving new problems that were similar, but not exactly the same. Harlow called
this phenomenon “learning set” or “learning-how-to-learn.” There is a distinct correlation
between cerebral development and the ability to establish learning sets; thus,
the great apes are more proficient at forming learning sets than are the prosimians.
Figure 1.1 compares learning set formation in three species of primates: Old World
rhesus monkeys, New World squirrel monkeys, and marmosets. Chimpanzees are

even more adept at learning set formation than are rhesus monkeys. These data and
the collective body of information 45 clearly support the hypothesis that, at least
within the order Primates, there is a relationship between cortical development and
complex learning skill. The generalizations exemplified by learning set formation
are foundational to the conceptualizations shown by human children and adults.
Learning set formation is only one example of cognitive processing that can discriminate
between species of primates. The ability to abandon previously learned,
but no longer appropriate behavior (extinction), is another behavioral characteristic
that varies naturally throughout the order Primates.
It is normal and adaptive to extinguish a response pattern when it is no longer
rewarded. The ability to perform this basic cognitive skill also distinguishes the
various primates. Since cortical development correlates with the cognitive skills
involved with learning sets, it would be expected that more developed animals would
be able to more rapidly extinguish non-rewarded response patterns as well. This was
confirmed in a study comparing highly developed Cercopithecus monkeys (includes
African green monkeys and mangabeys) with less developed Lemuridae (lemurs)
(Arnold and Rumbaugh, 1971). As seen in Figure 1.2, Cercopithecus monkeys
abandoned the non-rewarded choice patterns more rapidly than did the Lemuridae,
although the overall pattern of extinction was the same for both groups. This is only
one example of the correlation between cognitive skill and cortical development
with the order Primates.
Other behavioral test techniques also show this positive correlation between
cortical development and cognitive skill among primate species. Rumbaugh 61,57 has
developed a productive technique termed the Transfer Index (TI), which reflects
transfer of training skill and the ability to abandon previously learned object choices.
A typical procedure for the TI is to first train animals at two different criterion levels
of accuracy on a discrimination task. Then, the cue valences of the stimulus objects
are reversed. Thus, selection of the object that previously was rewarded with food
resulted only in an empty food well and the reward was shifted to the previously
incorrect object. The TI measures the percent accuracy on the reversal task (transfer).
Transfer skills are measured in terms relative to amount of learning accuracy on the
first discrimination task — not in terms of absolute percent accuracy. The TI was
designed to be a species-fair measurement of cognitive functioning. The implication
of the TI is that it allows a measurement of behavioral and cognitive flexibility. The
ability to transfer small amounts of knowledge to a new situation can be an important
FIGURE 1.1
Learning set curves for marmosets, squirrel monkeys, and rhesus monkeys.
FIGURE 1.2
Reversal performance plotted as a function of pre-reversal performance.
Rhesus Monkeys Squirrel Monkeys Marmosets
% Correct
Block of Problems
100
80
60
40
20
0
0 1-200 201-
400
600 801-
1000
Lemur Cerophitecus

Reversal % Correct
Training Sessions in Blocks
80
60
40
20
0
1234567
advantage. Conversely, a negative transfer indicates an inflexibility that represents
a disadvantage. Figure 1.3 compares transfer index of several non-human primates.
There is a high correlation between the TI and cortical development. Primitive
primates with less developed brains, such as lemurs, micro-cebus, and phaner, show
negative transfer. Negative transfer indicates a behavioral inflexibility that is
a handicap for the adaptation of these species. Macaques are intermediate to the
New World monkeys and chimpanzees as measured by the TI. This cognitive flexibility
of the Old World macaques facilitates adaptation to their environment, and
the overall cognitive skill of the macaques accounts for their popularity as behavioral
research subjects.
XII. Discussion
There are many variables to consider when selecting animals for behavioral experiments.
Not only are there important species differences, but there is considerable
variance among different individuals — even within the same lines. Although variance
requires large subject populations for behavioral studies, these individual variances
are not altogether without benefit. For example, the range of response to some
treatments may be manifest most broadly in outbred groups. One example of this
point is that there are numerous strains of laboratory rats available. 64,65,67 These
strains arose from different stocks and have different behavioral characteristics.
Importantly, these strains react differently to many common physiologic and pharmacologic
treatments. Thus, many important research strategies should involve
FIGURE 1.3
Transfer index of several non-human primate species.
Chimpanzees
Macaques
Vervets & Squirrel Monkeys
Cebus Monkeys
Lemurs
Phaner, Microcebus, &Talapoins
-25 -20 -15 -10 -5 0 5 10 15
careful selection of the animal strain or species that is to be selected, and perhaps
more consideration should be given to the replication of research findings across
multiple species or strains. The classic standardization studies of Harrington 29,30-39
and Harrington and Hellwig40,41 remain as immensely valuable standards by which
to select rats for behavioral studies.
One practical implication of the variation among species and strains is that
it is important for experimenters to personally spend some time observing the
behavior of various species, both in the home cage and in the experimental test
situation. Another practical implication of the variation among species and strains
is that various populations may exhibit high or low values of a behavioral factor
of interest. This has continued to be a valuable behavioral tool. There is important
interaction between many treatments and species-specific behavior. Because the
treatment is the usual focus of interest, it is tempting to accept experimental data
values blindly as they are presented by an automated device or research technician.
In such situations, the behavior is often viewed only as a dependent variable,
and thus of lesser interest. Much valuable information can be lost by such lack

of attention to the actual behavior of the animal. When the experimenter in charge
spends some time observing the animals as they perform their particular task,
the effort is often rewarded by increased insight into the behavioral intent of the
animal. If we pay attention, they will teach us much and, as a bonus, we will
learn much more about our treatments. Remember that animals can perform
certain tasks for many different reasons — not only the ones originally construed
by human planning.
Various animal species are uniquely endowed with characteristics that distinguish
them from other plausible animal subjects. Animal species or strains often
have specialized capabilities that allow them to cope with a narrow environment.
Thus, so-called lower animals may have certain capabilities that exceed the ability
of higher animals. A classic example of this can be seen in the olfactory functions
of the amygdala and other limbic structures of macrosmatic species, such as rats,
vs. the emotional/learning specialization of the same structures in microsmatic
animals, such as primates. There are important differences among the behaviors of
the species and different reactions to certain treatments are to be expected.
An interesting question that arises in this connection is whether phylogeny
recapitulates the ontogeny of human development with respect to higher cognitive
function.61 Certainly, we are familiar with the idea that animals with greater
cortical development have capacities that allow new functions to emerge. 62,63
Some highly developed primate species may have capabilities that are not represented
in animals with less developed brains. Rumbaugh et al. 61 have presented
convincing arguments that “emergents,” which are new capabilities that were
never directly rewarded by past experimental experience, are cognitive products
of highly developed cortical structure. Transfer of training is an example of
emergent behavior.66 Emergents resemble human concept formation, and are
present in many species, but are striking in primates, such as macaques, and
especially striking in the great apes.
The selection of animal subjects is an immense topic, and additional information
is available from many sources. Keying a search on any of the major search engines
of the world wide web will identify animal suppliers and additional clearinghouses
of information. Such key words as laboratory rat, laboratory mouse, primate, or
mouse behavior will prove productive. The NIH-funded multicenter effort for comparison
of mouse strains through the Psychology Department at SUNY at Albany
should be a good source of up-to-date mouse behavioral data. Two publications of
interest are Laboratory Primate Newsletter and Rat News Letter. The Oregon
Regional Primate Center provides a valuable collection of primate-related journal
citations, which are available for a modest fee. Many of these citations have been
organized into databases of specific behavioral topics. The reference section of this
chapter contains citations of papers written during the infancy of animal behavior
research. These papers are rich in the justification of animals as models for human
cognition and are worthy of rediscovery.
There are many reasons for conducting research upon animal subjects, but the
decision to do so should be given some thought. Species selection is among the first
problems to confront the researcher who hopes to pursue animal modeling of human
behavior. Often, new researchers may select a model that has been commonly
utilized, simply because it has become established and accepted. Although there is
utilitarian value to this approach, it is valuable to recall that there is considerable
difference between and among species and strains. Many opportunities are presented
by these differences.

U

More Related Content

U