Functional and anatomical development of medial prefrontal cortex in the Syrian hamster


Material Information

Functional and anatomical development of medial prefrontal cortex in the Syrian hamster
Physical Description:
x, 123 leaves : ill. ; 29 cm.
Crandall, James E., 1953-
Publication Date:


Subjects / Keywords:
Behavior, Animal   ( mesh )
Brain -- anatomy & histology   ( mesh )
Hamsters -- anatomy & histology   ( mesh )
Neuroscience Thesis Ph.D   ( mesh )
Dissertations, Academic -- neuroscience -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1980.
Bibliography: leaves 113-122.
Statement of Responsibility:
by James E. Crandall.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000891187
oclc - 24896136
notis - AEJ9633
System ID:

This item is only available as the following downloads:

Full Text







For my special scout

Every new adjustment arises out of the total of the
individual's experience as registered in structural
counterparts by means of the correlation of growth and
excitation in the developing neurones.

Coghill, 1929
"Anatomy and the problem of behavior"


I would like to thank the members of my supervisory

committee, Drs. Carl Feldherr, Marieta Heaton, Christiana

Leonard, William Luttge, Charles Vierck and Stephen

Zornetzer, for their interest and advice throughout the

course of this dissertation.

Special thanks are well deserved by the following

people for their special help during the last six years in

assisting with skill and cheer--Joan, Mike, Bob and Mark;

sharing lab life--Marjorie, Denise, Tim, Tom, Jim and

Dottye; listening and consoling--Joe, Jan, Dick and Mary

Margaret; teaching the language of computers--Jim Fleshman;

and encouraging my fledgling artwork--Carla Lehnkey.

I would especially like to mention the constant support

given freely at all times by Cliff Abraham. Thank you, Wick,

for introducing me to one of the finer arts for staying sane


My deepest gratitude and appreciation goes to my

adviser, Christiana Leonard. I know, Tiana, all your

patience and wisdom and effort has provided me with many

meaningful experiences. Thank you for not giving up on a

writer such as myself.

I have been supported by a Graduate School Fellowship

for the first year, an individual NSF predoctoral fellowship

for three years and a traineeship from the Center for

Neurobiological Sciences for the remaining two years. This

research was supported by NIH grant NS 13516.



ACKNOWLEDGEMENTS....................................... iv

ABSTRACT.................................. ............. ....... vii

ONE GENERAL INTRODUCTION...........................1

IN THE GOLDEN HAMSTER....................... 12

Methods................ ................... 13
Results................................... 17
Discussion........ ........... ... ........... 22

IN THE GOLDEN HAMSTER.......................27

Introduction............................. 27
Results................. .................. 33
Discussion.. .............................52


Introduction.............................. .58
Methods................... ................ .60


FIVE GENERAL CONCLUSIONS..........................109

REFERENCES.................................. .......... 13

BIOGRAPHICAL SKETCH...................................123

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



James E. Crandall

December 1980

Chairman: Christiana M. Leonard, Ph.D.
Major Department: Department of Neuroscience

The maturation of behavior, the ultimate, observable

endproduct of central nervous system activities, is often

assumed to be dependent upon the structural development of

the central nervous system. Substantiation of this tenet

would require the identification of specific alterations in

developing neuroanatomical morphology that correlate in time

with specific changes in developing behaviors of the

maturing organism. In this combined

neuroanatomical-behavioral study, development of a complex,

species-specific behavior, hoarding, and potential


underlying structural changes in medial prefrontal cortex

(MPFC) connections were examined in the Syrian (golden)

hamster (Mesocricetus auratus). The ontogeny of hoarding

behavior, which is dependent on the integrity of MPFC in the

adult, was initially determined to provide the basis for

selecting critical ages for the anatomical studies. Pups

were weaned and tested at different ages for food pellet

hoarding behavior using a modified in-cage method. Hoarding

was first seen on postnatal Day 17. The incidence of

hoarding doubled while the consistency of hoarding

quadrupled between Days 21 and 25.

To assay the development of MPFC afferents, horseradish

peroxidase (HRP) was injected into MPFC in Day 10, 21 and 25

hamsters. Labeled neurons were located in all of the brain

regions in which labeled neurons were found in the adult. A

substantial developmental change, however, occurred in the

pattern and quantity of mediodorsal (MD) thalamic labeling

between 21 and 25 days. Following small HRP injections on

Day 21, a large number of labeled neurons were found in MD.

In contrast, small injections at Day 25 produced few labeled

cells in this nucleus. After large injections, many MD

neurons were labeled at both ages. It appears that the

maturation of consistent hoarding behavior correlates with a

reduction in the size of the prefrontal cortical region to

which individual MD neurons project.


Surprisingly, a projection to prefrontal cortex from

"pararhinal" cortex was discovered and confirmed with

anterograde degeneration methods. Layer V neurons of this

periallocortical region were located in a narrow patch above

and below the rhinal sulcus from rostral thalamic to caudal

midbrain levels. The chief cytoarchitectonic features were

the absence of a distinctive band of layer II fibers and the

absence of a distinctive granular layer IV.

The development of MPFC efferents was studied using

anterograde degeneration methods after MPFC lesions at the

three ages used for the HRP studies. The hypothesis was

tested that the onset of long-lasting degeneration

argyrophilia (LLDA) would be a neuroanatomical marker of

functional maturity. In animals matched for lesion extent

and location, there were no differences between LLDA

patterns of MPFC efferents at Day 21 and Day 25. Short

survival degeneration argyrophilia was not evident at Day

21, but is clearly present on Day 10. On the other hand, no

LLDA was found on Day 10.

It appears that the occurrence of LLDA in MPFC

efferents does not correspond with the ontogeny of reliable

hoarding behavior. Dramatic changes, however, in the

afferents from the mediodorsal thalamic nucleus to HPFC do

correlate with the behavioral transition. Whether the

correlation between changes in developing afferents and the

emergence of behavior is coincidental or causal remains to

be established. The developmental organization of MPFC

afferents is hypothesized to involve sequential periods of

sparse but widespread terminal arborization, profuse

terminal proliferation and selective terminal segregation by

the growing axons. The findings that the efferent

projections of a brain region are established before its

afferentation is complete support a theory of retrograde

development of the neuron.


The frontal cortex attains its pinnacle of development

in the human brain. The great increase in the size of this

region in humans as compared to other mammals has stimulated

much speculation on its functional significance. Like most

studies of human brain function, the data base for

speculation arises from neurological and neuropsychological

investigations of patients with frontal cortex damage. Among

the complicated variety of symptoms characterizing the

frontal syndrome in man are dramatic mood and personality

changes, perseverative tendencies (Luria, 1973) and

perceptual impairments (Teuber, 1972), as well as specific

deficits on neuropsychological tests involving card-sorting

and maze-solving (Milner, 1964). From the pattern of

behavioral pathologies, various hypotheses have been

constructed concerning the function of the normal frontal

cortex. Teuber (1972) suggests that the frontal lobes

provide corollary discharges which preset sensory systems

for anticipated consequences of simultaneously initiated

motor patterns. Nauta (1971) characterizes the major action

of frontal cortex as an organization of sensory and effector

processes involved in perceptual mechanisms and behavioral

programming. Similarly, Lucia (1973) maintains that the

importance of the frontal lobe is manifest not as much in

elementary receptor processes and motor acts as in the

complex forms of self-regulating activity. Thus, the major

general functions of the frontal lobes would appear to be

coordinating, executing, and monitoring behavior.

Comparative neuropsychological investigations of Rhesus

monkeys with frontal cortical damage have basically

supported the findings in humans. The monkeys show a general

"response guidance" deficit in tasks such as delayed

alternation, delayed response, "go-no go" and object

discrimination (Warren and Akert, 1964). Social interactions

(threat gestures, facial expressions and vocalizations) are

decreased in monkeys with frontal cortex damage (Nyers,

1972). A decreased aggressive reaction and an increased

aversive reaction to socially threatening situations also

occur in these animals (Butter and Snyder, 1972).

In spite of the detailed, controlled behavioral data on

frontal lobe-damaged primates, neuroanatomical and

neurophysiological progress in unravelling correlates of

frontal lobe function has been painstakingly slow due to the

complications of human and primate experimentation. The

frontal cortex of rodents is more accessible both

technically and practically and could provide insights into

primate frontal cortex function, given the assumption that

homologous areas have have homologous functions. Lesions of

the frontal cortex of rats primarily result in the following

deficits: learning tasks such as delayed alternation

(Wikmark, Divac and Weiss, 1973) and delayed response (Kolb,

Nonneman and Singh, 1974); activity (Kolb, 1974a); social

behaviors--low physical contact and increased aversive

reactions (Kolb, 1974c); as well as complex species-specific

behaviors such as regulation of ingestion patterns (Kolb,

1974b; Kolb and Nonneman, 1975) and food pellet hoarding

(Kolb, 1974c).

The pattern of behavioral deficit evident in hoarding

behavior after medial frontal cortex lesions in the Syrian

hamster is strikingly similar to the behavioral disorders of

human patients with frontal lobe dysfunction. Although the

hamsters were observed performing each of the different

motor acts (grabbing, carrying, pouching and stacking food

pellets) necessary for competent hoarding behavior, they did

not seem capable of chaining or sequencing the motor acts in

a competent fashion (Shipley and Kolb, 1977). Rather, "the

lesions made it difficult for the animals to engage in the

correct behavior at the correct time and place"(p.1064).

The controversy over whether hoarding is an instinct

(Morgan, 1947) or a learned response (Marx, 1950) has never

been adequately resolved. Nevertheless, hoarding behavior

may be regarded as a complex series of stereotyped motor

acts dependent on an intact neocortex (VanderWolf, Kolb and

Cooley, 1978). The hoarding disturbances observed in

neocortical decorticated adult rats were not primarily due

to a loss of specific movements. Walking, climbing, and

rearing, as well as the usual posture movements involved in

hoarding behavior, were not disrupted. Rather, the motor

acts did not occur with the normal order or frequency. The

first attempt at anatomical localization of a neocortical

region involved in hoarding resulted in unaltered or

enhanced hoarding behavior of rats with large cortical

lesions (Zubek, 1951). In contrast, smaller bilateral

cortical lesions along the medial longitudinal axis of the

hemispheres correlated significantly with a decrease in

hoarding activity in rats (Stamm, 1953). More recently,

lesions of the ventral mesencephalic tegmentum (source of

the dopaminergic input to medial frontal cortex in rodents)

have resulted in disruption of hoarding behavior to the

extent that the animal's behavior is described as completely

disorganized (Stinus, Gaffori, Simon and LeMoal, 1978).

The similar nature of behavioral deficits in primates

and rodents following frontal cortex damage is paralleled in

the neuroanatomical similarities between rodent and primate

frontal cortex. Whereas the frontal cortex of monkeys

contains a distinct granular layer 4, the homologous region

in the rat cortex lacks a distinct fourth layer (Leonard,

1969). In spite of cytoarchitectural differences, the

thalamic projection to frontal cortical regions has been

used to demonstrate that all mammals have a mediodorsal (MD)

thalamic projection region in prefrontal brain areas (Rose

and Woolsey, 1948). Using this criterion, Leonard (1969)

suggested that the prefrontal cortex of the rat was composed

of two spatially separate cortical regions: 1) a sulcal

prefrontal cortex that receives input from the central core

of MD, and 2) a medial prefrontal cortex receiving input

from the lateral surround. This orderly pattern of SD

projections to separate frontal cortical regions corresponds

to a bipartite MD projection pattern in the Rhesus monkey.

That is, orbital prefrontal cortex receives projections from

the medial region of MD and dorsolateral cortex receives

projections from the lateral region of MD (Pribram, Chow and

Semmes, 1953).

Medial prefrontal cortex of the rat sends projections

to the following structures: 1) forebrain regions

(caudate-putamen and substantial perforata anterior), 2)

thalamic nuclei (reticular, midline, anterior, ventromedial

and mediodorsal), 3) mesencephalic regions (subthalamus,

pretectal area, deep layers of superior colliculus, central

gray substance, lateral tegmentum and prerubral fields), and

4) pontine regions (interpeduncular nucleus, Bechterew's

nucleus and medial pontine areas). The projections of the

sulcal cortex include: 1) forebrain regions (ventral

caudate-putamen, olfactory tubercle, substantial innominata

and lateral hypothalamus), 2) thalamic nuclei (reticular,

midline, commissural, paracentral, reunions, ventromedial

and mediodorsal), and 3) mesencephalic areas (subthalamus,

central gray substance and ventrolateral tegmentum).

Recent work using anterograde transport of

radioactively-labeled amino acids has generally confirmed

the subcortical projections of rat prefrontal cortex, and

has demonstrated cortico-cortical projections from both

prefrontal regions (Beckstead, 1979). Medial prefrontal

cortex sends fibers to contralateral homotopic cortical

regions, cingulate cortex (granular and agranular

retroplenial cortex), hippocampal formation (dorsal

presubiculum and parasubiculum), lateral entorhinal cortex

and medial sulcal cortex. Sulcal prefrontal cortex sends

fibers to the entire length of the medial hemispheric

cortex, presubiculum, perirhinal and lateral entorhinal

cortex, as well as to its homotopic contralateral cortex.

Thus, the cortical connections of medial and sulcal

prefrontal cortex of the rat correspond fairly well with the

efferents of dorsolateral and orbital prefrontal cortex of

the monkey (Nauta, 1964). The major difference would seem

to be the massive nature of the cortico-cortical projections

to broadly expanded regions of temporal, cingulate and

parahippocampal cortices in the monkey.

Corresponding to the spatial dissociation in brain

regions of sulcal and medial prefrontal cortex in rats is

the demonstrable double dissociation of lesion-induced

behavioral changes with respect to the prefrontal region

involved. Deficits in delayed-response tasks, but not in

object discrimination reversal, are observed in monkeys with

dorsolateral prefrontal damage. The reverse pattern of

behavioral deficit is exhibited by monkeys with orbital

prefrontal damage. Similarly, the effects of medial or

sulcal prefrontal cortex lesions in rats may be doubly

dissociated with respect to activity (Kolb, 1974a), social

behavior (Kolb, 1974b), spatial learning (Kolb et al.,

1974), or regulation of ingestive processes (Kolb and

Nonneman, 1975).

Another resemblance between monkey and rodent

prefrontal cortex lesion-induced behavioral changes involves

the relationship of recovery of function to the age at which

the cortical damage occurs. Prefrontal cortex lesions in

young animals yield sparing or recovery of function, whereas

lesions in older animals result in permanent behavioral

deficit (Goldman, 1974). No deficits are observed in delayed

response or delayed alternation following prenatal (but not

postnatal) removal of presumptive dorsolateral prefrontal

cortex in monkeys (Goldman and Galkin, 1978). The

interpretation of these experiments are complicated by the

use of the lesion method and recovery of function problem.

Still, this work suggests that complex behaviors dependent

on intact prefrontal cortex develop over a considerable

postnatal period in both monkeys and rodents.

The considerable effort expended on the determination

of the functional capacities of animals without intact

prefrontal cortex stands in surprising contrast to the

amount of work concerned with the normal neurobehavioral

development of this cortical region. Recently, Goldman and

Alexander (1977), using reversible cryogenic depression of

monkey dorsolateral prefrontal cortex, demonstrated that a

functional maturation of the region in delayed response

tasks occurs at 34 to 36 months of age. One of several

possible functional indices of prefrontal cortex maturation

in rodents might be the normal development of hoarding

behavior. The age at which rodents first develop the

sequencing of motor acts necessary for hoarding behavior has

not been examined.

Since hoarding behavior in the adult hamster has been

demonstrated to be dependent on the integrity of medial

prefrontal cortex (Shipley and Kolb, 1977), the conjoint

study of the normal development of hoarding behavior and

neuroanatomical changes occurring in medial prefrontal

cortex provides an intriguing model to correlate normal

anatomical and behavioral ontogeny. One advantage of using

the hamster as an experimental subject is the regularity and

reliability of its mating and gestation periods. This would

allow the reliable determination of date of birth, which is

necessary for developmental studies. The major drawback of

using the hamster as a subject concerns the scarcity of

anatomical data available on any cortical region of the

hamster at any age. An additional benefit from the attempt

to correlate anatomical and behavioral maturation would be

comparisons of medial prefrontal cortex organization between

common rodent species, the hamster and the rat.

It would be ideal to correlate maturational changes in

the medial prefrontal cortex with the development of

hoarding behavior, if there were some clues to what types of

anatomical changes could be used, to narrow down the

plentitude of possibilities. The only anatomical study on

the development of monkey dorsolateral and orbital

prefrontal cortex used animals of 2, 6 and 24 months of age.

An age-dependent increase in the density of the projection

to the caudate nucleus with no age-dependent changes in the

thalamic or subcortical fiber projection paths was

demonstrated (Johnson, Rosvold, Galkin and Goldman, 1976).

Unfortunately, those fiber pathways which might be expected

to mature later and might be expected to underlie complex,

postnatally developing behaviors, i.e., associational and

commissural corticocortical fibers, were not investigated.

Neuroanatomical work on neocortical development in

rodents and primates has emphasized the prenatal or

perinatal periods and formative events such as cell

proliferation, migration, dendritic and axonal

differentiation and synaptogenesis (Berry, 1974). The sparse

postnatal data available on cortical development in rodents

deals mainly with the primary somatosensory cortex. For

example, by the fourth postnatal day, the thalamocortical

fiber distribution to the somatosensory cortex is similar in

pattern to the adult, while the density of the commissural

projection increases throughout the first postnatal week

(Wise and Jones, 1978).

The somatosensory cortical region in the rat attains

its adult-like characteristics early in development. From a

functional viewpoint, it might be expected to be among the

first of cortical regions to mature, in order to process the

quantity of tactile stimulation that the infant rat

receives. The neuroanatomical development of associative

cortical regions, such as the prefrontal cortex, has not

been investigated. As the integrative and associative

abilities of the immature rodent pup develop, corresponding

maturational changes in prefrontal cortical structure might

be expected to occur.

So many behavioral and neuroanatomical changes occur in

prenatal and postnatal development that attempts to

understand the neural mechanisms underlying behavior may be

frustrated by the rapidity and complexity of changes, as

well as by the difficulties in experimentally assessing the

function and structure of the tiny and very vulnerable

developing brain. An appropriate alternative approach to

studying neuroanatomical correlates of behavioral

development would be to detect a later-developing change in

behavior and investigate its possible neuroanatomical basis

in a more stable system which is physically accessible for

easier experimental manipulation. Certainly, a study of the

ontogeny of hoarding behavior in the Syrian hamster,

together with a thorough investigation of the

neuroanatomical organization of medial prefrontal cortex,

would provide an advantageous system to identify the


specific neuroanatomical changes which may underly the

development of a complex species-specific behavior.



Food storing or hoarding behavior in rodents was

intensively investigated by comparative psychologists in the

1940's and 1950's. As initially described by Wolfe (1939),

hoarding consists of a series of behaviors. The animal

emerges from its home to repeatedly secure food pellets to

some depositing site without eating the food. Many

experimental variables such as previous food experience

(Stellar and Morgan, 1943), deprivation state (Licklider and

Licklider, 1950), cage size (Viek and Miller, 1939),

"frustration" (McCord, 1941) and "shyness and security"

(Bindra, 1948) have been hypothesized to have effects on

hoarding behavior (for review, see Ross, Smith and Woessner,

1955). Only the experimental study of Porter, Webster and

Licklider (1951) suggests any developmental trend in

hoarding. Thirty-minute tests of seven separate age groups

ranging from 26 to 318 days of age revealed that older rats

hoarded more pellets. The quantity of pellets hoarded was

related to age in an increasing but negatively accelerated

fashion. The occurrence of food storing behavior in the

golden hamster develops in a natural habitat between 30 and

40 days of age (Dieterlen, 1959). However, at least a week

earlier, the stereotyped individual motor acts which

comprise the complex behavior of hoarding (grabbing,

carrying, pouching and stacking pellets) are evident.

The dependent variable in typical hoarding experiments

is the number of pellets hoarded in a specific period of

time, usually 30 minutes. In a developmental study, the

interpretation of this type of test could be complicated by

maturational improvements in muscle strength and stamina.

Thus, the classic hoarding test requires alterations in

order to assess adequately the emergence of hoarding

behavior. The present study examines the normative

development of hoarding behavior in the golden hamster using

a modified in-cage test procedure which circumvents these

age related variables.



Fifty-three golden hamster (Mesocricetus auratus) pups

whose mothers were bred at Lakeview Breeding Laboatories

(Cambridge, Massachusetts) were the subjects. All animals

were maintained on a 12 hour light-dark cycle, with the

lights off from 0700 to 1900 hours. Pups and their mothers

received water ad libitum and standard Purina laboratory

food pellets, with bits of carrot and apple. On the day of

weaning, pups were individually housed in polyethylene cages

(45 X 26 X 15 cm). Each cage contained a one liter volume

of fresh mixed hardwood shavings that were not changed until

the end of testing. Pilot experiments indicated that a

greater volume of shavings resulted in a decreased

percentage of pellets hoarded. The supply of food pellets

was not regulated until hoarding tasting was initiated.

Water, sunflower seeds, and apple bits were available daily.

Treatment Groups

Four litters were weaned on Day 20 (day of birth = Day

0). Hoarding tests for two of these litters (n=21) began

immediately (Group WT). The other two litters (n=19) which

were weaned on Day 20 began tests later, on Day 24 (Group

WLT). Both WT and WLT pups were evaluated for ten

consecutive days. Since several hamsters hoarded when first

tested after the typical weaning age (Day 20), a group of

three litters of pups (n=13) was weaned at an earlier age,

Day 16. Hoarding tests began immediately after weaning for

these pups (Group EWET) and continued for 14 consecutive


Test Procedures

In order to minimize age-related differences in

musculoskeletal maturation, an in-cage test was developed.

Latency, defined as the time from the start of the test

until the animal began to hoard, as well as time to complete

hoarding, varied considerably from animal to animal and from

day to day for each animal. Preliminary work indicated that

24 hour food deprivation did not enhance hoarding as

expected (Stellar and Morgan, 1943), but resulted in a

decreased quantity of pellets hoarded by the adults. The

most consistent data were collected when there was unlimited

access to food. Other experimental work on weanling pups

revealed a preference for hoarding smaller pellets over

larger pellets when both pellet sizes were present in equal

numbers. The presence of the home environment in the testing

situation has been shown to lessen the deficiencies in

learning tasks of immature rats (Smith and Spear, 1978). The

in-cage procedure used to evaluate hoarding behavior in

hamsters in this experiment differs from the classic

hoarding test in the following respects: 1) allowing

continuous access to food pellets, thus eliminating the

arbitrary time limit; 2) eliminating extra-cage alleys; 3)

measuring the percentage of pellets rather than the absolute

number of pellets because a variable number remained at the

the end of testing due to consumption and/or breakage of the

pellets; 4) adjusting the size of the pellets for the age of

the animal. Twelve food pellets, weighing 4-6 gm each, were

scattered in the center of the floor of the home cage. The

pup had continuous access to the pellets. From Day 16 to

Day 26 smaller pellets, approximately one-third the size of

the standard pellet, were used. The pattern of food pellet

distribution in the cage was evaluated daily during the dark

cycle between 0800 and 1100 hours. The number of food

pellets, the location of the pellets on the cage floor and

the spatial arrangement of the pellets were recorded.

Hoarding was scored as having occurred when more than one

pellet had been rearranged such that pellets were tightly

grouped in close clusters in a restricted area of the cage,

usually in one of the four cage corners. A line of adjacent

pellets along a cage side was also scored as a hoard. The

food pellets were collected and a replenished supply was

returned to the center of the cage each day. Consistent

hoarding was defined as hoarding which occurred daily

without failure throughout the remaining days of testing. A

pup was classified as a consistent hoarder on the day in

which this errorless trend first started.

Data Analysis

Differences in percentages of animals hoarding and

percentages of consistent hoarders between any two ages

within a treatment group were statistically evaluated using

the nonparametric McNemar test or the Cochran Q test for

related samples, depending on the size of the groups.

Differences between treatment groups were statistically

assessed using the Chi-Square test or the Fisher test, if

the expected frequencies were less than five (Siegel, 1956).

The mean percentage of pellets hoarded by different groups

was compared using the Student's t test.


Hoarding behavior was first observed on Day 17, with

three out of fifty-three pups hoarding on this day. By the

typical weaning age, Day 20, two-thirds of the EWET pups had

hoarded at least once. All of the EWET and WT pups and all

but two of the WLT pups exhibited hoarding behavior by Day

25. Only one pup failed to hoard throughout the testing

period (through Day 33). This pup also failed to hoard when

tested at five months of age. No sex differences in the age

of initial hoarding were observed. Litters within a test

group did not differ in the average age of initial hoarding.

In the early period of hoarding behavior development,

20 day old pups explored the cage in rapid, jerky bursts of

movements. Pups usually gathered and carried pellets one at

a time. Seldom did the pups pouch an individual pellet and

immediately unload it upon arrival at a storage corner. The

series of motor patterns appeared easily disrupted, and

frequently the pups did not carry out the procuring of all

of the pellets at one time. The older pups (Day 25) hoarded

more deliberately in a routine, repetitive fashion, and they

initiated hoarding after a shorter latency. Frequently the

pups moved the pellets in a continuous stereotyped fashion.

The older pups stacked the pellets in a more compact manner.

The percentage of hamsters displaying hoarding behavior

increased sharply on every day until Day 25 and remained at

asymptotic levels throughout the remainder of testing

(Figure 1). Whereas only 42% of the EWET pups hoarded on Day

21, 92% hoarded on Day 25. Similarly, only 47% of the WT

pups hoarded on Day 21, compared to 97% on Day 25. The

percentage of pups hoarding in both EWET and WT groups

increased significantly between Days 21 and 25 [EWET: Q=6,

p<.02; WT: Q=10, p<.01]. Equivalent percentages of EWET and

WT pups hoarded on Day 21 [XI=4.75, p>.10]. Likewise,

similar percentages of pups in the three groups hoarded on

Day 25 [Fisher test, p>.05].

In contrast to the increase in percentage of animals

hoarding between Days 21 and 25, the percentage of available

pellets hoarded by pups displaying hoarding behavior did not

increase with age after the first day (Table 1). The overall

mean percentages of pellets hoarded were 85.9, 84.9 and 86.8

for EWET, WT and WLT hoarders, respectively.

Just as there were no appreciable differences in

quantity of pellets hoarded, there were no detectable trends

in the location of individual animals' hoards within their

home cages. Pups from all groups arranged the majority of

hoarding piles in a corner of the cage; the specific corner

varied some days and on other days remained the same as the

wean test
day day (N)

0o- 16
e-- 20

16 (12)
20 (21)
24 (19)

16 18 20 22 24 26 28 30 32

AGE (days)
Figure 1. The ontogeny of hoarding behavior for three groups
of hamster pups weaned and tested at different ages. Open
circles represent group EWET, closed circles represent group
ST and closed squares represent group WLT.




Table 1. Mean and the standard error percent of pellets
hoarded by hamsters of different weaning and testing ages.



17 70.3 16.0

18 87.2 = 4.7

19 82.5 + 12.5

20 95.6 4.4 64.8 7.0

21 96.6 3.4 81.2 + 6.7

22 86.4 6.6 81.9 + 5.2

23 83.4 5.9 92.3 3.6

24 82.3 7.3 76.4 5.3 32.3 4.1

25 74.8 7.6 81.6 3.9 38.0 3.7

26 89.3 4.7 80.2 4.7 35.4 4.4

27 90.1 = 4.5 84.7 4.8 93.1 4.3

28 87.1 3.7 87.3 4.6 37.6 4.2

29 90.9 3.8 87.5 4.9 35.0 4.8

30 35.7 4.1

31 93.5 3.6

32 32.7 + 5.2

33 34.9 4.9


r, 60


L 20-



16 18 20 22 24 26 28 30 32

AGE (days)

Figure 2. The ontogeny of consistent hoarding with groups
combined across weaning and testing ages.

previous day. Occasionally, the pellets were stacked in two

separate piles or along the side of the cage.

The percentage of consistent hoarders increased

dramatically between Day 21 and Day 25 (Figure 2). Only 18%

of the animals which hoarded on Day 21 continued to hoard

throughout the remainder of testing. On the other hand 80%

of the animals hoarding on Day 25 hoarded every day

thereafter until the end of testing. This sharp rise in

hoarding consistency was statistically significant for both

the EWET pups [Q=8, p<.01] and WT pups [Q=12, p<.01]. The

percentage of consistent hoarders among the EWET pups did

not differ from the WT pups on either Day 21 [Fisher test,

p>.05] or Day 25 [Fisher test, p>.05]. Likewise, the

percentage of WLT consistent hoarders compared similarly to

the other two groups on Day 25 [Fisher test, p>.10]. The

only discernable effect of early weaning and test experience

is that the EWET pups achieved consistent hoarding two days

earlier. The percentage of EWET consistent hoarders on Day

23 was significantly greater than the percentage of WT

consistent hoarders [X =4.89, df=1, p<.05].


Hamsters demonstrate well-established hoarding behavior

by postnatal day 25. When pups hoard, the percentage of food

pellets hoarded tends to be constant after the first day of

hoarding. The percentage of animals hoarding rises between

Days 21 and 25. This increase in hoarding occurs

irrespective of whether the animals were weaned at the

typical age (Day 20) and tested immediately or weaned four

days earlier and tested immediately. The primary effect of

early weaning and test experience is the age at which pups

become consistent hoarders. By Day 25, 28 of 33 pups in the

EWET and WT groups hoarded reliably. Pups weaned on Day 20

and tested later (Day 24), group WLT, show similar hoarding

incidence and consistency on Day 25. This indicates test

experience is not as important a factor in the development

of hoarding in late-weaned animals as it is for those

animals weaned earlier. Of course, the mother may act as a

substitute for test experience through the example of her

hoarding actions in the home cage environment as her pups

mature. Experiments in which the type of pretest experience

is varied (e.g., Bevan and Grodsky, 1958) will be necessary

to further assess the effects of early weaning and

experience on the ontogeny of hoarding. Specifically,

animals raised in a home environment with powdered food

could be weaned early and tested for hoarding immediately

with the introduction of pellets or after a delay.

The simple test procedure employed in this

developmental study is substantially different from the type

classically used, and this needs to be taken into account

when comparing the present data with the literature on

rodent hoarding. The use of a standard number of small food

pellets along with in-cage testing and the removal of time

limitations in the present study lessens the contributions

of age-dependent physical factors such as muscular strength,

body weight and stamina. The present finding of constant

numbers of hoarded pellets differs from the previous report

of an increased number of pellets hoarded with increased age

(Porter et al., 1951). Their group of young rats (26-38 and

33-45 days of age) hoarded almost no pellets until deprived

of food; after 24 hour deprivation, they hoarded an average

of ten pellets in a 30 minute test. Of course, interspecies

comparison of hoarding behavior should be cautiously

interpreted. Adult hamsters hoard twice as many pellets as

rats, or four and one-half times per gram body weight

(Waddell, 1951). The present data indicating no sex

differences in the ontgeny of hoarding agree with data of

Koski (1963) on the adult hamster, but not the previous

report that adult female hamsters hoard better than males

(Smith and Ross, 1950). Perhaps a later maturational change

in hoarding behavior occurs around puberty, as implied by

the recent finding of an inverse correlation between

estrogen and progesterone levels and hoarding in estrus

female adult hamsters (Estep, Lanier and Dewsbury, 1978).

The present experiment demonstrated an earlier onset of

hoarding behavior than that found by Dieterlen (1959). In a

semi-natural environment in which hamster pups were not

weaned from their mothers, food storing behavior began

between 30 and 40 days of age. This difference may be due to

the different environmental demands in the two situations.

In one case the pup is solely responsible for food

procurement and in the other the family can provide

nutritive support. Thus, the hoarding behavior might be

already developed but dormant, awaiting the necessary

environmental situation to be expressed. Compared to other

motor behaviors such as sniffing, suckling, rooting and

pouching, hoarding is a relatively late-developing behavior.

It would be interesting to investigate further the ontogeny

of the different individual motor actions and to demonstrate

the sequence in which they are chained together.

Early indications of a possible neurological basis for

hoarding behavior were contradictory. Neocortical lesions in

adult rats increased (Zubek,1951) or decreased (Stamm, 1953)

hoarding postoperatively. In the present study the one

exceptional hamster that did not hoard even as an adult was

found to have damage to prefrontal cortex upon postmortem

examination of its brain. Damage to subcortical areas such

as the mediodorsal thalamus (Kolb, 1977) or the ventral

tegmental area of Tsai (Stinus, Gaffori, Simon and LeMoal,

1978) also disrupts hoarding behavior. Since both of these

regions project to the prefrontal cortex in the rodent

(Berger, Tassin, Blanc, Moyne and Thiery, 1974; Leonard,

1969; Krettek and Price, 1977a), it is not surprising

Shipley and Kolb (1977) found hoarding behavior to be

disputed after damage to the medial prefrontal cortex in

adult male hamsters. These adults were described as capable

of performing the different motor acts necessary for

competent hoarding behavior, but they appeared unable to

chain or sequence the motor acts. Thus, an intriguing

parallel can be drawn between the hoarding behavior of the

developing hamster pup and the adult hamster with a

bilateral medial prefrontal cortex lesion (as has been done

for feeding in rats with lateral hypothalamic lesions;

Teitelbaum, 1971). Both the brain-damaged adult and the

young weanling can perform the various motor skills that

constitute hoarding behavior (grabbing, carrying, pouching

and stacking food pellets). Neither animal, however, can

adequately chain the motor acts in a fashion necessary for

competent hoarding. Normal developmental changes in the

medial prefrontal cortex afferents from the mediodorsal

thalamus and the pararhinal cortex between Day 21 and Day 25

(Crandall and Leonard, 1979) may provide the neuroanatomical

substrate responsible for the emergence of hoarding. Further

studies which alter the developmental course of these two

important afferent systems could determine whether this

developmental correlation between connectional changes and

hoarding ontogeny is coincidental or causal.



The suggestion that the prefrontal cortex be considered

as the projection field of the mediodorsal nucleus of the

thalamus in mammals (Pose and Woolsey, 1948) was initially

based on retrograde degenerative changes in the mediodorsal

nucleus after prefrontal cortex removal. Subsequently, much

anatomical and behavioral work has been directed toward

identifying the homologue of this important cortical region

in primates and nonprimates, especially in rodents (Warren

and Akert, 1964; Konorski, Teuber and Zernicki, 1972;

Markowitsch and Pritzel, 1977). The findings of Leonard

(1969), based on anterograde degeneration methods,

established two mediodorsal-cortex projection regions in the

rat: the anterior medial wall of the hemisphere dorsal and

rostral to the genu of the corpus callosus (medial

prefrontal cortex), and the rostral depths of the dorsal

bank of the rhinal sulcus (sulcal prefrontal cortex).

Subcortical afferents to these prefrontal areas were similar

to those described in the primate. These connections of

medial prefrontal cortex have been largely confirmed and

elaborated with the more recently developed hodological

methods of autoradiography (Krettek and Price, 1977a;

Beckstead, 1979), catecholamine histofluorescence (Berger,

Thierry, Tassin and Moyne, 1976; Lindvall, Bjorklund and

Divac, 1978; Divac, Bjorklund, Lindvall and Passingham,

1978a) and the retrograde transport of horseradish

peroxidase (HRP) (Beckstead, 1976; Divac, Kosmal, Bjorklund

and Lindvall, 1978b).

The catalogue of similarities between primate and

rodent prefrontal cortex has been expanded by the inclusion

of the corticocortical associational efferents of prefrontal

cortex. Beckstead (1979) has demonstrated prefrontal

efferents to cingulate, entorhinal, perirhinal and

hippocampal cortices in the rat. No work has been reported

on the cortical afferent input to rodent prefrontal cortex.

During investigation of the development of afferents to the

medial prefrontal cortex in the golden hamster, horseradish

peroxidase techniques revealed labeled neurons in an

unexpected, poorly defined cortical area which we have

designated the "pararhinal cortex," mainly because of its

location deep to the middle and caudal extent of the rhinal

sulcus. In this report we describe the afferents to medial

prefrontal cortex in the golden hamster, emphasizing this

hitherto undescribed cortical region.


Horseradish Peroxidase Experiments

Thirteen golden hamsters (Mesocricetus auratus) between

30 and 92 days of age were anesthetized with intraperitoneal

injections of Chloropent (0.04 mg/kg) and placed in a

stereotaxic instrument. A small burrhole was drilled midway

between the coronal and frontal sutures adjacent to the

sagittal suture. Fine glass micropipettes (inside diameter=

30-60 ja) containing 30% (W/V) HRP (Sigma type VI) in

physiological salene were lowered 0.15-1.50 mm below the

exposed pia. Using an adjustable 2.0 ml Burett pipette

connected with polyethylene tubing to the micropipette, a

pressure injection of 0.01-0.05 4l was made over a period of

five to ten minutes. Following a waiting period of five

minutes, the micropipette was slowly removed, the skin

sutured and the animal was returned to its cage for 24 to 48

hours before sacrifice. Under ether anesthesia animals were

perfused intracardially with a brief rinse of physiological

saline, 100 al of fresh 1.25% glutaraldehyde-1%

paraformaldehyde in 0.18 phosphate buffer at pH 7.2,

followed by 100 ml of cool 30% sucrose in the same buffer.

Brains were removed immediately and stored overnight in the

sucrose-phosphate buffer at 40C. Coronal sections of 50 pm

were cut on the freezing microtome and were immediately

treated with the tetramethyl benzidene procedure of Mesulam

(1978). The concentration of hydrogen peroxide was adjusted

to minimize the amount of crystalline artifact and

standardized for all tissue to be 0.25 ml of 0.15% hydrogen

peroxide per 100 ml substrate. Sections were mounted onto

chrome alum slides, lightly counterstained with neutral red,

rapidly dehydrated, coverslipped and systematically examined

with both bright and dark field illumination under the


Structures containing labeled neurons were subjectively

classified as "lightly" (+) or "heavily" (++) labeled. This

estimation was based on the relative number of labeled cells

in the section through the structure that appeared to

contain the most labeled neurons. Obviously, this partial

categorization of labeling as "heavy" or "light" depends on

the number of sections through the structure and the number

of cells per structure as well as additional factors that

may affect the degree of labeling such as histochemical,

injection and histological variables. Although the present

criteria for the degree of labeling of a structure are not

precisely quantified, this type of classification attempts

to depict broad differences between cases as a first level

of analysis that may be subsequently pursued in a more

quantitative fashion.

Anterograde Degeneration Experiments

Eight hamsters between 30 and 56 days of age were used

in this part of the experiment. Anesthetized animals were

positioned with their incisor teeth tightly clamped in a

rotatable nose bar. The skull was angled to facilitate a

lateral approach to the ventrolateral surface of the cortex.

Under view of a Zeiss Epitechnoscope a bone flap was gently

removed caudal to the zygoma where the middle cerebral

artery exits. A thin silver wire was heated by attaching it

to a fine point soldering iron. The hot wire was briefly and

repeatedly applied to the cortex just above the rhinal

sulcus. The heat lesion technique proved to be superior to

suction in that little hemorrhaging accompanied the

procedure. Following survival times of four or ten days,

animals were intracardially perfused with physiological

saline followed by 10% Formalin. After one week in the

fixative and two to three days in 30% sucrose-Formalin, the

brains were embedded in gelatin-albumin.

Alternating coronal series of one 50 pm section and two

consecutive 25 pm sections every 0.15 am were cut into 10%

foraalin on the freezing microtome and stored at 4 C. The 50

pm series was stained with cresyl violet and used to

determine the extent of the lesion. One series of the 25 p m

sections was stained according to the Fink-Heimer I

procedure (1967). The adjacent series was processed

similarly, except for the omission of the potassium

permanganate and bleach suppression steps (Leonard, 1974).

The non-suppressed silver-stained section provided a more

abundant evidence of argyrophilic particles indicative of

anterograde degeneration, although this process resulted in

a higher background level. The relative density and

distribution of terminal and fiber degeneration were charted

onto enlarged tracings of the sections with the aid of the

light microscope.

Prefrontal Cortex Terminology

The nomenclature of Krettek and Price (1977a) for the

subdivision of anteromedial cortex in the rat proved to be

appropriate for demarcation of similar cytoarchitectonic

regions in the hamster, i.e. prelimbic, infralimbic,

anterior cingulate and medial precentral regions. The

patterns of labeling, however, following involvement of

different cytoarchitectonic regions did not show appreciable

differences dependent on the cortical subdivision involved

in the injection. The subdivisions of prefrontal cortex have

been classified in many ways (see Krettek and Price, 1977a;

Markowitsch and Pritzel, 1978; Divac et al., 1978b). The use

of Brodmann's numerical system for species comparisons

relies solely on cytoarchitectural criteria in primates that

are difficult, yet not impossible, to apply to the rodent.

On the other hand, terms like "cingulate" and "prelimbic"

connote functional interpretations, many of which have not

yet been adequately established (e.g., frontal eyefields).

With the intention of providing a conservative basis for

comparative studies of prefrontal cortex and avoiding the

problems of numerical or functional terminology, we have

opted to divide the medial pregenual region on a

topographical basis into dorsal, intermediate and ventral

pregenual regions (Figure 3A). Our terminology can be

converted to the cytological and functional nomenclature in

the following way: 1) dorsal pregenual to medial precentral

and anterior cingulate (areas 8 and 24); 2) intermediate

pregenual to prelimbic (area 32); and 3) ventral pregenual

to infralimbic (area 25).



Horseradish peroxidase injections into pregenual cortex

resulted in labeled neurons distributed throughout the

cortex, forebrain, thalamus, midbrain and brainstem. The

unexpected finding of labeled neurons in a cortical region

adjacent to the rhinal sulcus which we have termed the

pararhinal cortical area led to further investigation of

this area. The cytoarchitecture of the pararhinal area in

normal adult hamster brains was characterized through the

use of Nissi, silver and myelin stains. Lesions of the

pararhinal cortical area produced degenerating fibers and

terminals located primarily in rostral sulcal and

dorsomedial cortical fields of presumed prefrontal cortex.

Thus, both anterograde and retrograde methods were

consistent in demonstrating a corticocortical afferent

system from pararhinal cortex to medial pregenual cortex.

Subcortical Afferents

The results from a representative case are

schematically charted in Figure 3. The relative density of

retrograde labeling as well as a limited number of

anterogradely labeled fibers are shown. There were no

apparent differences between this 34 day old hamster and

older animals with similar injection sites. Almost all of

the total area of anteromedial pregenual cortex is involved

in the injection site (Figure 3A, parasagittal

reconstruction). The injection site wasedelineated by an

area containing dense particulate HRP reaction product

equally distributed over cell bodies and neuropil. The size

of the injection site evaluated in this manner did not

correlate with the volume of HRP solution injected.

In the forebrain there were labeled neurons in the

rostral claustrum and the basal forebrain, a vaguely defined

region beneath the anterior commissure and among the globus

pallidus and internal capsule (Figures 3C, 4F). In the

thalamus labeled cells were found in the mediodorsal nucleus

as well as the paratenial, reticular, anteromedial (Figures

3D, 4A), intralaminar nuclei reunionss, rhomboideus,

paracentral, centrolateral, centromedial and parafascicular)

and ventromedial nuclei (Figures 3E, 4C, 4D). The most

heavily labeled regions were in the mediodorsal and

ventromedial nuclei. The mediodorsal nucleus contained a

central zone relatively free of labeled cells surrounded by

Figure 3. Distribution of HRP-labeled neurons in case 192.
A. The extent of the injection site is enclosed by the
dashed line in the parasagittal reconstruction through
medial neocortex. The rostrocaudal levels of coronal
sections B-G are shown with vertical lines. B-G. Dots
denote the relative density of labeled neurons and dashes
denote labeled fibers. Circled letters designate boxed
regions photographed for Figure 4. Arrows mark the extent of
pararhinal cortical area. Abbreviations---am:anteromedial
thalamic nucleus; bf:basal forebrain; CC:corpus callosum;
cem:centromedial thalamic nucleus; cla:claustrum;
CP:cerebral peduncle; DPG:dorsal pregenual cortex; dr:dorsal
raphe nucleus; RP:habenulopeduncular tract; IC:internal
capsule; IL:intralaminar thalamic nuclei; IPG:intermediate
pregenual cortex; md:mediodorsal thalamic nucleus;
LH:lateral hypothalamus; MT:maamillothalamic tract;
PRC:pararhinal cortex; pt:paratenial thalamic nucleus;
PT:pyramidal tract; ret:reticular thalamic nucleus;
reu:reuniens thalamic nucleus; RF:reticular formation;
vm:ventromedial thalamic nucleus; VPG:ventral pregenual
cortex; vta:ventral tegmental area of Tsai: ZI:zona incerta.




Figure 4. Darkfield photomicrographs of labeled neurons in
brain regions after HR? injection in Figure 3. See squares
in Figure 3 for location. A. Rostral thalamus (calibration
bar = 100 unm, also for C and D). B. Contralateral pregenual
cortex. C. Ventromedial and centromedial thalamic nuclei.
D. Mediodorsal thalamic nucleus. 9. Lateral amygdala
(calibration bar = 100 pri, also for B, F, G and H). F.
Basal forebrain. G. Ventral tegmental area. H. Dorsal
raphe nucleus.


a ring of heavily labeled neurons. Other forebrain regions

containing labeled cells were the lateral and basolateral

amygdala (Figures 3D, 4E), zona incerta and caudal areas of

the lateral hypothalamus. In the brainstem labeled cells

were sparsely distributed in the ventral tegmental area of

Tsai (Figures 3F, 4G), the dorsal raphe (Figures 3G, 4H),

the parabrachial nucleus and the central superior nucleus.

Cortical Afferents

Labeled neurons in the cortex contralateral to the

injection site were primarily restricted to the pregenual

cortex contralateral to the area encompassed by the

injection site (Figures 38, 48). Ipsilaterally, numerous

labeled neurons were seen surrounding the borders of the

injection site. The number of neurons labeled in the

ipsilateral medial cortex declined gradually caudal to the

injection terminus. They were located primarily in cingulate

cortex rostral to the beginning of the retrosplenial

granular and agranular cortical areas.

The finding of labeled neurons deep to the rhinal

sulcus along its mid-caudal extent prompted further analysis

of the characteristics of this poorly defined cortical

region. Labeled neurons were seen above and below the funds

of the rhinal sulcus from rostral thalamic to caudal

midbrain levels (Figure 5).

Large injection site areas that spread to the caudal

portions of dorsal and intermediate pregenual subdivisions

resulted in the greatest number of labeled cells in the

rostral portions of the pararhinal area as well as some

labeled neurons in the caudal extent of the pararhinal area

(Table 2). Smaller injection site areas that were restricted

primarily to the rostral aspects of dorsal pregenual cortex

produced fever labeled neurons in the rostral portions of

the pararhinal area and no labeling in the caudal portions.

The general labeling pattern did not differ after partial

inclusion of the ventral pregenual region. Overall, smaller

injection sites resulted in decreased labeling in most


Cytoarchitecture of Pararhinal Cortex

The extent of the pararhinal region was more clearly

delimited through examination of silver and Nissl stained

sections (Figure 6). The pararhinal cortex can be subdivided

into a rostral division, the insular area and a caudal

division, the perirhinal area. Although it is difficult to

separate the pararhinal region solely on the basis of

cytoarchitecture, the presence of the claustrum is a

distinguishing feature that may be used as a dividing

landmark. The narrow wedge-shaped insular cortex contains

both agranular cortex (Brodmann's area 13) and, more

dorsally, sparsely granular cortex (Brodmann's area 14). The

lack of a definitive granule cell layer (layer IV)

Figure 5. Center of pregenual cortex injection site area
and pararhinal cortex labeling. A. Light-field
photomicrograph of the center of the injection site of the
case diagrammed in Figure 3. B. Darkfield montage of the
cortical label in the insular region (Calibration bar = 100
is). The closed arrowhead denotes the pial surface adjacent
to the rhinal sulcus. The open arrowhead denotes the
external capsule. The white arrows point out several of the
labeled neurons. C. Darkfield montage demonstrating
cortical label in the perirhinal region. D. Black arrows
point to the granular HRP reaction product in the pararhinal
cortical neurons (Calibration bar = 100 pm).


,,-" Q



Table 2. Variation in the distribution of HRP-labeled
neurons as a function of injection site and size.

? P C



I lti1i


4 -'-+



---+ +HF

; ;

A-K + -F
+1- K



-E+|4+ + ++J+L+ ;j +I ++
- +++++ +

Note: "0" = no labeled cells, "+" = "lightly" labeled and
"++" = "heavily" labeled. See Figure 3 for abbreviations.



Figure 6. Topography and cytoarchitecture of pararhinal
cortex. A. Lateral view of hamster brain demonstrates the
location and extent of pararhinal area by the slashed lines
above and just below the rhinal sulcus. B-E. Low power
photomicrographs of the pararhinal area in Nissl-stained
sections (Calibration bar = 10 pm). B and C. Insular
region. D and E. Perirhinal region. Arrows denote agranular
and granular subdivisions, stars denote the rhinal sulcus.
F. Low power photomicrograph of pararhinal cortex,
Fink-Schneider stain. Dark arrows mark the dense layer II
and sparse layer VI fibers ending at the edge of the
pararhinal cortical area.



distinguishes the insular pararhinal area from the more

dorsally located temporal region. In the perirhinal area,

agranular area 35 and sparsely granular area 36 replace

areas 13 and 14 respectively. On the basis of

cytoarchitecture, the pararhinal cortical region can be

classified as a type of transition cortex, the

periallocortex (Sanides, 1970), situated between allocortex

and neocortex.

The broad layer V of the insular area contains widely

spaced, medium size pyramidal cells (Figure 6, B and C).

There is no obvious distinction between the relatively small

cells of layers II and III. The large cells of the claustrum

contrast sharply with the smaller ones in layer VI. In the

dorsally contiguous granular insular area, a few scattered

granule cells emerge to form layer IV. Tangential and

oblique fibers course through the lower portions of layers V

and VI. The denser layer V cells are somewhat smaller than

in the agranular insular area. Layers II and III can be

subtly distinguished based on the larger size of the layer

III cells. As in the agranular insular area, layers II and

III lack the prominent band of primarily tangential fibers

that are present in the bordering temporal region (Figure

6E) .

The insular and perirhinal regions merge gradually as

the cortical cytoarchitecture changes. The perirhinal region

(Figure 6, C and D) narrows in its overall width. The small

cells in the outer layers become less dense than before,

while the oval cells of layer VI spread out. There is still

a marked absence of layer II fibers, while the layer VI

fibers are also less evident and eventually disappear.

Anterograde Degeneration

Lesions of the pararhinal cortex resulted in a pattern

of degeneration which confirmed the HRP data by

demonstrating a projection to the medial pregenual cortex.

An example of the pattern of anterograde degeneration is

diagrammed in Figure 7. A lateral surface reconstruction of

the cortical extent of the lesion in this case demonstrates

that the lesion was restricted to the pararhinal area and

involved both the rostral and caudal subdivisions. All

cortical layers were damaged with minimal, if any, damage to

the underlying extreme capsule and corpus callosum. Other

cases with larger lesions included damage to white matter

and neighboring cortical areas as well as the caudatoputamen

and hippocampal complex. Degenerating fibers and terminals

stream medially from the lesion and continue obliquely along

the corpus callosum. In the cases involving white matter

damage, a denser and more widespread pattern of degeneration

was observed in layers V and VI and the callosum. Fine

irregular argyrophilic particles indicative of terminal

degeneration are most dense in the rostral sulcal cortical

region (Figure 8D) and the dorsal pregenual cortex (Figure

8A). Cortical regions exhibiting less dense degeneration are

Figure 7. A-D. Chartings of anterograde degeneration after
pararhinal cortex lesion. The lesion is reconstructed onto a
surface view in the upper right inset. The blackened area
represents the lesion extent. Small dots represent terminal
degeneration and a series of large dots represent fiber
degeneration. Lover case letters a and b refer to location
of the photomicrographs in Figure 8.
Abbreviations---AC:anterior commissure; CC:corpus callosum;
CPU: caudoputamen.



cc .



Figure 8. High power photomicrographs of Fink-Heimer
degeneration taken at the a and b loci of Figure 7. A.
Medial pregenual cortex. B. Contralateral medial pregenual
cortex from same animal. C. Medial pregenual cortex of the
field shown in A stained only with Fink-Heimer I including
the suppression steps. D. Rostral sulcal cortex.
(Calibration bar = 25 v m for all figures.)




~E I r ~L tl

located in intermediate pregenual cortex. Degeneration is

absent from the contralateral medial pregenual cortex

(Figure 8B). When the routine suppression steps are included

in the Fink-Heimer I procedure to stain sections from the

same brain where obvious degeneration is present, terminal

degeneration is not evident in the medial pregenual cortex

(Compare Figure SA and 8C). At the rostral pole,

degeneration spreads over the dorsal and lateral cortical

regions. In all of these cortical regions the degeneration

is restricted to the inner half of layer I and the outermost

portion of layer II.


Following 4HP injections into medial pregenual cortex

of the adult golden hamster, the pattern of subcortical

labeling was similar to that previously reported for the rat

(Divac et al., 1978b) with a few exceptions. No labeled

cells were seen in the lateroposterior thalamic nucleus when

the injection site was restricted to medial pregenual

cortex. In other cases where the injection site extended

caudally into more supragenual regions, labeled neurons were

seen in the lateroposterior thalamic nucleus as well as the

posterior and ventroanterolateral thalamic nuclei.

Injections sites that are restricted to caudal pregenual

cortical regions cortex are necessary to confirm the


meaningfulness of the pregenual cortical subdivisions

distinguished in this study.

The pattern of labeling in the mediodorsal thalamic

nucleus confirms previous reports of the existence of MD

subdivisions in the rodent (Leonard, 1969; Krettek and

Price, 1977a; Divac et al., 1978b). Extensive labeling of

the ventral and lateral regions of the nucleus and the

relative scarcity of labeled neurons in a central core

region support previous findings that MD is divided into a

core region projecting to sulcal cortex and a peripheral

region projecting to the medial wall.

A less clear but equally intriguing pattern of labeling

was found for the area generally referred to as the basal

forebrain area after Kievit and Kuypers (1975). A sparse

number of these relatively large HRP-labeled neurons were

diffusely distributed in ventral forebrain areas such as

substantial innominata, diagonal band and lateral preoptic

area extending dorsally to intermingle with the penetrating

fibers of the internal capsule and the globus pallidus. As

in the monkey (Mesulam and Van Hoesen, 1976) and hamster

(Crandall and Leonard, unpublished observations) these

neurons contain acetylcholinesterase. In the hamster as in

other species, medial pregenual (prefrontal) cortex is

clearly involved with the limbic system since it receives

afferents from the hypothalamus, basal forebrain, amygdala

and raphe.

The discovery of the ipsilateral pararhinal association

projection to medial pregenual cortex is the most striking

finding of the present study. Based on the gradation of

cytoarchitectural changes in this type of transition cortex,

it has been difficult to determine the precise boundaries of

the cortical region containing HRP-labeled neurons. Because

the area is located deep to the rhinal sulcus, we have

chosen the topographic, descriptive term "pararhinal",

subdividing this long, narrow gegion into a rostral, insular

region and a caudal, perirhinal region. From Nissl and

silver-stained material the cardinal features appear to be

the absence of a layer II-III fiber plexus and an absent or

sparse granular layer IV. It is difficult to make species

comparisons, even among rodents, because of the lack of

lengthy descriptions available of this cortical region (See

Table 3), and because of the lack of prominent landmarks.

Perirhinal cortex in general has been defined only by its

location deep to the caudal portions of the rhinal sulcus.

There have been few previous descriptions of this

region's connections and its relationship to function. In

the rat, the perirhinal region receives projections from the

basolateral, lateral and anterior cortical amygdaloid nuclei

(Krettek and Price, 1974; 1977b), the ventromedial (Krettek

and Price, 1977a; Herkenham, 1979) and reunions thalamic

nuclei, (Herkenham, 1978), the mesencephalic dopamine cells

of A 10 (Lindvall, Bjorklund, loore and Stenevi, 1974) and

areas 24 and 32 of medial prefrontal cortex and medial,

lateral and insular divisions of rostral cortex (Beckstead,

1979). It is interesting that medial pregenual and

pararhinal cortex not only appear to be reciprocally

connected, but that both are innervated by dopaminergic and

amygdaloid afferents. Medial pregenual cortex may definitely

be considered part of medial prefrontal cortex in the

hamster as the retrograde transport HRP method has confirmed

the existence of a mediodorsal thalamus projection to the

anteromedial cortical wall rostral to the genu of the corpus


The lack of knowledge concerning pararhinal cortex

efferents in the rodent hinders functional comparisons with

other species where study of such connections exist. In the

cat, insular cortex projects to auditory cortex (Diamond,

Jones and Powell, 1968), as well somatosensory, visual,

orbitofrontal and cingulate regions (Cranford, Ladner,

Campbell and Neff, 1976). Subcortical regions receiving

insular projections include the caudate (Avanzini, Mancia

and Pellicioli, 1969), the posterior thalamic nucleus

(Diamond, Jones and Powell, 1969), the

lateroposterior-pulvinar complex (Cranford et al., 1976),

the mediodorsal thalamus, the corpora quadrigemina, central

gray and amygdaloid nuclei (Siegel, Sasso and Tasson, 1971).

The primate perirhinal cortex receives projections from a

broad spectrum of cortical regions including orbitofrontal

(Van Hoesen, Pandya and Butters,1975), temporal (Jones and

Powell, 1970; Van Hoesen and Pandya, 1975a) and somatic

associational cortices (Jones and Powell, 1970). Primate

perirhinal cortex sends its efferents primarily to the major

source of hippocampal afferents, the entorhinal cortex (Van

Hoesen and Pandya, 1975b).

The function of the transition type of cortex found in

the pararhinal cortical region in rodents is unknown. In

carnivores and primates, Markowitsch and Pritzel (1978) have

hypothesized that the role of the insular region is that of

an integration center of multisensory information for higher

order perceptual abilities. Other speculations regard the

insular region as a sensory-visceral, multimodal sensory

discrimination region (Cranford et al., 1976). Clearly,

additional behavioral and electrophysiological work remains

to be done to elucidate further the functional roles of this

small but important convergent cortical region. Indirect

evidence from a developmental study of medial pregenual

cortex afferents in the hamster (Crandall and Leonard, 1979)

suggests a temporal correlation between the development of

pararhinal cortex afferents to medial pregenual cortex and

the development of the species-specific behavior of

hoarding. Future studies will directly investigate the

relationship between the ontogeny of this complex behavior

and the intricacies of the neurological development of the

pararhinal cortex.

Table 3. Pararhinal cortical areas in rodents.



(Itagaki, 1957)






Insular division

Insularis (caudal)

Insularis (caudal)

Anterior agranular
and granular insular

Insular fields 13
and 14 (caudal)

Perirhinal areas

13 and 20

Perirhinal division


Insularis and

(ventral) and

field 35 and
temporal field 36

Areas 20,35,36



The maturation of the cerebral cortex encompasses a

plentitude of morphological changes that transform the thin,

bilaminated embryonic cortical shell into the complex,

multi-layered adult cortical expanse. Cell proliferation,

migration, and differentiation, as well as synaptogenesis,

dendritic growth and axonal arborization, all shape the

growing neocortex. In the developing rat somatosensory

cortex, the pattern of thalamic afferents matures by Day 5

(Wise and Jones, 1978). Commissural afferent terminals

appear adult-like in their distribution pattern two days

later, even though the commissural neurons do not reorganize

from a continuous to a patchy pattern until the end of the

second week (Ivy, Akers and Killackey, 1979; Wise, Fleshman

and Jones, 1979).

The callosal efferent neurons in kitten primary

somatosensory and visual cortices undergo a somewhat similar

cortical postnatal regrouping (Innocenti and Caminiti,

1980). Developmental changes in cortical connectivity in

nonprimary sensory cortex, e.g., prefrontal associative area

cortex, have not been investigated in rodents. In the

monkey, however, Goldman and associates have studied the

anatomical and functional maturation of prefrontal cortex.

Corticocortical columnar connections are demonstrable with

autoradiographic methods at four days postnatal (Goldman and

Nauta, 1977). Prefrontal efferents exhibit degeneration

argyrophilia at their subcortical targets by at least 2.5

months (Johnson, Rosvold, Galkin and Goldman, 1976).

Following prenatal removal of presumptive prefrontal cortex,

mediodorsal thalamic neurons survive and do not undergo

retrograde degeneration as they typically do in postnatal

and older animals with prefrontal cortex damage (Goldman and

Galkin, 1978). The sparing of mediodorsal thalamic neurons

corresponds to a dramatic behavioral sparing in these

prenatally operated monkeys. Age dependency of recovery of

behavioral function has also been reported for prefrontal

cortex-damaged rats (Kolb and Nonneman, 1976; 1978).

The previously mentioned studies of prefrontal cortex

functional and anatomical development involved the use of

lesions to establish age-dependent behavioral phenomena. The

relationship between the maturation of a

cortically-dependent behavior and the development of

prefrontal cortex in normal animals remains unknown. Food

pellet hoarding behavior of the Syrian (golden) hamster

emerges during the third week (see CHAPTER TWO), and is

disrupted following damage to the prefrontal cortex in the

adult (Shipley and Kolb, 1977). In the present study, the

development of prefrontal cortex connections in the golden

hamster is investigated with respect to their relationship

to changes in hoarding behavior maturation. More

specifically, the development of afferent and efferent

connections of medial prefrontal cortex are examined with

horseradish peroxidase histochemistry and anterograde

degeneration methods. The pattern of efferent projections as

demonstrated by argyrophilic degeneration did not change

during the period in which hoarding behavior acquired its

mature characteristics. The afferent projections, however,

did appear to change and undergo reorganization during the

critical time period.


Horseradish Peroxidase Experiments

Golden hamsters (Mesocricetus auratus) were obtained

from the laboratory breeding colony whose stock came from

Charles River Lakeview Breeding Laboratory (Cambridge,

Massachusetts). Hamster litters were culled to eight pups at

birth and weaned to individual cages on postnatal day 20

(day of birth = postnatal Day 0). The 69 subjects included

18 Day 10, 16 Day 21, 19 Day 25, ten Day 33, and six adult

hamsters. Weanling and older hamsters were anesthetized with

intraperitoneal injections of Chloropent (0.04mg/kg) and

were placed in a stereotaxic instrument. Day 10 pups were

anesthetized with a combination of ether and ice-induced

hypothermia and were stabilized in an adaptable head holder.

A small burrhole was drilled midway between the coronal and

frontal sutures adjacent to the sagittal suture. A glass

micropipette with inside tip diameter 30-60 Pm containing

30% (W/V) HRP (Sigma type VI) in physiological saline was

lowered 0.1-1.5 mm below the brain surface. An adjustable

2.0 ml Burette pipette was connected to a three-way

luer-lock stopcock from which polyethylene tubing was

attached with dental cement to the large end of the

HRP-containing micropipette. A volume of 0.01-0.05 pl of HRP

was injected over a period of five to ten minutes. The

micropipette was slowly removed after a five minute waiting

period, the skin was sutured, and the animal returned to its

home cage for 24 hours before sacrifice.

Under deep ether anesthesia, the animals were perfused

intracardially with a brief rinse of physiological saline,

100 ml of fresh 1.25% glutaraldehyde-1% paraformaldehyde in

0.1 phosphate buffer at pH 7.2, followed by 100 ml of cool

30% sucrose in the same buffer. Brains were removed

immediately and stored overnight in the sucrose-phosphate

buffer at 40C. Coronal sections of 50 im were cut on the

freezing microtome into cool phosphate buffer and

immediately treated according to the tetramethyl benzidene

procedure of Mesulam (1978). The concentration of hydrogen

peroxide was adjusted to minimize the amount of crystalline

artifact, and was subsequently standardized to be 0.25 ml of

0.15% hydrogen peroxide per 100 ml of substrate. Sections

were mounted onto chrom alum slides, lightly counterstained

with neutral red, very rapidly dehydrated, coverslipped and

systematically examined under the microscope with both

bright and dark field illumination. Initially, in every

fourth section, each structure containing labeled neurons

was subjectively classified into either a no-label, lightly

labeled, or heavily labeled category.

Quantitative Analysis

All quantitative analyses were performed on coded

material so that the ages of the animals were not revealed

until all data were collected. The extent of each injection

site was reconstructed onto a parasagittal drawing of the

medial cortical wall. The injection site was defined as the

region in which the density of reaction product was similar

over cells and neuropil (see Figure 9). Two sets of animals

(one from each age group--Day 10, Day 21, Day 25 and adult)

were matched for injection site extent. One group consisted

of animals with similar, relatively large injection areas;

the other group consisted of animals with relatively small

injection sites. Camera lucida drawings of the mediodorsal

thalamic nucleus and the lateral and basolateral amygdaloid

nuclei in each section containing these structures were used

to mark the location of each labeled neuron. Seven sections

at specific landmark levels of pararhinal cortex were

analyzed similarly.

Anterograde Degeneration Experiments

one hundred and fifteen hamsters ranging from ten days

of age to adults were used in this part of the study (see

Table 4). Anesthetized animals were placed with their

incisors firmly clamped into a rotatable nose bar attached

to a stereotaxic frame. The skull was angled to facilitate

the surgical approach to anteromedial cortex. Under view of

a Zeiss Epitechnoscope a thin bone flap was removed from the

skull midline extending from the frontal to the bgegma

suture. After removing the pia, medial cortex was gently

aspirated through the tip of a drawn glass pipette, taking

care to stay within the visual landmarks: rostrally, the

olfactory bulb membrane, medially, the cortical wall pia,

and caudally, the white matter of the genu. Following the

cessation of bleeding, the wound was packed with gelfoam,

the skin sutured, and the animal was returned to its home

cage. At the appropriate survival period, animals were

anesthetized with other, and were intracardially perfused

with physiological saline followed by 10% Formalin.

After one week in the fixative and two to three days in

30% sucrose-Formalin, the brains were embedded in

gelatin-albumin. Alternating coronal series of one 50 um and

two consecutive 25 Um sections every 0.15 mm were cut on the

freezing microtome into 10% Formalin and were stored at 40 C.

The 50 im series was stained with cresyl violet and was used

to determine the extent of the lesion. Several different

central border caudal

figure 9. Darkfield photomicrographs of sections from
different regions of the medial prefrontal cortex involved
in (central and border) or adjacent to (caudal) the typical
Hi? injection site. All sections are taken from the same
animal. Ihe dark shaded area indicates the extent of the
entire HPP injection site on the parasagittal reconstruction
of redial neocortex.

Table 4. Number of hamsters at operation ages and survival
times for anterograde degeneration experiments.

(hou s)

18 24 72 240

10 2 17 14 5

21 4 9 17 6

25 1 6 13 3

1 4


methods were tried on the 25 pm series to demonstrate

terminal and fiber degeneration. The Fink-Heimer I (1967),

the de Olmos (1969) cupric silver, the Eager (1970) modified

Fink-Heimer, and the Fink-Schneider (1969) procedures were

tried in instances where clear evidence of degeneration

argyrophilia was not present when stained with the

non-suppressed Fink-Heimer I method (Leonard, 1974). In all

cases in this study, the alternative procedures did not

produce evidence contrary to that provided by the

non-suppressed Fink-Heimer procedure. At least one series of

25 um coronal sections was stained for each animal with the

non-suppressed Fink-Heimer method. For the adults,

additional series of sections were stained using varied

suppression times. A three minute permanganate step proved

to be optimal for differentially staining degenerating and

normal fibers. For each animal, silver-stained sections were

systematically scanned with the aid of the light microscope.

The locations of terminal and fiber degeneration were noted.

Subsequently, animals were matched for lesion site extent

and topography, and those structures which contained

degeneration were compared. The relative density and

distribution of terminal and fiber degeneration were charted

onto enlarged tracings of representative sections.



At all ages studied, HRP-labeled afferent neurons to

MPFC were distributed throughout the forebrain,

diencephalon, midbrain and brainstem. The labeling patterns

of the mediodorsal thalamic nucleus (MD), pararhinal cortex

(PRC), amygdala (AMY) and ventral tegmental area (VTA) were

quantitatively analyzed in animals from each age group. The

pattern of MD labeling changed dramatically between Day 21

and Day 25, with small HRP injections on Day 21 resulting in

many more labeled neurons than similar HRP injections on Day

25. The correlation of injection size and amount of ND

labeling that was found on Day 25 continued into adulthood.

In contrast to the developmental shift in the pattern of MD

afferents, the efferents of medial pregenual cortex did not

change between Days 21 and 25. At Days 21 and 25,

degeneration argyrophilia after long survival periods ( L72

hours) was well established in principal fiber tracts, as

well as in major terminal regions of the forebrain, thalamus

and midbrain. In Day 10 animals, degeneration had

disappeared by 72 hours, but was clearly evident after short

survival periods (24 hours). Conversely, degeneration

argyrophilia was not present after short survival periods in

the Day 21 and Day 25 animals.

Distribution of Afferents

Similar regions in cortex and subcortex contained

HRP-labeled neurons at all ages (Day 10, 21, 25 and older

animals) following relatively large HRP injections. The

animals included in this study had HRP injections restricted

to the anteromedial cortex rostral to the genu. The large

injection sites involved portions of dorsal, intermediate

and ventral pregenual subdivisions. Labeled neurons were

found in rostral subcortical regions in the claustrum, basal

forebrain, numerous thalamic nuclei (paratenial, reticular,

anteromedial, intralaminar, ventromedial and dorsomedial

nuclei), amygdala (lateral and basolateral regions), zona

incerta and caudal hypothalamus. Labeled neurons were

located in more caudal brain areas in the ventral tegmental

area, dorsal raphe, parabrachial nucleus, and central

superior nucleus. Cortical afferents were found in the

ipsilateral cortical region surrounding the injection site,

the homotypical cortical region contralateral to the

injection site and the ipsilateral pararhinal cortical


Developmental Changes in Afferents

Only in Day 10 animals were presumed glia labeled

bilaterally in the corpus callosum, anterior commissure and

internal capsule. These stained cells appeared similar in

type and distribution to those described by Ivy and

Killackey (1978) in the developing rat brain. Another

qualitative developmental difference was first noticed

between Day 21 and Day 25 animals in the pattern of MD

labeling after relatively small HRP injections (Figure 10).

After comparable HRP injections confined principally to

dorsal pregenual cortex, MD appeared heavily labeled at Day

21, but lightly labeled at Day 25 at both the rostral and

caudal levels (Figure 10A vs. 10D). Large HBP injections of

comparable extent and location at the two ages resulted in a

heavily labeled pattern (Figure 10A vs. 10C) similar to the

example of Day 21 small injection (Figure 10B). This obvious

qualitative developmental change in MD labeling pattern was

further analyzed in two sets of four animals each (one

animal from each age group) which were matched for injection

site location and extent (Figures 11 and 12). The location

of each labeled MD neuron in each section throughout the

extent of MD was charted onto camera lucida outlines of the

nucleus, each neuron being represented by a dot. In the set

of large injection sites, many MD neurons were labeled at

all ages, including Day 10. In contrast, after small HRP

injections, a comparable quantity of labeled neurons was

seen only on Day 21.

Labeling of the pararhinal cortex following large

injections of HRP increased with age. Figure 13 depicts

darkfield montages of the insular division of pararhinal

cortex following similar large HRP injection sites at each

age. These sections were taken from the same animals that

Figure 10. Low power darkfield photomicrographs of the
mediodorsal thalamic nucleus. Large and small refer to the
relative size of the HRP injection (inj.) site (See Figures
11 and 12). A. Day 21, large inj. B. Day 21, small inj.
C. Day 25, large inj. D. Day 25, small inj. (calibration
bar = 100 Pm, all pictures at same magnification).

DAY 21 DAY 25

.,, uJ,- .

Figure 11. Pattern of HRP labeling in mediodorsal thalamus
(MD) after large HRP injection extent. The MD drawings of
coronal sections 200 vm apart are shown in rostral to caudal
order for one animal at each age with similar injection
sites. Mh designates the medial habenula. Each dot
represents one HRP-labeled neuron. Differences in the size
and shape of MD probably reflect both developmental
differences and differences in the plane of section between
various brains.



0 (i




..-..,.~ i .~

Figure 12. Pattern of HRP labeling in mediodorsal thalamus
after small HRP injection extent. The MD drawings of coronal
sections 200 um apart are shown in rostral to caudal order
for one animal at each age with similar injection sites. Ah
designates the medial habenula. Each dot represents one
HRP-labeled neuron.





DAY 10








Figure 13. A.-D. Low power darkfield photomontages of
HRP-labeled neurons in pararhinal cortex at four ages
following similar large HRP injections. Solid arrowhead
marks the pial surface; open arrowhead marks the external
capsule. White arrows point out several labeled neurons. All
pictures were photographed at the same magnification
(calibration bar = 100 pm).

DAY 10

DAY 21

DAY 25


Figure 14. Pattern of HRP labeling in pararhinal cortex
after similar large HRP injections. Landmark coronal
sections through PRC in rostral to caudal order are shown
for one animal at each age. Cla designates the claustrua.
The star marks the rhinal sulcus. The single arrow
demarcates the pia; the double arrow points to the corpus













were analyzed for MD labeling after large HRP injections

(Figure 12). Labeled neurons throughout the rostral and

caudal extent of pararhinal cortex were further analyzed on

each of seven sections containing specific cortical

landmarks. Layer V neurons are more numerous and are more

widely distributed throughout pararhinal cortex in Day 25

and adult animals than in Day 21 animals (Figure 14).

Similarly, PRC labeling on Day 21 is more extensive than the

labeling seen on Day 10. The age-dependent increase in

labeling that occurred in MD following small injections at

Day 21 was not observed in nRC of the same animals.

The two sets of animals at four ages with matched

injection sites were analyzed by two quantitative methods.

For each of the three structures--MD, PRC and ANY--each

labeled cell in each section through the structure was

counted. The animals were ranked according to the total

numbers of labeled neurons. A separate ranking was

determined by the number of labeled cells on the most

heavily labeled section through each structure for each

animal. The two rankings had a very high degree of

correlation. The Spearmans rank correlation coefficient

equalled .97 for MD and PRC, and .93 for AMY (for each of

the three structures, If=6, p<.001). Subsequently, all

subjects were analyzed by counting the number of neurons on

the most heavily labeled section through the structure. The

four structures analyzed were the pararhinal cortex,

mediodorsal thalamus, amygdala and ventral tegmental area. A

frequency distribution was constructed for each region, and

the median number of labeled neurons was selected as the

criterion for designating structures as lightly labeled (+)

or heavily labeled (++). Structures containing a greater

number of labeled neurons than the median number were

designated as heavily labeled (++); structures containing

less than the median number of labeled neurons were

designated as lightly labeled (+).

Figure 15 summarizes the labeling patterns of each

animal in the four age groups. In general, as the injection

size increased, the labeling pattern of PRC, AMY, and VTA

changed from light to heavy. This was not the case for MD.

Both small and large injections on Day 21 resulted in

heavily labeled MD. In contrast, small injections in Day 25

and older animals resulted in lightly labeled MD, while

large injections in Day 25 and older animals resulted in

heavily labeled MD. Thus the earlier qualitative

observations were confirmed by these more rigorous

quantitative measures of degree of labeling. The dramatic

change in MD afferents to MPFC between Day 21 and Day 25 is

summarized in Figure 16. Each square contains the number of

animals exhibiting a particular labeling pattern of MD and

PRC. In the adult cases, labeling in PRC and MD was either

heavy in both structures or light in both structures. On Day

25, the pattern is similar, with eight having having light

label and six having heavy label in both structures. Only

one animal had heavy label in one structure and light in the

Figure 15. Summary of IRP injection sites for each age
group and the degree of labeling in PRC (pararhinal cortex),
MD (mediodorsal thalamic nucleus), AMY (amygdala) and VTA
(ventral tegmental area). "0" indicates non labeled, "+"
indicates lightly labeled and "++" indicates heavily labeled
categories. See text for the quantitative explanation. The
blackened areas in the parasagittal reconstructions of
medial prefrontal cortex mark the extent of the HRP
injection site.

DAY 10
Pm a v
r m t
c d ya

+ + +

+ ff-+ +

DAY 21

c d y a
0 + + +


+ i+ +

-t+f 4+ +H4

++4+ + +

DAY 25
c d y

S 04 0+

S + + +o

+ + + +

+ + 04+

+ + + +

+++ +

+ ++ ft
+ ++ ++

+ + -+ +

-4-++++-H +

+ 1+ 4+ +

+4++ -44

Pma v
r m t
c d y 0


++- +



4+ 1+ ++ +

4+H-T 4- ++

+-H- ++


Figure 16. Summary of the labeling patterns in mediodorsal
thalamus (MD) and pararhinal cortex (PRC) by age group.
"0/+" indicates non or lightly labeled; "++" indicates
heavily labeled. The number of animals exhibiting a
particular pattern of labeling for the two structures are
shown in each square. The range of HRP injection site size
for each age was between 30 and 100 percent involvement of
medial prefrontal cortex.





DAY 10

DAY 21


o/+ ++

1 6

DAY 25




other. On Day 21, however, the correspondence between the

degree of labeling of the two structures breaks down. Four

of eleven animals contain both heavily labeled MD and PRC,

but almost half (five of eleven) have a heavily labeled MD

with a lightly labeled PRC. This is the only age where this

labeling pattern is prevalent. This developmental shift in

the MD labeling pattern from Day 21 to Day 25 involves a

decrease in the number of labeled MD neurons following HRP

injections into medial prefrontal cortex.

Distribution of Efferents

The distribution of fiber and terminal degeneration

following damage to the medial prefrontal cortex in a Day

25, 72 hour survival case is charted in Figure 17. This

lesion involved much of the dorsal pregenual division and

part of the intermediate pregenual division sagittall

reconstruction, Figure 17A). Coarse argyrophilic fibers were

clearly seen in the corpus callosum, internal capsule,

cerebral peduncle and pyramidal tract. Numerous brain

regions contained the finer, dustlike silver particles

indicative of terminal degeneration. Contralateral to the

lesion (Figure 17B), terminal degeneration was located in

all layers of the homotypical pregenual cortical region.

Layers II and V contained the most dense distribution of

terminal degeneration. As the degenerating fibers of the

internal capsule passed through the caudoputamen, terminal

degeneration was evident throughout most of the rostral

extent of the caudoputamen. More caudally, the area

containing terminal degeneration became restricted to the

medial third of this structure (Figure 17D). Terminal

degeneration was most dense in the areas of caudoputamen

closest to the degenerating internal capsule fibers. The

internal capsule fibers converge in fascicles lateral to the

rostral thalamus. A small portion of the fibers can be seen

in the ventral caudoputamen, coursing toward the amygdala,

where sparse terminal degeneration is evident (Figure 17E).

More ventrally, the basal forebrain region contains sparse

terminal degeneration. In the rostral thalamus (Figure 17E),

degenerating fibers pass through the reticular thalamic

nucleus. Signs of sparse terminal degeneration are present

in the anteromedial and reunions thalamic nuclei. Terminal

degeneration is most dense in the mediodorsal thalamic

nucleus, particularly the ventrolateral (Figure 17F) and

caudal (Figure 17G) areas. The ventromedial thalamic nucleus

contains much terminal degeneration as well. A sparser

distribution of terminal degeneration is seen in

centrolateral and centromedial thalamic nuclei. Pascicles of

degenerating fibers bundle into the ventral portions of the

cerebral peduncles. Some degenerating fibers pass dorsally

into rostral portions of the midbrain tegmentum (Figure 17H)

and continue dorsally into the pretectal regions. Along this

dorsal pathway, terminal degeneration is evident. More

caudally, degenerating fibers enter the deep layers of the

superior colliculus (Figure 171). Sparse patterns of

terminal degeneration are present in the widespread

tegmental areas and the central gray area.

Developmental Changes in Efferents

The pattern of degeneration argyrophilia described

above was found after 72 hour survivals in both Day 21 and

Day 25 hamsters with comparable MPFC lesions. At 240 hour

post-operation, degeneration was still present in the major

fiber and terminal regions, although it appeared slightly

less coarse and less dense. After short survival times (24

hours), no signs of degeneration argyrophilia were evident

in any terminal or fiber areas in Day 21 or Day 25 hamsters

with similar lesions. In contrast, short latency

degeneration was clearly seen in MPFC efferent terminal

areas and fibers after Day 10 lesions with but had

disappeared by 72 hours. The nature of the short latency

degeneration at 10 days was less dense and more irregular

than the long latency degeneration in the older animals.

Examples of the age differences in degeneration

argyrophilia of MPFC efferents after 24 hour and 72 hour

survival times are pictured for three brain regions: the

forebrain area of the caudoputamen (Figure 18), the

diencephalic region of the mediodorsal thalamus (Figure 19),

and the midbrain tegmentum (Figure 20). These examples are

taken from animals with lesions similar to the one shown in

Figure 17A.

Figure 17. A.-I. Chartings of anterograde degeneration
medial prefrontal cortex lesion in Day 25, 72 hour survival
case. The lesion is shown by the shaded area on the
parasagittal map of the medial cortex. Small dots represent
terminal degeneration. A series of larger dots represent
fiber degeneration. Small dashes represent fascicles of
fiber degeneration. Circled letters A, B and C designate
boxed regions photographed for Figures 18, 19 and 20,
respectively. Abbreviations---ac:anterior commissure;
AD:anterodorsal thalamic nucleus; AM:anteromedial thalamic
nucleus; AMY:amygdala; AV:anteroventral thalamic nucleus;
bc:brachium conjunctivum; BF:basal forebrain; cc:corpus
callosum; CEM:centromedial thalamic nucleus; CG:central gray
area; cp:cerebral peduncle; fx:fornix; GE:gelatinosus
thalamic nucleus; hp:habenulopeduncular tract; ic:internal
capsule; LG:lateral geniculate; LHA:lateral habenula;
LP:lateroposterior thalamic nucleus; MD:mediodorsal thalamic
nucleus; MHA:medial habenula; mt:mamillothalamic tract;
NCL:centrolateralis thalamic nucleus; ot:optic tract;
pc:posterior commissure; pt:pyramidal tract; RET:reticular
thalamic nucleus; REU:reuniens thalamic nucleus;
RH:rhomboideus thalamic nucleus; SC:superior colliculus;
SM:stria medullaris; TEG:tegmentum; VB:ventrobasal thalamic
nucleus; VM:ventromedial thalamic nucleus.


sc .. RH

L Ofx





0of M

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ES9IG74ZJ_IOI7S4 INGEST_TIME 2012-03-02T22:44:24Z PACKAGE AA00009121_00001