Physiology &Behavior, Vol. 43, pp. 47-55. Copyright PergamonPress plc, 1988. Printed in the U.S.A. 0031-9384/88 $3.

0031-9384/88 $3.00 + .00

Plasma Corticosterone Levels During

Repeated Presentation of Two
Intensities of Restraint Stress:
Chronic Stress and Habituation
D A V I D L. P I T M A N , ~ t J O H N E. O T T E N W E L L E R * t AND BENJAMIN H . N A T E L S O N * t ~ tl

*Primate Neuro-Behavioral Unit, VA Medical Center, East Orange, N J 07019

tDepartment of Neurosciences, New Jersey Medical School, Newark, N J 07103
~lnstitute o f Animal Behavior, Rutgers University, Newark, N J 07123

R e c e i v e d 10 A u g u s t 1987

PITMAN, D. L., J. E. OTTENWELLER AND B. H. NATELSON. Plasma corticosterone levels during repeated
presentation of two intensities of restraint stress: Chronic stress and habituation. PHYSIOL BEHAV 43(1) 47-55,
1988.--This study measured plasma corticosterone levels in male rats during repeated daily presentations of two intensities
of restraint stress. The corticosterone response to a stress session was defined as the change from pre-stress levels to levels
after 60 minutes of restraint. With the relatively intense stress imposed by four limb prone restraint, the corticosterone
response partially habituated over seven days due to increasing basal corticosterone levels. However, even on day 7, there
was still a large corticosterone response. With the milder stress of immobilization in a tube, the corticosterone response did
not habituate at all over 21 days of repeated stress despite rising basal levels. Stress levels of corticosterone did not show
significant change over days in either of the two restraint groups. Further, rising basal corticosterone levels suggest that
repeated restraint produced a chronic stress state in these rats which may vary in some qualitative way with stressor
intensity. Control rats placed in the same room as the stressed rats during the two stresses initially had increased
corticosterone levels that matched the levels achieved in the stressed rats. The responses in control rats for the intense
stress did not habituate completely in 7 days, whereas those in the control rats for the mild stress habituated completely
within 3 days. These data suggest intraspecific communication of the intensity of stress.

Habituation Stress Corticosterone Intraspecific communication

F A C T O R S related to associative learning (e.g., Pavlovian
and instrumental conditioning) can alter the visceral re-
sponse produced by stress. For example, the effects of stress
on gastric pathology [26] and adrenal hormone secretion in
rats [2] vary with the reliability with which a conditional
stimulus is repeatedly paired with an unconditional stressor.
However, our ability to understand how associative learn-
ing impacts on the stress response depends to a large extent
on our knowledge of the importance o f non-associative fac-
tors. Specifically, repeated exposure of an animal to a stress-
or can result in either habituation (a decrease) or sensitiza-
tion (an increase) in responsiveness to that stressor, both of
which reflect non-associative learning. Because studies of
associative learning also require repeated presentation of
stressors, a better understanding of habituation and sensiti-
zation is required before the effects o f associative learning
on the response to stress can be fully appreciated. But, stress
researchers have not yet established rules for non-associa-
tive learning which could predict how the response to a par-
ticular stressor will change with time, i.e. will it increase,
decrease, or stay the same. There are suggestions in the
literature that such rules may exist. For example, one study
using mild stressors showed partial habituation of the corti-
costerone response [2], whereas another using a more in-
tense stress found no habituation [8]. While these studies
suggest that there is less habituation to more intense stress-
ors such a conclusion would only be justified from a com-
parison of the rates of habituation to stressors of different
intensity in the same laboratory. The data from such studies
are presented in this paper.
In the laboratory environment, neurophysiologists have
been the ones most interested in studying an organism's re-
sponse to repeated presentation of the same stimulus. Be-
cause of this, it may be possible to use the neurophys-
iological literature as a framework for better understanding
the role of nonassociative learning in chronic stress. That
literature shows that systematically manipulating stimulus
frequency, duration and intensity can predictably produce
either habituation or sensitization [24]. Since stressors may
be similar to such stimuli, it is likely that the approach of

'Requests for reprints should be addressed to B. H. Natelson, Primate Neurobehavioral Unit (127A), VA Medical Center, East Orange, NJ
07019.

47
 48
12.0
A
m
o
¢LI
t-
Q
m
8
O
m
E
W
Q
8.0'
4.0"
1 2
FIG. 1. Basal plasma corticosterone concentrations before the stress sessions• Sam-
ples were drawn immediately upon removal of the rats from their cages. Day 1 was
the first stress day. Both groups had significantly elevated basal corticosterone levels
on day 7, and the levels were also elevated in the restrained rats on days 2 and 3. The
prone groups consisted of 12 rats, whereas there were 6 control rats. The
means_+standard errors are presented in all figures.
neurophysiologists can be extended to try to understand an
organism's response to repeated presentation of a stressor.
The methodologies used by the neurophysiologists also
provide a guide for the best ways to study habituation
and sensitization. Our current information about these
phenomena has been derived largely from studies measur-
ing an individual subject's responses to repeated presenta-
tion of the same type of stimulus over time. Additionally,
response in these studies is defined as the change between
control levels of the respondent and its levels following pre-
sentation of the stimulus [23]. In contrast, the difficulty of
repeatedly assessing visceral responses to stress has often
forced experimenters to measure the stress response in
groups of animals, each of which has a different history of
exposure to the stressor [ 17]. The current studies applied the
methods and approaches used by the neurophysiologists to
determine how stressor intensity affected corticosterone re-
sponses when the stressor was repeatedly administered
over time.
Finally, characterizing habituation or sensitization during
repeated exposures to stressors is relevant to the question of
what constitutes an adaptive response to stress. Most re-
searchers would agree that the physiological response to a
particular stress is adaptive in that it allows the individual to
cope with the stress and return to a state of normalcy (e.g.,
[ 19]). However, it is clear that repeated exposures to stress-
ors can be debilitating and produce disease ranging from
hypertension [7] to gastric pathology [15] to adrenal hyper-
trophy [1]. The ultimate issue in research on chronic stress is
to determine how the organism can best adapt to it. Are
individuals healthier if they continue to respond to each
PITMAN, O T T E N W E L L E R AND N A T E L S O N
l ....Contzols
I
f
3 4 5 6
Stress Day
occurrence of the stressor or are they healthier if they cease
responding? Do increasing tonic levels of stress respondent
prepare an individual to cope with a subsequent stressor? On
the other hand, these changes could also be responsible for
the pathology that develops with chronic stress. These issues
are ones we are beginning to address in our laboratory.
EXPERIMENT 1
Young adult male Sprague-Dawley rats (175-200 g) were
obtained from Taconic Farms (Germantown, NY) and
allowed to acclimate to our laboratory conditions for 2
weeks. These included a 12 hour photoperiod with the onset
of light at 0600 hr and maintenance of temperature at
22-2°C. Rats were housed individually in plastic cages with
bedding for 1 week before the experiment was begun. Except
during experimental manipulations, rats were allowed ad lib
access to tap water and Purina Rat Chow.
The stressor used in this experiment was immobilization,
which has been described in detail elsewhere [16]. Briefly,
rats were transferred to another room and restrained in a
prone position by taping their outstretched limbs to a board.
To control for handling and the novelty of moving the rats to
another room during restraint, control rats were placed into
large opaque plastic pails (approximately 12 inches in diam-
ter and 14 inches high) with bedding material. They were
transferred to and remained in the restraint room for the
same time as the restrained rats.
Between 0800 hr and 0900 hr, rats were removed from
their home cages, sampled by tail clip, and then either re-
strained or placed in buckets. Rats were resampled one hour
after being moved to the restraint room. After two hours of
C H R O N I C STRESS A N D H A B I T U A T I O N
40.0'
30.0'
4)
e-
p
_o
0
0 20,0'
m
E
m
m 10.0 '
x-
1 2
FIG. 2. Corticosterone levels 1 hour after the initiation of prone restraint or 1 hour
after being placed in the buckets and moved into the same room as the stressed rats.
Corticosterone concentrations reached the same levels in the restraint group over
the 7 days, but the levels in the controls declined.
restraint, the restrained rats were released, and all rats were
returned to their home cages. This procedure-was repeated
for 7 consecutive days. This duration was selected on the
basis of behavioral observations which were made as the
study was being conducted. Over the course of the first few
sessions of restraint, the rats showed a pattern of decreasing
behavioral arousal, defecation and vocalization. Thus, the
number of restraint sessions was limited to seven in this ex-
periment. Rats were weighed on the day before the initial
restraint and after 5 sessions. Initial sampling was begun
between 0800 hr and 0900 hr to minimize the influence of
circadian plasma corticosterone rhythms. The basal blood
sample was obtained within 90 seconds of removing the
rats from their home cages in o r d e r to obtain unstressed
corticosterone levels. Blood samples were collected into
heparinized microhematocrit tubes and centrifuged, and the
plasma was stored frozen at -40°C until assayed for corti-
costerone. The corticosterone radioimmunoassay has been
described elsewhere in detail [14]. The minimal detectable
dose was 0.5 /zg/dl while the intra- and inter-assay var-
iabilities were both less than 10% in this assay procedure.
The large number of samples generated in this work required
multiple assays. In each assay, all of an individual animal's
samples were run. Additionally, an equal number of restraint
and bucket control samples were included in each assay.
Plasma corticosterone data were analyzed using a re-
peated measures Analysis of Variance (ANOVA) for un-
equal group sizes [27]. A priori comparisons were made
using Dunn's tD statistic [9], and post hoe comparisons were
made using Fisher's Protected Least Significant Differences
procedure (PLSD) [9]. The individual comparisons were
considered significant if p<O.05. The significance o f re-
sponses on day 7 was assessed with paired t-tests between
the basal and one hour corticosterone levels. The body
49
3 4 5 6 7
Stress Day
weight gains per day were calculated for both groups and
compared with a Student's t-test. Because we expected the
weight gain to be less in the stressed animals, a one-tailed
t-test was used. All values are presented as means +- stand-
ard errors of the mean.
Results
Basal corticosterone levels. Basal corticosterone levels in
restrained and control rats are presented in Fig. 1. There was
a significant interaction between sampling day and treatment
group; F(6,93)=3.17, p<0.01. Basal corticosterone levels
in the restrained rats on days 2 and 3 were elevated over
those seen in the controls on these days and over basal levels
prior to restraint on day 1; PLSD, p<0.05. Although the
levels on days 4--6 were not significantly elevated over those
on day 1, by day 7 both restrained and control rats had
significantly elevated basal corticosterone levels compared
with their respective basal levels on day 1; P L S D p<0.05.
Post-stress corticosterone levels. Plasma corticosterone
concentrations after one hour of restraint or one hour after
the controls were placed in buckets are presented in Fig. 2.
On day 1, these levels were similar in the two groups. Addi-
tionally, the corticosterone levels achieved during restraint
were similar over the 7 stress sessions; F(6,66)= 1.26, p >0.1.
However, the corticosterone levels in the controls declined
significantly as they were repeatedly placed in the buckets;
F(6,41)=2.94, p<0.05). This decline occurred after one ex-
posure to the novel situation (PLSD, p<0.05), and there was
no further decline from days 2 to 7. On day 7, the one hour
corticosterone levels in both restrained and control rats were
greater than the basal corticosterone levels on this day.
Corticosterone responses. Corticosterone responses were
calculated by subtracting basal levels from those measured
 50
40.0
30.0
G)
o
._8 20.0 '
1::
o
(.1
CU
E
(0 10.0 "
Ca
O.
1 2
FIG. 3. Corticosterone responses in prone restraint and bucket control rats over 7
days of stress. The corticosterone responses in both groups habituated, but there
were still significant responses on day 7.
one hour into stress or one hour after the controls had been
placed in their buckets. These data are presented in Fig. 3.
The corticosterone responses declined in both groups over
the course of the experiment [Restrained: F(6,66)=5.92,
p <0.001; Controls: F(6,30)=4.38, p<0.01]. Because the one
hour corticosterone levels in both groups were relatively
stable over the 7 days (Fig. 2), the changes in response over
the course of the experiment reflected primarily changes in
basal corticosterone levels (Fig. 1). Responsiveness declined
smoothly in the controls, whereas the responses in the re-
strained group were the mirrored image of the changes in
basal levels seen in Fig. 1, i.e., there were significantly lower
responses in the restrained rats on days 2, 3, and 7 than on
days I, 4, 5, and 6; PLSD, p<0.05. Thus, both groups
showed significant decrements in responsiveness over the 7
days of treatment, but they still showed significant cortico-
sterone responses on day 7 [Restrained: t (11 ) = 8.31, p < 0.001;
Controls: t(5)=3.45, p<0.02].
Body weight. Stressed rats gained significantly less body
weight (1.44_+0.77 g/day) than the controls (3.79_-_0.77 g/day)
over the first 6 days of this experiment, t(16)= 1.92, p<0.05.
Control rats showed a normal pattern of weight gain for rats
of this size with about 4 grams being gained per day during
the experiment. However, the stressed rats gained less than
half as much weight as the controls.
Discussion
The stressed rats did not show measureable changes in
the stress levels of corticosterone after a week of repeated
prone restraint, even though basal levels increased. On the
other hand, the controls showed a decline in their corticoste-
rone levels over the 7 sessions in the buckets. However,
both groups continued to show responses to their treatments
for the full 7 days. The controls showed the same corticoste-
PITMAN, O T T E N W E L L E R AND NATELSON
P=one I
Control8
3 4 5 6 7
Stress Day
rone response as the prone restrained rats on the first day of
stress. It is not clear if such a large response would have
occurred in the control rats had they been kept separate from
the restrained rats. However, by the second day, some of
this response in the controls had attenuated, and the one
hour corticosterone levels on day 2 in the controls were less
than those on day 1. The response in the restrained rats also
declined, but this was due to increased basal levels on day 2
and not to decreased restraint levels. Thus, the declines in
responsiveness shown in Fig. 3 are similar for the 2 groups
on days 2 and 3, but they are due to different effects of
repeated stress. The controls showed declining one hour
levels, whereas the restrained rats showed increased basal
levels. Although the restrained rats showed greater respon-
siveness on days 4 and 5, the 2 groups showed similar re-
sponses on day 7.
There are 2 ways to interpret these data. First, it is possi-
ble that the corticosterone response to even this relatively
intense stressor may have been primarily a novelty response.
This is supported by the fact that we found no greater re-
sponse to restraint than to novelty until days 4 and 5, and
even this difference in responsiveness disappeared by day 7.
This may mean that the decrements in responsiveness in the
restraint groups were due to declines in the novelty of the
restraint procedure. If this was the cause, then the restraint
was no more stressful than novelty in these rats. This con-
clusion is unlikely because of the obviously more excited
behavior engendered by restraint, as opposed to that caused
by the novelty of being placed in buckets.
However, the second, more likely possibility is that the
bucket controls were stressed by more than just novelty be-
cause they were in the same room as the restrained rats. The
restrained rats vocalized and urinated during stress which
may have provided intraspecific signals that resulted in in-
creased stress in the control rats. This is supported by the
CHRONIC STRESS AND H A B I T U A T I O N
TABLE 1
PLASMA CORTICOSTERONELEVELS IN CONTROLRATS
Day 1 Day 3 Day 5
pre post pre post pre post pre
Mean 2.5 4.0 3.6 3.8 1.4 2.5 1.4
(/zg/dl)
_+SEM 0.42 0.83 1.47 0.38 0.37 0.67 0.39
fact that there were gradually increasing basal levels of corti-
costerone in the novelty controls as they were exposed to
more sessions with the stressed rats, whereas the stressed
rats showed immediate increases in basal levels as early as
day 2.
Because there was a decline in corticosterone levels and
responses in the bucket controls and this stress was milder
than the restraint stress, we reasoned that by decreasing the
intensity of restraint as a stressor we might find even greater
declines in these glucocorticoid measures. In addition, basal
corticosterone levels were still rising on the last sampling
day in this experiment, which suggested that further changes
might have occurred if the experiment were carried on
longer. Thus, we performed a second experiment in which a
milder stressor, tube restraint, was used, and the rats were
exposed to this stressor for 3 weeks. In this experiment, we
again included novelty controls in buckets, but there was
also a sampling only control group to determine how much of
the response in bucket controls was simply due to drawing
repeated blood samples via tail clip.
EXPERIMENT 2
Young adult male Sprague-Dawley rats (200-275 g) were
maintained as described for the first experiment. The re-
straint for these rats consisted of placing them in tubes (PVC
tubing, outside diameter=2 inches) and restricting their for-
ward and backward movement with Plexiglas partitions.
Animals which served as novelty controls were again placed
in buckets with bedding material in the same room as the
restrained rats. At the same time, a third group of rats was
kept in their home cages and only sampled. For 21 consecu-
tive days, the restrained rats were placed in tubes and trans-
ferred to the restraint room, and the novelty controls were
placed in buckets and transferred to the same room. The
sampling only rats remained in their home cages except dur-
ing sampling. Blood samples were drawn from all animals on
the initial stress day (day 1) and on days 3, 5, 7, 10, 14, 17,
and 21. The basal blood sample was drawn by tail clip im-
mediately after each animal was removed from its home
cage, and the second sample was obtained one hour after the
rats had been restrained, placed in buckets, or the initial
blood sample had been drawn in the sampling only group.
After 2 hours in restraint tubes or buckets, the rats were
returned to their home cages. Body weights were obtained
on all rats before the repeated stress sessions began and after
5 to 20 sessions.
Because the results from the first experiment were avail-
able before this experiment was run, we planned a priori
statistical comparisons to analyze the data from the second
experiment. These comparisons involved determining
whether the basal or stress corticosterone levels changed
from day 1 and whether the stress levels were higher than
51
Day 7 Day 10 Day 14 Day 17 Day 21
post pre post pre post pre post pre post
2.6 3.4 6.2 1.3 3.3 6.4 4.2 1.0 2.8
0.57 1.63 4.64 0.40 0.61 4.33 1.03 0.22 0.67
controls on any day. D u n n ' s tD statistic was used as it is
appropriate for a priori multiple comparison of means [8].
Body weight changes were again analyzed with Student's
t-tests.
Results
Sampling only controls. The results for the sampling only
control group are presented in Table 1. It can be seen that
this sampling protocol produced no elevations in corticoste-
rone by itself; neither an increase in the one hour values due
to the basal sample nor an increase in either basal or one hour
values occur over the 21 days of the experiment. All means in
Table 1 are in the range of basal levels for rats in our labora-
tory. They are also well below the range of stress values that
we have measured, and thus they indicate that our sampling
protocol was not, in itself, stressful to the rats.
Basal corticosterone levels. The corticosterone concen-
trations for the tube restrained rats and their bucket controls
are presented in Fig. 4. By day 10, the tube restraint group
had higher basal corticosterone levels than their bucket con-
trols; tD =2.5, p<0.05.
Post-stress corticosterone levels. The corticosterone
levels after an hour of tube restraint or an hour after the
controls had been placed in buckets are presented in Fig. 5.
Corticosterone levels during tube restraint did not change
over the entire 21 days; F(7,35)= 1.00, p>0.1. In the bucket
control rats, there was a significant decline in one hour corti-
costerone levels between days 1 and 3 (tD =8.5, p <0.01), and
no further decline occurred during the rest of the experi-
ment. As in the first experiment, the corticosterone levels
were similar during tube restraint and the novelty of the
buckets on the first sampling day. At this time, the corticoste-
rone levels were approximately half those seen during the
initial prone restraint in the first experiment. However, after
the third day the levels in bucket controls were not elevated
above basal levels on day 1 even after being placed into
buckets and transferred to another room. This is in contrast
to the results for the first experiment, in which the bucket
controls continued to respond throughout the 7 days of the
experiment.
Corticosterone responses. The corticosterone responses
to mild tube restraint and being placed in buckets are pre-
sented in Fig. 6. This figure emphasizes the point that there
were no significant responses in the bucket controls (not
different from 0) except on day 1. In this experiment, there
were also no differences in the responses of corticosterone to
tube restraint over the 21 days, F(7,35)= 1.12, p >0.1. This is
in contrast to what was seen in the flu'st experiment where
there was a decline in responsiveness over 7 days (see Fig.
3). The responses on day 7 were similar in the stress groups
from both experiments, and this level of response (about 20
/xg/dl) was maintained until day 21 in the second experiment
52 PITMAN, OTTENWELLER AND NATELSON

12.0

a) 8.0"
¢::
o
m
o
.u_
1::
O
O
4.0"
1=
t~

D,.

, , , , , , , , , , . , , , . . . , , , ,

1 3 5 7 10 14 17 21

Stress Day

FIG. 4. Basal plasma corticosterone concentrations before the tube restraint sessions. The
tube restrained rats showed elevated basal corticosterone levels with repeated stress ses-
sions, w h e r e a s there were no differences in the controls. Both groups consisted o f 6 rats.

Controls

40.0

30.0

.o

20.0

E
m(~ lO.O
x"

, , , , , , , , , , , , , , , , , , , , ,

1 3 5 7 10 14 17 21

Stress Day

FIG. 5. Corticosterone levels 1 hour after the initiation of tube restraint or 1 hour after
the basal samples were drawn in the control group. There were no differences in the
corticosterone levels in the tube restrained rats over the 21 d a y s of stress, w h e r e a s the
levels in the controls declined significantly b e t w e e n days 1 and 3. After day 3, t h e s e
levels reflected basal corticosterone values and the a b s e n c e o f stress.
C H R O N I C STRESS A N D H A B I T U A T I O N
30.0
10.0
J 0,0
-10.0 , i , • , , , ,
1 $ 8 7
FIG. 6. Corticosterone responses in tube restrained rats and their bucket controls over 21
days of stress. The corticosterone response disappeared completely (was 0) in the controls
by day 3, whereas there was no change in the response to tube restraint over the 21 days.
with tube restraint. Since the responses appeared to have
reached an asymptote, the data may indicate that declines in
corticosterone responses following exposure to moderately
strong stressors does not occur. This is in contrast to other
hormonal stress respondents such as prolactin which
habituate quickly and completely to even intense stressors
[28]. The difference in rates of habituation between these
two stress respondents suggests the existence of different
control mechanisms.
Body weight. In this experiment, the tube restrained rats
had similar weight gains (3.48+_1.13 g/day) as the controls
(5.33-+1.15 g/day) over the first 6 days of the experiment,
t(10)=1.15, NS. both these weight gains were similar to
those in the controls during the first experiment. However,
over the last 2 weeks of this experiment, the controls contin-
ued to gain weight at this rate (3.50+-0.79 g/day), whereas
the tube restrained rats gained significantly less weight
(1.92--_0.26 g/day) during this time, t(10)= 1.9, p<0.05. This
level of weight gain in the tube restrained rats was similar to
that seen during the week of prone restraint stress in the first
experiment.
Discussion
In the second experiment, we demonstrated that our
sampling procedure itself was not stressful even when used
repeatedly. It should be noted that we take much care not to
disturb the rat during this sampling by being very quiet in the
animal rooms, handling the rats gently, and using minimum
pressure on the tail to draw the blood. In addition, only 1-2
mm of tail are clipped, and blood can often be made to flow
after it has stopped by flicking the scab off the tip of the tail.
Often, a rat will only have to have its tail clipped once,
unless it is sampled less frequently than twice a week. This
protocol has allowed us to repeatedly sample rats to assess
53
Contro3.8
i
,L 4
, , , , , t t , , , , • i
10 14 17 |1
their individual responsiveness to repeated presentations of
a stressor without the implantation of chronic indwelling
catheters or using one of the other methods which in and of
itself is a real stressor (e.g., cardiac puncture, orbital sinus
sampling).
GENERAL DISCUSSION
From neurophysiological studies of habituation we know
that the less intense the stimulus, the more pronounced is the
habituation [23]. And when intense stimuli are used, there
may be no habituation at all [25]. A goal of these experiments
was to determine how well the same rules applied when the
stimuli were stressful and when plasma corticosterone levels
were used to monitor habituation. When only the post-stress
levels of corticosterone were used as the measure of habitu-
ation, neither of the restraint groups showed evidence of
habituation, In another experiment, less stressful procedures
were used to immobilize rats, and we found habituation of
the corticosterone response in these rats [12]. Thus, we be-
lieve our not finding decreases in post-stress hormone con-
centrations in this study was due to the intensity of the stim-
uli used here. However, when we used the change from basal
corticosterone levels to levels obtained after 60 minutes of
stress as our measure, i.e, the corticosterone response; the
four limb restraint group habituated but the tube restraint
group did not. Thus our data indicate that the interpretation
of an experiment designed to monitor habituation using cor-
ticosterone levels depends on which measure is used. Be-
cause it is not clear which of these measures is better for
assessing habituation, we feel that examining both hormone
response as well as post-stress hormone levels will be impor-
tant in future studies designed to understand decrements in
the stress response over time.
Not finding habituation in the rats receiving the less in-
54
tense restraint stress and finding partial habituation in the
more intensely stressed rats is the opposite of what one
might expect. Others [6] have also not found decrements in
corticosterone levels following mild stressors. However, had
our tube stressed rats showed less variable responses, we
believe habituation would have been seen in that group also.
(Note that the clear cut decreases depicted for the animals in
Fig. 6 from days 1 to 5 are offset by more variable responses
later).
The discussion above ignores the obvious descrepancy
we found between the corticosterone responses in our 2
groups of bucket controls. In the controls accompanying the
less intense stressor of tube restraint, the corticosterone re-
sponse completely habituated. But in those accompanying
the more intense stressor of four limb restraint, the corticoste-
tone response persisted for 7 days. The difference between
these 2 sets of results suggests that some type of communi-
cation existed between the four limb restraint rats and their
controls. Experimental designs in stress research usually
separate effects of stressor from confounding influences of
extraneous variables by incorporating a set of control
animals which are treated like experimental ones except they
are not stressed. Our results indicate that this method is
potentially confounding because the stressed animals them-
selves can influence controls and appear to do so in a
treatment-dependent fashion.
This influence likely represents pheromonal and/or vocal
communication between stressed and control rats.
Pheromones play a role in sexual readiness and sexual be-
havior as well as in communicating the stress state in mice
[4,18]. Additionally, a stress pheromone can produce
changes in activity and impairments in learning in otherwise
normal controls [10,11]. Rodents also appear to use intensity
of vocalization to communicate their level of arousal [3];
these vocalizations are aversive in that rats will work to
avoid them [21]. Finally, communication from stressed con-
specifics can produce an adrenocortical stress response in
nonstressed rats [5]. Because stressed and control rats were
maintained in the same room, out data were probably influ-
enced by conspecific communication of stress.
Smotherman et al. [22] have shown that the levels of cor-
ticosterone in mothers reunited with pups is related to se-
verity of stress imposed on the pups. Our data also suggest
that stressor intensity can also be communicated. The
bucket controls for the rats subjected to the more intense
four limb restraint showed a greater corticosterone response
and one which did not fully habituate, whereas the bucket for
the rat subjected to the less intense stressor fully habituated.
This observation is relevant because it suggests that intra-
specific interaction may be an important variable in charac-
terizing habituation or sensitization in chronic stress
paradigms, both in controls and stressed animals. Finally, this
observation may have practical importance to stress
physiologists as animals simply housed in the same room as
stressed rats could be used to study stress in a situation
where the experimental animal has never been subjected to a
physical stressor.
The data from these experiments also relate to the prob-
lem of defining chronic stress. Adrenocortical activation has
been associated with stressors since the early writings of
Selye [20], but Selye's data primarily address initial re-
sponses to stressor. And surprisingly, when Selye studied
the morphological appearance of the adrenal gland following
repeated presentations of the stressor, he found that the
PITMAN, O T T E N W E L L E R AND N A T E L S O N
gland's appearance had returned to normal [20]. Thus, a crit-
ical remaining question concerns the association between
adrenocortical activation and chronic stress. One definition
might be continued glucocorticoid activation during each
presentation of the stressor. Our own data and those of
others [8] show a continued adrenocortical activation consis-
tent with this definition. However, a question that remains is
whether this response would eventually disappear if enough
stress sessions were presented.
Because the process of habituation appears to affect the
immediate response of the adrenal cortex to the repeated
presentation of a stressor, a more compelling definition of
chronic stress might be finding elevated glucocorticoid levels
when the stressor was absent. That was the case in this
work. After several days of exposure to the more intense
stressor of four limb restraint, plasma corticosterone levels
prior to stress were elevated over the levels measured before
the first exposure to the stressor. These elevated basal levels
were measured more than 20 hr after the prior stress and
demonstrated chronic elevation of corticosterone when no
stressor was present. We have shown in previous work [14]
that these elevated basal levels cannot be attributed to a shift
in the peak of the circadian corticosterone rhythm.
Additionally, our data suggest that the more intense the
stressor the quicker chronic stress develops. Rats exposed to
four extremity restraint developed elevations in basal plasma
corticosterone quickly, rats exposed to tube restraint did so
less quickly, bucket controls for the four extremity group
even lesss quickly and the bucket controls for the tube re-
straint not at all. This interpretation of the data opens again
the question of the relation between adrenocortical hor-
mones and stressor intensity. Natelson et al. [13] have
shown that these hormones do not vary linearly with the
intensity of singly presented stressors. However, the data
presented in this paper suggest that adrenocortical function,
when assessed in the basal state, may relate in some quan-
titative way to the degree of chronic stress. Further research
must be done to substantiate this hypothesis.
These experiments suggest that habituation and chronic
stress must be viewed as related processes. An animal which
has habituated to repeated stressor presentation is thought to
be better "adapted" to cope with stress [19]. But we found
evidence for both partial habituation and chronic stress in
the same animal in that the intensely stressed rats showed
habituation of the corticosterone response at the same time
that baseline levels were tonically elevated. This is different
from the neurophysiologist's notion that habituation is an
adaptive mechanism for shutting out repetitive, redundant
stimuli [23]. Nonetheless, a crucial issue that requires further
study is how habituation and chronic stress interact with one
another. Apart from this issue, our data make it clear that
chronic stress can develop; this fact may have clinical rele-
vance. Using an animal model such as ours to study the
development of chronic stress and habituation may be im-
portant in defining who falls victim to the complex clinical
anxiety disorder known variably over the years as shell
shock, combat fatigue, and most recently as the chronic
stress syndrome.
ACKNOWLEDGEMENTS
This research was supported by VA Medical Research Funds.
We wish to thank Dr. Walter Tapp for his helpful and informative
comments regarding this manuscript.
CHRONIC STRESS AND HABITUATION
REFERENCES
1. Bassett, J. R. and K. D. Cairncross. Morphological changes
induced in rats following prolonged exposure to stress. Phar-
macol Biochem Behav 3: 411-420, 1975.
2. Bassett, J. R., K. D. Calrncross and M. G. King. Parameters of
novelty shock predictability and response contingency in corti-
costerone release in the rat. Physiol Behav 10: 901-907, 1973.
3. Bell, R. W. Ultrasounds in small rodents; arousal-produced and
arousal-producing. Dev Psychobiol 7: 39-42, 1974.
4. Carr, W. J. amd W. F. Caul. The effects of castration in rats
upon discrimination of sex odors. Anita Behav 10: 20-27, 1962.
5. Fuchs, E., G. Flugge and H. D. Hutzeimeyer. Response of rats
to the presence of stressed conspecifics as a function of the time
of day. Horm Behav 21: 245-252, 1987.
6. Hennessy, M. B. and S. Levine. Effects of various habituation
procedures in pituitary adrenal responsiveness in the mouse.
Physiol Behav 18: 799-802, 1977.
7. Huddack, W. J. amd J. P. Buckley. Producing hypertensive rats
by experimental stress. J Pharmacol Sci 50: 263-264, 1961.
8. Kant, G. J., B. N. Bunnel, E. H. Moughey, L. L. Pennington
and J. L. Meyerhoff. Effects of repeated stress on pituitary
cyclic AMP, and plama prolactin, corticosterone and growth
hormone in male rats. Pharmacol Biochem Behav 18: 967-971,
1983.
9. Kirk, R. E. Experimental Design: Procedures for the Behav-
ioral Sciences. Monterey, CA. Brooks/Cole, 1982.
10. Makay-Sim, A. and D. G. Laing. Rat's response to blood and
body odors of stressed and non-stressed conspecifics. Physiol
Behav 27: 503-510, 1981.
11. Minor, T. R. and V. M. LoLordo. Escape deficits following
inescapable shock: the role of contextual odor. J Exp Psychol
10: 168-181, 1984.
12. Natelson, B. H . , J. E. Ottenweller, J. A. Cook, D. L. Pitman,
R. McCarty and W. N. Tapp. Effect of stressor intensity on
habituation of the adrenocortical stress response. Physiol Behav
43: 41-46, 1988.
13. Natelson, B. H., W. N. Tapp, J. E. Adamus, J. C. Mittler and
B. E. Levin. Humoral indices of stress in rats. Physiol Behav
26: 1049-1054, 1981.
14. Ottenweller, J. E., D. L. Pitman and B. H. Natelson. Repeated
stress at the same time each day does not mimic timed feeding in
its effects on the plasma corticosterone rhythm in rats.
Chronobiologica 14: 1-6, 1987
55
15. Par6, W. and L. J. Temple. Food deprivation, shock stress and
stomach lesions in the rat. Physiol Behav 11: 371-375, 1973.
16. Pitman, D. L., J. E. Ottenweller and B. H. Natelson. Method-
ological problems in the study of classical aversive conditioning
of adrenocortical responses. Physiol Behav 38: 677--685, 1986.
17. Quirce, C. M., M. Odio and J. M. Solano. The effects of predic-
able and unpredictable schedule of physical restraint upon rats.
Life Sci 28: 1897-1902, 1981.
18. Rottman, S. J. and C. T. Snowdon. Demonstration and analysis
of an alarm pheromone in mice. J Comp Physiol Psychol 81:
483--490, 1972.
19. Sakellaris, P. C. and J. Dernikos-DaneUis. Increased rate of
response of pituitary-adrenal system in rats adapted to chronic
stress. Endocrinology 97: 597-602, 1975.
20. Selye, H. Studies on adaptation. Endocrinology 21: 169-188, 1937.
21. Simonov, P. V. Avoidance reactions to tape-recorded pain cry
in rats. Neurosci Behav Physiol 13: 87-88, 1983.
22. Smotherman, W, P., S. G. Weiner, S. P. Mendoza and S.
Levine. Maternal pituitary adrenal responsiveness as a function
of differential treatment of rat pups. Dev Psychobiol 10:113-
122, 1977.
23. Thompson, R. F. amd D. L. Glansman. Neural and behavioral
mechanism of habituation and sensitization. In: Habituation,
edited by T. J. Tighe and R. N. Leatron. Hilldale, NJ: Lawrence
Erlbaum Associates, 1976.
24. Thompson, R. F., P. M. Groves, T. J. Teyler, R. A. Roemer. A
dual process theory of habituation: theory and behavior. In:
Habituation: Behavioral Studies, edited by H. V. S. Peeke and
M. J. Herz. New York: Academic Press, 1973.
25. Thompson, R. F. and W. A. Spencer. Habituation: A model
phenomenon for the study of neuronal substrates of behavior.
Psychol Rev 173: 16--43, 1966.
26. Weiss, J. M. Somatic effects of predictable and unpredictable
shock. Psychosom Med 32: 397-408, 1970.
27. Winer, B. J. Statistical Principles in Experimental Design. New
York: McGraw-Hill, 1971.
28. Yelvington, D. B., G. W. Weiss and A. Ratner. Habituation of
the prolactin response in rats to stress. Psychoneuroendocrinol-
ogy 10: 95-102, 1985.