EFFECTS GE: RE95ATED §E$$EGNS 03%
RESPQNSE EQ-‘E‘ZQEMENT CF A
HEAD‘SHAKE RESPQRSE

”Elma fee the Dogma 0‘; M. A.
MICHEGAH STATE UNWERSETY

Bruce C. Leibrecht

1967

2: "J LIBRARY
THESIS Mnchigan State
University

  
    

ABSTRACT

EFFECTS OF REPERTED SESSIONS UN RESPONSE DcCREflENT
OF A HEAD-SHAKE RESPCHSE

by Bruce C. Leibrecht

Habituation (defined as response decrement accompanying repeated
stimulation) is a widc5pread phenomenon occurring across the entire
phylogenetic range, and as such it is an extremely useful behavior pro-
cess For the comparative method of analysis. Since habituation is gen-
erally regarded as a Form of learning, retention is a characteristic
of prime concern in the study of habituatory reaponse decrement. The
bulk of the existing evidence indicates that habituation of a wide
variety of reaponses is retained For at least 24 hrs., and often For
periods up to a week or more. At the same time, many of the Findings
pertaining to habituation are somewhat contradictory, indicating the
need For an adequate preparation for the detailed analysis of the process.

The present study was designed with the foregoing considerations
in mind. The major problem investigated was the retention of habitua-
tion of the head-shake response (a rapid rotation, or twisting, of the
head about the front-to-rear axis) in the laboratory rat. A second-
ary aim was the further exploration of a promising preparation For the
study of habituation.

Eight male and eight Female Sprague-Dawley rats, 30 to 60 days of
age, served as subjects. Following an initial pretest to screen out
"non-reSponders", gs were randomly assigned to an EXperimental or Con-

trol group. Subjects in the Experimental Group received 14 standard

oruce C. Leibrecht

test sessions, 12 and 24-hr. intervals and the last two at 6-hr. inter-
vals. Each test session consisted of 40 30-sec. presentations of the
eliciting stimulus (a stream of pressurized air) in the left ear, 30
sec. of no stimulation intervening between presentations. Control gs
were simply observed for 12 sessions comparable in every reSpect to
those for the EXperimental Croup except for the absence of the elicit-
ing stimulus. The Control Group served to establish the base rate of
the head-shake response.

The results for the EXperimental Croup indicated that, while the
head-shake reaponse exhibited substantial decrement within each
session, the decrement was not retained for 24 hrs. Only a slight
suggestion of retention was observed with a 6-hr. inter-session inter-
val. The finding of no retention, although not unique, is contrary to
a considerable body of previous findings. whether or not the reaponse
decrement was habituatory in nature is therefore questionable.

The head-shake response was found to have a definite base rate,
approximately one-sixth as great as the stimulated response level.
Females in the Control Group had a significantly higher reaponse rate
than males, although there were no significant sex differences for the
Experimental CrOUp. Scores for both grOUpS exhibited a high degree of
reliability across sessions. While head-shake reSponses usually occurred

within the context of other on—going behaviors, no predominant pattern

flQzflfie W

C irman I

of responding emerged.

mam)
I

Dates

EFFECTS OF REPEATED SESSIDRS ON RESPONSE DECREMENT

OF A HEAD-SHAKE RESPONSE

By

Bruce C. Leibrecht

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1967

Cams:

2~1!‘C\K

ACKNOWLEDGMENTS

The author wishes to eXpress his gratitude to Dr. S. C. Ratner,
uhose inspiration and guidance were of inestimable value during all
stages of research and writing. Thanks are also due to Drs. David
Raskin and Ralph Levine for their careful reading of the manuscript
and their critical suggestions. Finally, the author wishes to thank
Henry Askew, whose thoughtful suggestions and criticisms were greatly

appreciated.

ii

INTRODUCTION

[SE-[HUD o o 0

RESULTS . .

DISCUSSION .

REFEREHCES .

APPENDIX . .

OF CONTENTS

iii

Page

16

42

53

LIST OF TABLES

Table Page

1. lntercorrelations (rankuordcr) among total head-

shake reaponses per session, Group E . . . . . . . S2
2. Intercorrelations (rank-order) among total head-

shake reSponses per session, Group C . . . . . . . 33
3. behaviors preceding and following head-shake

reopenses for sessions 1 and 12 . . . . . . . . . SS
4. Amount of time (in sec.) sp,nt moving the head,

gnas ing, and VOSHLHQ, sessions 1 and 12 . . . . . 37
S. Intercorrelations {rank-order) among behavioral

measures, sessions 1 and 12 . . . . . . . . . . . 41

Figure

4.

(.11

LIST OF FIGURES

umber of stimula

hake responses
n81,7,128nd14.........

t

ed and non-stimulated
by S-min. blocks for

Hedian number of sti mulate ed and non-s imulated
ses per session . . . . .

head-

shake re apon

Wdian number of st mulate d head-shake resoonses
during the first and last stimulated S-min.

blocks

per sec

'7
oO-sec.

of each s or

ber 01 s
sion (C
stimul

L
F‘
O
3
.
0

ed head-s haLe responses
for each half of each

ZO

K)
h)

INTRODUCTION

Habituation, defined operationally as response decrement accom-
panying repeated stimulation, has for some time been recognized as a
process of considerable significance. Humphrey (1933), Thorpe (1963),
and Ratner and Denny (1964) regard it as the simplest form of learn-
ing. One of its most striking features is its generality. Humphrey
(1933), in the first comprehensive discussion of habituation, cites
examples from a wide range of phylogenetic levels. In the first sys-
tematic review of literature dealing with habituation, Harris (1943)
includes cases ranging from one-celled animals to the highest verte-
brates. Amoeba, hydra, planaria, earthworms, snails, spiders, turtles,
rats, and humans represent only a few of the Species in which habitua-
tion has been observed. The broad occurrence of the phenomenon has
led Thorpe (1963) to make the statement that "something like it (ha-
bituation) is universal in animals."

The most widely accepted definition of habituation has been sum-
marized by Harris (1943) as "response decrement as a result of repeated
stimulation." In spite of the appealing simplicity of this statement,
a number of qualifications have been deemed necessary by various in-
vestigators. Both Thorpe (1963) and Hinde (1954), for example, main-
tain that response decrement must be relatively permanent in order to
qualify as habituation. However, Thompson and Spencer (1966) take
exception, charging that such a qualification is arbitrary since the
time course for recovery of a habituated reaponse depends on a num-

ber of variables. Nevertheless, the notion of relative permanency

goes along with the idea that a change in behavior which is learned
should have somewhat lasting effects. Changes which disappear quickly
are generally termed either receptor adaptation or effector fatigue.
Relatively long-term retention of habituation has been reported by a
number of investigators. Hinde (1954), studying the mobbing response
of the chaffinch; Rodgers, Melzack, and Segal (1963), working with the
"tail flip response" in goldfish; and Keen, Chase, and Graham (1965),
dealing with neonatal heart rate acceleration to a tone, found that
habituation was retained 24 hours. Numerous other studies report re-
tention for longer periods of time: two to three days for certain com-
ponents of the orienting reflex in humans (Galbrecht, Dykman, Reese,
and Suzuki, 1965); several days for EEG arousal in cats (Sharpless

and Jasper, 1956); four days for the earthworm's response to vibra-
tion (Gardner, 1966); six days for the startle response in rats (Noyer,
1963); two weeks for nystagmic responses in cats (Brown, 1965; Capps
and Collins, 1965); and three weeks for the nystagmic reaponse in hu-
mans (Guedry, 1965).

Another qualification is generally considered basic to the phe-
nomenon of habituation. In order for decrement of a response to be
considered habituatory, the reaponse involved must be unlearned.
Otherwise, the decrement is labeled extinction. However, Thompson
and Spencer (1966) again have taken exception with this requirement,
choosing to equate extinction with habituation. It is thus apparent
that the various investigators concerned with habituation have occa-
sionally found themselves at odds.

Controversy, of course, is not uncommon in the study and theory

of behavior of organisms. Nevertheless, apparent contradictions and

discrepancies should be cause for concern. One of the sources of dis-
crepancies may be the great variety of preparations (animals and sit-
uations employed in the study of a specific process) and their rela-
tive differential adequacies. The problem of identifying an adequate
preparation is underscored as central to the comparative method of
analysis (Ratner and Denny, 1964). Thus, a major step in the compar-
ative method is the establishment of good preparations, these being
necessary for detailed analysis of a process. The essential elements
in a good preparation include: Species, sex and age of the organism,
the response to be studied, characteristics of the stimulus, and a
base-rate control (Ratner, 1967). Considering the relatively meager
amount that is known about habituation, plus the seeming discrepan-
cies among the findings that do exist, there appears to be a definite
need for an adequate preparation for its study.

The considerations underlying the present investigation are
therefore two-fold. The major problem addressed is the retention of
habituation: what are the effects of repeated sessions on the course
of habituation? A secondary aim is the further eXploration of a pro-
mising preparation for the study of habituation. The head-shake re-
Sponse of the rat, elicited by a stream of pressurized air, serves
as the basis of the preparation. A 30-seo. stimulus duration plus a
fixed inter-stimulus interval (151) are utilized. One group of ani-
mals is observed as a base-rate control group. In order to test for
retention, a series of test sessions are administered at 24-hr. in-
tervals. The rationales underlying these choices are presented below.

The rat was chosen as the Species for study primarily because it

is an extremely common laboratory animal. However, the rat offers

certain other advantages: a great deal is already known about the
Species, and rats are quite adaptable to laboratory study. Animals
between 30 and 60 days of age are used to insure sufficient biologi-
cal develOpment, yet limit the Opportunities for habituation prior

to the formal experiment. Both males and females are studied, Since
possible sex differences are an important consideration in establish-
ing a good preparation. The finding of sex differences, for example,
may limit the papulation from which subjects can be drawn and/or con-
tribute to the understanding of the process of habituation.

The response to receive primary attention can best be described
as a rapid rotation, or twisting, of the head about the front-to-rear
axis. However, a complete analysis of the sequence, which would in-
volve cinemagraphic techniques, has not yet been attempted. The re-
sponse, as far as is known, is unlearned, and it appears to be rather
stereotyped within and between animals. It is easily identifiable,
occurring in all-or-none fashion; rarely is there a question as to
whether or not the animal has emitted the response. The first formal
work involving the response (Askew, 1966) indicated that it has con-
siderable promise for the study of habituation.

A stream of pressurized air is employed as the eliciting stim-
ulus. It is somewhat less than 100% effective: Askew (1966) found
that it did not elicit the response in approximately 25% of the ani-
mals tested. However, an air stimulus allows the maximal degree of
control and can be manipulated with respect to intensity. Since the
occurrence of "non-reSponders" is likely, a pretest is given to all
subjects as an initial screening measure.

The value chosen for the duration of stimulus presentations

(30 see.) is by usual standards somewhat large. However, it might be
pointed out that the concept of "trials", as well as the values cho-
sen to delimit their length, is rather arbitrary. what an eXperi-
menter considers a "trial" may not represent any such thing to the
organism being tested. Studies of the effect of stimulus duration

on habituation have produced inconclusive results. Askew (1966) and
Koepke and Pribram (1966) found no significant effects when stimulus
duration was varied, although Hinde (1954), Keen (1964), and Keen,
Chase, and Graham (1965) found greater habituation with long dura-
tions. For the purposes of this study the length of the stimulus pre-
sentation is not considered critical. Rather, it is intended that
employing a 30-sec. duration will allow the process under study as
much freedom to SXpress itself as possible.

As mentioned above, the interval between stimulus presentations
will remain constant. Fox (1964) reported that while habituation oc-
curred with a fixed 181, it did not occur when the 151 was variable.
However, Askew (1966) reported that habituation was observed in each
of four conditions involving a variable 181.

As was also indicated earlier, study of the behavior of a base-
rate control group constitutes a major part of the eXperiment. Such
study is a critical element in any preparation. Before effects can
be attributed to the chosen stimulus (variable), one must establish
what occurs when the Stimulus (variable) is omitted, but the animal
is otherwise in the same test situation. Askew (1966) found the
head-Shake response to have a zero base rate. However, since the
present preparation differs in a number of reSpects from his, a base-

rate control group is necessary.

A number of behaviors in addition to the head-shake reSponse are
relevant for study. It should, of course, be obvious that any lab-
oratory eXperiment takes place against a background of on-going be-
haviors. Yet it is often the case that all but a narrow, selected
behavior are completely ignored. Conceivably, much can be gained by
sampling a larger portion of the activities of the animals under study.
For example, Gardner (1966) found that as the withdrawal response of
the earthworm to vibration habituated, the subject concurrently en-
gaged in other activities, such as feeding. The present test appara-
tus is therefore constructed so as to eliminate the necessity of re-
straining the animals. This allows the subjects to remain in a state
as close to natural as possible, giving them the opportunity to en-
gage in normal activities. Pilot studies indicate that animals in
the preposed test situation engage not only in head-shaking, but in
several other kinds of activities as well. It is of interest to ask
what effects the test sessions will have on these behaviors. At the
same time, it is relevant to ask if any of the observed behaviors cor-
relate with the occurrence of head-shake reSponses.

The selection of the length of the interval separating test ses-
sions (24 hrs.) was based on previous findings relating to the reten-
tion of habituation. The studies cited earlier reported retention
lasting from 24 hrs. to three weeks. However, it should be noted that
a few studies have failed to find retention for 24 hrs. While Meyer
(1963) reported 24-hr. retention of habituation of the startle re-
sponse in rats, using gross body movement as the response measure,
several other investigators found no such retention in dealing with

the startle reSponse. Berg and Beebe-Center (1941), using the cardiac

reSponse in humans, Lehner (1941), measuring the respiratory response
in rats, and Presser and Hunter (1936), studying the twitch of the
gastrocnemius muscle in rats, all found no retention of habituation
at 24 hrs. Finally, Askew (1966), working with the head-shake re-
Sponse of the rat, reported no retention of habituation for a 24-hr.
period. This last finding is particularly relevant, since the same

response is studied in the present investigation.

METHOD

Subjects

Eight male and eight female Sprague-Dawley rats (obtained from
Spartan Research, Haslett, Michigan) served as subjects (is). The
mean age was 49 days at the beginning of the eXperiment, with a range
mean age from 40 to 60 days. The average weight of the females was
150 g. (range from 109 to 206 g.), of the males, 185 g. (range from

r"

6 to 239 g.). All subjects were born and reared at Nichigan :tate

LO

University, and all were echrimentally naive. They had remained with
their reopective litters until approximately 30 days of age, at which
time they were transferred to small cages housing five or six animals

of the same age and sex.

0

Asgaratu
The test stand, on which S rested during all test sessions, con-
sisted of a small platform atop a narrow length of wood (see Fig. 1).
The platform, constructed of é-in. hardware cloth, was 1% in. by 4
in. by & in., open at the top such that short sides extended upward
on all edges. Attached to and hanging vertically downward from the
edges of the platform were sections of sheet aluminum, forming a "col-
lar" 9 in. long to discourage escape attempts. The platform with its
attached cellar was fastened to the top of the wooden column with two
screws, so that the entire unit could be removed for cleaning. The
wooden column was mounted on a round wooden base, 7 in. in diameter,
which in turn was attached to a 4-in. "lazy susan" ball bearing. The

bearing was fastened by its lower edge to a rectangular hard-wood

g placed here

 

 

 

 

 

 

 

29“

 

 

 

 

 

Fig. 1. Test stand. (Scale: %" = 1')

10

base, 5 in. by 11 in. Thus the entire test stand could be rotated
on its base. To facilitate rotation by hand, brass "spokes", 3% in.
long, extended outward from the wooden column at its approximate
midpoint. The total height of the stand was 29 in.

The test stimulus consisted of a stream of pressurized air from
a Silent Giant aquarium pump, Model 120. The air was delivered through
a hand-held rubber tube, 1/0 in. in diameter. A l/B-in. long section
of smaller rubber tubing was inserted in the end of the hose, effect-
ively making the epening 1/16 in. in diameter. The pressure maintained
by the pump was sufficient to diSplace a column of mercury, measuring
14.0 cm. on both sides of a manometer, a distance of 13.4 cm. It
should be noted that the sound produced by the pump was carried
through the tube, providing a complex air-tone stimulus.

An Esterline Angus ink recorder was used to record selected be-
haviors. It was eperated by a control panel, consisting of plastic
tabs, or "keys", which depressed push-button electrical switches con-
nected to the pens of the recorder. The tabs, attached to a wooden
structure, were arranged so that all could be operated simultaneously
by the fingers of one hand.

A round metal container with slightly sloping sides, 9 in. in
diameter at the bottom and 10 in. deep, was used to convey-SS to and
from the test room. A stopwatch was used for timing intervals.

All test sessions were conducted in a small test room, 8 ft. by
0% ft. (see Fig. 2). A single door with a double glass window led
into the room.. Light was provided by two panels of ever-head fluor-
escent tubes; the light intensity at the point at which.§ rested was

50 ft.-candles. The temperature during all experimental sessions

11

 

 

fl- fi_~—W

 

 

 

Desk-chair——

Keyboard K?

_.- 8| 6|! ~92
m “mlimmxsw- 1...-“ ______.--, .11, __
Vent ( 1”", 1r K
, i ! IRecorderE

"/ \ rm] 1 i

‘f'h \‘ : '

= I t

/ l ’ m ‘

(g/ H \ E lable f

' (3 ii ;
;/,//\ -m/’ 1

1

,fjiilpmfiTest g

.//////\\ EC, Stand i

\\fhair‘ 8'

/) - 7

Stool/(2::T i ;

Inverted ‘ 3

Box 1

__1lm-

Fig.

2.

 

 

  
  

 

 

 

 

 

 

 

 

 

 

Test room.

(Scale: ln = 2')

 

12

remained relatively constant at 74-76 degrees Fahrenheit. The amount
of noise filtering in from the outside was slight. An intake open-
ing for the building's ventilation system was located near the floor
of the back wall. This was the source of a constant level of back-
ground noise.

All gs remained in their regular home cages except for experi-
mental sessions. Conditions were maintained according to the estab-
lished routine of the colony room. Food, consisting of Wayne Lab
Blex, and water were available at all times. The lights in the cel-
ony room were generally off, except for occasional periods when rou-
tine maintenance tasks were being perfermed or when an animal was be-
ing procured for an experiment. The temperature remained at approxi-

mately 65 degrees Fahrenheit.

Procedure

Pretest. The initial phase of the experiment involved a pretest
for the selection os_§s. The experimenter (E) first selected at ran-
dom a cage of animals of the appropriate age and sex and marked each
animal on the tail with a felt-tip marker. Then-E chose an animal at
random and held it Up by the tail for several seconds to test for dam-
age to the inner ear. Upon finding no evidence of damage, E conveyed
the animal to the test room for administration of the pretest. After
allowing five minutes uninterrupted on the test stand, §_delivered
five 30-sec. presentations of the test stimulus directly in the ani-
mal's right ear by hand. The stimulus was moved rhythmically back
and forth, left-to-right, across the center of the ear at the rate of
approximately three complete cycles per sec. and at a distance of 3/R

to 1/2 in. The path of the stimulus lay within the borders of the

l3

outer ear and remained as constant as possible. Stimulus presenta-
tions were separated by 30-sec. periods, during which E held the end
of the tube eight to ten in. away from and out of the visual field

of the animal. During the pretest-E continuously recorded occurrences
of head-shake responses. After the animal had been on the test stand
a total of ten minutes, §_returned it to the home cage. If the ani-
mal had given less than ten head-shake responses during the five stim-
ulated periods, it was rejected. Otherwise, it became anufi and was
randomly assigned to one of two grOUps: (1) Control (base rate - no
test stimulus), or (2) Experimental (test stimulus in left ear). Four
males and four females were assigned to each group.

Tests. Test sessions began on the day following the pretest.

 

.E securedlfi by hand and placed it in the metal container. (Handling
during all phases of the eXperiment was kept to a minimum.) §_then
conveyed S to the test room, a distance of approximately 50 ft. through
an artificially lighted corridor. Upon arrival in the test room, E
placed.§ on the test stand, then turned on the recorder and simulta-
neously started a stopwatch, which was used to measure all intervals
during the session. Then_§ administered one of the following treat-
ments, depending on S's assigned group:

(1) Groug G (Control): §_remained undisturbed en the stand for
45 min. while_§ simply observed.

(2) Group E (EXperimental): After allowing S a 5-min. period
undisturbed, E delivered 40 30-sec. presentations of the
test stimulus in the left ear, in the same manner and under
the same conditions as were described above for the pre-
test. Stimulus presentations were again separated by 30
sec. on no stimulation by.E. At the end of the session S
remained on the stand for an additional 5 min. without
stimulation bylg. whenever §,meved its head an appreci-
able distance during stimulation,_§ moved his hand accord-
ingly to keep the stimulus directly in 5's ear. If an.§
turned around on the stand, E rotated the round wooden

14

base with his foot until'S was again in the proper orien-
tation, at which time the "normal" stimulation pattern
was resumed. In the event anl§ succeeded in escaping at
any point during the session, E immediately discontinued
stimulation and placed S back on the stand. Stimulation
was resumed after the next regular 30-sec. period of non-
stimulation. If.§ escaped repeatedly, E terminated the
session and returned S to its home cage.
At the end of each normal session.E returned S immediately to its
home cage.
For the entire duration of all sessions.g, operating the key-
board with his left han, recorded occurrences and/or durations of
the following behaviors via the mecaanical recorder: (l) head-shake
responses, (2) head movement, (3) gnashing, (4) washing, and (5) turn-
ing around. Head-shake responses, which were clearly identifiable,
have been described earlier. Head movement was defined as any ob-
servable movement of the animal's head, except the head-shake response;
very brief as well as extended movements were recorded. Often head
movements were accompanied by gross body movement; such behavior pat-
terns were recorded only as head movements. Gnashing consisted of
grinding together of the teeth, accompanied by movement of the lower
jaw and occasionally by audible noise. Such behavior frequently oc-
curred along with head movement, in which case both behaviors were
recorded Simultaneously. Washing, or grooming, consisted of bringing
the front limbs up to the mouth and sweeping them back and forth ever
the head and shoulders. This behavior usually involved head movement
as well as what appeared to be gnashing, the entire pattern being re-
corded singly as uashing. Turning around consisted of.§'s completely
turning its body so that it faced the apposite direction on the test

stand. When turning around occurred, head movement was recorded

simultaneously.

All gs received 12 sessions, approximately 24 hrs. apart. In
addition, the experimental gs received two further sessions, the first
at 6 hrs. and the second at 12 hrs. after the end of the twelfth ses-
sion. It was necessary to run SS in three cycles, sixugs being run
during the first cycle and five during each of the latter two. Within
each cycle the random running order remained the same for all sessions
except 12, 13 and 14. During these last three sessions the experi-
mental §s were run in the first positions. Each-S was weighed following

completion of the third session.

RESULTS

The presentation of results will proceed in the following man-
ner. Primary attention will be devoted to the head-shake (H-S) re-
sponse. The second major section will deal with the remainder of
the reSponse measures - head movement, gnashing, washing, and turn-
ing around.

Scores involving the H-5 reSponse will reflect frequency of oc-
currence. Data for all response measures within sessions will be
organized in successive blocks of 5-min. length, and nearly all with-
in session comparisons will therefore necessarily be made in terms of
5-min. units. .Nherever comparisons between groups involve session
totals, only data for the first 45 min. of each session for the g;
in Group E will be considered in order to keep the amount of time
for both groups equal. Finally, a number of comparisons between
Groups E and C will involve reSponse measures taken during "stimu-
lated" periods - those during which the test stimulus was being de-
livered. Since is in Group C never received the test stimulus after
the pretest, the term "stimulated" does not truly apply to them.
Nevertheless, when applied to Group C it will be used to designate
periods corresponding to those which were actually stimulated for
GrOUp E.

Preliminary examination of the distributions of H-S scores in-
dicated that they were skewed. Distributions for Group E were gen-
erally negatively skewed, while those for Group C were positively

skewed. In addition, the shape of the population distributions is

16

17

unknown, and the number of SS in each group is quite small (H = 4

for sex comparisons). Therefore, non-parametric techniques were used
throughout, particularly the sign test, the Mann-Whitney U-test, and
Wilson's (1956) "distribution-free" procedure for analysis of variance
hypotheses. Two-tailed levels of significance were used unless other-
wise indicated.

During the course of the eXperiment an occasional session had to
be terminated early due to repeated escapes from the test stand. This
occurred for both groups, although more often for Group E. In analyz-
ing the data, gs which had escaped early were assigned scores of zero
for the remainder of the session. When the analysis dealt strictly
with either of the last two 5-min. blocks of the session, SS which
had escaped were dropped from the particular analysis. It should also
be pointed out that the scores for one male §_in Group E on sessions
8 through 11 have been dropped completely from all analyses. This 5
apparently develOped an infection during this period ( as evidenced by

mattered eyes) and his reSponse rate nearly doubled.

Protest

 

A total of 21 animals were pretested in order to obtain the final
16.58. Five were rejected because they gave less than 10 H-S reSponses
during the stimulated periods of the pretest. Of the 5 rejected, 4
were males; 3 were 40 days of age, 2 were 48 days - none of the 5 ani-
mals 60 days of age which were pretested was rejected. The range of
stimulated reSponses for the 5 rejects uas 4-9, with a median of 8.5.
The range of non-stimulated responses for this same group was 0-21,

with a median of 3.

18

For the 16 animals which became 58, the range of stimulated H-S
reSponses was 12-43 (mdn = 19.5). For non-stimulated reSponses, the
range was 0-25 (mdn = 4). Se in Groups E and C were fairly evenly
divided in terms of total number of responses on the pretest: Group

C, median = 28; Group E, median = 27.5.

Head-shake reSponses

Response level. Scores for the total number of H-S responses per
session are presented in Appendix A. It can be seen that all animals
from both Groups E and C made H-S reSponses on all sessions. The range
for Group C was 2-177, while for Group E it was 50-278. In order to
determine if there were differences between groups and sexes at the
beginning of the series of sessions, a 2 by 2 "analysis of variance"
(Wilson, 1956) was performed on the total number of H-S reSponses for
session 1. The results indicated a significant difference between
groUps (X2 = 9, df = l, p < .005) but not between sexes (X2 = 0).

Another 2 by 2 analysis of variance was performed on the total
number of responses summed across sessions 1 through 12 to see if the
same relationships were reflected in the total number of reSponses a-
cross sessions as in session 1. Again, there was a significant dif-
ference between grOUpS (X2 = 4, p < .05) but not between sexes (X2 = O).

A series of Mann-Whitney U-tests on total reSponses (sessions 1
through 12 combined) was then used to compare response levels during
stimulated and non-stimulated periods. No significant sex differences
for either group were found. For Group E, the response level during
stimulated periods was significantly higher than during non~stimulated
periods (p = .000, sign test). For Group C there was no significant

difference between "stimulated" and "non-stimulated" response levels

19

(p = .36, sign test). Between-group comparisons revealed that the
reSponse level during stimulated periods was significantly higher
for Group E than for GrOUp G (g = 0, p‘< .001), but there was no sig-
nificant difference for non-stimulated periods (3; = 22, p = .27).
within-session habituation. Medians for the number of H-S re-
Sponses per S-min. block (stimulated and non-stimulated periods sep-
arate) for 4 representative sessions (1, 7, l2 and 14) appear in Fig. 3.
In order to determine if within-session decrement occurred, stimulated
reSponses during the first and last stimulated 5-min. blocks were com-
pared for sessions 1, 7, l2 and 14. For each of these sessions eXper-
imental SS gave significantly fewer responses during the last block
than the first, as indicated by sign tests (session 1, p = .004; ses-
sion 7, p = .03; session 12, p = .03; session 14, p = .008). Control
.is during comparable periods showed no systematic changes on sessions
1, 7 or 12. Likewise, no significant changes in the number of non-
stimulated reSponses (first vs. last stimulated block) were observed
for either Group E or C on sessions 1, 7, 12 or 14.

As an estimate of the extent of the decrease in stimulated re-
Sponses for Group E on session 1, the number of stimulated reSponses
during the last stimulated 5-min. block of session 1 was divided by
the number of stimulated reSponses in the first block. A percentage
score was thus obtained for each.§ comparing the stimulated response
level at the end of the session with that at the first. The range of
these scores was 33-92%, the median was 49%. To check the possibil-
ity of sex differences in extent of decrease, a U-test was performed
on these percentages. No significant difference was found (Q = 5,

p = 024).

20

.CH Uce NH .b .m vcewuucw be;
miueHL .thlm >3 ecccecucp exwLmItme Te¢em3eepetcoc use emyefleEHue mo nemesc cewtcg .K .c:

L
O

«H cvwcmrw NH comets. a cewweem H some

wgeeam opocwCIo>Hw

C
r—{
C u
p-
k -
a
F")
N
r—T

angwwmvmmaofimwfiomemmaoaomwmntnme

 

 

 

T iisA‘su1. III. .I' It- I. v-1’D .,.s: I O

Hiya: eecocnee cwwefireHuezre:

ea c , a o ,N v a” N H OH 0 m. a @ mm a mm N H OH 9 MW e «W m 6 mp N H 3H 0 mi r o m“ a n, N w

 

O O
-b.nO .0:
0.: .9: -.Q 0.101.. .p/

o d. 0 .V

.o
0 10:Ax\o
- Oille;AY!¢3xv
\b.Ar!o

IQ. O i

 

 

 

 

3. 13C 1 3 d..c..l-v.suo..0lnj‘..qo 6.”..-044 ....6..4|HI.1?:0.I.. .omuruio _l
.4 E (kw TIIII. MEN/(H. CLCLCLCL Erwumcpuifug

(3;?

r x (’1 i
L) L.) J '4

21

In order to determine if base-rate reSponse levels changed within
a session, the number of H-S reSponses in the first 5 min. was com-
pared with the number during the last 5 min. for both sessions 1 and
12. No significant changes were found for Group E or C using Sign

teStS 0

U]

Changes across ses ions. Fig. 4 presents the median number of

0')

stimulated H-S response per session for Groups E and C by sex. Hon-
stimulated reSponses are also represented similarly in the same figure.
A Wilson (1956) analysis of variance en the data for Group E showed no
significant sex differences nor session effects for either stimulated
or non-stimulated responses. Two apparent depressions in the curve for
experimental males, one at session 6 and one centering around session
10, can be seen in Fig. 4. The reduced numbers of reSponses at these
points were caused largely by early escapes. 0n session 6, one male
repeatedly escaped after only 15 min., necessitating termination of the
session. Omitting this S's score, the median becomes 171, as opposed
to 123. On each of sessions 9, 10 and 11 the same male escaped after
approximately 35 min.; since the scores for one male were already dis-
carded for these sessions due to an apparent infection, the medians

may not be representative. The curves for all of the sessions for both
males and females in Group E show no systematic decrease in response
level as a function of sessions. What shifts there are appear to be
closely aligned with premature terminations of sessions for a few_§s.
In order to test for changes in the extreme case, the number of stimu-
lated responses (males and females combined) for sessions 1 and 14 were
compared. No significant changes were found (p = .14, sign test).

A Nilson analysis of variance on the session totals (stimulated

er of head—shake reSponses

7:
J

Fedian num’

Stimulated responses

 

140

100

00

60

4O ,

20

 

22

O——-—HD 'rcu3 -, Isl _
0"‘"--0 'Sreup E, Icrdlre
°-———-<> Group C, Tale‘

 

 

I—-‘
[J
(23
.f_\

60

 

10 11

Ca
KO
)-—l
I\)
t—'
(,1

l4

 

 

Fin. 4. =edian

responses per

number of at;

and non-stimulated periods combined) for Group C revealed significant
sex differences (X2 = 0.17, df = l, p < .01), females exhibiting the
higher reSponse level.

Another Nilson analysis of variance was used to compare stimulated
responses per session for GrOUp E with the number of responses during
comparable periods for Group C. Sexes were combined for each group.
Only 9 scores for Group E fell below the median fer the combined dis-
tributions. This difference between groups is highly significant

(x2

= 126.6, df = 1, p < .001).

In order to compare GrOUps E and C on base-level responding, non-
stimulated responses per session (sexes combined within groups) were
subjected to a Wilson analysis of variance. No significant difference
between the groups nor effects of sessions were found (between grOUps,

2 = .08, p > .75; sessions, X2 = 3.5, p > .97).

X
In Fig. 5 the number of stimulated reSponses during the first and
last stimulated 5-min. blocks for each session (Group E and C) are rep-
resented graphically. The difference between the starting and ending
points for each session represents the amount of drop in stimulated
H-S responses for that session. In order to test the extreme case for
systematic cross-session changes in initial reSponse level, stimulated
reSponses during the first stimulated block for sessions 1 and 14 were
compared. No significant differences were found (p = .66, Sign test).
The same test was made on the data for the last stimulated block;
again, no significant change was found (p = .34, sign test). Since
there were no systematic changes in frequency of H-S responses from

the first to last 5-min. block across sessions, the median of each S's

14 scores was then taken for both the first and last stimulated blocks.

80.5

reopen

ad-shake

O

ber of h

18!”! HUD

,.

r
L

_ \
A
I

24

o————o first stimulated blocks, Group E
o---o Last stimulated blocks, Group 5
O——-—0 First stimulated bloc 5, Group C

0----O Last stimulated blocks, Group C

24
2O

5 A A r”* / \

 

 

 

\ I
12 0”, \ ’ \VY’ \
\ / \
s / b
\
E’ \g’
4
1 2 3 4 5 6 7 0 9 10 ll 12 13 14

Fig. 5. Anoian number of stimulated head-shake responses durinc
first and los stimulated S—min. hlnckn of each session.

the

25

The range oF these median scores For the First block was 9.5-39, For
the last block, 6-19.5. Cut oF a total oF 103 cases, in only 7 was
the reSponse level at the end higher than at the First. In order to
obtain an estimate oF the drop in stimulated response level, the median
score For the last block For eachug was divided by the median score For
the First block. The resulting percentages ranged From 40% to 64%,
with a median 0F 61%. Thus, using medians as the central indicators,
the terminal response level was approximately 60% 0F the initial level.
The Foregoing analysis indicated that response level during the
First stimulated block does not decrease across sessions. However,
the units oF time involved were no Finer than 5 minutes. It is pos-
sible that residual eFFects oF prior habituation might be exhibited
at a more subtle level. ThereFore, the number 0F reSponses during the
First stimulus presentation was investigated comparing responses on
session 1 with session 12. When scores For these two sessions were
compared, no signiFicant diFFerences were Found (p = .66, sign test).
The curves For sessions 1, 7, 12 and 14 (see Fig. 3) indicate
that the amount oF drop in stimulated reSponses From the First to the
second stimulated 5-min. block increased across sessions. The slope
0F the curve connecting these blocks becomes progressively steeper,
while the portion oF the curve between blocks 3 and 9 appears on ses—
sions 12 and 14 to be approximating a straight line with a slope oF
zero. This suggests that response decrement within a session may have
become more rapid as sessions increased. In order to test this possi-
bility, the amount oF drop between the First 2 stimulated blocks was
compared with the amount oF drop between the First and the last blocks.

The number 0F H-S responses on the First block minus the number on the

26

second block (A) was divided by the number oF responses on the First
block minus the number on the last block (B). The resulting ratio
indicated how much 0F the total within-session decrement occurred be-
tween the First and second stimulated blocks. Such scores were ob-
tained For each §_in Group E on each session. Since either A or 8
could be zero or negative, the Following rules were observed in as-
signing scores in such cases. IF A was zero or negative, a score 0F
zero was assigned, regardless oF the value oF 6. IF A was positive
but B was zero or negative, an arbitrary score oF 1.0 was assigned.
The scores oF two male-is were not included in the Formal analysis;
one-i had develooed an inFection during sessions 8 through 11, and
another had escaped prematurely on several sessions. The remaining
scores were subjected to a Friedman analysis 0F variance (Siegel,
1956). No signiFicant eFFect due to sessions was Found (X2 = 10.84,
dF = 13, p > .50).

In order to determine iF there were systematic changes across
sessions in beginning and ending base levels, comparisons (sign tests)
between sessions 1 and 12 were made on the number oF H-S reSponses oc-
curring during the First and last 5 min. (these blocks being entirely
non-stimulated). In no case were signiFicant diFFerences Found (First

.14, Group G, p = .36; last 5 min.: Group E,

5 min.: Group E, p

.66). In order to make sex and group compari-

p = .11, Group G, p
sons, the median 0F each §fs 12 scores was taken For the First and
last S-min. blocks respectively. U-tests were then used to compare
the resulting scores. In no case were signiFicant sex or group diF-
Ferences Found.

EFFects oF reduced inter—session interval. In order to determine

27

1F the reduced interval (6 hrs.) separating sessions 12, 13 and 14

had any speciFic eFFect, a number oF indices were examined. The stim-
ulated responses For each oF the three sessions were compared with

sign tests. No signiFicant diFFerences were Found (12 vs. 13, p = .64;
12 v8. 14, p = .64- 13 vs. 14, p = .36). When both beginning and end-
ing reSponse levels (First and last stimulated S-min. blocks) For the

3 sessions were compared, no between-session changes were Found except
between the terminal levels on sessions 12 and 14 (p = .03, one-tail
sign test). Finally, when percentage drop in reSponse level (last
stimulated block divided by First) was compared, no signiFicant diFFer-
ences were Found between sessions 12 and 13 or 13 and 14. However, the
percentage drop For session 14 was signiFicantly greater than For ses-
sion 12 (p = .03, sign test).

Reaponse decrement within 30-sec. stimulus presentations. It was
observed early in the course 0F the eXperiment that, within any given
30—sec. stimulus presentation, more reSponses were generally given dur-
ing the First 15 sec. than during the second. This relationship can
be observed in Fig. 6; to obtain the data For this Figure, each stim-
ulated period For sessions 1 through 12 was split in halF and the num-
ber oF H-S reSponses counted separately. To determine iF this "within-
stimulus" decrement was signiFicant, scores For the First and second
halves on session 1 were compared. SigniFicantly Fewer resoonses oc-
curred during the second halves than during the First (p = .03, one-
tail sign test). Next, data For sessions 1 and 12 were compared to
determine iF changes occurred across sessions. No signiFicant changes
in reSponse level were observed For either the First (p = .14) or sec-

ond halves (p = .36, sign tests). ThereFore, the median oF each g's

rs -.
a f! f)

reopen

r.
k4

Median number oF head—shak-

110

100

50

40

30

20

 

20

r- (.‘L

 

o——————o :irst halves
o—--—-o Sec-ond halve:
," W
I, \
\
\ I, ‘x I \\ ’D’
V V \ /’
b’
l 2 3 4 6 6 7 0 0 10 ll 12
Sessions
Fig. 6. iedian number cF stimulated head-shake responses cor
session (Group a; For each ialF 0F each 30-sec. crie-
ulatod period.

12 scores was taken For the First and second halves respectively. In
order to compare response levels For the two halves, the median For the
second halves was divided by the median For the First, resulting in a
distribution oF percentages ranging From 51 to 90%, median = 64.5%.
Thus, the response level during the second halves oF stimulated periods
decreased to approximately 65% 0F the level during the First halves.

Temporal conditioning. Since the interval between stimulus pre-

 

sentations was constant, there exists the possibility that some temporal
cue might have served to keep the reSponse level higher than it would
otherwise have been. That is, a high rate oF responding may have be-
come conditioned to the temporal schedule oF stimulus presentation.

IF this were the case, it would be expected that any eFFects built Up
during stimulation would extend into the last 5 min. oF the session,
which were not stimulated. In addition, §s might be eXpected to make
more reSponses toward the end oF each 30-sec. period separating stimulus
presentations, since the time For another delivery oF the stimulus was
drawing near.

In order to test the First prediction, the last 5-min. blocks
(non-stimulated) For each session were divided into 30-sec. periods
corresponding to the schedule oF stimulation established in the
previous 40 min. The number oF reSponses during "stimulated" periods
(those corresoonding to periods during which the stimulus had been
presented) was compared with the number oF responses occurring during
"non-stimulated" periods. No signiFicant diFFerences were Found For
either session 1 (p = .34, sign test) or session 12 (p = .19). Since
there were no diFFerences in the extreme cases, scores were summed

across the 12 sessions For each 5 and the resulting totals compared.

30

No signiFicant diFFerence was Found between "stimulated" and non-
stimulated periods (p = .36, sign test).

To test the second prediction, the 30-sec. non-stimulated periods
separating stimulus presentations were divided into halves (First and
last non-stimulated S-min. blocks excluded). The number oF reSponses
in the First halves was then compared with the number occurring in the
second halves. No signiFicant diFFerences were Found For either ses-
sion 1 (p = .50, sign test) or session 12 (p = .23). Again, since
there were no diFFerences in the extreme cases, scores were summed
across the 12 sessions For each,§ and the resulting totals were com-
pared. No signiFicant diFFerence between the First and second halves
was Found (p = .14, sign test).

Intercorrelations among head-shake response measures. In order
to establish the reliability oF the H-5 reSponse as a stable measure,
rank-order correlations were computed between a number oF the obtained
measures. With N = 8, a correlation coeFFicient 0F .64 or greater is
required For signiFicance at the .05 level (Edwards, 1963).

Correlations were First calculated between responses on the pre-
test and session 1. 0n total reSponses For these periods, r = .12
(Group E) and r = .47 (Group G). For GrOUp E, stimulated responses
alone correlated non-signiFicantly (r = .14); however, For non-stimu-
lated responses alone, r = .62. For GrOUp G, non-stimulated responses
on the pretest correlated signiFicantly with non-stimulated responses
on session 1 (r = .03). When the number 0F reSponses From the First
10 min. oF session 1 were considered For GrOUp E, the correlation with
total reSponses on the pretest did not increase a great deal (r = .31).

Rank-order correlations were next calculated between total numbers

oF H-S responses For all possible combinations oF sessions. Tables 1
and 2 present these coeFFicients For GrOUps E and C reSpectively. It
should be noted that, For GrOUp E, correlations involving sessions 8
through 11 are based on an N 0F 7, since scores oF one.§ For these
sessions were discarded. With N = 7, an r 0F .68 or greater is needed
For signiFicance. The range oF r's For Group E was .23 to 1.00; For
Group C, it ass .30 to 1.00. For Group E, only 10 out 0F 91 correla-
tions were not signiFicant (i.e., below .64 or .60). For Group G, 14
out 0F 66 were not signiFicant.

Three remaining reSponse measures were correlated For Group E:
non-stimulated response level, stimulated response level, and percent-
age drop in reSponse level within sessions. Scores For the First two
measures were obtained by taking the median oF each §}s 14 scores.

For the third measure, percent drop scores derived earlier under the
section "Changes across sessions" were used. base rate (non-stimulated)
and stimulated reSponse levels were not correlated (r = 0). Base rate
level and percent drOp were signiFicantly correlated (r = -.7l). How-
ever, stimulated response levels did not correlate signiFicantly with

percent drop (r = .10).

Frequencies oF behaviors preceding and Following head-shake responses
It became apparent in the early stages oF the experiment that H-S
responses usually did not occur when the animal was inactive. In an
attempt to discover iF there were certain patterns oF activity within
which H-S responses occurred, the behaviors which preceded and Followed
H-S reSponsos were examined For sessions 1 and 12. For each response
which occurred, preceding and Following behaviors were recorded (Fre-

quencies only) according to the Following categories: head movement,

32

Table 1. Intercorrelations (rank-order) among total head-shake
responses per session, Group E.

 

 

1 .72 .72 .82 .71 .35 .75 .81 .52 .69 .71 .94 .70 .73

2 .83 .78 .74 .23 .74 .86 .93 .36 .82 .71 .78 .83
3 .88 .90 .37 .86 .96 .82 .82 .82 .86 .80 .74
4 .90 .60 .98 .93 .86 .71 .96 .83 .92 .91
5 .49 .91 .96 .82 .61 .93 .78 .92 .86
6 .64 .43 .50 .93 .61 .31 .62 .55
7 .93 .86 .71 .96 .76 .97 .93
8 .86 .54 .86 .89 .85 .86
9 .50 .93 .61 .88 .89
10 .68 .54 .67 .68
11 .64 .99 1.00
12 .66 .64

13 .97

Table 2.

Sessions

10

ll

33

Intercorrelations (rank-order) among total head-shake
responses per session, Group G.

 

.85 085 074 091 078 .86 071 071 .74 .64 038

.80 .70 .71 .57 .83 .70 .70 .79 .62 .52

.96 .86 .65 .60 .65 .65 .61 .65 .41

086 .74 062 069 069 064 076 .62

088 081 081 081 076 074 052

.91 .91 .95 .83 .62

1.00 .98 .93 .61

.98 .93 .61

.71

.5

gnashing, head movement plus gnashing, washing, and inactivity. Turn-
ing around was not included because oF its inFroquent occurrence. The
category oF head movement plus gnashing was included because in many
cases the two overlapped or were in such close temporal sequence that
it was diFFicult to determine which came First or which was predominant.
There were occasional cases oF overlap involving other combinations oF
behaviors e.g., head movement plus washing), but these were so inFre-
guent that placing them in separate categories would have added nothing
to the analysis. These latter cases were recorded under the category
oF the behavior which was predominant or was temporally closer to the
8-5 response. That is, the categories were mutually exclusive: each
instance was recorded under one and only one category. Gases where a
recorded behavior did not occur within 3-4 see. oF a H-S response were
recorded as "inactivity". However, the 3-4 see. criterion was used
only as a general guideline; it was usually possible to establish a
"normal” pattern For each 5, which was then used as the criterion.

Preceding and Following behaviors were examined separately For
stimulated and non-stimulated periods For Group E. Final scores were
derived For each 5 by dividing the number oF cases in each category by
the total number oF cases, resulting in percentages. Each g in Group
E, then, produced Four scores For each category, correSponding to pre-
ceding and Following behaviors For stimulated and non-stimulated per-
iods. .58 in GrOUp G produced only two scores - For preceding and Fol-
lowing behaviors. These scores are summarized in Table 3.

It can be seen From Table 3 that the predominant behavior proced—
ing H-S reSponses was head movement; this applied to sessions 1 and 12

For both groups. The least Frequent preceding behavior was washing.

Table 3.

P

35

Behaviors preceding and Following head-shake reSponses For

sessions 1 and 12 (percentage scores).

Precedino behaviors
‘2‘

 

 

—_.

Stim'd Periods

Experimental Group

Non-Stim'd Periods

Control Group

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Session 1 Session 12 Session 1 Session 12 Session 1 Session 12
Behavior Range Mdn Range Mdn Range Mdn Range Mdn Range Mdn Range Mdn
Head 24-82 47 18-67 36 44-75 62 26-85 60 49-68 61 13-79 71
movement
Gnashing 4-22 15 6-24 8 3-19 11 0-19 5 0-13 11 0-30 5
Head “Vt 6-21 9 3-50 13 6-24 12 10-37 22 9-33 16 7-55 20
& Gnash
Washing 0-16 1 0-6 3 0-15 0 0-29 3 0-15 0 0-9 1
Inactiv. 8-55 24 13-62 23 0-31 7 0-15 6 0-25 4 0-23 1
Following behaviors
Head 31-87 57 18-68 36 20-82 57 21-56 33 11-66 45 12-38 27
movement
Gnashing 2-22 9 2-37 11 6-33 10 10-38 26 7-54 21 0-60 28
”Fad ”Vt 9-34 19 19-56 35 9-28 17 15-43 21 12-42 21 18-38 26
a Gnash
washing 0-18 1 0-6 3 0-24 0 3-18 6 0-21 0 0-16 3
Inactiv. 2-28 8 4-13 9 0-19 10 0-15 7 0-24 5 0-21 7

 

The predominant behavior Following H-S reSponses For session 1 was

head movement; however, there appeared to be no predominant Following
behavior on session 12. The least Frequent Following behavior, in
general, was washing. Within-group comparisons were made using sign
tests to determine iF signiFicant diFFerences existed between sessions
1 and 12. None oF these comparisons revealed a signiFicant diFFerence.
U-tests were then used to compare Groups E and G. Only one signiFi-
cant diFFerence was Found: the percentage oF H-S reSponses preceded
by inactivity on session 12 (stimulated periods) was lower For Group C
than For Group E (0 = 4, p = .002).

In spite oF the group trends, considerable individual diFFerences
were apparent. No animal appeared to reapond in an extremely rigid
pattern; the highest proportion oF reaponses accounted For by any one
behavior For a single 5 was 85$. At the other extreme, the highest
preportion oF responses accounted For by a single behavior was as low
as 33; in a Few cases. Thus, while a single behavior category dominated
preceding and Following behaviors For a Few-So, in most cases there was

substantial variability.

Gurations oF behaviors. other than head-shake, durigg sessions 1 and 12
The remaining behaviors which were recorded will be dealt with in

the Following order: head movement, gnashing, washing and turning around.

All behaviors during sessions 1 and 12 were examined. These sessions

were chosen as representing the extreme cases. Scores (ranges and medi-

ans) For these two sessions on all behaviors except turning around ap-

pear in Table 4. Scores For turning around were excluded From Table 4

because they reFlected discrete occurrences. All other scores represent

durations rather than discrete occurrences, since the associated beha-

37

Table 4. Amount oF time (in sec.) spent moving the head, gnashing, and
washing, sessions 1 and 12.

 

Experimental Group Control Group
Session 1 Session 12 Session 1 'Session 12
Behavior Range Ndn Range Mdn Range Mdn Range Mdn
Head 171-1265 625 105-1002 333 170-509 420 160-667 342

movement

Gnashing 225-554 305 137-824 301 298-548 412 220-2219 286

Hashing 42-357 223 22-231 142 61-251 150 59-250 148

 

 

 

 

 

 

[A
63

viors generally lasted several seconds or longer.

{cad movement. Ranges and medians oF the scores For this response

 

measure appear in Table 4. SigniFicant sex diFFerences (indicated by
U-tests) were Found in only two cases. In Groue E, scores For stimu-
lated periods oF session 12 were higher For Females than males (g = 0,
p = .028). And in Group C, scores on session 1 were higher For Females
than males (0 = 0, p = .028). In order to determine iF head movement
decreased within a session, the scores For the First stimulated period
were compared (sign tests) with those From the last stimulated period.
He signiFicant diFFerences were Found in any oF the comparisons. For
Group E on both sessions 1 and 12 signiFicantly more time was Spent
moving the head during stimulated periods than during non-stimulated
periods (both sessions, p : .004, sign tests). 0n session 1, Group E
Spent signiFicantly more time moving the head during stimulated periods
than Group G (g = 13, p = .025); there was no diFFerence, however, For
non-stimulated periods (0 = 31, p = .40). By session 12 there were no
signiFicant diFFerences between groups. When session 1 was compared
with session 12, no signiFicant changes were Found For either group.
Read movement during the two sessions was not highly correlated For
either group (Group E, r = -.60; Group G, r = .45).

Gnashing. The scores For gnashing are summarized also in Table
4. No signiFicant diFFerences between sexes were Found. Scores For
the First and last stimulated periods oF each session were compared
to determine iF gnashing decreased or increased within a session. No
signiFicant diFFerences were Found. On both sessions 1 and 12, Group
E spent signiFicantly more time gnashing during stimulated periods than

during non-stimulated periods (session 1, p = .03; session 12, p = .004;

39

sign tests). There was no signiFicant diFFerence between Groups E
and C during stimulated periods oF session 1 (Q = 19, p = .10); how-
ever, scores during non-stimulated periods oF session 1 were signiFi-
cantly lower For Group E than For Group G (g = 9, p = .007). By ses-
sion 12 there were no signiFicant diFFerences between groups, although
For non-stimulated periods the diFFerence nearly reached signiFicance
(g = 17, p = .065). Comparing session 1 with 12 revealed no signiFi-
cant changes For either group. In terms oF total amount oF time spent
gnashing, session 1 did not correlate highly with session 12 (Group E,
r = -.40; Group G, r = -.26).

flashing. Ranges For amount oF time Spent washing on session 1 and
12 are presented in Table 4, as are medians. SigniFicant sex diFFerences
occurred in only one case. Control males washed more than Females on
session 12 (Q = 0, p = .028). To determine iF washing decreased within
a session, scores For the First and last stimulated periods were com-
pared. No signiFicant diFFerences were Found For Group E. However,
scores For Group G decreased signiFicantly on both sessions 1 and 12
(session 1, p = .008; session 12, p = .016). On neither session 1 nor
12 did Group E Spend signiFicantly more time washing during stimulated
periods than during non-stimulated periods (session 1, p = .14; session
12, p = .36; sign tests). There were no signiFicant diFFerences be-
tween Groups E and C on stimulated or non-stimulated periods For either
session 1 or 12. when session 1 was compared with session 12 (sign
tests), no signiFicant changes were Found For either group. In terms
oF total amount oF time spent washing, session 1 did not correlate
highly with session 12 (Group E, r = .29; GrOUp G, r = .48).

Turning around. Subjects turned around relatively Few times on

40

the test stand. For Group E, the range on session 1 was 0-32 (median,
7.5); For Group G, the range was 0-4 (median = 2). Scores For ses-
sion 12 ranged From 0-4 (median = 0) For Group E, and From 0-9 (medi-
an = 2) For Group C. There were no signiFicant sex diFFerences For
either group. On neither session 1 nor 12 did Se in Group E turn
around more times during stimulated than non-stimulated periods (sign
tests). Between-group comparisons indicated that total scores on ses-
sion 1 were higher For Group E than Group G (g = 11, p = .014), but
there was no diFFerence on session 12 (g = 19, o = .10). Between-
session comparisons indicated a signiFicant decrease From session 1

to 12 For GrOUp E (p = .008, sign test) but not For Group G (p = .50).

Intercorrelations among behavioral measures

Rank-order correlations among the various scores (totals) on ses-
sions l and 12 were computed primarily For descriptive purposes. Turn-
ing around was not included here because oF its extremely low Frequency
oF occurrence. The correlation coeFFicients appear in Table 5. CF par-
ticular relevance are the correlations between H-S reSponses and the
other behavioral measures. These correlations range From -.52 to .42
For Group E, From -.43 to .48 For GrOUp G. None oF the coeFFicients
in Table 5 indicate more than a low or moderate degree oF correlation

between any oF the measures involved.

41

Table 5. Intercorrelations (rank-order) among behavioral measures,

sessions 1 and 12.

 

 

Session 1

Experimental Group

Control Group

 

 

 

 

 

 

 

 

 

 

 

Behavior Head mvt Gnashing washing, Head mvt Gnashing Washing
Head-shake .42 .01 -.34 .31 -.43 -.10
Head mvt -.40 -.07 .02 -.57
Gnashing .57 -.12
Session 12

Head-shake -.52 .31 -.12 -.10 .48 -.l7
Head mvt -.57 .29 -.15 .26
Gnashing .30 -.29

 

Head-shake reSponses

Peseonse level. All gs, both experimental and control, gave H—S
responses on every session. The Finding oF a substantial base rate For
Group C runs counter to Askew's (1966) results. It is assumed that
diFFerences in the test conditions or in the age oF the subjects are
reSponsible For the discrepancy. The animals were approximately 100
days old in Askew's study and were restrained in small wire cages, con-
siderably restricting the animals' movements.

The Fact that-gs in Group C emitted H-S reSponses at all is curi-
ous, since no immediate stimulus can be identiFied. The stimulus in
this case is possibly an internal one, but why the base rate should be
as high as 4-5 per min. is unclear. The biological Function oF the re-
sponse seems apparent: the H-8 reSponse Functions to remove particles
oF debris or living organisms (e.g., parasites) From the ear oF the
animal. But its Function when there is apparently no Foreign matter
in the ear is not obvious. A useFul concept to invoke here may be that
oF vacuum reSponses (Ratner and Denny, 1964), reFerring to the occur-
rence oF instinctive acts when drive level is heightened and only min-
imal or inapprOpriate stimuli are present. The experimental setting
may have increased the drive level oF the animals, leading to the oc-
currence oF "vacuum" head-shake reSponses.

There were no signiFicant sex diFFerences For either group on
total number oF responses For session 1 or on summed scores For all

the sessions combined. However, when uncombined session totals were

42

43

subjected to "analysis oF variance", Females were Found to have a sig-
niFicantly higher reSponse level than males in Group C. 0n the other
hand, males in Group E had a higher reSponse level during stimulation
than Females, although the diFFerence was not signiFicant. The Find-
ing that sex diFFerences apparent under control conditions (Females
higher than males) disappear (For non-stimulated periods) or are per-
haps even reversed (For stimulated periods) under test conditions must
remain For the moment without eXplanation. It is, oF course, possible
that such an interaction is a unique artiFact oF the test situation
and the H-8 reSponse or is due to sampling error.

That the H-8 response level was substantially modiFied during
application oF the test stimulus is clear. The rate oF reSponding dur-
ing stimulated periods was 6 to 7 times higher than it was during non-
stimulated periods. However, repeated application 0F the test stimulus
did not inFluence the base rate (non-stimulated responding) oF Group E;
Groups E and C did not diFFer signiFicantly on this measure.

within-session decrement. Two types oF decrement in H-S reSponses
within sessions were demonstrated For Group E. The First oF these oc-
curred within each 30-sec. stimulus presentation: signiFicantly Fewer
H-S reSponses occurred during the second 15 sec. than during the First.
This "within-stimulus" reSponse decrement took place within each 30-sec.
period 0F stimulation, and recovered substantially during the 30-sec.
non-stimulated periods separating presentations oF the test stimulus.
The rapid recovery time suggests that this decrement may have been due
to Fatigue eFFects.

The second type oF reSponse decrement consisted oF a signiFicant

drop in the number oF stimulated H-S reSponses From the First stimulated

44

S-min. block to the last For each test session. This decrease within
the session appeared to Follow a negatively accelerated Function, gen-
erally held to be characteristic oF habituatory decrement (Thompson
and Spencer, 1966). The terminal number oF H-S reSponses was approxi-
mately 60% 0F the initial number. While the level oF responding to
the test stimulus decreased within a session, the base rate (non-stim-
ulated) remained unchanged. Thus, the reSponse decrement appeared to
be SpeciFic to the test stimulus. IF Fatigue oF the reSponding mecha-
nism had been primarily responsible For the decrement within each ses-
sion, the number oF base-level responses should have decreased along
with the number oF stimulated reSponses, at least during periods sep-
arating stimulus presentations. Since this did not occur (only stim-
ulated responses decreased), Fatigue can be ruled out as the cause oF
within-session decrement.

The results For Group C show that the base rate oF the H-8 reSponse
is Fairly constant when no test stimulus is applied. No signiFicant
diFFerences in the number oF reSponses during the First and last S-min.
blocks oF each session were observed For this group.

Changes across sessions. No measures oF the H-5 reSponse For

 

Group E showed systematic changes across sessions. The total number
oF stimulated reSponses per session For GrOUp E showed no systematic
changes across the 14 sessions. Initial and terminal stimulated re-
Sponse rates remained at approximately the same levels For all 14 ses-
sions. In addition, the habituation curves oF Group E are similar For
all sessions, showing no apparent change in the course oF habituation
within a session. It must thereFore be concluded that, under the con-

ditions employed, habituation oF the H-8 reSponse in the rat is not

45

retained over a period 0F 24 hrs. Further discussion oF this point
will be Found under a later section.

The base-rate occurrence oF H-S responses is apparently quite
stable. No signiFicant changes in total number oF reSponses For 12
consecutive sessions were Found For GroUp C. Nor were any changes
observed in non-stimulated reSponse rates For Group E, in Spite oF
the large numbers oF reSponses which occurred in the presence oF the
test stimulus.

EFFects oF reduced inter-session interval. when the interval
separating test sessions was reduced From 24 to 6 hrs. For Group E,
some evidence For retention oF habituation appeared. SpeciFically,
the amount oF decrease in the H-8 response level during session 14
'was signiFicantly greater than during session 12 (the latter session
preceding the First 6-hr. interval). It appeared that, with only 6
hrs. separating sessions, eFFects oF previous habituation sessions
may have begun to cumulate slightly. However, the total number oF
stimulated responses did not diFFer signiFicantly For sessions 12, 13
or l4. Thus, signiFicant eFFects were apparent only in the rate oF
habituation, not in the total number oF reSponses.

Temporal conditioning. The results indicate that the Fixed inter-
stimulus interval did not provide a temporal cue to keep the stimu-
lated H-S reSponse level For Group E From decreasing across sessions.
The higher rate 0F responding during stimulated periods did not carry
over into the last S—min. block (non-stimulated entirely). Nor did
.§s make anticipatory H-S reSponses as the end oF each 30-sec. non-
stimulated period approached. ThereFore, the possibility oF a con-

Founding temporal Factor can be ruled out.

46

Intercorrelations among head-shake response measures. The corre-
lations between total numbers oF H-S resoonses indicate the response
is a highly reliable measure. This applies in gereral to both Groups
E and G. For Group E, correlations between session 6 and the remain-
ing sessions on total number oF H-S reSponses were noticeably lower
than For the other sessions. The same applied to correlations between
session 10 and the remaining sessions. This was due largely to the
Fact that two animals escaped early on each oF these sessions, one aFter
only 15 min. on session 6. Their scores were thereFore spuriously low
For these days, and the correlations were aFFected accordingly.

Base-rate and stimulated reSponse levels For Group E were not sig-
niFicantly correlated. That is, $3 with high base levels did not also
have high response levels to stimulation. The extent oF within-session
drop in stimulated responses did not correlate signiFicantly with stim-
ulated reaponse levels, but the Former did correlate signiFicantly with
base levels (r = -.7l). In other words, For animals with high base

levels, degree oF within-session habituation tended to be small.

Behaviors preceding and Followine head-shake responses

No single, dominant pattern oF behaviors preceding and Following
H-S responses emerged. In Fact, in only a Few cases was a single be-
havior class strongly dominant at all For a given individual. Hot un-
commonly, H-S responses were preceded and Followed by two or even three
diFFerent behaviors with about equal Frequency. hevertheless, it was
clear that head movement preceded and Followed H-S reSponses more Fre—
quently than any other behavior. The least Frequent behavior associated
with H-d responses was washing. Ho diFFerences were Found between Groups

E and G with reapect to behaviors preceding and Following H-S responses.

 

47

Behaviors other than head-shake responses

The various behaviors measured in addition to H-S responses be-
haved diFFerentially, without consistent patterns. 0nly scattered sex
diFFerences were observed; no systematic trend was Found. The only be-
haviors which appeared to be aFFected by the test stimulus were head
movement and gnashing. The gs in Group E spent signiFicantly more time
engaged in these two behaviors during stimulated than during non-stim-
ulated periods. This may have been due to aversive properties oF the
test stimulus.

There were only occasional diFFerences between groups on any oF
these behaviors, and these occurred in session 1. 0y session 12, no
signiFicant diFFerences between groups were to be Found. Scores For
the behaviors showed no signiFicant within-session changes, except For
washing in Group G. Likewise, no signiFicant changes From sess'on l
to 12 occurred, except For a decrease in turning around For GrOUp E.
Finally, scores For the respective behaviors on session 1 did not cor-
relate highly with those on session 12. The lack oF consistent diF-
Ferences plus the low stability oF the measures preclude meaningFul

generalizations at this time.

General comments on retention

A substantial number oF studies which reported retention oF habit-
uation For 24 hrs. or longer were cited in the Introduction. The pre-
sent Finding that habituation oF the H-8 response in the rat is not
retained over a period 0F 24 hrs. stands in apparent contradiction to
these prior studies. However, it does substantiate Askew's (1966) re-
sults. That the execrimental conditions were at Fault seems an unlikely

possibility. Each g in Group E received a total 0F 20 min. stimulation

 

48

in relatively massed presentations on 14 consecutive occasions. Within-
session decrement oF the H-5 response was clearly demonstrated under
these conditions. Indeed, indications oF retention at 6 hrs. were ob-
served in the Form oF slightly Faster habituation within session 14.
However, the residual eFFects even at 6 hrs. were by no means great.

The data point to the conclusion that the recovery time oF the habit-
uated H-S response in the rat is probably not greater than 6 hrs., per-
haps less. This is considerably shorter than the recovery times re-
ported For many other habituated responses.

The eXplanation For the apparent discrepancy may lie in the nature
oF the H-5 reSponse itselF. Tinbergen (1951) has divided the activi-
ties oF organisms into two classes - appetitive and consummatory. Ratner
(in press) has suggested that these two classes oF behavior are subject
to diFFerential habituation. Consummatory responses "may only show
reFractory phases (brieF decrements with repeated elicitation)", while
appetitive responses show longer term decrements. It is quite possible
that the H-5 response oF the rat is a consummatory reSponse. It appar-
ently belongs in the broader class oF care oF the body surFace (Scott,
1958). It appears to be quite stereotyped, a prime characteristic oF
consummatory reSponses (Ratner and Denny, 1964). However, detailed
analysis would be necessary in order to establish whether or not the
H-5 reSponse is truly consummatory in nature. For the moment, the pos-
sibility is advanced as a tentative explanation For the lack oF long-
term retention oF habituation.

Another possibility presents itselF in accounting For the rela-
tively rapid recovery: the H-5 reSponse may be a simple reFlex (much

like the knee-jerk in humans) which becomes Fatigued with repeated

 

49

elicitation. However, this approach is ill-equipped to handle the
Fairly substantial base rate oF the reSponse when no apparent elicit-
ing stimulus is present. In addition, it will be recalled that al-
though the H-5 response showed a decrement with repeated elicitation
within a session, the base rate during non-stimulated periods remained
unchanged. This seems to rule out Fatigue as the mechanism reaponsible

For the observed decrement.

Adequacy oF the preparation

 

P

species. The laboratory rat appears to be a quite adequate sub-
ject For the study oF habituation. The availability and tractability
oF this species enhance the preparation considerably. The existence
oF several strains presents the possibility oF within-Species compari-
sons - a powerFul, Fine-grain tool oF the comparative psychologist.
The relatively large amount that is already known about the rat readily
aFFords the opportunity For comparing the characteristics oF habitua-
tion with those oF other behavior processes. Animals 30 days oF age
or even younger can be studied with no diFFiculty, the behavioral rep-
ertoire being suFFiciently extensive by this age. The diFFerence be-
tween the sexes at this age constitute no handicap, although it is
possible that diFFerences appearing with sexual maturity would have
more serious eFFects on the preparation.

The response. The H-S reSponse possesses several desirable char-

 

acteristics. It is apparently unlearned, easily identiFiable, and oF
short duration. Its rate oF occurrence to stimulation is relatively
high, but shows deFinite decrement with repeated stimulation. While
the reaponse does have an above-zero base rate (at least under the con-

ditions employed in this study), it is apparently quite stable across

50

time. The high reliability oF the reSponse measure, For both base-
level and stimulated rates, is a particularly desirable attribute. The
relatively rapid recovery oF the habituated response would Facilitate
intensive analysis oF the recovery Function. Finally, should Further
study indicate the H-5 reSponse to be consummatory in nature, it would
provide a good preparation For investigating diFFerences in habitua-
tion oF appetitive and consummatory reaponses. Appetitive reSponses

oF the rat could certainly be identiFied For use in such studies.

Test stimulus. The general eFFectiveness oF the test stimulus
is demonstrated by the Fact that stimulated reSponse rates were 6 to
7 times as great as base rates. However, the test stimulus was an
eFFective eliciting stimulus For the H-5 response in only 75% 0F the
animals initially tested. It did elicit at least a Few responses even
in those animals which were rejected. It is thus possible that modi-
Fication oF the test situation would eliminate "non-reSponders", al-
though exactly what type oF modiFication is not clear.

Pretest. The 10-min. pretest was successFul as a screening mea-
sure - no non-reSponders were encountered during any oF the test ses-
sions. However, the H-5 reSponse measures From the pretest were oF
very limited value in predicting later H-S activity. Stimulated and
non-stimulated reSponse rates during the pretest were not signiFicantly
correlated with the corresponding measures on session 1 For Group E.
The correlation For base-level rates For Group G was signiFicant, indi-
cating that later reSponse levels were consistent with those on the

pretest.

 

REFERENCES

Askew, H. R. The eFFects oF stimulus duration on the habituation oF

a head-shake response. h. A. Thesis, Michigan State University,
1966.

Berg, R. L., and Beebe-Center, J. G. Cardiac startle in man. J. exp.
pS‘ZChOlo, 1941, 2.9., 262-279.

Brown, J. H. Acquisition and retention oF nystagmic habituation in
cats with distributed acceleration eXperience. J. comp. physiol.
Psychol., 1965,_§Q, 340-343.

Capps, m. J., and Collins, W. E. EFFects oF bilateral caloric habitu-
ation on nystagmus reSponses oF the cat. FAA Civil Aeromedical
Research Institute ReportL No. 64-14, 1965, 22 p.

Fox, 5. S. Evoked potential habituation rate and sensory pattern pre-
Ference as determined by stimulus inFormation. J. comp. physiol.
pSYChOlo, 1964, E, 225-2320

Galbrecht, C. R., Dykman, R. A., Reese, w. G., and Suzuki, T. Intra-
session adaptation and intersession extinction oF the components
oF the orienting response. J. CXp. Psychol., 1965, 20, 585-589.

Gardner, L. E. Habituation in the earthworm: Retention and overha-
bituation. Ph. D. Thesis, Michigan State University, 1966.

Guedry, F. E., Jr. Habituation to complex vestibular stimulation in
man: TransFer and retention oF eFFects From twelve days oF ro-
tation at 10 rpm. Percept. mot. Skills, 1965,‘gl, 459-481.

Harris, J. D. Habituatory response decrement in the intact organism.
DSYChOlo BUllo, 1943, $9., 385-4230

Hinde, R. A. Factors governing the changes in strength oF a partially
inborn reSponse, as shown by the mobbing behavior oF the chaFFinch
(Fringella coelebs): II. The waning oF the response. Eggs. Roy.
Soc. Lon., l954,_142, 331-359.

Humphrey, G. The nature oF learning in its relation to the living
system. New York: Harcourt, Brace and Co., 1933.

Keen, R. E. EFFects oF auditory stimuli on sucking behavior in the
human neonate. J. exp. child Psychol., 1964, 1, 348-354.

Keen, R. E., Chase, H. H., and Graham, F. K. Twenty—Four hour reten-

tion by neonates oF an habituated heart rate reSponse. Psycho.
.§9;., 1965,.2, 265-266.

51

 

Koepke, J. E., and Pribram, K. H. Habituation oF GSR as a Function oF
stimulus duration and Spontaneous activity. J. comp. physiol.
Psychol., 1966,_§l, 442-448.

Lehner, G. F. J. A study oF the extinction oF unconditioned reFlexes.
J. exp. Psychol., 1941, 29 435-456.

_‘;.;_9

 

Meyer, K. E. Startle response: Habituation over trials and days, and
sex and strain diFFerences. J. comp._physiol. Psychol., 1963,
_56, 863-065.

Presser, G. L., and Hunter, w. 5. The extinction oF startle reSponses
and spinal reFlexes in the white rat. Amer. J. Physiol., 1936,
117, 609-618.

 

Ratner, S. C. Annelids and Learning: A critical review. In N. C.
Corning and S. C. Ratner, Chemistry oF learning. New York: Plenum
Press, (in press).

Ratner, S. C. Comparative aspects oF hypnosis. In J. E. Gordon (Ed.),
Clinical handbook oF experimental hypnosis. New York: The Mac-
millan Co., 1967, pp. 550-587.

Ratner, S. G., and Denny, m. R. Comparativegpsychology. Homewood,
Illinois: Dorsey Press, 1964.

Rodgers, w. L., Melzack, R., and Segal, J. R. "Tail Flip reSponse" in
goldFish. J. comp.#physiol. Psychol., l963,_§§, 917-923.

Scott, J. P. Animal behavior. Chicago, Illinois: University oF
Chicago Press, 1958.

Sharpless, 5., and JaSper, H. Habituation oF the arousal reaction.
Brain, 1956,.19, 655-600.

Siegel, S. Nonparametric statistics For the behavioral sciences.
New York: McGraw-Hill, 956.

Thompson, R., and Spencer, w. Habituation: A model phenomenon For
the study oF neuronal substrates oF behavior. Psychol. Rev.,
1966, 13, 16-43.

Thorpe, H. H. Learning and instinct in animals. Cambridge, Mass.:
Harvard University Press, 1963.

Tinbergen, N. The study oF instinct. CyFord, England: Claredon
Press, 1951.

Uilson, K. V. A distribution-Free test oF analysis oF variance hypoth-
eses. Psychol. Bull., 1956, 23, 96-101.

 

53

APPENDIX A

Total number oF head-shake responses per session For each subject.

Experimental Grogp

 

 

Sessions
Subject 1 2 3 4 5 6 7 8 9 10 11 l2 13 14

 

1 240 183 175 212 195 185 195 186 183 197 176 188 216 175
2 165 186 153 133 131 146 175 156 200 164 158 166 152 166
3 157 155 201 186 197 185 180 170 156 148
4 240 278 245 213 271 97 227 223 209 154 186 229 220 189
5 195 202 237 255 243 189 285 207 268 186 198 187 229 210
6 201 185 212 194 189 158 178 194 197 184 156 201 146 100
7 121 80 91 90 112 116 91 121 98 91 103 112 87 88

8 73 86 99 87 78 67 68 68 65 50 70 71 87 81

Control Group

 

 

1 99 101 123 135 82 104 99 94 88 127 155 135
2 23 9 34 27 56 66 48 54 57 36 53 39
3 25 22 40 30 44 30 34 29 20 19 35 69
4 21 19 31 24 22 24 24 13 9 9 6 5

5 177 117 98 84 100 118 163 131 113 135 117 90
6 103 126 135 102 90 89 152 147 125 154 118 49
7 33 33 31 24 38 49 57 47 51 50 50 42

8 2 19 24 24 34 57 55 72 61 76 89 77

(For each group, 5s 1-4 are males, gs 5-8 are Females.)

 

MIC

\

 

ll 111311121111 1111111111) 11111111“