—""'VVT—VV—vvw"vv " I
- .

4..“ AAA L~-‘- .

vv vv.

 

 

--o.o.-90>oco«"o‘qo' '¢-00§9¢~~O\90Qo 2’ ’ O qn-aq.9mmflohIQFv-w—v

a
o

-,
o...—.O.--o...‘.yo‘.. ....

OUO'I'.C-'b l'

'ugro

C

0:97...

. .«I‘O. . .ofl'

6.. O .“

 

N, AND

HABITUAUON, DISH

ABITUATIO

' 3

.

SANA

NAL

.

TY

LAST ABDOMI

P. AMERI
W. s. .

ROACH,

eo

Q

‘ PHILUP WILLIAM SINGER

   

e Degre

MICHIGAN STATE UNIVERSI

OF THE 00

Thesisrfor th

HABITUATION IN THE-

'GANGUON

 

.
.
.
_
g
.
..
...
_
.
.
..
o
.-
— o
.
..r .
'-
_.
.‘
u
.. .
a I
.
v
.
.
. o. :0, no,
'.I
. .71.“... O“
.-
u
.
.
.
.
_'
. .
_
_
_.
.
_ ..
..
. . . .
. ..‘......r.vo.
..
o
_ .
. a
u
_
.
.
.
. .. .. . . . c
. c. .
. . . b... ‘.
. s ,a. o . u.
. . . ..... ..
. _ . u
.. . r..
.0. u v.
. .. ....' 0 .
.. o C .n. .o..
. . .. ro‘ . o .
.
.. . o ..' . u 0 ..
.. .
c . . u .H. .,
.
o . sno .O'o . u
.
‘ .. . .u. . . VD. .. u
. y .o . . . u
. . I. . a
.. , ....
.. v..o. .
. c n .3...) .
. . .
. . t.. . .v
. .. l. . t . .
. _. on on. v.
.. .
.9
r . ,. . .x. ‘ ......t\.t
- . v 1.. .0
_
. ...-.. q 25. . .o .5. .uo..o......oIs.n
1" L» I I'D

 

 

 

 

I; ‘ h..-

W» “Inn“ 3.
H 9 9 h 1'9 - 7"

N ‘1." {2,14 ‘ '
JHP—wb Universiy ,. «.

. J

“AI-3‘ ‘ J.) .-l

 

.-.Al 5—. K -.A

 

ABSTRACT

HABITUATION, DISHABITUATION, AND TRANSFER OF
HABITUATION IN THE LAST ABDOMINAL GANGLION
OF THE COCKROACH, E. AMERICANA

 

BY

Phillip William Singer

This investigation was performed to demonstrate a
model system suitable for the study of habituation. The
number of spikes recorded in axons of the connectives

anterior to the last abdominal ganglion of the cockroach,

 

g. Americana, in response to deflection of the cercus, were
counted as a function of both number and frequency of
deflections. The number of spikes decreased as the number
of deflections increased, but returned to a higher level
after a rest period. The decrease in number of spikes was
moderated by increasing the frequency of deflection; this
decrease was also moderated by deflection of the contra-
lateral cercus. After a series of deflections, deflection
of the contralateral cercus yields fewer spikes than if not
preceded by a series of deflections. These results are
taken as examples, respectively, of habituation, with the

usual properties of increased frequency of the habituating

Phillip William Singer

stimulus; dishabituation by a novel stimulus, and transfer
of habituation; in a simple neural system.

The preparation was prepared by dissecting out the
last abdominal ganglion, cutting the anterior connectives
and cereal motor nerves, and attaching a device for
mechanically deflecting a cercus medially. Recordings
were made with suction electrodes, platinum hook electrodes,
and glass insulated tungsten microelectrodes.

Deflection of a cercus results in a train of 15-30
spikes, lasting 40-70 msec., being recorded in the anterior
connectives. Repeating the stimulus at a rate of 1.6/sec.
results in a reduction of the response in both number of
spikes and length of train. The reduction is complete
within 10 trials, and consists of a 50% reduction of spikes
and a reduction of train length to 20-25 msec. The number
of spikes and train length are highly correlated.

Dishabituation was found to require a stimulus
upon the contralateral cercus different from that on the
ipsilateral cercus. Dishabituation does not result from an
increase in intertrial background discharges.

Transfer of habituation is apparent after a single
training trial. The degree of habituation transferred was
on the order of a 10% decrement from the initial value,
as compared to a 50% decrement on the trained cercus.

Close analysis of the spike train revealed that

there are three classes of spikes in the train: Giant

Phillip William Singer

Spikes, which did not habituate; medium spikes, which did;
and small spikes, which are scattered throughout the
train. An afterdischarge was found in those axons
responsible for the medium spikes. The theory is
advanced that the habituation mechanism is a reduction of

the afterdischarge, reducing the train to 25 msec.

HABITUATION, DISHABITUATION, AND TRANSFER OF
HABITUATION IN THE LAST ABDOMINAL GANGLION

OF THE COCKROACH, B. AMERICANA

 

BY

Phillip William Singer

A THESIS

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

MASTER OF SCIENCE

Department of Biophysics

1974

ACKNOWLEDGMENTS

This research was supported by NSF grant GBZ3371
to E. M. Eisenstein. P. W. Singer was supported by funds
from the colleges of Osteopathic Medicine, Human Medicine,
and Natural Science. Thanks are due Norman St. Pierre
for technical assistance. Thanks are also due Drs. E. M.
Eisenstein, J. I. Johnson, and C. Tweedle for helpful

suggestions.

ii

TABLE OF CONTENTS

Page

LIST OF TABLES . . . . . . . . . . . . iii
LIST OF FIGURES O O O O O O O I O O O 0 iv
INTRODUCTION 0 O I O O O O O O O O I O 1
Definition of Learning . . . . . . . . . 1
Habituation C O O O O O O N O O O O O O 3
How to Study Habituation . . . . . . . . . 5
Last Abdominal Ganglion . . . . . . . . . 6
Giant Fibers O O I O O O I O O O O O O 7
MATERIALS AND METHODS . . . . . . . . . . 10
Experimental Preparation . . . . . . . . . 10
Experimental Design . . l . . . . . . . ll
Habituation . O O O O O O O O O I O 13
Dishabituation . . . . . . . . . . . 13
Transfer of Habituation . . . . . . . . 14
Data Analysis . . . . . . . . . . . . 14
RESULTS 0 O O O O O O O O O O O O O O 16
Single Deflection O 0 C O O O O O O O O 16
Habituation O O ‘ O O O O I O O O O O O 16
Transfer of Habituation and Dishabituatio . . . 25
Microelectrode Studies . . . . . . . . . 30
DISCUSSION 0 O 0 O O C O O O O O O C O 33
Model . . . . . . . . . . . . . . . 34
SUMMARY 0 O O O O O O O O O O O O O O 36
APPENDIX . . . . . . . . . . . . . . . 37
LIST OF REFERENCES . . . . . . . . . . . 41

iii

LIST OF TABLES

Table Page

1. Data on dishabituation. The results with
the pulse stimuli are not significant;
those with the brush are significant
(p < .05, two-tailed sign test) . . . . 27

iv

Figure

1.

10.

11.

LIST OF FIGURES

Diagram of the neuroanatomy of the last
abdominal ganglion . . . . . . . .

Top: Diagram of preparation used; Bottom:
Placement of recording electrodes . . . .

Results of repetitive stimulation, recording
as in Figure 2a . . . . . . . . .

Correlation between length of spike train and

the number of spikes in the train . . . .

Habituation at different stimulation rates .

Effects of a two minute rest upon subsequent
habituation runs . . . . . . . . .

Typical response in cercal nerve, showing no
habituation . . . . . . . . . . .

Relation of transfer of habituation to
habituation O O O O O O O O O O 0

Example of dishabituation with a brush
stimulus '. . . . . . . . . . . .

Examples of background activity on connectives
prior to testing for dishabituation . . .

Typical recordings with micro- and hook
electrodes, as in Figure 2c . . . . . .

Page

12

17

19
20

22

23

26

28

29

31

INTRODUCTI ON

This study is in the area of learning and memory.
This problem has many facets, one of which is the difficulty
of even formulating the questions. Evidently, before any
precise work can be done, the underbrush must be cleared

away.

Definition of Learning

 

The subject area itself, learning and memory, is
poorly defined. Some authors (Glaser 1966) are so restric-
tive as to exclude all animals, except some mammals, from
the capability of learning. Others (Eisenstein 1972) want
to generalize it enough to allow protozoa to be said to
learn. This distinction can roughly be said to center
around the amount of "thought" believed to be used in the
learning operation by the animal involved.

Glaser provides a suitable example of a restricted
definition of learning. He defines learning as the produc-
tion of new responses; the animal now responds to a stimulus
in a manner which it had never used previously in response
to any stimulus. Glaser feels that this forces learning to
be a by—product of what he believes to be "thought." This
definition excludes phenomena such as Pavlovian conditioning,
and Glaser thinks this desireable.

l

The exact content of "learning" is an epistemologi-
(cal question that each person must ultimately answer for
himself. But there are good methodological reasons for
defining learning so as to incorporate such phenomena as
habituation and conditioning. "Thought" is not a quantity
that can be measured, either quantitatively or qualita-
tively, nor is there any agreement among philosophers as to
what "thought" is. Phenomena such as conditioning are the
easiest to examine experimentally and interpret theoreti-
cally. This is because all possible experimental outcomes
can be listed in advance; the system is self-contained.
Moreover, the patterns of stimuli are as simple as possible,
and the stimuli can be made as simple and explicit as
possible.
A general definition is that of Eisenstein. The
formal statement is as follows:
A system is said to demonstrate learning when the

output (response) to a given test input (stimulus) is

a function of the total previous input—output pattern

of which the test input was a part.
Such a formulation includes instrumental learning as well
as Pavlovian conditioning and habituation.‘ It applies to
any case in which the organism's behavior (output) changes
over time, in response to some repeated input (stimulus),
and where the change is dependent on the particular context

in which the stimulus and response occurred. One advantage

is that it allows one to sensibly discuss learning related

phenomena in isolated parts of the nervous system, e.g., a

single ganglion.

Habituation

 

Habituation was chosen as the specific type of
learning for this study for the following reasons. It
seems to be the most primitive type of learning, as well as
one of the most widespread. It is also involved in most
other kinds of learning. Finally, the theoretical analysis
is straightforward.

The following is the most common definition of
habituation:

A subject (or preparation) is said to exhibit
habituation to a stimulus when, after repeated pre-
sentations of the stimulus (l) the response shows a-
decrement, and (2) after a rest period, the response
level demonstrates recovery.

However, the theoretical description of habituation is
currently unsatisfactory. Theories of habituation are,
without exception, ad hog postulates which do little more
than shift the unknown quantity into something less amenable
to experimentation than habituation itself. The definition
is refined in the methods section; elsewhere I attempt
further clarification (Singer, in preparation). I there-
fore digress slightly to show what is wrong.

First, there is no generally accepted definition of
habituation. Some workers (Gardner 1968, Baxter 1957,

Hughes 1965) use our definition, while others (Kupferman

et a1. 1970) require all of Thompson and Spencer's (1966)

nine criteria to be present before they are willing to
call a response decrement "habituation." The root of the
problem is ignorance about what is happening theoretically.

Sokolov (1963) and Glaser (1966) illustrate one
extreme of the confusion. Their theories explain habitua-
tion by postulating a processing of neural information in
the brain. Of course, no one knows how the brain processes
information, and thus these theories explain nothing. But
more importantly, there are too many instances of habitua-
tion in brainless preparations for these theories to be the
answer.

Stein (1966) has a theory intermediate in confusion.
Rather than leave the explanation in terms of brain pro-
cessing, he reduces habituation to what he takes as a
primitive process, Pavlovian conditioning. However, it
seems much more likely that Pavlovian conditioning is a
more complicated phenomenon. Furthermore, no attempt is
made to explain conditioning, so habituation is not truly
explained.

Groves and Thompson (1970) illustrate the opposite
extreme to Sokolov. Their theory purports to be a molecular
mechanism. Close examination reveals that they do nothing
but postulate that the neural elements involved act in
such a way as to produce the experimentally observed

response. Moreover, they have to postulate a neural

process, "facilitation," which they admit is not directly
observable experimentally.

Dispite the confusion, habituation is worth study-
ing. Instances of habituation are occurring almost con-
stantly in sensory systems of animals. For example, one
is not normally aware of temperature, background noises,
or the clothes on one's back even though these stimuli are
constantly there. Any organism must have a way of subordi-
nating continuous stimuli to novel, and thus potentially

dangerous stimuli, and the process is that of habituation.

How to Study Habituation

 

Having decided on the problem to study, it is time
to decide on the preparation as well as the stimulus,
input, and response output to be measured. Many different
systems have been exploited in habituation studies, such
as the startle response of rats (Korn and Moyer 1966),
recognition in infants (Frantz 1964), the escape response
of earthworms (Gardner 1968), the wiping response of the
frog (Kimble and Ray 1963) and contractility in Protozoa
(Osborn et a1. 1973).

The difficulty with most of these preparations is
that the results are not unambiguous. It is difficult to
exclude other stimuli from the animal, and most of these
other stimuli are capable of causing or modifying the

desired behavior. Furthermore, even if the technique of

the experiment is correct, the data are difficult to
interpret. This is because often the interactions of the
entire nervous system are acting on the reaponse.

A more tractable approach then, is to use a phylo-
genetically simpler or surgically simplified preparation.
Such studies run the gamut from single cell studies in
isolated mammalian spinal cords to protozoa. Molluscs
anthropods are convenient preparations as integrative func-
tions tend to be localized in ganglia near the site inner-
vated by the ganglion (Bullock and Horridge 1965) . These
ganglia can be isolated from the rest of the nervous
system With little apparent loss of integrative capacity.
Moreover, the surgery is considerably simpler than with
animals having highly developed respiratory and circulatory

systems.

Last Abdominal Ganglion

 

The preparation we have chosen to exploit is the
isolated last abdominal ganglion of the cockroach. Besides
the above mentioned technical advantages, this preparation
has some methodological advantages. The sensory stimuli
on the cerci form a well defined input. Moreover, asso-
ciated with the cerci is a behavior (escape) seen in the
intact animal.

The central nervous system is organized around 11

ganglia joined together by a pair of connectives. These are

two ganglia in the head, three in the thorax (one for each
pair of legs), and six in the abdomen (Guthrie and Tindall
1968) (for simplicity, we are ignoring the stomatogastric

nervous system, which does not seem to be involved in what
we will study).

The classical picture of the microanatomy of the
last abdominal ganglion and associated structures was drawn
by Pumphrey and Rawdon-Smith (1937) and by Roeder (1948a),
using extracellular recording. They concluded that each
cercal nerve contains hundreds of small nerve fibers.

These fibers pass into the last ganglion, where some

synapse onto cell bodies in either the isolateral'or contra-
lateral side, while others, the so-called through fibers,
continue up the connectives uninterrupted, again in either
side. However, it has recently been shown (Richard et a1.
1973) that this ganglion is more complicated in design and
function than this simple picture would suggest, and care
must be taken to insure that the output observed is

entirely due to the experimental input.

Giant Fibers

 

An important class of cell bodies in the ganglion
are the ones which give rise to giant fibers. First
Roeder (1948a), and later Pipa et a1. (1959) showed that
there are six to eight of these fibers in each connective,

lying in three tracts. They fire reliably to electrical

 

 

 

 

 

 

 

 

 

 

Figure l.--Diagram of the neuroanatomy of the last
abdominal ganglion. S: Sensory fibers;
G: Giant fibers; M: Fiber receiving synaptic
input from giant fiber, possibly medium fiber
of text; T: Through fibers.

stimulation of the cercal nerve at rates up to 20/sec.

By histologically examining lesioned specimens at varying
degrees of degeneration, it was found (Hess 1958, Farley
and Milburn 1969) that each giant fiber had several cell
bodies associated with it. However, dye injection studies
have only been able to stain one cell per fiber (Harris
and Symth 1971, Milburn and Bently 1971).

Giant fibers are important for two reasons. The
first is technical; they have a large action potential,
easily distinguished from both background noise and other,
smaller, axons (Bullock and Horridge 1965). This makes
them convenient preparations for electrophysiological
studies. The second is behavioral. By tradition, they
have been ascribed as mediating escape behavior (but see
Appendix 1).

These details can now be related to the preparation
used experimentally. For simplicity of analysis, extraneous
components, such as input from other ganglia, were removed
from the system. Next, a suitable method of stimulation
the cerci was derived. Finally, a system for recording

from the nervous system was constructed.

MATERIALS AND METHODS

Experimental Preparation

 

Thirty-nine adult, male cockroaches were used.

They were anesthetized with CO gas, and the legs and wings

2
removed. Next, they were pinned, dorsal side up on a clay
block, the dorsal cuticle, the gut, and the fat body
removed. The last abdominal ganglion was isolated by cut-
ting the motor nerve to the cercus and also cutting the
interganglionic connectives just posterior to the fifth
abdominal ganglion. The exposed nervous system was covered
with Roeder's saline (1948b). When using suction electrodes
the system was next covered with a thin film of Vaseline.
For stimulation, we wanted to stimulate the sensory
receptors as naturally as possible. Initially, air puffs
on the cerci were tried. This resulted in very few spikes
being recorded on the connectives and was difficult to
analyze. However, a well defined train of spikes could be
generated if a blunt insect pin was attached to the voice
coil of a 2" p.m. loudspeaker and fastened to the fifth or
sixth joint of a cercus with wax so as to produce a medial

deflection of the cercus whenever a pulse was fed to the

loudspeaker.

10

11

To prevent the deflection from vibrating the entire
animal and possibly stimulating other receptors, the base
of the cercus was fastened to the block with clay. No
spikes were recorded on the connectives when the sensory
nerve was cut and the cercus thus stimulated. Occasionally,
the wax joint would break. No response was then recorded.

Three msec. square wave pulses provided by a Tek-
tronic 161 & 162 modular pulse generating system were used
to drive the loudsPeaker, a homebrew amplifier being
inserted between them as an impedance metch. For recording,
platinum hook electrodes, suction electrodes (consisting
of silver wire in a polyethylene tube), and glass insulated
tungsten microelectrodes were used, depending on the
specific experiment. The preparation and recording pro-

cedures are diagrammed in Figure 2.

Experimental Design

 

The experimental designs were based upon the follow-
ing considerations. First, they are immediately adapted to
a simple statistical test, the sign test. Second, they
permit unambiguous definitions of habituation, dishabitua-
tion, and transfer of habituation. Finally, they are
delicate enough to detect effects hidden within a large
amount of scatter. For example, the magnitude of the trans-
ferred habituation is less than the normal variation in the
initial response. However, the response reduction is con-

sistent enough to be detected by a sign test.

12

fl Au

A6

 

 

 

P'n I‘ Speaker

Wax

 

Figure 2.--Top: Diagram of preparation used. Dorsal
view. A4 - A6: Last three abdominal ganglia.
M: Cercal motor nerve; S: Cercal sensory
nerve.

Bottom: Placement of recording electrodes.

A: Suction electrode over both cut connectives
of A6; B: Suction electrode over cut sensory
nerve; C: Hook electrode and microelectrode
simultaneously on cut connectives.

l3

Habituation

 

The following operational definition was used for
habituation in this study. A response decrement was said
to occur in a given series of stimulus trials (called runs)
whenever the number of spikes recorded on trial 10 was less
than that of trial 1; and, after a two minute rest, the
response demonstrated at least partial recovery, defined as
a response increase over that of trial 10. Habituation
was said to have been demonstrated in a given preparation
if six or more of the eight runs showed such a decrement.
See Bradley (1968) for details of statistics used in this

paper.

Dishabituation

 

Two different stimuli were used, a single pulse
from a stimulator of the contralateral cercus or a brushing
of this cercus. The stimulus was applied between trials 10
and 11, with a five minute rest between runs. A response
increase was said to occur if the number of spikes on
trial 11 (the "test stimulus") was greater than the number
on trial 10. A given preparation was said to exhibit dis-
habituation to a given stimulus if three or more of the
five runs made demonstrated a response increase. Both
stimuli were used on each animal, one being used in five
consecutive runs, and then the other for five runs, the

order being varied on different animals.

14

Transfer of Habituation

 

A pulse stimulator was attached to each cercus.
An experiment on a given preparation consisted of 10 runs,
one each of 1, 2, 4, 6 or 10 trials on either the right or
left side, which was then followed by a "test" trial on
the contralateral cercus at what would have been the time
of the next regular stimulus. Thus there were five runs
which began on the right cercus and five which began on
the left. Transfer was defined to have been demonstrated
in a given run if the number of spikes to the "test" was
less than the expected response to the first stimulus on
that side. The expected response to the first stimulus on
a side was computed as the median number elicited by the
first stimulus of the five runs beginning on the side in
question. The order of the runs was determined randomly
for each experiment. A five minute rest was allowed

between runs.

Data Analysis

 

The data were analyzed in the following manner.
The number of spikes on the first trial of a run was scored
as 100 percent. The number of spikes on subsequent trials
were scored as a percentage of this first trial. Multiple
runs were averaged by medians. Scatter was measured by

the interquartile range.

15

The following variation was used for Figure 6,
which compares different initial trials. For a given
animal, trial 1, run 1 was scored as 100 percent. Subse-
quent trials, including trial 1 on subsequent runs, were

scored as a percentage of trial 1, run 1.

RESULTS

Single Deflection

 

The deflection of a cercus resulted in a train of
spikes being recorded in the preparation (Figure 3A).
There were three classes of spikes in the train; large
spikes, somewhat smaller (or medium) spikes, and very small
spikes. The large spikes, probably from giant fibers,
appeared primarily in the first 20 msec., and the medium
ones appear primarily thereafter. The small spikes were
scattered randomly throughout the trace.

Throughout this paper, when "spike" is mentioned,
it is a large or a medium spike that is being considered.
The small spikes were scattered randomly throughout the
trace, and were easily distinguished from the other Spike
classes. The spike train, so defined, elicited by a
single deflection of a well rested preparation, contains
15-30 Spikes and lasts 40-70 msec., depending on the

animal.

Habituation

 

Repeating the stimulus lead to a progressive
decrease in length of train and quantity of spikes
(Figure 3B, C). A two minute rest resulted in recovery.
This was noted in all 39 preparations, usually in all

16

17

 

10 msec.

Figure 3.--Results of repetitive stimulation, recording
as in Figure 2a. A: Response to first trial;
B: Response to trial 10 at a stimulation
frequency of l.6/sec.; C: Response to trial
10 at a stimulation frequency of 1/10 sec.

18

runs. The spikes did not "drop out" uniformly throughout-
the train; rather they progressively "dropped out" from
the left as is shown by the high correlation between train
size and duration (Figure 4). Little, if any change
appeared in the first 15-25 msec. of the train. Thus,
habituation occurs mainly in the class of "medium size"
fibers described above.

Reducing the frequency of stimulation resulted in
a reduction of the degree of habituation, the differences
appearing in the center portion of the train (Figures 3c
and 5). Intervals longer than 30 seconds did not lead to
noticeable habituation over a 10 to 20 trial period.
Individual variations tended to mask the frequency effects
for the early trials; however by trial 10 the different
frequencies were seen to fall into clearcut rate classes
(Figure 5).

There was no deterioration of the preparation
during the experiments; as for the 34 preparations for
which spikes were counted, the initial trial on the last
run contained more spikes than the initial trial of the
first run in 15 cases. Fifteen preparations showed the
reverse, and there were four ties.

We had planned to compare subsequent runs with the
first run to See if repeated training results in faster
habituation. Preliminary work had suggested that recovery

was complete within two minutes, following 20 trials at

19

30

25
Number
of 20
Spikes

IS

IO

 

3

Length of Train (in mueL.‘

Figure 4.--Correlation between length of spike train
and the number of spikes in the train. Study
made on-4 runs of one preparation at the rate
of 1.6/sec. Correlation significant at .001
level (Hotelling and Pabst' s Spearman rank
order correlation test).

20

 

1 .
I \
I \
, \
4 , \ ’ Frequency
\ I
’l \ I, \s I "’A\“ (S?C.)
'00 : A ‘v-. a” “\‘ l/jo
q
° l/IO
z
m
U)
'0 ~.
0 ~~1~
Z ‘s ‘
m ~~_~
m ‘~+ l/S
\‘ .
“““ '
‘ ° ----- ° L6

 

 

 

Figure 5.-—Habituation at different stimulation rates.
The difference between trials 1 and 10 is
significant at the .05 level for all runs
except 1/30 sec. (two tailed Sign test).
Error bars are interquartile ranges. The
same seven animals used for all frequencies
except 1/30 sec., where data on four animals

were collected. Median percent reSponse
plotted.

21

the rate of 1.6/sec. A more detailed examination (Figure 6)
revealed that there is still some retention at the start

of the second run. There was no change with further train-
ing; trial 1 or runs 3, 4 and 5 did not differ signifi-
cantly from trial 1 of run 2. Also, there is no increase
in the degree of habituation, all five curves asymptoting
at the same level (N = 12 for all of these statements about
Figure 6; p < .05, two tailed sign test). Recovery is
complete after five minutes, as is shown in the transfer of
habituation series. In 40 instances where the same cercus
was trained twice in sequence, fifteen times the initial
response to the second run was greater than the first, 23
times it was less, and there were two ties. The difference
is not even significant at the .10 level.

Habituation was not due to a reduction of the
sensory input in the cereal nerve, but rather to a reduc-
tion of output from the ganglion. This is seen by looking
at the response in the isolated cercal nerve to stimulation
of the cercus. It consists of 8-12 large action potentials,
which appear to be compound potentials, lasting typically
20-25 msec (Figure 7). That habituation is not due to a
reduction of activity in the cercal nerve is inferred from
the following two observations: (1) using the above
criterion for habituation (6 or more of the 8 runs must
Show a decrement) four animals out of seven were found to

demonstrate a response decrement between trials 1 and 10,

36 RESPONSE

22

IOO‘

90‘

80‘

70'

 

60-

 

50‘

 

 

 

1; SI
TRIAL

“1

N
wd
o‘d
N
m.
KO.
C)

Figure 6.--Effects of a two minute rest upon subsequent
habituation runs. 100% = trial 1, run 1.
Stimulus rate 1.6/sec. Median percent
response plotted.

23

 

Hlnsec.

Figure 7.--Typical response in cercal nerve, showing no
habituation. Recorded as in Figure 2b.
Stimulation rate 1.6/sec. A: Trial 1;

B: Trial 10.

24

two showed an increase, and one showed no change. A two-
tailed sign test is not even significant at the .50 level,
whereas the decrement from the connectives is highly sig—
nificant (p < .001); (2) 80% of all runs on all animals
show a response increase on the sensory nerve between
trials 1 and 2, while more than 80% of the recordings

from the connectives show a decrease between trials 1 and 2.

Although others (Hughes 1965) have also noted that
the cercal nerve response is independent of the number of
stimuli, Zilber-Gachelin and Chartier (1973a) dissent.

They find a partial reduction in cercal activity, in degree
amounting to about half of that found on the connectives.
This discrepancy may be partly explained by our stimulating
primarily one type of receptor (joint receptors) while
Zilber-Gachelin and Chartiers and Hughes stimulated hairs.
At the rates Hughes used, 1/30 second and slower, Zilber-
Gachelin and Chartier found no decrement in the cercal
response either.

In comparing this work to that of Hughes, Constan-
tine (1972), and Zilber-Gachelin and Chartier, it is found
that they did not remove the upper abdominal ganglia. When
we repeated our habituation experiments, but sectioned the
nerve cord at the level of the first, rather than the last,
abdominal ganglion it resulted in well over sixty spikes
being recorded at the level of the last ganglion in a

fifty msec. time frame (as compared with thirty in the

25

isolated ganglion). Neither Constantine nor Zilber-
Gachelin and Chartier were able to find evidence for
transfer of habituation, most likely because the change
is now too small relative to the total response, to
measure.‘ Similar remarks pertain to Zilber-Gachelin and
Chartier's failure to find dishabituation.

Transfer of Habituation
and Dishabituation

 

 

The number of Spikes elicited was reduced if the
contralateral cercus was first stimulated. This transfer
of habituation was evident after the very first ipsilateral
stimulus (Figure 8). In transfer of habituation, no dif-
ference was observed between the left and right sides, and
the results are pooled in the following discussion.

The decrement was significant at the .01 level
for test trials 4 and 10, the .05 level for trials 1 and 2
and the .10 level for trial 6, using a two-tailed sign test
on the 15 animals. The extent of habituation (on the "test"
side) is less than that of the fully habituated side.
Indeed, the test side is habituated only to the extent of
8-15%, compared to approximately 50% for the trained side.
No trend is seen in the magnitude of transfer as a function
of level of habituation training.

Interpolating a suitable stimulus on the contra-
lateral cercus during a run resulted in a response increase

to the next regular stimulus trial (dishabituation). The

IOO

90

80

70

60

% RESPONSE

 

26

  

 

~

(I) (’1) (6)

 

12 31:5 6 7 8 9 10!!
TRIAL

Figure 8.--Relation of transfer of habituation to

habituation. Dotted line is the habituation
curve for the 15 preparations; solid line
signifies transfer to the contralateral cercus.
The point at the end of the solid line is the
median response to the "test" at that point.
Numbers in parentheses refer to the number of
prior training trials before the test for
transfer. Median response plotted. Error
bars are interquartile ranges.

27

typical increase in responsiveness was two spikes (Table 1
and Figure 9). Unlike habituation,'dishabituation was not
present on every run. Critical to the mechanism of the
dishabituation is the determination of whether the
increased portion of the response occurs to the test
stimulus, or is due to an increase in background discharges
as a result of the dishabituating stimulus. No correlation
whatever was found between background activity before trial
11 and the presence or absence of dishabituation on trial
11 (n = 51, p < .95, x2 test). Indeed, on many runs no
background activity whatever was noted. Figure 10 is an

example.

TABLE l.--Data on dishabituation. The results with the
pulse stimuli are not significant; those with
the brush are significant (p < .05, two-tailed
Sign test).

 

No. of Experiments No. of Experiments

 

Disgzg;figiging with without
Dishabituation Dishabituation

Brush 10 0

Pulse

Stimulator 7 3

 

28

 

IO msec.

Figure 9.--Example of dishabituation with a brush stimulus.
A: Trial 10 (before dishabituating stimulus);
B: Trial 11 (after dishabituating stimulus).

29

IO msec.

 

Figure lO.--Examp1es of background activity on connectives
prior to testing for dishabituation. The start
of the response to the test stimulus is visible
at the far right. A: Background prior to last
habituation trial (10); B: Background prior to
(successful) test for dishabituation (11):

30

Microelectrode Studies

 

Harris (1969) stimulated the cereal nerve with
brief electrical pulses and recorded after-discharges from
single fibers in the anterior connectives of the last
ganglion. This after-discharge was located in fibers 10-15
microns in diameter. This probably corresponds to the size
range of our medium sized spikes, the class of fibers which
showed habituation.

The significance of Harris' work can be seen by
considering the following question: What triggers the final
35 msec, of spikes in our spike train? The response in the
cereal never lasts only 25 msec. The known synaptic delays
and conduction velocities can account for almost another
5-10 msec. However, the final 35 msec. of the spike train
is left without a direct stimulus. An after-discharge
would explain the presence of spikes without a stimulus.

Since the problem was to record from single axons,
but not necessarily to make intracellular potential measure-
ments, intracellular electrodes were not necessary. It was
found that glass microelectrodes did not penetrate readily,
nor were they easily inserted at a specific place in the
nerve cord. The tungsten electrodes avoided these problems.
The small tip size (~10u) insured that only a one or two
axons would be recorded from (Figure 11).

The results of the recordings confirm that there is

an after-discharge present in the ganglion, and that indeed

31

 

Figure ll.--Typical recordings with micro- and hook
electrodes, as in Figure 2c. Stimulus rate
l/sec. Upper trace is the hook electrode,

lower trace microelectrode. A: Trial 1;
B: Trial 10.

32

it is responsible for the spike train. No giant fiber
was recorded from in this study, so comparisons between
giant and medium fibers cannot be made. Other classes of
fibers were observed, such as fibers stimulated only by
the contralateral cercus, but they were not studied in

detail.

DISCUSSION

The results just described show that the isolated

last abdominal ganglion is thus a suitable model system

for the study of habituation. This is important, not 7
because of any relationship the model phenomena may have fl
to a behavior in the intact animal (see Appendix 1 for more

discussion on this point) but rather because of the possi-
bility for understanding the details of a neural circuit
capable of producing habituation. This is because the
ganglion is highly amenable to single cell studies.

For example, the work of Callec and Boisted (1971)
demonstrated habituation of the post synaptic potential in
giant cells of the last abdominal ganglion. These results
cannot presently be applied to the spike train, as the
cell bodies involved are unknown; however, the difficulties
in advancing the work to the point where these results could
be applied is much less than in mamalian systems.

Dye injection studies seem to be the quickest way
to determine the microanatomy of the circuit involved in
the habituation. Fortunately, the cockroach is one of the
favorite animals used in such studies. Thus far, efforts

have concentrated on injecting the giant axons, possibly

33

34

because they were the only ones thought to be important.
After this study, there should be some effort made to
inject the medium axons.

The advantages of the definitions employed are
shown by this study having found transfer of habituation,
whereas Constantine reported no evidence of it. Applying
our criteria to her data would indicate that she did
obtain it. The difficulty probably was that in Constan-
tine's preparation the transfer is not dramatic and, having
no clear—cut decision procedure, she preferred to err on
the side of caution. To avoid ambiguities on this point,
it is best to give a detailed, explicit statement of the
criteria one is using, as was done in the Methods section

of this paper.

Biz-12.1.

It seems paradoxical that both transfer of habitua-
tion and dishabituation can be induced by the same input
channel. However, if one postulates that the two cercal
inputs converge and effectively overlap on the same post-
synaptic fiber, and that habituation occurs post synapti-
cally, then habituation and transfer of habituation are
easily explained. Transfer of habituation is explained by
the overlap as the contralateral input finds its post-
synaptic membrane already partially depressed. Dishabitua-

tion might be due to temporal summation. That is, the

35

extra dishabituating stimulus may produce some post-
synaptic depolarization which is not sufficient to produce
spiking. However, the depolarization produced by the

next regularly occurring stimulus in the habituation
sequence adds to the first and produces a larger after-
discharge than would have otherwise occurred (mechanism of

dishabituation suggested by F. Krasne to E. Eisenstein).

SUMMARY

1. Brief mechanical deflection of a Cockroach's
cercus results in a spike train in the anterior connectives
of the isolated last abdominal ganglion which consists of
15-30 spikes and lasts 40-70 msec.

2. Habituation of the spike train is complete
within ten trials at a stimulation range of l.6/sec., and
is retained for at least two minutes.

3. The extent of habituation decreases as the
frequency of stimulation is decreased.

4. Both transfer of habituation and dishabituation
have been observed.

5. The habituation process is hypothesized as the
progressive reduction of after-discharges in "medium“
sized fibers.

6. A possible model for the neural elements

involved is discussed.

36

APPENDIX

37

APPENDIX

GIANT FIBERS AND ESCAPE

For a great many years it was believed that the
giant fiber system in the cockroach mediated the escape
response. Consequently, any work on the last abdominal
ganglion contained a discourse on escape behavior. Very
recent developments have changed the picture, however,
and these events are sketched here for the sake of complete-
ness.

A puff of air to the roaches' cerci results in
escape behavior (Baxter 1957). The receptors on the cerci
are extremely sensitive to air vibrations (Wozniak et a1.
1967), and produce responses in the giant fibers. By the
mid 1960's the correlation between behavior and physiology
seemed complete, and the search was on to trace the giant
fiber response to the legs. The fibers were found to
extend to the brain (Spira et al. 1969a, b; Parnas et a1.
1969). But a funny thing happened on the way to the fore-
brain; the bug turned out not to act like the book said it
was supposed to.

Dagan and Parnas (1970) investigated the role of

giant fibers in escape behavior with various techniques.

38

39

First, they stimulated a single giant axon intracellularly,
with no response in the leg. Next they blocked conduction
of giants; first by cutting the nerve cord at the last
ganglion, allowing the giants to degenerate, and stimulat-
ing anterior to the lesion; secondly by applying nicotine
sulfate to the ganglion which blocks ACH medistal synapses,
and stimulating a cereal nerve. In both cases they were
able to obtain a response in the leg. Finally, they
checked the threshold of fibers in the connectives and
found the threshold for leg activity higher than that of
giants. This has been confirmed in the roach by Iles
(1972) and in the locust by Seabrook (1970). Thus giants
are not the sole factor in escape behavior.

Exactly which fiber class is responsible for escape
behavior is presently an open question. (Indeed, it may
be that many classes of fibers are involved, making it a
meaningless question). The best guess seems to be that a
class of "medium fibers," intermediate in size between the
gianst and the small through fibers is responsible for
conducting the escape message (Iles 1972; Seabrook 1970;
Harris 1969).

Surprisingly, there are apparently giant fibers
which synapse onto other ascending axons in the last
ganglion in addition to sending their axons upwards (Harris
and Smyth 1971; Milburn and Bently 1971). Obviously,

stimulating a giant fiber in the connectives need not affect

40

these fibers. Nor will they degenerate when the connec-
tives are transected. Indeed, if these synapses are not
blocked by nicotine sulfate then these fibers could be

the medium fibers of this paper. It would be of interest
to know whether nicotine sulfate blocks spike train we
have observed in the last ganglion. If it did, then
escape behavior must be mediated in the thoracic ganglia,
and the spike train and escape behavior must be handled in
different neural circuits.

It is interesting to speculate about the relation-
ship of the spike train to escape behavior, but little is
known about the neurophysiology of the last abdominal
ganglion in the intact animal. Most studies done so far
(Zilber-Gachelin and Chartier 1973b; Dagan and Parnas 1970;
Illes 1972) have been on restrained animals, which have
also frequently been surgically modified. However, it
has been found that both the behavior and the neuro-
physiology of such a preparation differs significantly
from the intact animal. The leg movements observed vary,
depending on how free the animal is, and similar stimuli
produce different giant fiber responses depending on
whether the animal is intact or surgically modified

(V. Chen, personal communication).

LI ST OF REFERENCES

41

LIST OF REFERENCES

Baxter, C. 1957. Habituation of the roach to puffs of
‘ air. Anat. Rec. 128, 521.

Bradley, J. V. 1968. Distribution-Free Statistical Tests.
Englewood Cliffs, N.J.: Prentice Hall.

Bullock, T. H., Horridge, G. A. 1965. Structure and
Function in the Nervous System of Invertebrates.
San Francisco and London: W. H. Freeman and Co.

Constantine, S. B. 1972. A neural concomitant to habitua-
tion in the cockroach, P. Americana. Ph.D. Thesis.
Columbia University.

 

Dagan, D, Parnas, I. 1970. Giant fibre and small fibre
pathways involved in the evasive reSponse of the
cockroach, P. Americana. J. Exp. Biol. 52, 313-332.

 

Eisenstein, E. M. 1972. Learning and memory in isolated
insect gangliaiJ1"Advances in Insect Physiology"
vol. 9.

Eisenstein, E. M., Peretz, F. 1973. Comparative aspects
of habituation in invertebrates in "Habituation,
vol. II." H.V.S. Peek and M. J. Herz, ed., pp. 1-34.
New York: Academic Press.

Farley, R. D., Milburn, N. S. 1969. Structure and
function of the giant fiber system in the cockroach,
P. Americana. J. Insect Physiol. 15, 457-476.

 

Frantz, R. L. 1964. Visual experience in infants:
Decreased attention to familiar patterns relative
to novel ones. Science 146, 668-670.

Gardner, L. E. 1968. Retention and overhabituation of a
dual-component response in L. terrestris. J. Comp.
Physiol. Psychol. 66, 315-318.

 

Glaser, E. M. 1966. Physiological Basis of Habituation.
Oxford: Oxford University Press.

42

43

Groves, M, Thompson, R. F. 1970. Habituation: a dual-
process theory. Psych. Rev. 11, 419-450.

Guthrie, D. M., Tindall, A. R. 1968. The Biology of the
Cockroach. London: St. Martins.

Harris, C. L., Smyth, T. 1971. Structural details of
cockroach giant axons revealed by injected dye.
Comp. Biochem. Physiol. 40A, 295-303.

Harris, C. L. 1969. Existence of a neural circuit for
delayed excitation in the American cockroach. Ph.D.
Thesis, Penn. State University.

Hess, A. 1958. Experimental anatomical studies of path-
ways in the severed central nerve cord of the
cockroach. J. Morphol. 103, 479-502.

Hughes, G. N. 1965. Neuronal pathways in the insect
nervous system in "The Physiology of the Insect
Central Nervous System." J. E. Treherne and
J.W.L. Bennent, ed. New York: Academic Press.

Iles, J. F. 1972. Structure and synaptic activation of
the last coxal depressor motoneuron of the cock-
roach. J. Exp. Biol. 56, 647-656.

Kimble, D. P., Ray, R. S. 1965. Reflex habituation and
potentiation in R. pipiens. Anim. Behav. 13,
530-533.

)

Korn, J. H., Moyer, K. E. 1966. Habituation of the
startle response and of heart rate in the rat.
Can. J. Psychol. 20, 183-196.

Kupferman, I.; Castellucci, V.; Pinsker, H.; Kandel, E.
1970. Neuronal correlates of habituation and
dishabituation of the gill-withdrawal reflex in
Aplysia. Science 161, 1743-1745.

Milburn, N. 5., Bentley, D. R. 1971. On the dendritic
topology and activation of cockroach giant inter-
neurons. J. Insect. Physiol. 11, 607-623.

Osborn, D.; Blair, H. J.; Thomas, J.; Eisenstein, E. M.
1973. The effects of vibratory and electrical
stimulation on habituation in the ciliated
protozoan S. ambiguum. Behav. Biol. 8, 655-664.

44

Parnas, I.; Spira, M. E.; Wermaz, R.; Bergman, F. 1969.
Non-homogeneous conduction in giant axons of the
nerve cord of 2, Americana. J. Exp. Biol. 50,
635-643. —_

 

Pipa, R. L.; Cook, E. F.; Richard, G. 1959. Studies on
the Hexapod nervous system. J. Comp. Neurol. 113,
401-433.

Pumphrey, R. J., Rawdon-Smith, A. F. 1937. Synaptic
transmission of simple nerve impulses through the
last abdominal ganglion of the cockroach. Proc.
Roy. Soc. Lond. 3222, 106-118.

Richard, G.; Zilber-Gachelin, V. F.; Chartier, M. P. 1973.
What are the so called 'cercal nerves' in the
cockroach 2. cranifer? J. Insect Physiol. 22.

Roeder, K. D. 1948a. Organization of the ascending
giant fibres system in the cockroach, 2, Americana.
J. Exp. Zool. 108. 243-261.

 

Roeder, K. D. 1948b. The effect of potassium and calcium
on the nervous system of the cockroach, 2. Americana.
J. Cell Corp. Physiol. 22, 327-338.

 

Seabrook, W. D. 1971. An electrOphysiological study of
the giant fiber system of the locust 2. gregaria.
Can. J. Zool. 22, 555-566.

Singer, P. W., Eisenstein, E. M. 1972. Habituation in the
last abdominal ganglion of the cockroach. Abstract
for 1972 Am. Biophys. Soc. Meeting.

Sokolov, Y. N. 1963. Perception and the conditioned
reflex. Oxford: Pergamon.

Spira, M. E.; Parnas, I.; Bergmann, F. 1969a. Organiza-
tion of the Giant axons of the cockroach,
' 2. Americana. J. Exp. Biol. 22, 615-627.

 

Spira, M. E.; Parnas, I.; Bergmann, F. 1969b. Histological
and electrophysiological studies on the giant
axons of the cockroach, 2. Americana. J. Exp.
Biol. 22, 629-634.

 

Stein. L. 1966. Habituation and stimulus novelty: A
model based on classical conditioning. Psych. Rev.
22, 352-356.

45

Thompson, R. 6., Spencer, W. A. 1966. Habituation: a
model phenomenon for the study of neuronal sub-
strates of behavior. Psych. Rev. 22, 16-43.

Wozniak, A.; Alvarez, R.; Wilson, E.; Austt, E. G. 1967.
Cercal potentials in the 2. Americana. Acta
Physiol. Lat. Amer. 21, 102-111.

 

Zilber-Gachelin, N. F., Chartier, M. P. 1973a. Modifica-
tion of the motor reflex responses due to repeti-
tion of the peripheral stimulus in the cockroach I:
Habituation at the level of an isolated abdominal
ganglion. J. Exp. Biol. 22, 359-381.

Zilber-Gachelin, N. F., Chartier, M. P. 1973b. Modifica-
tion of the motor reflex responses due to repeti-
tion of the peripheral stimulus in the cockroach II:
Conditions of activation of the motoneurones.
J. Exp. Biol. 22, 383-403.

 

R

 

wymmmigwmunnnumt'mmn

00000 91

William)”