Citation
An Analysis of spontaneous and conditioned eyelid response magnitude in Macacca mulatta

Material Information

Title:
An Analysis of spontaneous and conditioned eyelid response magnitude in Macacca mulatta
Creator:
Dunn, Richard Sherwood, 1940- ( Dissertant )
Pennypacker, H. S. ( Thesis advisor )
Webb, W. B. ( Reviewer )
Gunnell, B. M. ( Reviewer )
Krug, F. A. ( Reviewer )
Burns, James F. ( Reviewer )
Jones, E. Ruffin ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1969
Language:
English
Physical Description:
viii, 144 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Blinking ( jstor )
Educational conditioning ( jstor )
Experimentation ( jstor )
Eyelids ( jstor )
Frequency distribution ( jstor )
Integrated magnitude ( jstor )
Mathematical dependent variables ( jstor )
Monkeys ( jstor )
Observational research ( jstor )
Primates ( jstor )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Eyelid conditioning ( lcsh )
Monkeys ( lcsh )
Psychology thesis Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )
theses ( marcgt )

Notes

Abstract:
Two experiments on a group of four rhesus monkeys studied spontaneous and conditioned eyelid behavior, and evaluated Blink Magnitude as a dependent variable. In the first of these a standard conditioning treatment was employed, consisting of one free recording session, five acquisition sessions and two extinction sessions. Fifty trials were administered in each daily session. The results showed no systematic effect on spontaneous blinking related to either the conditioning treatment or the overall experimental procedure. Conditioning was observed and Response Magnitude was seen to compare favorably with Frequency of Responding as a measure of response strength. The second experiment obtained differential conditioning to auditory CSs , to visual CSs , and to mixed blocks of auditory and visual CSs. When auditory and visual cues were presented together in a special test sequence, summation was observed; and the advantages of response magnitude over frequency in detecting this effect were demonstrated. The results of both experiments were discussed in relation to other studies of eyelid conditioning In monkeys and humans and some Implications for later research were considered.
Thesis:
Thesis - University of Florida.
Bibliography:
Bibliography: leaves 140-143.
Additional Physical Form:
Also available on World Wide Web
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000559388 ( AlephBibNum )
13483476 ( OCLC )
ACY4844 ( NOTIS )

Downloads

This item has the following downloads:


Full Text









AN ANALYSIS OF SPONTANEOUS AND CONDITIONED

EYELID RESPONSE MAGNITUDE IN Macacca mulatta












By
RICHARD SHERWOOD DUNN


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN -PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY














UNIVERSITY OF FLORIDA
1969













ACKNOWLEDGMENTS


The writer gratefully acknowledges his indebtedness

to all members of the committee, and especially to Dr. H. S.

Pennypacker whose assistance and support were indispensable

to the study.

Special thanks are also extended to Mr. Hunter Jackson,

of the University of Florida's Center for Neurobiological

Sciences, for his capable technical assistance and to the

University of Florida Primate Facility for generously

allowing the use of the animals.

The study could not have been completed without the

continued encouragement and assistance of the writer's wife,

Marie. Her conscientious application to the tedious scoring

tasks could not have been replaced.

This research was jointly supported by Public Health

Service Grants, MH-08887 and HE-06379, to Dr. H. S. Pennypacker;

NB-02131 to Dr. F. A. King; FR-00421 to Dr. A. F. Moreland;

and by MH-10320 to the University of Florida's Center for

Neurobiological Sciences where the author was a pre-Doctoral

trainee.














TABLE OF CONTENTS

Page
ACKNOWLEDGMENTS ........... .. .. ... ... .... .. ... ii

LIST OF TABLES . . ... ..... ........ .... ........... iv

LIST OF FIGURES ...........v........ ....... ..... .. . v

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

INTRODUCTION ................. ..... .... ................. 1

METHOD . . . . . . . . .......... . . .

RESULTS .. ........... ..... .... ..................... 22

DISCUSSION .............................. .. .......... . 114

SUMMARY . ............ ................ ...... .. ..... 121

APPENDIX A .... ... ...... .. .... .... .......... .... ..... 122

REFERENCES .. .................. ... .... .. . . . 140

BIOGRAPHICAL SKETCH ................................... 144


iii













LIST OF TABLES


Table Page

1 Validity Coefficients of Magnitude Measure
for Subjects .............. ....... ...... .... 23

2 Average and Standard Deviation Values of Inter-
Blink Interval and Magnitude Distributions ..... 27

3 Intercorrelation Matrix for Frequency
Measures .... ... .. .. ......... ............. 29

4 Intercorrelation Matrix and Multiple Correlations
of Average Blink Magnitude ..................... 36

5 Correlation of Blink Magnitude Scores and
Inter-Blink Intervals .......................... 46

6 Alpha Blink Magnitude and Latency
Statistics ..... ............. ...... .......... 52

7 Intercorrelation of On-Trial Response Measures 80

8 Correlation of Off-Trial Measures with On-Trial
Measure s....... .......... ..... . . .... .... 82

9 CR Magnitude and Frequency Distribution
Statistics ........... .................... ... 83

10 State Transition Matrix for Latency and Magnitude
for Days 1 through 7 ...... ........... ....... 86

11 Statistical Tests of CS+CS- Differences
within Days ..... ..... ....... ....... .......... 107

12 Statistics for Experiment 2 CR Latency and
Magnitude Distributions ............ ........... 108

13 Experiment 2 Overall Group Outcomes and
Comparison Data l........................ .. 109

14 Experiment 2 Test Trial Outcomes .............. 113













LIST OF FIGURES


Figure Page

1 Total Number of Off-Trial Blinks as a
Function of Days ............ ..... ............. 31

2 Total Magnitude of Off-Trial Blinks as a
Function of Days .................. ....... 33

3 Average Blink Magnitude of Off-Trial Blinks
as a Function of Days ........*............... 35

4 Off-Trial Spontaneous Blink Measures as
Functions of 10 Trial Blocks .................. 40

5 Frequency and Average Blink Magnitude of
Off-Trial Blinks as a Function of Time
since Last UCS ...... ............ ............. 43

6 Frequency and Average Blink Magnitude of
RM 117 Alpha Blinking as a Function of Days ... 49

7 Frequency Distributions of Alpha Blink
Latency and Magnitude for RM 117 and for
Rm 174, 180, and 182 Combined ................. 51

8 Frequency of CRs as a Function of Days *....... 54

9 Total Magnitude of CRs as a Function of Days .. 56

10 Average Blink Magnitude of CRs as a
Function of Days ............................ 58

11 Average Latency of CRs as a Function of Days .. 60

12 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day A ........ 64

13 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 1 ........ 66

14 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 2 ........ 68

15 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 3 ........ 70







LIST OF FIGURES (CONT.)


Figure Page

16 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 4 ......... 72

17 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 5 ........ 74

18 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 6 ........ 76

19 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 7 ........ 78

20 Frequency of CRs as a Function of Days ........ 89

21 Total Magnitude of CRs as a Function of Days .. 91

22 Average Blink Magnitude of CRs as a
Function of Days ............................. 93

23 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 1 ........ 96

24 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 2 ........ 98

25 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 3 ........ 100

26 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 4 ........ 102

27 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 5 ........ 104

28 Frequency and Total Magnitude of CRs as a
Function of Blocks of 10 Trials, Day 6 ........ 106

29 Frequency and Magnitude Measures of Test
Trial Blinks as a Function of CS
Combinations ....... .. ... ...... 112







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


AN ANALYSIS OF SPONTANEOUS AND CONDITIONED EYELID
RESPONSE MAGNITUDE IN.Macacca mulatta

By

Richard Sherwood Dunn

June, 1969

Co-chairman: Dr. H. S. Pennypacker and
Co-chairman: Dr. W. B. Webb
Major Department: Psychology

Two experiments on a group of four rhesus monkeys

studied spontaneous and conditioned eyelid behavior, and

evaluated Blink Magnitude as a dependent variable. In

the first of these a standard conditioning treatment was

employed, consisting of one free recording session, five

acquisition sessions and two extinction sessions. Fifty

trials were administered in each daily session. The results

showed no systematic effect on spontaneous blinking related

to either the conditioning treatment or the overall experi-

mental procedure. Conditioning was observed and Response

Magnitude was seen to compare favorably with Frequency of

Responding as a measure of response strength.

The second experiment obtained differential condition-

ing to auditory CSs, to visual CSs, and to mixed blocks of

auditory and visual CSs. When auditory and visual cues were

presented together in a special test sequence, summation

was observed; and the advantages of response magnitude over

frequency in detecting this effect were demonstrated.

The results of both experiments were discussed in


vii'







relation to other studies of eyelid conditioning in monkeys

and humans and some implications for later research were

considered.


viii












INTRODUCTION


The extension of eyelid conditioning techniques to

non-human primates has been pursued because it offers many

potential advantages over research with humans. These fall

primarily into the general category of experimental control.

Employing animal subjects, the experimenter can have a much

greater degree of control over all the variables and factors

influencing behavior and can administer a much wider variety

of experimental treatments. But there are several potential

gains in the use of primates for eyelid research in addition

to the usual benefits of animal subjects. Numerous studies

have shown major effects in eyelid performance due to

personality variables (Spence & Spence, 1966) (Prokasy &

Whaley, 1962) (Runquist & Spence, 1959), instructions and

other verbally mediated language effects (Hartman & Grant,

1962) (Ross, 1965) (Hilgard & Humphreys, 1938), social

interactions (Spence & Goldstein, 1961), and previous

experience with the conditioning paradigm (Meiselman &

Moore, 1965). It is reasonable to expect that the use of

monkey subjects will largely eliminate these confounding

factors and allow a more direct investigation of the

conditioning process.

In the human research field agreement has not been

achieved on the issue of voluntary responders. On the basis







of form and latency criteria many experimenters have

separated the results of subjects whose blinks appear to be

voluntary avoidance closures rather than conditioned responses

(Spence & Ross, 1959) (Hartman & Ross, 1961) (Ross, 1965).

Generally these results are eliminated from consideration as

experimental data, a practice with which other researchers

are not in agreement (Gormezano, 1965) (Hickok et al., 1965).

Monkeys do not show typical voluntary responses so this issue

can also be avoided, although, in the long run, they may

provide the final solution to the problem of voluntary

responders.

The conditioning process at the primate level is of

interest by and for itself, of course, but there are also

significant benefits to be gained in terms of comparative

data. The voluntary response problem is an example. If

voluntary responses are in fact cases of instrumental

avoidance behavior then the results of systematic research

on eyelid conditioning in monkeys, both instrumental and

classical, should provide a basis for interpretation of this

type of behavior as part of an orderly system. Pairing lower

primate subjects with the relative convenience of the eyelid

paradigm makes practical the type of systematic research called

for. Results of such a program would have theoretical and

systematic importance as well as comparative relevance to

issues in the human research field.

Recent efforts in this direction have met with some

substantial difficulties. A series of experiments at the

University of Florida gave only fair to moderate conditioning







figures and several anomalous experimental outcomes were

observed. Mourant (1965) tested two levels of CS intensity

using squirrel monkeys. He found the lower intensity CS to be

more effective, contrary to expectation, and the highest level

of performance observed occurred in the first block of trials

reaching about 60 per cent CRs. This is the highest general

level of performance in the series of reports to be reviewed

here. It is probable that the relatively long CS-UCS interval

Mourant employed, 3 sec., accounts for this; especially in

view of the spontaneous blinking factors mentioned below.

Pennypacker & Cook (1967) tested four CS-UCS intervals

in squirrel monkeys with paired and unpaired trials. Reliable

conditioning in terms of differences between the paired

group and the unpaired (pseudoconditioning control) group

was found only for the first 200 of 550 trials. Higher

rates were observed for greater CS-UCS intervals, with per

cent CRs rising only to about 45 per cent for a 4 sec.

interval group. This finding is in contrast to the usual

superiority of interstimulus intervals in the neighborhood

of 0.5 sec.

Cook (1966) employed cebus monkeys in a factorial ex.

periment on UCS type and intensity. The direct relationship

between UCS intensity and performance which would normally

be expected failed to emerge. For both air puff and para-

orbital shock conditions the lowest intensity UCS was most

effective and, again, conditioning performance was discourag-

ing; rising only to 50 per cent CRs for the overall group

performance in the highest treatment group.







A very significant finding did emerge from the Cook

study, however. He measured pre- and post-block Spontaneous

Blink Rate (SBR) for each 25 trial block. Examination

revealed that SBR and Per Cent Conditioned Responses correl-

ated .90 for acquisition and .95 for extinction. This

observation identified spontaneous blinking factors as an

important consideration and even suggested that outcomes of

previous studies may have been entirely the result of altera-

tions of SBR rather than limited changes in responsiveness

to the CS due to CS-UCS pairing.

Pressing this issue in the final study of the series,

Cook (1968) performed a factorial experiment which carefully

coordinated independent and joint effects of conditioning

stimuli upon spontaneous behavior and conditioning per-

formance in acquisition and extinction. An additive linear

model was constructed to express the effects on SBR of CS

only and UCS only presentations as well as unpaired CS/UCS

and no-stimulation groups. Applied to his conditioning data

the model appears to demonstrate that the frequency measure

of conditioning reflects only. the additive sum of positive

and negative general effects on SBR of the UCS and CS,

respectively. Notably absent was any interaction increment

that could be termed associative learning in the CS-UCS

paired conditions of the study.

The overall picture that emerges from these studies

is gloomy, indeed. The performance figures are discouragingly

low, ranging from 35 per cent to 60 per cent CRs, and are

accompanied by anomalous experimental outcomes as compared








to traditional findings. Finally, there is the strong

suggestion that performance figures are severely contaminated

by or determined entirely by irrelevant spontaneous behavior.

The only other study employing monkeys as eyelid

conditioning subjects is that of Hilgard & Marquis (1936).

They found reliable evidence of conditioning to a light CS

in four of five rhesus monkeys. Their results seem straight-

forward and conventional in most respects. Response frequencies

rose to the 80 per cent to 90 per cent range after six daily

50 trial sessions and control measures for random responses

remained low, in the neighborhood of 5 per cent. Several

features of their report, however, raise questions about the

generality and applicability of their findings. They found

extinction hard to demonstrate. Typically, performance was

superior during three extinction days when compared to the

last acquisition day. After nine days of extended extinction

sessions, one subject fell to 36 per cent CRs and rose

again to 60 per cent CRs by day 12. The authors concluded

that extinction in monkeys was "so slow as to be practically

undemonstrable."

All of the Florida studies mentioned earlier observed

similar extended performance in extinction with occasional

increases in frequency of CRs after UCS presentation ended.

In general, such a finding is sufficiently anomalous to call

into question the validity of a conclusion in favor of

conditioning. In most such situations non-associative

factors such as sensitization or alpha conditioning are found

to be operating.A further inspection of the Hilgard & Marquis








study indicates that a cautious approach is warranted. They

used a 400 msec. interstimulus interval because it had been

effective in human research. The overall mean latency of CRs

was observed to be 226 msec. If comparative data are relevant,

then up to half of their observed CRs fell in the range of

alpha responses since the time range of these responses in

man is reported variously up to 200 msec. (Spence & Ross,

1959). Furthermore, they observed blinks as complete closures

and the trend in latency was decreasing as conditioning

continued. Considering these results together, the possibility

that Hilgard & Marquis observed primarily sensitized alpha

blinks cannot be discounted. It is notable that the studies

performed in the Florida series observed typical alpha

blinking in squirrel and cebus monkeys ranging in latency

up to slightly over 200 msec.

Although the use of primate subjects has several

potential benefits, the research to date has not been able

to capitalize on any of them. They present instead a variety

of technical and methodological problems and there is even

the suggestion that classical conditioning of the eyelid

response has not been demonstrated in monkeys at all.

Probably the most significant single outcome is the identi-

fication of serious contamination of the dependent variable,

Frequency of CRs, by spontaneous blinking. This is a methodo-

logical difficulty which requires correction before research

can continue. Beyond this, the experimental results, if they

are allowed to stand,pose questions about the generality

of the conditioning process and threaten the usefulness of








comparative data.

A number of observations, interpretations, and sug-

gestions have emerged from the Florida studies and bear

mention here. One feature common to all of this research

including the Hilgard & Marquis study is the technical

difficulty encountered in restraining the subjects during

the experiment and the direct effects of restraint upon

performance. Recording equipment requires a nearly rigid

mount between the transducer and the subject's head. Hilgard

& Marquis made use of a restraining box with a head yoke,

plus various pads and clamps. Their remarks on the topic are

not specific but indicate that restraint sufficient for

recording was difficult to achieve and actively resisted

by their untamed subjects.

The Florida restraint procedure has gone through

several revisions but consists mainly of a modified primate

chair fitted with specially constructed head retaining devices.

Cook (1968) rated subjects for accommodation to the conditioning

situation for each of 14 days and found no reduction in

struggling or emotionality. One interpretation offered for

the generally negative results emphasizes the stressful

nature of restraint and the highly emotional state which

results. Cook suggested that failure to observe typical

conditioning in his study and in previous ones may be con-

strued as support for the existence of an optimal drive level

above which eyelid conditioning either does not occur or

becomes obscured by extraneous behavior. This interpretation

has much to recommend it. A state of emotional excitement







cannot be expected to contribute to finely localized

associations. In eyelid conditioning this is particularly

important because spontaneous blinking often reflects arousal;

it is also increased as a consequence of generalized strug-

gling activity.

South American monkeys have only one natural enemy,

the various species of constricting snakes. While it is a

fanciful thought, perhaps, it is not unreasonable to suppose

it natural for close restraint to cause a state of utter

panic in these animals. Cook's observationsin this area

indicate that the technique of restraint and measurement

employed may be inappropriate for the subjects used and cause

violent behavior patterns that mask or eliminate conditioning.

He suggests the use of chronic chair techniques or free

ranging equipment so that the process can be studied in a

situation to which the subjects are habituated..

Cook also questions the suitability of the Frequency

of CR measure in this application. Since it is greatly

influenced by SBR factors stemming from emotional effects

and struggling artifacts, it may be that the Per Cent CR

measure cannot detect conditioning or discriminate it from

the background spontaneous behavior. He suggests that

measures other than rate and latency which would not be so

greatly masked by high spontaneous rates would be of value

in determining the presence or absence of conditioning.

Incidental observations from unreported studies

(Milan, 1966) have provided support for this idea. Technical

problems prevented conclusive demonstration but there appeared







to be a basic difference between On-trial "true" CRs, so to
1
speak, and Off-trial spontaneous blinks. The difference is

one of form and duration, CRs being larger and longer with

prolonged or irregular shapes as opposed to the arrow like

up-down record of a typical spontaneous blink. It is clear

that a valid and suitable correction procedure is required if

the Frequency of CR measure is to be useful. Beyond this,

refinement of the dependent variable may provide a solution

in the form of a different measure of response strength.

The present research was undertaken as an initial effort

to resolve these technical and methodological difficulties.

It had three main objectives:

1. To compare cebus (Cebus albifrons) to rhesus

(Macacca mulatta) monkeys as subjects, under optimum conditions,

and demonstrate the presence or absence of conditioning for

each species.

2. To make a complete analysis of spontaneous blinking

behavior for each species.

3. To evaluate the suitability of a magnitude measure

of response strength in eyelid conditioning.

A number of procedural alternatives remain to be ex-

plored before it can be concluded that cebus monkeys are not

appropriate subjects for this kind of research. In the present

study every effort was made to optimize the experimental para-

meters in terms of the collective experience of the Florida

laboratory. Previous practice has revolved around brief

1
The terms On-trial and Off-trial are used in this
paper to refer to the 'interstimulus interval and the intertrial
interval, respectively.







habituation of essentially wild animals to the conditioning

situation followed by severely massed training sessions.

Even without the prohibitively expensive resources required

for the chronic chair or free ranging techniques which Cook

suggested, a number of procedural variations are possible

which can minimize stress placed on the subjects in the

conditioning apparatus and improve the likelihood of successful

conditioning. The seeming contradiction between Hilgard &

Marquis' results and the Florida studies may be resolved by

a comparison of rhesus and cebus monkeys under similar

conditions.

Reasons for the second goal, an analysis of spontaneous

behavior, should be clear. Previous investigations have not

recorded all of the behavior of the eyelid system during

conditioning. Cook's technique (Cook, 1968), the most thorough

to date, incorporated a counting device and a timer-printer.

From thesedata the Average Blink Rate for extended intervals

could be determined. Having established the need for valid

rate correction procedures, it became necessary to gather

extensive and detailed information about spontaneous blinking

so that frequency distributions and time course curves could

be constructed. These will provide the basis for corrections

to the Frequency measure, if it is to remain in use, and will

provide accurate information on the effects of experimental

stimuli on the overall system behavior.

Finally, an alternative response strength measure,

Response Magnitude, was selected for evaluation. It has much

to recommend it as the next step in developing an appropriate







dependent variable. One can say that Pavlov's basic measure,

the number of drops of saliva in a given time interval, was

a magnitude measure (Pavlov, 1927, 1928). Defined as the

definite integral of response amplitude over the time course

of a blink, Response Magnitude incorporates information from,

or related to, response occurrence (Frequency), Latency,

Amplitude, and Form. It has been employed with favorable

results in human eyelid conditioning (Pennypacker, 1964).

In human research the workhorse Frequency measure has

generally been found more reliable and more useful than measures

of Latency or Amplitude (Campbell, 1938) (Lumsdaine, 1941). In

the only clear test of the Magnitude measure (Pennypacker,

1964), Frequency odd-even reliability was .97 while that for

Magnitude was .99. One other study appears to have employed

a Magnitude measure (Grant & Schneider, 1948) and quoted

reliability for both Frequency and Magnitude as .96. The

uncertainty here stems from the early suggestion that Average

Amplitude over several trials, with zeroes included for non-

response, be termed Magnitude to distinguish this scoring

procedure from Average Amplitude (Humphreys, 1943). Insofar

as it has been tested, Magnitude appears to equal or exceed

Frequency in reliability and has been found suitable as a

response strength measure.

A Magnitude measure values responses on at least an

interval scale as contrasted with the essentially binary

nature of Frequency measures. This offers the potential of

increased sensitivity to changing trends since single responses

can be compared to one another. This would be very useful in







single trial test situations such as spontaneous recovery

tests wherein the first trial of a session is the only one

which truly meets the requirements of the test. Alternatively,

one could take advantage of the increase in dependent variable

sensitivity by performing research at the same level of

precision but with fewer subjects.

Another illustration of the value of interval or ratio

scaling over the binary nature of the Frequency measure is

encountered when attempts are made to show differences in

groups performing at the same levels, particularly at high

Frequencies. In one such study (Spence & Platt, 1966) high

rate responders were removed to demonstrate that the rest

continued to improve. The artificial nature of the 100 per

cent upper bound of the Frequency measure made this in-

efficient technique necessary. In contrast, the Magnitude

measure has an upper bound determined by physiological and

physical properties of the eyelid system itself. In principle,

it could detect differences in response strength for groups

performing at identical rates. Not only is its upper bound

natural rather than artificial, butin an overall sense, as

Pennypacker (1966) has pointed out, it has construct validity.

Eyelid closure is effected by a sphincter muscle and a score

such as Magnitude is logically appropriate for describing its

responsiveness and the extent of its action over a given time

interval.

From all of this it can simply be said that Response

Magnitude appears to have many advantages as a dependent

variable. Its evaluation in monkeys was selected as an




13


objective of this study with the specific hope that its

increased precision would allow a definite conclusion to be

made as to the presence or absence of conditioning in both

species.

Three experiments were performed; two on rhesus

subjects and one on cebus subjects. Because of scoring delays

and the extensive volume of data to be covered, the latter

will be reported elsewhere.













METHOD


Subjects.--Four adolescent rhesus monkeys (Macacca

mulatta) were selected from the colony-bred stock of the

University of Florida's Primate Facility. Three of these were

female, one male,and they were moved to the conditioning

laboratory one month before the experiment began. In the ad-

olescent stage the large supra-orbital ridge of the adult

is not fully developed and does not interfere with the

recording apparatus.

Apparatus.--The laboratory and its facilities have been

described in detail elsewhere .(Pennypacker, et al., 1966)

(Cook, 1968) but briefly consisted of a conditioning chamber,

and air delivery system, a restraining chair, and recording

and control equipment.

The restraining chair required minor alterations to its

fixtures to suit the larger head size of the Ss, and a small

rhesus chair replaced the cebus seat. Using this system, the

S's head position is stabilized rostro-caudally by a grooved

pad above, and a split metal collar below. Anterior-posterior

motion is prevented by a u-shaped hard rubber stop beneath the

occiput, and a smaller u-shaped stop bearing on the mandible.

The latter is aided by a metal tooth bar secured behind the

lower canines. The various fixtures can be adjusted to suit

individual Si and can be secured without excessive pressure. :








The system renders the head immobile without causing injury

or indications of discomfort. The equipment does not inter-

fere with the S's vision or hearing.

The recording system was altered by the addition of

an AC coupled amplifier and an electronic integrator, Grass

models 5P5 and IU-1 respectively. The recording system

employs a microtorque potentiometer in a battery excited

bridge circuit to generate a DC voltage analog of eyelid

position. The eyelid is coupled to the potentiometer by a

lever, a fine wire and a small triangular cloth patch

secured with surgical appliance cement.

The first channel of a model 5 Grass four channel

polygraph was used to record the DC eyelid position infor-

mation. A second channel recorded time in seconds and the

presentation of CS and UCS. Because frequent changes in

baseline are present in the DC information it has not been

possible to properly use an electronic amplitude scoring

device in conjunction with this system. Cook (1968) noted this

problem by indicating that his Schmitt trigger circuit, used

to count blinks, often scored baseline shifts. To obtain a

signal suitable for integration, that is, one without a

changing baseline, an AC coupled Grass 5P5 amplifier was

used to re-record the eyelid information on a third channel

of the record. With an essential time constant of 0.8 sec.,

baseline shifts decay to zero in about 4 sec. (5xt). Since

blinks are completed within approximately 0.1 to 0.15 sec.,

they were not altered greatly in form. Occasional blinks

of extended duration were reduced in size, however.







The output of the 5P5 was fed directly to a Grass

IU-1 electronic integrator.. Its output, on the fourth

channel, gave the definite integral of amplitude of eyelid

response as indicated by the 5P5. Integration began when the

eyelid position voltage rose above zero, or baseline, and

continued until the baseline was crossed again; at which

time the integrator reset was automatically triggered.

Integrator pen deflection indicated the integral of the

blink voltage signal over its period of action. The simplest

way to regard this is as measuring the area between the

blink trace and the baseline. The resulting records contained

time and stimulus indications, DC eyelid position information,

a re-recorded, AC coupled representation of the eyelid posi-

tion, and magnitude of the AC coupled response record. The

paper grid lines allowed precise location and measurement of

blink times and magnitudes.

Procedure.--Two experiments were performed. The first

was a straightforward conditioning procedure applied to gather

extensive data on spontaneous and conditioned behavior of

the eyelid system. Experiment 2 was a differential condition-

ing procedure performed in lieu of the traditional multiple

control group design. Its purpose was to demonstrate the

presence of conditioning and to provide a fine test of the

Magnitude measure as a dependent variable.

The four rhesus Ss served in both experiments. They

were housed in a special colony room in individual cages.

Each S was fitted with a chain collar and leash which re-

mained in place throughout the experiment. The leashes







were an aid to handling the animals, and facilitated the

daily chair installation procedure. The animals were main-

tained on liberal quantities of Purina Chow supplemented

with fruit and vitamins. Daily rations were given 30 min. to

1 hr. after the conclusion of each daily session. Water

was available ad lib.

Every effort was made to acclimate the So to the

laboratory, E, and the restraining devices with the least

possible discomfort. While efforts to tame the animals were

not entirely successful in that they could not be handled

as pets, the Ss did show very good adjustment to the situa-

tion. Threat postures were never seen in the cage room; there

were little or no vocalizations or facial threat expressions,

and in the conditioning chair the Ss remained calm and did

not struggle.

Experiment l.--Five successive days of chair adapta-

tions sessions were administered. On the last of these a

complete recording session was conducted similar in all

respects to the conditioning sessions to follow except that

no stimuli were presented. This session, termed Day A, pro-

vided background information on spontaneous behavior before

conditioning stimuli were presented. There followed five

dialy acquisition sessions and two extinction sessions

termed Days 1 through 7 respectively.

Each session was composed of 50 trials. The intertrial

interval averaged 60 sec. and ranged from 30 sec. to 90 sec.

Trials were controlled by a tape timer with a random sequence

of 21 intervals falling at each 3 sec. interval from 30 sec.







to 90 sec. inclusive. The tape was started in the same place

for every session in order to achieve equal session lengths

and control order effects. The CS was a 1000 Hz tone pro-

duced by an audio oscillator and delivered through two 6 in.

speakers mounted on either side of the monkey chair. Sound

pressure level measured at the center of the S's head space
2
was 68 db (re. 0.0002 dynes/cm. ). The UCS was a 2.0 psi

puff of air with a nominal 0.1 sec. duration. It was de-

livered to the temporal cornea of S's eye through a 15-gauge

square tipped hypodermic needle with the tip 0.25 to 0.40 in.

from the eye. Careful placement of the needle tip was required

to prevent S's eyelashes from contacting it. The interstimulus

interval was 1.0 sec. with CS offset coincident with UCS onset.

Subjects were run on successive days in the same order.

Each daily conditioning session was preceded by careful

calibration of the electronic components of the recording

system. Calibration was standardized at 5 mm. pen deflection

per mm. of eyelid movement. The IU-1 was calibrated to 12

units of magnitude for each 30 sq. mm. of eyelid trace

area. The paper ran at 30 mm./sec.; thus, one mm. of eyelid

closure operating over 0.2 sec. would produce a Magnitude

score of 12. The results are reported directly in these

Magnitude units.

Each S was installed in the chair, fitted to the

recording device, and allowed 5 to 10 min. to adapt to the

situation before recording began. An entire conditioning

session required 60 to 75 min. Late in this period, subjects







often showed signs of drowsiness, sometimes appearing to be

completely asleep. No effort was made to prevent this during

Experiment 1; once placed in the recording chamber, Ss were

not disturbed unless adjustments to the equipment were

required.

Experiment 2.--Experiment 2 entailed a somewhat more

complex design, although animal handling and recording pro-

cedures remained the same. The purpose was to demonstrate

differential conditioning to auditory and visual stimuli and

to subject the Magnitude measure to a more demanding test.

Six daily differential conditioning sessions were

administered. The intertrial intervals were halved so that

they averaged 30 sec. and ranged from 15 sec. to 45 sec.

Interstimulus interval and UCS remained unchanged. Auditory

stimuli were 600 Hz and 1400 Hz tones presented at 70 db. SPL.

Day 1 and 2 sessions were comprised of 100 trials each

of differential conditioning, with 2 Ss receiving the 600 Hz

tone as the positive CS (CS+) and the 1400 Hz tone as the

negative CS (CS-). For the other 2 Ss, conditions were

reversed. Trials were presented in a random alternating se-

quence, constrained to prevent more than two consecutive

trials of the same type. Sequencing of CS+ and CS- was

accomplished automatically with control devices programmed

to produce the entire 100 trial sequence without attention

from E.

On Days3 and 4, visual stimuli were employed as CSs.

Red and green circles, 1 in. in diameter were presented by

an inline-readout stimulus projector mounted on the chair







approximately 3 in. directly in front of S, at eye level. The

stimulus projector remained in place throughout the experiment.

Again, 100 trials per day were administered. One S from each

auditory condition-was selected to receive red as CS+ and

green as CS-. Stimuli were programmed and administered as

before.

On Days 5 and 6, auditory and visual trials were

presented in a mixed counterbalanced sequence. At the outset

of each day, trials were alternated between visual and

auditory stimuli in groups of eight to twelve trials. Spacing

was gradually reduced so that by trial 50, and from there

onward, the alternation occurred every two to four trials.

This was counterbalanced carefully so that order effects

were cancelled and the changes did not interfere with

sequencing of CS+ and CS- trials. Switching between auditory

and visual stimuli was accomplished by E according to a

program prepared in advance.

Experiment 2, thus, consisted of two days of dif-

ferential conditioning to auditory CSs, two days of visual

CSs, and two days with mixed auditory and visual CSs. During

the last half of Day 6, twelve special test trials were

administered as a test for summation. They were presented

without UCS and scored separately to test the sensitivity of

the Magnitude measure. Each trial was composed of the paired

presentation of one CS from each modality. The four possible

combinations, CS+CS+, CS+CS-, CS-CS+, CS-CS-, were given

three times each by repeating this sequence three times.

The order of presentation was the same for all Ss, and each




21


test trial had the same location in the regular trial

sequence. The test trials were spaced out over the last 50

trials and were programmed by E.













RESULTS


Magnitude Measure

The last half of each Day A record (no stimulation)

was carefully hand scored. The area of each blink in the DC

record was determined by the dot grid procedure (Pennypacker,

1966). These scores were compared with the electronic

magnitude measure scores. The overall validity of the electron-

ic magnitude score thusly determined exceeds 0.97. Values

for individual Ss are given in Table 1. A check of this type

was required because there were two questionable features to

the electronic scoring process. First, blinks do suffer some

distortion, although slight for most, as a consequence of

the AC coupling circuit. Second, blinks, at the calibration

selected for this experiment, occasionally drive the inte-

grator pen to full deflection. A finite reset time then inter-

venes before integrator scoring resumes. Blink area falling

in this interval is not measured. One result of this

problem can be seen in the frequency distributions to follow;

blinks scoring from full scale values of 40 up to some

undetermined level, judged to be between 50 and 65, are all

scored as 40. Beyond this, scoring returns to normal with,

for example, 61 being scored 1, 62, 2, etc. Thus, a bunching

at 40 is seen in frequency distributions. For the purposes

of this study, the effect was ignored. For future research,

















r

Average r = 0


TABLE 1

Validity Coefficients of Magnitude
Measure for Subjects


RM 174 RM 117 RM 18


0.977 0.962 0.958

.972


0


RM 182


0.989







calibration procedures should be altered to eliminate the

problem.

Scoring procedures.--In scoring the data, a few basic

procedures were followed throughout both experiments. Off-

trial blinks were scored for time to the nearest tenth of a

sec. and for magnitude. If the blink closely followed a

baseline shift upward, it sometimes happened that IU-1 reset

occurred at the same point as the blink and no magnitude

score was available. In these cases magnitude was hand scored.

During On-trial periods, all blinks were scored for latency

and magnitude. As observed by Hilgard & Marquis (1936), a

frequent component of the response to the CS was an initial

opening of the eye. This occurred most often at the begin-

ning of the experiment and disappeared by the end of

Experiment 1. It frequently happened that CR blinks following

such an eye opening were mostly or entirely below the base-

line. In such cases, it was the consistent procedure to

employ the unadjusted magnitude score if the IU-1 responded

at all, even if the majority of the blink clearly fell below

the baseline. If the IU-1 showed no response, however, the

blink was hand scored.

All blinks beginning .025 msec. or more after CS

onset and preceding UCS onset were scored. Blinks beginning

before 300 msec. were judged as alpha or CR blinks on the

basis of form and other behavior which followed. Separate

records were maintained for each type. Alpha blinks are

rapid, full closures, generally in the 150 to 200 msec.

latency range. CRs, however, often began in this range with







partial closures, continuing as the CS elapsed to a full

closure or blink. In the case of multiple CRs, only the first

was scored for latency, whereas the entire sequence of

blinking behavior was scored for magnitude.

Since vertical eye movements introduce small signals

to the recording system, a criterion for minimum blinks was

necessary. To be counted as a CR, pen deflection was required

to exceed 1 mm. in the direction of closure and thereafter

be followed by an opposite movement within 200 msec. If the

UCR occurred before the 200 msec. period had ended, a CR was

also scored. By this procedure some CRs were too small to be

scored for magnitude by either the IU-1 or by hand. In these

cases a minimum magnitude score of 1 was assigned.

Experiment 1

The results of Experiment 1 will be given in two

sections, treating Off-trial spontaneous behavior first

and On-trial conditioned responses second.

Spontaneous blinking.--All Off-trial blinking, that

which occurred during the intertrial intervals, was scored

for time of occurrence and magnitude. These scores were

transcribed for computer input as sequences of blinks. Time

values were included for the start and end of each session

and each trial. Electronic processing allowed a variety of

information to be extracted from the raw data. For this

analysis On-trial behavior was not included. Blinks during

the CS and UCS were not counted. Trials were timed as begin-

ning and ending with UCS offset. After the UCS offset,

secondary:.blinks followed the UCR so frequently that they







were considered part of the UCR. These blinks were system-

atically eliminated from Off-trial scores by not counting

Off-trial behavior which occurred within 1 sec. of UCS

offset. Although the On-trial blinks were not scored, time

figures do include On-trial time intervals; that is, the

interstimulus interval, the UCS presentation time, and the

1 sec. scoring delay period. Time values therefore include

the entire period from the start of the session to the end.

Trial and session length.--To check on programming

and control procedures, trial and session lengths were

computed. Variation from nominal figures was introduced by

inclusion of intertrial intervals and minor inaccuracies in

the tape and clock control devices. Trial length averaged

61.14 sec. and ranged from 31.9 sec. to 91.3 sec. Results of

this analysis indicate that programming control was adequate

and session lengths can be considered equivalent for the

purpose of comparison between subjects and days.

Spontaneous blink rate.--With the data scored and an-

alyzed as indicated above, it was possible to construct

a rather complete picture of what happens to spontaneous

blinking as the conditioning paradigm is applied. Frequency

distributions were computed for the range of Blink Magnitude

values and Inter-Blink Interval (IBI) values. These are

presented in Appendix A for each subject on each day. Table

2 presents mean and standard deviation values for these

distributions. The entire set of data is included for ref-

erence in future research as well as for the sake of complete-

ness in explicating SBR effects in rhesus subjects.







TABLE 2

Average and Standard Deviation Values of Inter-Blink
Interval and Magnitude Distributions


Day RM 174 RM 117 RM 180 RM 182


Ave IBI 7.59 8.70 31.96 20.44
SD 7.07 13.79 72.90 33.45
Ave Mag A 15.71 10.70 9.92 17.51
SD 11.31 7.03 8.21 10.64

Ave IBI 8.49 13.22 21.80 17.16
SD 9.34 12.70 35.58 14.21
Ave Mag 1 16.95 14.38 11.89 18.54
SD 11.60 6.99 6.79 9.79

Ave IBI 10.16 7.99 18.48 14.03
SD 9.56 9.56 26.60 13.35
Ave Mag 2 14.53 10.17 7.93 15.78
SD 11.09 7.06 7.04 9.78

Ave IBI 9.60 11.27 22.51 16.20
SD 9.46 11.70 28.44 14.27
Ave Mag 3 11.52 6.43 7.65 12.76
SD 9.50 4.35 6.37 6.41

Ave IBI 9.95 9.26 15.13 15.91
SD 8.24 9.66 24.85 15.11
Ave Mag 4 13.70 7.34 8.33 13.45
SD 11.44 4.07 6.36 7.80

Ave IBI 11.59 14.11 28.60 17.62
SD 12.42 13.33 32.24 16.08
Ave Mag 5 11.50 9.12 6.27 15.03
SD 9.38 4.45 5.23 8.05

Ave IBI 9.36 10.49 32.62 16.86
SD 8.77 9.76 54.20 22.26
Ave Mag 6 14.23 9.30 9.54 14.86
SD 11.33 5.08 5.89 8.19

Ave IBI 9.13 12.38 18.41 17.71
SD 8.72 10.88 27.18 22.28
Ave Mag 7 10.96 10.41 9.54 14.47
SD 9.89 5.60 7.09 8.37







Considered as day values, the basic frequency measures

are strongly related due to equal session lengths. Table 3

shows correlations obtained for the computed values of

Number of Blinks Average IBI, and Blink Rate. Because the

curves for these parameters are very similar in shape, only

the curves for Total Number of Blinks are presented. As

Figure 1 indicates, Total Number of Spontaneous Blinks

ranged from 74 to 401 for the extreme subjects and days. SBR

appears to fall off slightly over the course of the experi-

ment.

An analysis of variance was performed, and here, as

well as in the analyses to follow, Winer's model for single

factor repeated measures designs (Winer, 1962) was used.

Throughout this study, the .05 level of significance was

required for results to be considered significant. This

first analysis of variance indicated that the daily values of

Number of Blinks cannot be considered to differ reliably,

F(7,21)1l.58,p<.25. Significant trends are also absent.

Similar curves were constructed for Total Magnitude

of spontaneous blinks and Average Blink Magnitude as shown

in Figures 2 and 3 respectively. These two measures are not

as strongly related as one might expect. Average Blink

Magnitude correlated .633 with Total Magnitude, whereas Total

Magnitude correlated .824 with Total Number of Blinks. Ref-

erence to Table 3 will bring these figures into perspective.

Off-Trial Average Blink Magnitudesshow a remarkable tendency

to vary together. Table 4 presents the intercorrelation

matrix for these values. Multiple correlations of Average








TABLE 3

Intercorrelation Matrix for
Frequency Measures


# Blinks Ave IBI Blink Rate


# Blinks 1 -.917 +.982

Ave IBI 1 -.902

Blink Rate 1






























O



44


















E81
,o





































43
0
ES
r4














Eo
01




























4*
0
0d
&<



i-t




Tl






cb^


















CL r- C* C

LD



I
I

I
10 6


-4 <


C0
CO
S~N
S)AN 18


CM
JO 'ON


_


.4 CY

























O



O
-4
0



-i
a





a







C
ed














EA
ui



0











0
4.
*ri












0

'I

d
ri




















cu




*L






























I I
C)



Ss2n2 l N/ //

i oc q u -




03

















I I I I I

0 0 O O O 0
0 0 0 0 0 0
0 0 0 0 0 0
(0 CD cr) CM
3nliiN9VI iVIOI


































rO




Od






o-
0

0ld



U
4












en


0*4
FO









*r0
r (
C







































































I B I I I I


I I a I


- C\ 0
T- T- 1


I I I I


CO % C\j o co


(10 c CN


3anilN9VlE ANII8 39V83AV


0 O Q -







TABLE 4

Intercorrelation Matrix and Multiple Correlations of
Average Blink Magnitude


RM 174 RM 117 RM 180 RM 182


RM 174 .658 .697 .801

RM 117 .754 .919

RM 180 .670

RM 182

R .897 .961 .859 .967







Blink Magnitude were also computed and the final row of

Table 4 indicates the multiple correlation coefficient, R,

for the rest of the group on each individual subject.

Analyses of variance for both Total Magnitude and

Average Blink Magnitude showed significant day effects.

Off-Trial Total Magnitude values differed over days, F(7,21)=

3.39,p<.05. The data were tested for first, second, and third

order trends and none of these was significant. If a

systematic effect were present due to effects of the condi-

tioning situation itself, one would expect a linear trend.

If an effect were present due to the conditioning treatment

then a second order effect would be expected since condi-

tioning trials were present on Days 1 through 5 whereas Days

A, 6, and 7 were composed of free recording or extinction

trials. It can be concluded, therefore, that Off-Trial values

of Total Magnitude differ from one another, but the differences

are not systematically related to either the conditioning

treatment or the experimental measurement situation.

Results of a similar analysis of Off-Trial Average

Blink Magnitude showed significant differences over days

beyond the .01 level, F(7,21)=12.25. Linear and Quadratic

trends were reliably present beyond the .01 level also,

F(1,21)=27.22, 21.47 respectively. As can be seen in Figure

3, the Linear trend is decreasing and minimum values are

roughly centered in Days 3 and 4. The interpretation of these

outcomes is unclear. The most important factor to the present

research is that the situation and the conditioning treatment

clearly do not have an increasing effect on either of the







Spontaneous Blink Magnitude measures; the effect of these,

if there actually is one, is a depressing effect. Beyond

this, the very strong relationship present between subjects

in the Average Blink Magnitude data suggests the influence

of an extraneous variable. For example, these shifts may be

an artifact of the calibration procedure although every effort

was made to standardize this over days. A second possibility

is humidity or temperature variables. Presumably, blinking

vets the cornea and would thus be related to the evaporation

rate and therefore to these factors. The last possibility,

of course, is that these are the systematic effects of the

conditioning treatment, an initial elevation on Day 1

followed by progressive decreases as conditioning appears and

a final return to normal as extinction takes its course.

Attention will return to this point in a later section.

Further analysis examined time course effects on

Spontaneous Blink measures within sessions and within trials.

The within sessions data are presented here in block form

for the entire experiment (Fig. 4). It is to be noted that

this is the same set of data just discussed but recast to

show the time course of effects within sessions.

In this figure, Total Magnitude and Number of Blinks

are shown as averages for all subjects over all days. That

is, Average Total Magnitude and Average Total Number of

Blinks. The Total Magnitude values over blocks are reliably

different, F(4,28)=10.7,p<.01, and Linear and Quadratic

trends are in evidence, F(1,28)-35.2,p<.1l; F(1,28)-

6.6,p<.05, respectively. Average Blink Magnitude also showed



































Fig. 4. Off-Trial Spontaneous Blink Measures
as Functions of 10 Trial Blocks.




40









800 80




700 70




600 MEAN MAGNITUDE 60


m

| 500 o 50
. .



400 NO. OF BLINKS 40









z
w 30

300 a 13- 30

F-


200 12- *MEAN BLINK MAGNITUDE 20

J

100 11- 10



---...I!-.. .. I --- I -- I -

1 2 3 4 5

BLOCKS







reliable differences, F(4,28)-8.l,p<.01, and Linear and

Quadratic trends, F(1,28)l19.l,p<.01; F(1,28)=12.6,p<.01

respectively. The block values of Total Number of Blinks were

significantly different, F(4,28)11.4,p<.01, but only a Linear

trend was in evidence, F(1,28)19.6,p<.01.

It can be concluded that the amount of spontaneous

blinking in terms of Number of Blinks, Average Blink

Magnitude, and Total Magnitude, decreased as the sessions

elapsed.

While these results are given in the form of whole

experiment averages by 10 trial blocks, the same measures

were also examined in a different form. Computer generated

values for Average IBI and Average Blink Magnitude were

obtained for each minute of each daily session. These were

printed as single minute curves and as five minute running

averages. In the interest of space the curves are omitted,

but it may be said that falling rates and falling magnitude

averages similar to the group curves discussed above were

not generally observed for individual subjects. Typical

results instead displayed almost perfectly stationary rates

and magnitude values, interrupted during Blocks 3 and 4 by

brief periods of unresponsiveness. Records in these periods

indicated sleep or drowsiness. Subjects were generally fully

alert by the middle or end of Block 5. This is probably the

only significance of the Quadratic effects present in the

magnitude curves.

In Figure 5, the spontaneous blinking data are again

recast; this time to show within-trials time order effects.




































0 to
0


m








M4
Sa
SE-
rC









40






ad
I,


















tio
C4

















04
U0

















*1-4
w a
o T<
SS
Q a







h0r
a)
t<











3














MEAN BLINK MAGNITUDE


'- CY) C\J -
t- t- t- q


O
Co

00









CW
CD)

CO 0
CO
C\j0 L


COf
co L


c a 00
,,;-.


r- CO L 0 0 C\1

SANI 1 JO 'ON







Trials were divided into 5 sec. intervals and collapsed

together. The Number of Blinks and the Average Magnitude of

all blinks falling in each 5 sec. interval (beginning with

UCS termination) were computed. For constant blink rate,

as a trial elapses, the expected distribution of Number

of Blinks has the same form as the distribution of 5 sec.

intervals. This is constant out to 30 sec. (period 6)

because all trials were at least 30 sec. long. Thereafter

it decreases linearly to 0 at 100 sec. (period 20). This

expected distribution, appropriately scaled for the overall

blink rate, is shown by the dashed line in Figure 5. Observed

blink frequency is seen to deviate strongly from the expected

values. A Pearson Chi-Square test for goodness of fit estab-

lished beyond the 0.001 level of significance that the ob-

served distribution differs from the expected distribution for

constant blink rate. Notably, the rate of blinking, or the

Number of Blinks, is greater than expected during approximate-

ly the 25 sec. to 60 sec. portion of the intertrial interval

and the effect is present in group curves for every day

except Days A and 1.

Average Blink Magnitude, on the other hand, did not

show this effect. In a similar test of goodness of fit,

variations in this measure showed no reliable differences

from the expectation of constant Blink Magnitude.

In concluding the presentation of spontaneous blink-

ing factors, the relation of IBI to Blink Magnitude was

examined for each subject in each session. It was hypothesized

that blinks separated in time by longer periods would be







larger in magnitude than blinks which followed one another

more closely. Thus a strong positive correlation was

expected between IBI and Magnitude values. Table 5 presents

the correlation values obtained for all of the IBI and

Magnitude scores of each S on each day. They are all small

in value, variable in direction positive and negative, and

nearly half of them do not reach a significance level of

0.05. The hypothesis was rejected and the conclusion is drawn

that if a general relationship exists between IBI and Blink

Magnitude, it is very small and of little interest.

Other aspects of spontaneous blinking behavior were

examined but curves are not presented here because the re-

sults are generally of little or no interest. These included

parameters such as median and mode, standard deviation, etc.,

of the frequency distributions given in Appendix A. The data

presented contain sufficient information to allow anyone

interested to reconstruct these figures if it is desired.

On-Trial behavior.--On-Trial measures revealed both

conditioned behavior and a substantial amount of alpha blink-

ing. Frequency, Magnitude, and Latency data are presented

below for each type. Once again, results in this section

will generally be given in both individual and group form.

This is necessary in any study which attempts to serve as

the basis of further research because the relation of subject

to group outcomes is not always direct. It is made doubly

necessary in this study by the fact that only four subjects

were used and the likelihood of sampling error is thus

increased. In such a situation, the critical reader may







TABLE 5

Correlation of Blink Magnitude Scores and
Inter-Blink Intervals


Days A 1 2 3 h 5 6 7


RM 174 +.00 +.20 +.26 +.06 +.06 -.03 +.18 -.06

RM 117 -.12 -.04 -.28 -.09 -.10 -.07 -.00 -.10

RM 180 -.12 +.08 -.Oh -.20 -.09 -.01 -.22 -.08

RM 182 -.17 -.16 +.12 +.13 +.05 -.06 -.06 -.10







justly remain unconvinced if subject data are absent.

In 1200 trials, subjects RM 174, 180, and 182 gave a

total of 37 alpha blinks. This number is small enough to

ignore. Subject RM 117, however, displayed a considerable

amount of alpha blinking. Frequency and Average Magnitude of

alpha blinking are given in Figure 6 for RM 117 only. It may

be noted that this S is the only male in the group. The rate

and average size of alpha blinks is actually seen to increase

in extinction (Day 6). RM 117 was subjected to over 100

extended extinction trials (not shown here) and no decrease

was observed below the overall 75 to 85 per cent rate of

responding. For the other Ss, alpha blinking Average Magnitude

curves run flat at mean vlaues with no indication of changing

trends. In Figure 7, Frequency distributions are presented

for alpha blink Latency and Magnitude. These are given

separately for RM 117; and for the other Ss, combined. Table

6 contains the essential parameters of the distributions

presented in Figure 7.

For conditioned blinking, much more uniform general

outcomes were observed. Descriptively, conditioned responses

(CRs) were highly variable in latency, magnitude, and form.

Alpha blinks tended to be uniform in latency and shape where-

as CRs exhibited a wide range of variations.

Daily values for Frequency, Total and Average Magnitude,

and Average Latency of CRs are presented in Figures 8 through

11 respectively. The Frequency and Magnitude measures reflect

effective conditioning and extinction. Only nine CRs were

scored on Day A so the Day A values of Average Blink Magnitude



















tio



a
r'l


rl




-H















C 0





Pq 0





1:1
a)




Cic
ci-











3anifi
C D o w
N V-1 T


N9Vi ANI78 NV31
CU o o


SANI18 VHd1V 1N33 83d

































Fig. 7. Frequency Distributions of Alpha Blink
Latency and Magnitude for RM 117 and for
RM 174, 180, and 182 Combined.
















80 -


701-


60 -


cn

50

m
S40
z


30 -


20 -


RM 117


101-


100 200
LATENCY (MSE


mlO
m10
L-











20


m
z
._.1










4-'










300
C. )


RM 174
RM 180
RM 182


BLINK MAGNITUDE


SRM 117





10 20 30
BLINK MAGNITUDE


RM 174
RM 180
RM 182


100 200 300
LATENCY (MSEC.)







TABLE 6

Alpha Blink Magnitude and
Latency Statistics


RM
RM 117 174 180 182
Combined

Total No. 289 37
Magnitude Mean 9.67 7.11
S.D. 5.38 6.11
Range 31 24




Mean 153.84 181.11
Latency S.D. 32.63 57.01
Range .250 290



































Fig. 8. Frequency of CRs as a Function of Days.













100

80
It
,60

( 40

a 20


innl


RM 117


A 1 2 3 4 5 6 7 A 1 2 3 4 5 6 7
DAYS
RM 18n RM 182


80L


S60
i-

S40

20


A 1 2 3 4 5 6 7 A 1 2 3 4 5 6 7
DAYS


100. GROUP


S80

z 60



a-
20


I rL.


I I


A 1 2 3 4 5 6 7
DAYS


I I t I I


I I I.


ILL


I I


- -- -


RM 174


1 uU, I,, ,

































*4



0


0
-H
4
a










ct

o


0
0






4)
co
Cd


to-

0


o
.-



































































3anllINVN iVIOI


- r- Co 00
SCO CO





























0
eV




0
*ri



0
0




P4





t-
0



d


a
0
















42)









H
ti
0





a>
13r

;3













a- c- D0 CJ
C I. -- c cO
D -- -- W-- t-
CDZ09


I (
I
I
I1h


O O-


3anl lNO9VJ AN118 NV31I
































at




0
~r
o
*1


3


U)

@3
a




U)
o
8.)




0

C,
0
I,
'S


a,
w



@3
4)











r4
*1l


b)
to
4




60













I O
\06
I










CD D CD C CD C-




C C C CD C C
----- i -- l -- i --- l -----\








o 0 Lo o C

(033SV) A3N31V1 83 NV3V1







and Average Latency should be regarded cautiously.

Conditioning is evidenced by reliable increases in

the Rate of CRs, Total Magnitude of CRs, and Average CR

Magnitude during training. In extinction, these measures

showed a rapid return to or below the level of Day 1

responses with the sole exception of the Magnitude measures

for RM 180. Two Ss, RM 174 and RM 117, who showed the

greatest amount of spontaneous blinking, are also seen to

be stronger conditioners. They display shorter average

latencies than the other two Ss as seen in Figure 11.

An analysis of variance was performed on each set

of data with the following results. The frequency data

reflect reliable day differences, F(7,21)=25.6,p<.01, and

Linear and Cubic trends are in evidence at the .05 level;

F(1,21)-6.22, 5.00, respectively. For the Quadratic trend,

a higher value of F is achieved; F(1,21)=161.29. From this

it can be concluded that conditioning and extinction

treatments were effective. Planned comparisons showed Day 1

reliably different from Day A, and similarly for Days A, 1,

6, and 7 as constrasted with Days 2, 3, 4, and 5.

The effect of Days in the analysis of Total Magnitude

was reliable, F(7,21)=46.5,p<.01. The Quadratic trend was

similarly significant, F(1,21)=46.5,p<.01, while Linear

and Cubic trends were absent. Virtually the same results

were obtained for Average CR Magnitude: the daily values

differ at the .01 level, F(7,21)-3.87, and the Quadratic

trend is reliable at or beyond this level also, F(1,21)=

20.5. Linear and Cubic trends were not significant.








An analysis of variance performed on the Average CR

Latency figures did not yield a significant value of F.

These results indicate that the daily values of Average CR

Latency cannot be considered to differ reliably.

Frequency and Total Magnitude of CRs were also plotted

by blocks for closer inspection. These curves are presented

for each day in Figures 12 through 19 with Frequency and

Magnitude data side by side for purposes of comparison. The

remarkable similarity of these two measures is notable. In

general, group curves reflect subject results accurately.

For all subjects, accelerated performance is seen on Day 1.

First-block values for all subjects on both measures are

larger on Day 2 than any on any Day 1 block. This finding

can be the result of either large overnight gains or in-

session inhibitory effects. This may resemble Hovland's

"inhibition of reinforcement" but in any case it is taken

to reflect inhibitory processes which depend upon session

length. Day 2 and the following days all show in-session

decrements which leave the curves looking very much like

the decreasing SBR session curves (Fig. 4) and probably

both sets of curves are the result of processes which

could be characterized as falling drive or alertness levels

and are systematically related to the loss of vigilance

which was discussed in the SBR section.

In extinction, the first-block.values on Day 6 for

all Ss on both measures are at or below the last previous

block values.

The increase in Day 7, Block 1 values over the last



































Fig. 12. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day A.

















RM 174


RM 174


w
S100


3 4 5 1 2 3 4 5
'=1O
I-


RM 182


-.J
C500
i-
400


1 2 3 4 5 1 2 3 4 5


S RM 180


RM 182


300

200 -

100 -


1 2 3 4 5 1 2 3 4 5
BLOCKS

DAY


SI I I I '2 5
1 2345 1 2 345


GROUP


5001


w

S400


-J
w 300
I-
w

g 200


GROUP


100l


S L Lx


, I


1 2 3 4 5


BLOCKS


RM 180


100

80

60

40

20


100

90

80

70
C-3
- 60

50

S40

30

20

10


1 2 3 4 5


= _


RM 117


RM 117



































Fig. 13. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 1.

















RM 174I I 5R1
- 500

400

300

200

V100
I--

1 2 3 4 5 1 2 3 4 5Lo
mJ


100

80

60

40

20




0loc
I-

w10E
alOC


RM 182


0500
40
400;


1 2 3 4 5 1 2 3 4 5


- RM 180


RM 182


200 -




1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
BLOCKS

DAY I


5001


GROUP


a
' 400

4s
-J


I-
300

La

200
La


a I I I


1 2 3 4 5 1
BLOCKS


GROUP


2 3 4 5


RM 180


I I I I


m


l


RM 117


RM 174


PD 117


mil "t A


I




































Fig. 14. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 2.

















RM 174 RM 117
500

400

300


RM 174


RM 117


200[

S100


100

80

60

40

20



C.



80

60
I









40

20


L I I 1 I 1 I I I I
1 2 3 4 5 1 2 3 4 5

RM 180 RM 182












1 2345 1 2345
1 2 3 4 5 1 2 3 4 5


BLOCKS


DAY 2


GROUP


500


400
i-


300
-J
i-
c
I-

200
Li


GROUP


1001


I I .-C I I
1 2 3 4 5
BLOCKS


1 0 I I -


1 2 3 4 5


RM 182
M 500,

400

300

200

100


1 2 3 4 5 1 2 3 4I 5
1 2 3 4 5 1 2 3 4 5-


RM 180


1 I I I 4 I I I 4
1 2 3 4 5 1 2 3 4 5


.


































Fig. 15. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 3.


















RM 174 RM 117
100t


500

400

300


RM 117


2001


Li
1100

I I I I. I I I I L 1
1 2 3 4 5 1 2 3 4 5
-j
RM 180 RM 182 500
\ -
400

300

200


1 I I 4
1 2 3 4 5


I 2 3 4 5 1 I I I4
1 2 3 4 5 1 2 3 4 5


S RM 180


RM 182


100 I


1 2 3 4 5 1 2 3 4 5


BLOCKS

DAY 3


- GROUP


GROUP


200 1



100 -


I IL 11 _


I


1 2 3 4 5


BLOCKS


I-
Li


Ca
810(
81
81


1 2 3 4 5


s


RM 174


































Fig. 16. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 4.















RM 174 RM 117
500

400

300

200

100
-1 2 3 4 5 1 2 3 4 5
S 2 3 5 I 1 I i 5z
1 2 3 4 5 1 2 3 4 5


100

80

60

40

20
C.,
i-
z
C-
10ooi

80

60

40

20


-J
RM182 .500

400

300

200

100


a I I I- I I I V I
1 2 3 4 5 1 2 3 4 5


RM 174


RM 117


1 2 3 4 I 1 2 I I 5
1 2345 1 2345


RM 180


1 2 3 4 5


RM 182


1 2 3I 4
1 2 3 4 5


BLOCKS

DAY 4


100o


I I I I


GROUP


I I 1 I I


1 2 3 4 5


RM 180



EW


S GROUP


100

90

80

70

60
I-
50

' 40

30

20

10


500



S400
z
-3


i-
t300



w-200
U.


1 2 3 4 5


BLOCKS


































Fig. 17. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 5.















RM 174 RM 117
500

400

300

200

-100

II
I 2 I 5 I I I I 5 -
1 2 3 4 5 1 2 3 4 5


RM 182


\1,


o500
C)
400

300


1 2 3 4 5 1 2 3 4 5
1 2 3 45 1 2345


RM 174


1 2 3I I I I I.
1 2 3 45 1 2 3 45


- RM 180


RM 182


v102



1 2 3 4 5 1 2 3 4 5
1 2345 1 2345


BLOCKS

DAY 5


S GROUP


a I I r~


1 2 3 4 5


cr
I-


Woc
a-10


GROUP


500


w

S400
-0


g 300
i-
U
cr
L 200


100l


1 2 3 4 5


BLOCKS


I I I I I


RM 117


I,


































Fig. 18. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 6.
















RM 174


RM 117


- 300

200

100

p I I I im
1 2 3 4 5 1 2 3 4 5

RM 180 RM 182 c500
i-

t 400

S\300

200

100

I I I I-I
1 2 3 4 5 1 2 3 4 5
BLOCKS

DAY


100

90

80

70

60
I-
W 50

S40

30

20

10


GROUP


500 1-


400


CD
-c

S300
-J
C2
0
I-
a 200
cc



I I I I
1 2 3 4 5 1 2 3 4 5

RM 180 RM 182











1 2 3 4 5 1 2 3 4 5


GROUP


-


100i


I I I I I
1 2 3 4 5


1 2 3 4 5
BLOCKS


RM 117


RM 174



































Fig. 19. Frequency and Total Magnitude of CRs
as a Function of Blocks of 10 Trials, Day 7.
















RM 174


500-


400

300

S200
"-i
S100


1 2 3 4 5 1 2 3 4 5
C.-,


- RM 180


RM 182


_J
500
400
400


-,V



1 2 3 4 5 1 2 3 4 5
BLOCKS


1 2 3 4 5 1 2 3 4 5


- RM 180


RM 182


1 2 3 4 5 1 2 3 4 5
BLOCKS


DAY 7


5001-


GROUP


a400
I-

tc
W300


200
u200
-c


100o


I 1 I I


1 2 3
BLOCKS


4 5


w100

80

60

40

20


GROUP


100

90

80

70

o 60
I.-
50

S40
a.

30

20

10


1 2
1 2


3
BLOCKS


4 5


RM 174


RM 117


RM 117







block of Day 6, which is present in both measures for all

Ss except RM 117, can be termed spontaneous recovery.

(There is no reason to think that this is anything other than

the result of the inhibitory process mentioned above, but

the same may be said of the process of spontaneous recovery.)

It is evident in the figures that the various On-

trial response measures are strongly related to one another.

The intercorrelations of the various day values are pre-

sented in Table 7. Average and Total Magnitude are both

closely related to the Frequency measure and to one another.

For all three measures, a negative correlation exists be-

tween response strength and latency. That is, stronger

response strength values are associated with shorter latency

values. Total Latency values are also strongly related to

the CR Frequency data; and, in fact, resembled the CR data

so strongly that curves for these data have been omitted from

this presentation as being of little additional interest.

Reliability of the primary response measures was

determined by splitting each day session into odd and even

trials. Unadjusted for session length, the resulting relia-

bility coefficients were: Frequency, 0.967; Total Magnitude,

0.969; Latency, 0.913. Magnitude is thus equal to or better

than Frequency and all three figures are in close agree-

ment with values reported for human research.

To illustrate the relationship between response

measures somewhat further, a few more correlation values

may be of interest. Employing a linear least squares fit,

Total Magnitude correlates with the Frequency measure, 0.906.











































H





0 4
4

t- 0 d


,- 0



H 0

0 4

0 m




H

OFi


+4





43
0









aS
EA







U












a)



US
t2


0

H)


0
O
E-4









O
+>


P-

*
H O

































H CO oH















oo of\






S* *

















0 a
&t u E-





*o 0 -

Z E- 'd E
ai 3~
Wa +>







For this procedure, Number of CRs = 0.0217 Total Magnitude +
2
12.2114 and the index of determination (r ) 0.8213. In

a Hyperbolic fit, Number of CRs = Total Magnitude/(0.0136

Total Magnitude + 9.5873) and the index of determination is

raised to 0.8772 (r = 0.937). These are very substantial

correlations and they indicate that Number of CRs and Total

Magnitude largely contain the same information. But the

other measures may be seen to hold contributing variance

also; multiple correlation of Total and Average Magnitude

with Frequency gave an index of determination of 0.8683 (R =

0.932); these values rose substantially when Total and
2
Average Latency were added R = 0.957; R = 0.978.

The relation of On-trial to Off-trial responding was

explored and it was found that there is little relation be-

tween SB measures and On-trial response values. As indicated

in Table 8, Total Magnitude measures On- and Off-trial are

independent. In addition, Number of CRs shows no systematic

relation to the various measures of spontaneous blinking.

It should be noted that if these data were analyzed by blocks,

a relation would be expected due to the inhibitory factors

discussed above. This is regarded as an appropriate topic

for later research.

Latency and Magnitude distributions for all CRs are

presented in Appendix A. The parameters of these distri-

butions are listed in Table 9. Comparison of these values

with those for spontaneous blinking (Table 2) will further

indicate that the Magnitudes of CRs represent a class of

responses different from spontaneous blinking. Average CR







TABLE 8


Correlation of Off-Trial Measures
with On-Trial Measures

On-Trial
Off-Trial No. CRs Total Magnitude


Total Magnitude


Average Magnitude


.071
non-significant
t(30)-.389
-.192


-.118
non-significant
t(30)=.651


non-significant
t(30)=1.069
No. Spontaneous Blinks .309
non-significant
t(30)=1.782































H

ra
c(

0
*w-t


4-')
V)






4*




0)











$4
0












Cd
C.)
a



o



^Q

F4

*rl

Q,

a
o


fr(
Tl
Ce
aj
0:
p;
u


tv\
I







N-


0\







r\















cv,







C'







H-


0 LI\ (Y

0\ O\t

NH00
0A 0




0H \D
OHH


HH





O\ 0







C\J.
HO

0 \0







C C









0 **\
-H















0 C\






C\H





O \D



c'j C',




C\J
( 0
0 **
































0 **


cOH (
LA\ O\ H

COC







0 \D






OH
coNo









HH



O\ 0-\O







m\O H









H.j












t- C\ \D








o \D\D
r-l f


*O

S.0
N --r





. (Yo
0 (M







m O\
0 0 Lo

.*






\ (1 \D
* \o








N \O
Mt






.\r0
rO N




O CO
*O
00 .0












m N CYU
. .0

M O\O

rIc *
(O




S0
M V














0
O









4





0 \0
U--)Ot












) Q ^
Q 0
a cou


C\ oUN
O\O\Uo
*\D

HNI






0
OO I







O 0
ON o o\
. 0




o~-
\0C1
\DCN


\o m



r\ O t













o C--






0 *
t- --







C\0

COO *
0 1 I
r\ No





0\ 0




0 t- H
CO I

















S 0








H
C0



O




S0
03 u


0

0


41)
rA

P
$4
4-)
0



0
4-,
m






4)
0
o





40




C
-0


He
Pk-
oa








Magnitude on the performance Days 2 through 5 is 30.91 and

this figure is considerably higher than the Average

Spontaneous Blink Magnitude of any S.

Information is also provided about the relation of

response latency to magnitude. The correlation of these

measures is given first for the data with zero values

inserted for missing CRs, and more importantly for the

regular data with no zeroes included. Blinks which occur

late in the interstimulus interval can not reach the

magnitude that earlier ones can because of the simple fact

that not as much time is available to the late blink for

building magnitude values. This results in an inverse re-

lationship between response magnitude and latency. The

effect is seen to appear on Day 1, grow strongest in Days

2 through 5, and disappear in Days 6 and 7.

At the start of this study, the hypothesis was held

that Latency and Magnitude measures would show inhibition

of delay or response differentiation. This would result

in increasing latencies and decreasing magnitudes as the ex-

periment progressed. The data presented thus far do not

provide support for this hypothesis.

The inhibition of delay concept involves a gradual

movement of the CR in the direction of the UCS in time and

an overall decrease in CR Magnitude before UCR occurrence

with the eventual overlapping of the two. This is a different

effect from an overall decrement in performance observed in

the session data. Incidental observation indicated that an

inhibition of delay effect might be hidden in a saw-tooth







sort of process. Responses appeared to grow independently

smaller and later until one was missed, whereupon respond-

ing resumed at high levels and low latencies. To test for

an effect of this sort, data were scored in six classes

for both Magnitude and Latency. A first order Markov state

transition matrix was constructed for each of the distri-

butions given in Table 9. Examples of these are included in

Table 10 (for Days 1 through 7).

If inhibition of delay were present as a repetitive

decay process one would expect higher totals in the upper

right half than the lower left half of each matrix. That is,

above and below the diagonals indicated in the table. Such

was not the case for the class intervals examined. These

findings, taken together with other results, have not

supported the hypothesis of inhibition of delay and the

conclusion is drawn that the effect is absent. In fact, no

systematic changes or trends in latency have been identi-

fied at all.

Experiment 2.--In Experiment 2, differential condi-

tioning was attempted with auditory stimuli on Days 1 and 2,

visual stimuli on Days 3 and 4; and mixed auditory and

visual stimuli on Days 5 and 6. In an initial series of

habituation trials nothing unusual was observed and it was

possible to assign auditory stimuli to each S so that CS-

was the frequency to which he was initially most responsive.

An experimenter error on Day 2 resulted in reversal of

stimulus conditions for RM 174. This is reflected in the

curves to be seen later. The error was unadjusted and the







TABLE 10

State Transition Matrix for Latency and Magnitude
for Days 1 through 7

Response (n+1)
Response n 0 1 2 3 4 5


LATENCY


[305]
6
30
45
61
42


6
[11]
39
8
2
1


28
46
[170]
49
31
16


47
1
58
(57)
29
17


55
3
28
31
[39]
19


48
0
15
19
13
[13]


MAGNITUDE


[441]
102
36
33
2
1


104
[79]
34
45
2
0


27
33
[73]
65
14
2


40
47
61
[79]
16
2


2
2
10
19
[8]
2


1

0
4
1
[0]







outcome was allowed to work against the hypothesis in the

analyses that follow. On-trial responses were scored as

before except that a 200 msec. criterion was adopted to

eliminate alpha blinks. Scorable blinks were classed as

CRs if latency exceeded 200 msec. In other respects, the

procedures and scoring were identical to Experiment 1.

Group and subject outcomes are displayed as day values

in Figures 20, 21, and 22. These are, respectively, Per

Cent CRs, Total Magnitude of CRs, and Average Blink Magni-

tude for CRs. Here and throughout the following figures,

solid lines represent responses to CS+, dashed lines to CS-.

Statistical comparison of these curves showed that respon-

siveness to the CS+ was reliably superior to the CS- for all

measures. One-tailed sign tests (Siegel, 1956 ) performed on

each data set resulted in significant differences at the

p<.001 level for the Frequency measure and the Total

Magnitude measure. The level of significance for the

Average Blink Magnitude values exceeds the .05 level.

Curves are not shown for the Total and Average

Latency values because the former strongly resemble the

Frequency data and add little, and the latter run.almost flat

without apparent trends or systematic effects. Sign tests

on these data showed the differences in Total Latency

significant at the .001 level and no reliable differences

for Average Latency.

From these results it can be concluded that the

differential conditioning treatment was effective, and the

Magnitude measures are seen to compare favorably with the



































Fig. 20. Frequency of CRs
as a Function of Days.











RM 174


100

80

60

40

o 20
I-
w
Lii
C-3


100

80

60

40


a1 I I I I AI


RM 182


i I I I


I I


1 2 3 4 5 6


100
lo
C 80
I-

S60

o 40

20


1 2 3 4 5 6

DAYS


GROUP


I I I I I


1 2 3 4 5 D
BAYS


F.


I I


---


L


s m -1


I


RM 117




































Fig. 21. Total Magnitude of CRs
as a Function of Days.










KM 11


2000

1600

1200

800

i 400
I-



-J
2000
0
1600

1200

800

400


RM 180


RM 182


I I I I I I


1 2 3 4 5 6 1 2 3 4 5 6
DAYS


1200

1000
Lii

_ 800

S600

_J 400
I-

- 200


- GROUP


'/ \
\ / \



-V

I I I I I l


1 2 3 4 5 6
DAYS


a a a I V I 1


I I I I


-


-- -


I -


L RM 174