Infants' responses to temporally regular events and their omission

MISSING IMAGE

Material Information

Title:
Infants' responses to temporally regular events and their omission
Physical Description:
ix, 93 leaves : ill. ; 28 cm.
Language:
English
Creator:
Davies, Margarete Boettcher, 1957-
Publication Date:

Subjects

Subjects / Keywords:
Biological rhythms -- Measurement   ( lcsh )
Cognition in children   ( lcsh )
Time perception in children   ( lcsh )
Newborn infants   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1985.
Bibliography:
Includes bibliographical references (leaves 89-92).
Statement of Responsibility:
by Margarete Boettcher Davies.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000874846
notis - AEH2378
oclc - 14638501
System ID:
AA00003398:00001

Full Text












INFANTS' RESPONSES TO
TEMPORALLY REGULAR EVENTS
AND THEIR OMISSION




By

MARGARETE BOETTCHER DAVIES


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


UNIVERSITY OF FLORIDA


1985





























To Tom, with love and gratitude












ACKNOWLEDGEMENTS

To thank Dr. W. Keith Berg only for his help and

support with this dissertation would be to understate the

extent of his efforts. From the beginning of my graduate

education, Dr. Berg has always brought his expertise,

kindness, and encouragement to bear on the seemingly

countless crises involved in studies such as this one.

Special thanks also go to Dr. James J. Algina for his help

with the statistics. This appreciation is extended to the

other members of my committee, Drs. Barbara G. Melamed,

Scott A. Miller, and Wilse B. Webb.

I am also grateful to Alonso Avendano, Cecile Chapman,

Robin Cohn, Myrna Jones, Sanjiv Patel, and Diane Pugh for

their help with data collection and reduction.

Furthermore, I would like to thank my subjects and

their parents for their time and effort.

Finally, I would like to thank my husband, Tom, for his

help in all phases of this dissertation.


iii













TABLE OF CONTENTS



Page

ACKNOWLEDGEMENTS....................................... iii

LIST OF TABLES .... ......... ........................... vi

LIST OF FIGURES..... .................. ................. vii

ABSTRACT.............................................. viii

CHAPTERS

I INTRODUCTION....... ................ ........... 1

The Importance of Rhythms and Timing......... 1
The Special Importance of the
Time Dimension for Infants................ 2
Focus of this Dissertation.................. 6
The Time-Locked Response................. 7
The Anticipatory Response................ 9
The Impact of Temporal Variables
on the Processing of Information....... 10


II METHODS...... ........ ..... ...................... 13


Subjects..................................... 13
Stimuli..................................... 14
Apparatus ... ............... ..... ............ 15
Procedure................ ........... ........ 17
Data Analysis............................... 18


III RESULTS........... ....... ....................... 20


Stimulus Onset and Offset
Trials 1 to 5. ..... ...................... 23
Stimulus Onset Trials
1, 3, and 5.............................. 24
Stimulus Offset Trials 2 and 4........... 27
Additional Offset Analyses
and Comparisons. ...................... 29
Pulsed versus Continuous Stimuli........... 30












Response to Stimulus Omission,
Trials 6, 7 and 25, 26...................
Stimulus Omission Trials 6 and 7........
Stimulus Omission Trials 25 and 26......
Dishabituation...............................
Anticipation .............................
The Anticipatory Response
after Stimulus Change....................


IV DISCUSSION.....................................


APPENDICES
A
B
C

D

E

F


INFORMED CONSENT FORM......................
MEDICAL QUESTIONNAIRE.......................
ANALYSES OF SEX EFFECTS
FOR TONE ONSET TRIALS..................
ANALYSES OF
PULSED VERSUS CONTINUOUS STIMULI........
ANALYSES OF
DISHABITUATION TRIALS....................
TABLES......................................


REFERENCES.............................................
BIOGRAPHICAL SKETCH ....................................














LIST OF TABLES


Table 1 Results not mentioned in the
text....................................... 77

Table 2 Main effects of sex and frequency............ 85













LIST OF FIGURES


Figure

Figure

Figure

Figure

Figure


Figure

Figure

Figure

Figure

F igure

Figure


Figure 12


Stimulus Pattern.............................

Response to Onsets versus Offsets............

Sex Effects to Stimulus Onset Trials........

Stimulus Offset Trials 2 and 4.............

Onset and Offset Responses
for Pulsed versus Continuous Stimuli.......

Response to the Initial Stimulus Omissions..

Responses to Initial Offset Omission........

Response to the Final Stimulus Omissions....

Response to the Final Offset Omission.......

Dishabituation ..............................

Frequency Specific Response on
Dishabituation Trial........................

Anticipatory Response......................


vii
















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




INFANTS' RESPONSES TO
TEMPORALLY REGULAR EVENTS
AND THEIR OMISSION

By

Margarete Boettcher Davies

August, 1985

Chairman: Dr. W. Keith Berg
Major Department: Department of Psychology

Infants one to two months of age were presented with

pulsed stimuli that came on and off at regular 20 second

intervals with omissions of these events occurring as test

trials. Following initial test trials, an uninterrupted

series of trials was provided to heighten the chance for an

anticipatory heart rate response to develop. Finally, there

was a frequency change to examine the phenomenon of

dishabituation and its effect on timing. The stimuli

employed were pulse trains consisting of 70 dB (A) sine

waves at either 1000 or 1600 Hz that came on and off at 500

millisecond intervals. This pulse train was alternately

turned on and off each 20 seconds.


viii










While the infants exhibited a large orienting response

to pulse train onset, there was virtually no response to

pulse train offset. The implications of this result are

discussed in the context of orienting theory. The

time-locked response, i. e., the response that occurs at and

following stimulus omission, was rather weak compared to a

previous experiment by Davies and Berg using continuous

stimuli; it occurred predominantly to the first of two

consecutive event omissions, and the deceleration was

rather irregular in shape. This outcome supports Thomas and

Weaver's theory which proposes that higher levels of

information processing disrupt timing. Evidence for

anticipatory responding was weaker than that for a

time-locked response. This is consistent with the

literature on this age group. The weak anticipatory

response was disturbed by the change in frequency.

Dishabituation was significant only for a frequency change

from high to low.
















CHAPTER I

INTRODUCTION

The Importance of Rhythms and Timing

For animals and humans alike, the ability to estimate

time intervals is adaptive. Examples ranging from mating

and predatory behaviors, where temporal coordination leads

to greater efficiency, to skills which usually depend on

high levels of temporal accuracy attest to the importance of

timing capability. Organisms' ability to judge time

intervals is often hypothesized to be based on physiological

oscillations (e. g. Ashton, 1976). Rhythmic oscillations

have been found at all physiological levels in such

astonishing numbers that three journals deal with them

exclusively (Chronobiologia, the International Journal of

Chronobiology, and the Journal of Biological Rhythms).

Though the most obvious example is the sleep-wakefulness

rhythm with its far reaching physiological effects,

oscillations have been found at all physiological levels

(Conroy & Mills, 1970, Palmer, 1978). The ubiquity of

physiological rhythms suggests that rhythmic phenomena

should also be apparent at a psychological level. Several

behavioral rhythms have already been found. Performance,

for example, cycles at approximately 90 minute intervals

1













during waking hours, appearing to be an extension of the 90

minute REM cycles occurring during sleep (Lavie, 1982).

Other examples include rhythms in play (Thomae, Erftmann,

Lehr, & Schapitz, 1972), intrinsic temporal patterning in

the spontaneous movement of neonates (Robertson, 1982), and

rhythms in skilled performance (Shaffer, 1982). Thus, the

diversity of physiological rhythms seems also to be evident

among psychological phenomena. The aim of the present

dissertation is to investigate timing mechanisms during

infancy and the possible influence rhythmic aspects of a

stimulus may have on simple information processing.

The Special Importance of the Time Dimension for Infants

A variety of findings suggests that the time-dimension

is more important during infancy than during adulthood.

Infants display rhythmic behaviors that disappear later in

development (Wolff, 1967); the infantile rhythmic sucking

pattern, for example, cannot be reproduced by adults, but

may reoccur in cases of degenerative central nervous system

disease (Wolff, 1967, p. 214). Other rhythms that occur

only in infants and have a periodicity similar to the one of

the sucking rhythm include crying (Wolff, 1967) and scanning

in the dark (Haith, 1980). This ontogenetic development

parallels changes across phylogenetical levels: lower

animals tend to exhibit more periodicities than higher

animals (Richter, 1965).










3

Data from conditioning studies indicate that young

infants perform at least as well if not better than older

children or adults on time estimation tasks. Two-month-old

infants gave a significant heart rate response to the

omission of a stimulus after only four 20 second intervals

had been presented while the qualitatively different

response of 7-month-old infants did not reach significance

with so few intervals (Davies & Berg, 1983). The response

of the younger group was time-locked; i.e., it occurred at

the time the stimulus would have occurred, while the older

group showed, later in the test session, anticipatory

responding. Since different mechanisms may underly the two

kinds of response, no definite conclusion can be drawn as to

which age group is more accurate in timing. Nonetheless,

the data make it evident that young infants are able to very

quickly detect and reflect an externally imposed interval in

their physiological activity. Brackbill, Fitzgerald, and

Lintz (1967) conditioned pupillary constriction and dilation

to time and to a tone in adults and in infants in an age

range from 26 to 86 days. To condition pupillary responses

to time, the off/on light was alternated in a 20 seconds

off, 4 seconds on (constriction training) or 20 seconds on,

4 seconds off (dilation training). To condition pupillary

response to tone, a delay conditioning paradigm was used

with tone onset preceding a 4 second light change by 1.5

seconds. Variable intertrial intervals were used. For the










4

infants, pupillary dilation and constriction were

successfully conditioned to time but not to the tone. The

adults, on the other hand, could be conditioned to the sound

but not to time. The authors' interpretation is that time

perception may be already fully developed in young infants,

and that this ability may be impaired rather than enhanced

during later development. A problem with this study is that

the pupillary diameter values of the 4 seconds immediately

preceding the test trials were used as baseline values. If

there was anticipatory dilation and constriction in the

adult group, as Davies and Berg's (1983) data suggest there

may have been, then a temporal conditioning effect would

have been masked. A comparison of the baseline values of

the time condition versus the sound condition from Table 7

of the Brackbill et al. (1967) monograph is suggestive of

such an effect. Although a superiority in timing ability on

the part of young infants cannot be unequivocally inferred

from either the Brackbill et al. (1967) or the Davies and

Berg (1983) studies, a study by Hoffman, Cohen, and DeVido

(unpublished manuscript) gives more conclusive evidence for

more optimal and, in some respects, more accurate timing on

the part of infants 7 to 60 weeks old. In a delay

conditioning paradigm, a tap to the glabella was preceded by

a 500 msec tone. Conditioned anticipatory blinks occurring

during the tone period had a longer latency with infants

than adults. As a consequence, for the infants but not










5

adults, the tap to the glabella occurred when the neural

mechanisms for a subsequent blink were still in a relatively

refractory state. This had the effect of reducing the

subsequent blink, elicited by the tap, which the authors

interpret as greater efficiency (presumably because less

energy was expended) on the part of the infants. The

authors make a claim for greater accuracy on the part of the

infants, since in their paradigm anticipation too far in

advance of the UCS may be seen as disadvantageous.

Caution is warranted when comparing the Hoffman et

al. results with the two temporal conditioning studies

described earlier since it differs substantially in the

temporal interval assessed and the age of the subjects

tested. Nonetheless, these few data available attest to the

remarkable timing capabilities of infants.

The temporal dimension spans a wide range of infant

behaviors, from simple physiological reflexes to highly

organized social behaviors. Not only do infants perform

well in timing tasks, they also seem to enjoy the temporal

aspects of their lives. The time dimension is important in

many infant games (Miller & Byrne, 1984) as well as in

soothing infants (Ter Vrugt & Pederson, 1973). Davies and

Berg's (1983) experiment indicates that young infants may

process time qualitatively differently than older infants

and adults. While older infants and adults react with an

anticipatory heart rate response to a predictable temporal










6

interval (Davies & Berg, 1983, Johnson & May, 1969), younger

infants tend to react with a time-locked response, i. e., a

response occurring exactly when the omitted stimulus should

have occurred (Clifton, 1974, Davies & Berg, 1983, Stamps,

1977). The response of the younger group may constitute a

precursor or a necessary component upon which the

anticipatory response of older infants and adults could be

based. The many different theories on timing that have been

proposed may be put to a better test with infant subjects

because the adult or even the preschooler exhibits behaviors

too complex to clearly reveal the underlying mechanisms

(Ashton, 1976).

Focus of this Dissertation

The main topic of this dissertation is to investigate

the characteristics of the time-locked response which

appears during the presentation of significant events

regularly spaced in time. Developmentally, the time-locked

response seems to be the earliest occurring response to time

(Clifton, 1974, Davies & Berg, 1983), and theoretically the

most primitive one: the only components necessary are an

oscillator and a counting mechanism. For the

developmentally later occurring anticipatory response

(Davies & Berg, 1983), on the other hand, at least a

subtracting mechanism is needed as well, even though the

possibility exists that anticipation works by a completely

different mechanism. Developmental and stimulus











7

determinants of the anticipatory response will be a

secondary topic of this dissertation. In addition, the

impact on the timing process of a rhythmic, pulsed stimulus

and of a qualitative change in that stimulus is explored.

Finally, cardiac orienting to onset and offset of pulsed

stimuli is assessed.

The Time-locked Response

The time-locked response has been shown to occur after

only very few stimulus presentations (Davies & Berg, 1983),

an impressive accomplishment on the part of the

2-month-olds. Since this finding is so extraordinary, the

first step of this dissertation is to replicate and extend

the 1983 results with a different stimulus event. To this

end, a pulse train--instead of a continuous stimulus as in

Davies & Berg (1983)--is employed. In the earlier study a

continuous tone was alternately turned on and off at 20

second intervals; in the present study the pulse trains

started and stopped at 20 second intervals, resulting in a

somewhat more complex temporal situation. The pulses within

the trains went on and off at 500 millisecond intervals.

Thus during the pulse trains there was a regular rhythm at a

frequency that occurs in several spontaneous behaviors of

young infants, as for example sucking or crying (Wolff,

1967). In younger infants (under approximately 5 months),

pulsed stimuli are more likely to elicit attention as

measured by the size of the orienting response (Bohlin,













Lindhagen, & Hagekull, 1981). The difference between the

stimuli that are used in the present study and those used in

Davies and Berg (1983) can be summarized as follows: in the

present study, the stimuli are more complex, since two

temporal intervals are superimposed upon each other (.5

seconds and 20 seconds); the rhythm that is superimposed on

the 20 second intervals matches the frequency of many

spontaneous behaviors in young infants (Wolff, 1967); and

the stimuli are more effective in eliciting young infants'

attention (Bohlin et al., 1981).

If the 2-month-old infants again exhibit a time-locked

response, this result would argue for the pervasiveness of

the phenomenon. If such a phenomenon occurs to two

different stimuli, the implication is that it may have

importance for a larger set of situations in the daily life

of the infant. If the time-locked response is disrupted by

this more complex stimulus, it would be an important

consideration in hypotheses concerning the underlying

mechanism. Disruption of the time-locked response by

stimuli that have a high information value would be

consistent with Thomas and Weaver's (1975) theory of a timer

and a cognitive processor between which attention is

shared. Presumably, a complex stimulus would put a greater

demand on the cognitive processor, and result in reduced

capacity for the timer. Disruption of timing by a more

complex stimulus would also be in agreement with the data of











9

Brackbill et al. (1967), where infants could be conditioned

very easily to one time interval, but it was impossible to

condition them to a pattern of intervals. Even though the

primary focus of this dissertation is the time-locked

response, an investigation of the time-locked response

cannot be entirely separated from considerations for the

anticipatory response.

The Anticipatory Response

The anticipatory response seems to be the mature

response, since it has been found in adults (Johnson & May,

1969) and older infants (Davies & Berg, 1983), while in

younger infants anticipatory responding either could not be

elicited (Clifton, 1974, Davies & Berg, 1983), or the

evidence for anticipatory responding was rather weak

(Leavitt, Brown, Morse, & Graham, 1976, Stamps, 1977, Stamps

& Porges 1975). Clifton (1974), Stamps (1977), and Stamps

and Porges (1975) used newborns in their studies, Leavitt et

al. (1976) used 6-week-olds, and Davies and Berg (1983)used

2-month-olds. In the Davies and Berg (1983) experiment, the

conditions for an anticipatory response to develop were far

from ideal since the test trials (stimulus omissions) were

inserted early and fairly frequently, which can disrupt the

overall perception of the timing pattern. While results

from previous experiments show that anticipatory responding

in newborns is questionable at best, for 2-month-olds the

issue is still unresolved. Thus the present dissertation











10

included a more extended and uninterrupted series of

intervals (see Figure 1) so as to more effectively address

the question whether 2-month-olds are capable of developing

an anticipatory response.

The Impact of Temporal Variables on the Processing of
Information

Young infants tend to exhibit a larger orienting

response to pulsed stimulation than to continuous

stimulation (Bohlin et al., 1981), an effect that disappears

with age (Graham, Anthony, & Zeigler, in press). Also,

pulsed stimulation seems to be advantageous for

dishabituation in young infants (Clarkson & Berg, 1983).

Reports on the response to the offset of pulsed stimuli are

somewhat conflicting. Clifton & Meyers (1969) inan

experiment with four-month-olds with 300 Hz square wave

stimuli found a significant decelerative offset response

only to continuous but not to pulsed tones. This finding is

in agreement with Rewey (1973) who used a pulsed 2-tone

combination and a continuous single sine wave. Berg's

(1972) data on 4-month-olds using pure tones of 1100

and 1900 Hz pointed in the same direction, but the effect

was not statistically substantiated. In a study on

6-week-old infants (Leavitt et al., 1976), no response to

the offset of 162 second presentation of pulsed sine waves

that alternated in frequency was reported, but a significant

heart rate deceleration to the offset of 60 second trains of










11

speech syllables (500 msecs on, 500 msecs off) was found.

Bohlin et al. (1981) did not find an offset response to

pulsed or continuous tones in 13 to 18 week olds; however,

an offset response to continuous but not to pulsed tones was

found in a group of 26 to 35 week olds. There was some

suggestion in the data that with repeated stimulus

presentations the younger infants may have tended to respond

to the offset of continuous stimuli. Because the stimulus

duration was only 10 seconds in this experiment, the

authors suggested that the minimal duration to produce

offset responses may be shorter for continuous than for

pulsed tones. This implies that offset responses to pulsed

stimuli might be observed if the pulse train is sufficiently

long. The present dissertation used pulsed pure tones

which stayed on for 20 seconds with a 20 second

interstimulus interval (ISI). At one point, the pulses

stayed on for 60 seconds (see Figure 1). Thus, the duration

hypothesis of Bohlin et al. was tested to a certain extend.

For comparison, data were available from a similar previous

experiment (Davies & Berg, 1983) using continuous stimuli.























II


1 2 3 4 5 6


7 8 9 10 11 12 13


14 15 16 17 18 19 20 21 22 23 24 25 26










Figure 1. Pattern of pulse train onsets and offsets for
the 26 trials. Cashed arrows indicate where onsets (upward
arrows) or offsets (Downward arrows) were omitted. Cross-
hatched areas indicate the change in tone frequency.

















CHAPTER II

METHOD

Subjects

Subjects were obtained through telephone solicitation

to the parents of infants born in Alachua County, Florida.

Upon the parents' arrival, the study was explained to

them in more detail, and they were asked to sign an informed

consent form (Appendix A). In addition, the parents)

filled out a medical questionnaire (Appendix B) to screen

the infants for health problems. Parents were paid $5.00

for each visit of their child.

Infants with an age range from 55 to 92 days and a mean

age of 68.1 days were tested. To obtain a total of 20

usable subjects, 70 had to be tested. Of these, 41 were

eliminated because of fussing or sleepiness, three for

health problems, two for pre- or postmaturity, three due to

equipment failure, and one for experimenter error. For an

analysis of the first seven trials, the sample consisted of

10 subjects aged 68 days or older and 10 subjects younger

than 68 days with average ages of 75.7 and 60.5 days

respectively. Each age group contained five males and five

females with two males and three females listening to 1000

Hz tones first and three males and two females listening to

13










14

1600 Hz first. Only eight subjects with an average age of

70 days maintained a satisfactory state for the remaining

trials (trials 8 through 26), of which seven were also

judged satisfactory for trials 1 to 7.1 Of the eight

subjects, four were male and four were female. Of the male

subjects, three started the session with 1600 Hz and one

with 1000 Hz, while one female subject started with 1600 Hz

and three female subjects started with 1000 Hz.

Stimuli

Sine wave tones with a pulse rate of 500 milliseconds

on / 500 milliseconds off and a 25 millisecond rise and fall

time came on at 40 second intervals and stayed on for 20

seconds. Onset and offset of a pulse train were considered

separate trials. Pairs of stimulus onset and offset

omissions were created by allowing the pulse train to either

stay turned on or remain turned off for one complete tone on

and off period (see Figure 1). Thus, the test period

resulted in a 60 second tone on period in the cases of the

initial test (trials 6 and 7) and a 60 second silent period

in the case of the final test (trials 25 and 26). The tones

were 70 dB (A) over a background level of 22 dB (A) and 1000

Hz or 1600 Hz in frequency. On trial 21, the first


1Data of the eighth subject were not used for analyses
of the first seven trials due to excessive movement during
this period.










15

dishabituation trial, the frequency was changed from

1000 Hz to 1600 Hz or from 1600 Hz to 1000 Hz.

Apparatus

The stimuli were generated by a Krohn-Hite Model 5300

function generator, a Wavetek Model 131A VCG Generator, or

a Hewlett Packard Model 200 CD Wide Range Oscillator. The

stimulus durations and ISIs were controlled by a Digital

PDP8/e laboratory computer. An Iconix 6837 electronic

switch set to a rise and fall time of 25 milliseconds gated

the onset and offset of the stimuli. The tones were

amplified by a Dynaco SCA-80 amplifier and delivered

free-field through an Analog and Digital Systems Model

810 speaker which was placed facing the infant at a distance

of approximately 50 centimeters inside a 1 by 1.5

meters sound attenuating chamber. Intensity levels were

measured free field on the a scale of a Bruel & Kjaer Type

2203 Precision Sound Level Meter at the site of the infant's

head. A Hewlett Packard 350D attenuator set was employed to

achieve the intensity matching (equating for a 5 dB

difference) between the two frequencies.

The laboratory assistant who stayed with the infant

during testing listened to approximately 95 dB white noise2

in a frequency range of 900 to 1700 Hz that completely

masked the tones. The noise was produced by a


2 Since nonstandard circumaural earphones were used,
the exact loudness could not be ascertained.











16

Grason-Stadler Model 455C noise generator, filtered through

a Krohn-Hite Model 3700R filter, and amplified by a Sansui

AU-517 Integrated Amplifier. It was presented through Koss

PRO/4AA earphones under which the assistant wore earplugs

(McKeon) in order to protect his/her hearing. Leakage of

the noise through the earphones increased the level of

background noise by about 2 dB.

The EKG was detected with Beckman Ag/AgCl 16 mm surface

electrodes that were attached with Micropore 3M

No. 1530 surgical tape, and Synapse electrode cream (Med-Tek

Corporation). Respiration was recorded with a mercury

filled strain gauge and a Parke Electronics model 270

plethysmograph. Both signals were amplified and recorded on

a Beckman polygraph model 5411B using Type 9806A A-C

couplers which also recorded a pulse marking 20 second time

intervals coincident with the stimulus pattern. The

polygraph output of the signals and the stimulus pulses were

stored on a Hewlett-Packard 3960 Instrumentation FM tape

recorder. For heart rate analysis, the EKG signals stored

on FM tape were replayed and, using a shop-built peak

detector to sense each R-wave, the beat to beat intervals in

milliseconds were computed by a Digital PDP8/e laboratory

computer. The interbeat intervals were stored on floppy

discs for later analysis.

The sessions were recorded with a SONY U-matic model

VO-2610 videocassette recorder with a Hitachi CCTV camera











17

Model HV62U and a Model ECM-150 SONY Electret

condenser microphone. They were displayed for the parents

on a Panasonic solid state TV Model No. TR-195MB.



Procedure

The electrodes were taped to the infant in a modified

lead II configuration with the two active electrodes located

approximately one centimeter above the right nipple and at

the bottom of the left rib cage and the ground electrode

approximately one centimeter above the left nipple. The

strain gauge was stretched over the infant's stomach at

waist level. The infants sat in a semi-reclining infant

seat on a table with colored Christmas lights suspended from

the ceiling. The laboratory assistant sat next to the

infant in an angle of approximately 60 degrees at a distance

of approximately 40 centimeters. A subject was tested only

if he or she was in a satisfactory alert state. If the

infant appeared to change to an unsatisfactory state, the

assistant intervened in a subdued manner by either gently

rocking the infant, stroking his/her arm, letting the infant

look at his/her face, or presenting a quiet toy. The

interventions were as few and as rare as possible. Since the

assistant listened to noise, his/her judgement of the

necessity of intervention was impaired due to the lack of

vocal information. The high drop out rate of the present

experiment is in part due to the resulting inability of the











18

assistant to adequately control the infant's state. The

decision on whether to use data from an infant depended on

two independent state raters who judged the infants' state

from the video and audio recording. State criteria for

eliminating infants were as follows: infants were rejected

for fussing, i.e. when they displayed excessive movement,

grimacing, restlessness, or distress vocalization, and they

were rejected for sleepiness, i.e. when they made little or

no movements, the breathing was very regular, and the eyes

were closed or nearly or occasionally closed. If an infant

became seriously upset, data from the subsequent trials were

not used. If an infant became seriously upset before trial

8, the subject was not used for the study. Two trained

state raters independently judged the infants' state on each

trial from the videotapes. The interrater reliability for

the final sample was 98% for whether a trial should be

accepted or rejected. A trial was used for data analysis

only if both raters judged the infant's state satisfactory.

Data Analysis

Interbeat intervals stored on floppy disks were edited

for skipped R-waves, T-waves, and movement artifacts. They

were then converted to average heart rate per second for all

seconds of each predetermined analysis period (see results

section). Analyses of orthogonal components of trends were

performed with the 2V version of the BMDP statistical

package. Trends of higher than cubic order are not











19

reported, because they are uninterpretable for these

data. When main effects were reported, the Huyhn and Feldt

(1976) estimator of the Box corrective adjustment for

degrees of freedom was used in the calculation of the final

p value in order to correct for violations of sphericity

(Huyhn & Feldt, 1970). In the text, the original degrees of

freedom are given.


















CHAPTER III

RESULTS

The analysis was begun by establishing the response to

the regular stimulus trials--the response to actual stimulus

onsets and offsets. Next the response to the stimulus

omissions, trials 6, 7, 25, and 26, was analyzed. Finally,

the dishabituation trials and the possibility of a

developing anticipatory response were examined.

Initial analyses up to and including trial 7 contained

the between subjects factors of age, sex, and frequency.

Age was included as a factor, because preliminary inspection

of the data suggested that it might have an effect. Due to

subject attrition, one of these between subject factors had

to be dropped for analyses of trials later in the sequence.

The factor "age" was chosen to be excluded, since the age

range in this study is rather narrow.

Only eight subjects remained in a satisfactory state

throughout all 26 trials. Analyses on dishabituation,

anticipation, and the later stimulus omissions were done

first on these eight subjects only in order to facilitate

comparisons across analyses and because these subjects' data

could not be influenced by earlier state changes. However,

20










21

the analyses were repeated using all subjects who were in an

acceptable state on the subset of trials of interest in

order to confirm the results with a more substantial N.

The length of the post-stimulus period to be examined

was based on an inspection of the responses on the initial

trials which revealed a large monotonic deceleration that

reached its nadir around 6 seconds after onset of the

stimulus trains and appeared not to entirely recover within

the 20 second ITI (see Figure 2). Therefore, it was

judged advisable to use a long analysis period; the longest

post-stimulus interval possible under the constraints

imposed by the heart rate processing software was the period

from .5 seconds preceding to 15.5 following stimulus

change. Anticipatory responding was analyzed over a

period of 6.5 to .5 seconds preceding stimulus change, as

in Davies and Berg (1983).

In the following analyses, trends of higher than cubic

order are not reported since they are not easily

interpretable. Results of lower order that are not directly

relevant to the questions at hand, that cannot be

interpreted unequivocally, or that are redundant are not

mentioned in the text but are listed on table 1 in Appendix

F. Several analyses yielded main effects for sex and

frequency as a result of consistently higher heart rates for

females than males and higher heart rates for subjects who

started with 1600 Hz than those who started at 1000 Hz.
































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


.5 1 3 5 7 9 11 13 15


SECONDS








Figure 2. Averaged response over onset trials 1, 3, and 5
(solid line), onset trials 3 and 5 (dotted line), and offset
trials 2 and 4 (broken line).










23

Since these results are unlikely to have much of an

influence on the heart rate responses of principal interest

here, they are only noted in table 2 in Appendix F.

Stimulus Onset and Offset Trials 1 to 5

Visual inspection of the results indicated substantial

differences between stimulus onset and offset (see Figure

2). The response to stimulus onset consists of a

substantial orienting response in form of a heart rate

deceleration, whereas there is little response to stimulus

offset, or possibly a slight acceleration. The first

analysis sought to establish whether this difference was

statistically significant. Trials 2 and 4 and trials 3 and

5 were combined into a block of offset trials and a block of

onset trials, respectively. Trial 1 data were excluded

from the onset block in order to provide a more conservative

comparison of onset and offset and to eliminate any bias

against offset that might result from habituation. If

habituation effects occurred, the trials selected would

favor offset responses. A significant on/off by quadratic

seconds interaction, F(l, 12) = 32.58, a < .00025, reflected

the substantially larger response to stimulus onset than

offsets. These results indicated the need for separate

analyses of onset and offset responses. There were also

interactions marking sex: on/off by sex by quadratic seconds

and also by cubic seconds interactions, F(l, 12) = 8.8, E <

.025, and F(l, 12) = 7.56, p < .025, respectively. These










24

results indicate a sex effect for either onset or offset

responses or both, and were further examined through

separate analyses for onset and offset responses.

Stimulus Onset Trials 1, 3, and 5

A quadratic seconds effect, F(l, 12) = 45.56, p <

.0001, which accounted for over 20% of the variance,

and a cubic seconds effect, which accounted for under 4% of

the variance, F(1, 12) = 22.10, p < .001 were statistical

evidence of the unexpectedly large heart rate deceleration

to stimulus onset.

Surprisingly, no significant trials by quadratic or

cubic seconds interaction were found, indicating that

significant overall habituation did not take place over

these 3 trials. A linear trials effect, F(1, 12) = 5.02, p

<.05 was found which in the absence of a trials by seconds

interaction suggested a change in prestimulus heart rate

level. Indeed, there was a slight rise in heart rate level

at second -.5 from 149.83 bpm at the beginning of trial 1 to

150.35 bpm at the beginning of trial 3 to 151.42 bpm at the

beginning of trial 5. Even though this difference was

statistically significant, the absolute change of 1.59

bpm from trial 1 to trial 5 is so small as to make it

of minor consequence with regard to effects on responses.

Some habituation effects were suggested by a

significant trials by cubic seconds by sex by frequency

interaction, F(1, 12) = 5.21, p < .05. It indicated that










25

the response developed differently over trials for different

subgroups, and along with a quadratic seconds by sex

interaction, F(l, 12) = 6.67, p < .025, was followed up by

analyzing each sex separately.

Analyses of sex differences for stimulus onset trials

1, 3, and 5. Based on the seconds by sex interactions in

the response to onset trials, these trials were reanalyzed

for each sex separately with no between subject factors and

only trials and seconds as within subject factors. For the

male subjects, the analysis yielded quadratic and cubic

seconds effects, Fs(l, 9) = 48.5 and 19.23, ps < .0025. The

analysis of the female subjects also yielded quadratic and

cubic seconds effects, Fs(l, 9) = 8.29 and 11.11, ps < .025,

but there was a linear trials by cubic seconds interaction,

F(l, 9)= 5.76, p < .05, not found for males. Thus the

analyses indicated that both sexes show the significant

orienting response to the onset of stimulus trains, but that

the response of the females changes over trials. As is

evident from figure 3, this interaction results from

habituation occurring in female but not male subjects.

Since there was no significant interaction of trials by sex

for trials 1, 3, and 5, the different response development

is unlikely to be due to changes in heart rate level for one

sex or the other. Further analyses of each trial (see

Appendix C) demonstrate that stimulus onset does not differ

between males and females to the initial stimulus, but only
















- -


._male subjects
-_female subjects


5 7.5 11.5 135


.5 35 7.5 11.5 15.5


N


.5 15.5


-.5 35 7.5 11
SECONDS


Figure 3. Response to stimulus onset trials 1 subfiguree
I), 3 subfiguree II), and 5 subfiguree III) for male (solid
line) as compared to female (borken line) subjects.


T 3

-1l
-5
-9

-13
-17


0.
E


-1

w5
z-9
-I
O


-1i

-5
-9
-13


=A










27

to trials 3 and 5, where the response of the females is

significantly smaller than that of the males (see Figure

3). Therefore, these results cannot be explained as a sex

difference in responding to pulsed auditory stimuli per se.

Rather, they may indicate a faster learning of the temporal

interval on the part of the females. There is a precedent

in the literature for this possibility (Stamps and Porges,

1975), as will be described in the discussion section.

Stimulus Offset Trials 2 and 4

In contrast to the response to stimulus onsets, the

response to stimulus offset trials looks rather erratic and

is slightly accelerative (see Figure 2). An analysis of

trials 2 and 4 yielded no significant effect for seconds by

themselves, but only an interaction of linear seconds by sex

by frequency, F(l, 12) = 5.09, p < .05, and a trials by

cubic seconds by sex by frequency interaction, F(l, 12)

5.95, p < .05. Follow-up analyses indicated these

interactions reflected a significant cubic seconds effect,

F(l, 8) = 6.62, p < .05 for the female subjects for trials 2

and 4 combined and a main effect for seconds, F(16, 80) =

2.54, p < .05, on trial 4 for only those male subjects who

had been tested at 1600 Hz (see Figure 4). Taken together,

the erratic significance pattern for trials 2 and 4 seems

likely to result from chance events and not consistent with

an orienting response to stimulus offset (e.g. Berg & Chan,

1972, Davies & Berg, 1983).



















-Trial 4, males at
1600 hz


-- "---

--Trials 2 and 4
females


.5 1 3 5 7 9 11


13 15


SECONDS












Figure 4. Response to stimulus offset: male subjects
tested at 1600 Hz on trial 4 (solid line) and the averaged
response for trials 2 and 4 for female subjects only
(broken line).









29

Additional Offset Analyses and Comparisons

The offset following the 60 seconds of pulsed stimuli

occurring during the initial test trial (see Figure 1) was

analyzed, since the long uninterrupted train of pulses might

have increased the chance for an offset response. However,

no significant seconds effect was found. Nonetheless,

inspection of this trial and the subsequent offset trials

suggested analysis of a pooled block of trials might be more

successful. Therefore, trials 8, 10, 12, and 14 were

combined for an analysis. Only those 15 subjects who

remained in a satisfactory state for all of these trials

were used in the analysis with sex, frequency, trials and

seconds as factors. The response to stimulus offset still

did not reach significance. The analysis was repeated for

trials 8, 10, 12 with the same result. For comparison with

responding to stimulus onset, an analysis was conducted on

stimulus onset trials 9, 11, and 13. The analysis indicated

that there was no longer a significant response to tone

onsets either, probably due to habituation, which may have

been facilitated by the occurrence of 60 seconds worth of

pulses during the test trials.

Based on the previous studies showing offset responses

to continuous stimuli, the lack of an orienting response to

stimulus offset for pulsed stimuli is unexpected. This

finding as well as the surprisingly large orienting response

to pulsed stimulus onset suggested the value of analyses to









30

directly compare the onset and offset responses to pulsed

and continuous stimuli.

Pulsed versus Continuous Stimuli

From a very similar previous experiment (Davies & Berg,

1983), data were available on heart rate responses to

stimulus onsets and stimulus offsets of continuous tones. A

complete comparison of these data with the data from the

present experiment is presented in Appendix D and summarized

here.

The subjects previously tested with continuous stimuli

were similar to those in the present study in most aspects.

On the average they were one day older than the pulsed

group. The timing and loudness of the stimuli was the same

for trials in question. In contrast to the present study,

parents) in the earlier study stayed with the infant

throughout the experiment and neither the parents nor the

experimenter listened to noise to mask the tones. However,

interactions between the adults and the infants were

minimized. Data for the group previously tested with

continuous stimuli were available only for 8.5 poststimulus

seconds, so the same analysis interval was employed for the

present pulsed stimulus data. The 10 subjects from the

pulsed group who were tested at 1000 Hz, the only frequency

used in the earlier study, provided data for comparison.

A comparison of the onset and offset responding to the

two stimulus types is shown in Figure 5. In contrast to the













pulsed tones


-- offset
-onset


35 .5 2.5 45 65 -5


E




UI
z


continuous tones


N


;'5 .5 2.5 4.5 6.5 85

SECONDS




Figure 5. Responses to stimulus onset trials 3 and 5 (solid
line) compared to stimulus offset trials 2 and 4 (broken
line) for pulsed tones subfiguree I) and continuous tones
subfiguree II).









32

previously shown differences in onset and offset responses

to pulsed stimuli, the onset and offset responses for those

subjects who listened to continuous tones appear very

similar, namely a quadratic deceleration in both cases (see

Figure 5). This contrast was confirmed in an analysis

comparing offset trials 2 and 4 to onset trials 3 and 5 by a

significant on/off by linear seconds by pulsed /continuous

interaction, F(l, 12) = 8.83, p < .025.

To follow up on this result, the response to stimulus

onset only was compared for pulsed versus continuous

tones.3 The fact that the response to pulsed stimuli is

larger and more sustained than the response to continuous

stimuli was substantiated by a significant linear seconds by

pulse/continuous interaction, F(l, 12) = 19.12, p < .001.

The interaction with linear seconds instead of quadratic

seconds was to be expected, since with this shorter interval

of 8.5 seconds the recovery period of the large orienting

response to pulsed stimuli is cut off (see Figure 2).

Analysis of offset responding led to the opposite

finding. While the offset of continuous tones results in an

orienting response, the offset of pulsed tones is not

accompanied by a clear response (Figure 5). Statistical


3 Only trials 3 and 5 were used for the analysis
because for the continuous group the response to trial 1 may
have been influenced by an insufficient baseline period.









33

evidence for this was a significant quadratic seconds by

pulse/continuous interaction, F(1, 12) = 6.29, o < .05.

Finally, stimulus onset responses were compared to

stimulus offset responses for the continuous group only.

There was no significant on/off by seconds interaction. The

response shape was a deceleration of quadratic order, F(1,

8) = 15.19, p < .025 (see Figure 5).

Response to Stimulus Omission,
Trials 6,7 and 25,26

For trials 6 and 7 all 20 subjects were available. The

analyses on trials 25 and 26 were initially done only on

those 8 subjects who were in a satisfactory state on these

and all previous trials throughout the experiment, but

subsequently an attempt was made to replicate the obtained

results with analyses involving all subjects who were in a

satisfactory state on trials 25 and 26, regardless of prior

interruptions of state.

Stimulus Omission Trials 6 and 7

The first omissions of a stimulus offset and subsequent

onset (trials 6 and 7) were analyzed separately because

the responses to stimulus onset and offset on trials 1

through 5 were markedly different. Also, the response to a

first and a second stimulus omission may not necessarily be

the same, and the response curves of these two trials look

rather dissimilar.









34

By the beginning of trial 6 (Figure 6), the heart rate

level had returned to 151.42 bpm, a value slightly above

what it was at the beginning of trial 1. Thus, the response

to an offset omission should not be influenced by an

unrecovered orienting response. The response to offset

omission (trial 6) is a heart rate deceleration that starts

about two seconds after the omission, reaches its nadir at

9.5 seconds and is not completely recovered at the end of

the 15.5 seconds interval (Figure 6). However, in an

analysis that included age, sex, frequency, and seconds as

factors, response effects became significant only in the

interaction of sex by age with linear and quadratic seconds,

Fs(l, 12) = 4.91 and 4.79, ps < .05, respectively.

Separate analyses for males and females, with age and

seconds as factors, yielded significant results only for the

females (see Figure 7): a linear seconds effect, F(l, 8) =

12.13, p < .01, and a linear seconds by age interaction,

F(l, 8) = 5.54, p < .05. To follow up on these results, the

older females and the younger females were analyzed

separately with only seconds as a factor. There was a

linear seconds effect, F(l, 4) = 38.14, p < .005, for the

older females (see Figure 7), but analyses on the younger

females yielded no significant results.

In summary, only a subgroup, the female subjects, gives

a significant response to the omission of a stimulus

offset. Even though the deceleration is larger for the
















C
0
._

E
o
\C
\


Z
0
U
in
CO


(wdq) 3ONVHO 8H






Figure 6. Response to the initial stimulus omissions:
offset omission on trial 6 (first omission, solid line),
and onset omission on trial 7 (second omission, broken
line).


CY C4 1- 0 r- C4 CI

































/




-. C


L
0

w
U,


CrJ 0 cY JC


(Ludq) 30NVH3 8H








Figure 7. Response to stimulus offset omission trial 6
by all female subjects (solid line), by the older female
subjects (broken line), and by the younger female subjects
(dotted line).









37

younger females, it is more consistent and therefore becomes

significant only for the older females. A high response

variability typical for the age group of the younger

subjects (Graham, Anthony, & Zeigler, in press).

The response to the onset omission, trial 7, looks

rather irregular and is generally accelerative (see Figure

6). This acceleration may in part be due to an incomplete

recovery of the heart rate deceleration on trial 6. An

analysis of trial 7, including age, sex, frequency, and

seconds as factors, yielded no significant results. Thus,

subjects responded to only the first of two test trials, and

this was restricted to a subgroup of female infants.

Stimulus Omission Trials 25 and 26

For trial 25, an onset omission, data of those 8

subjects who remained in a satisfactory state throughout the

experiment were analyzed. The response is a brief

acceleration over the first 3 seconds of the interval

followed by a deceleration that does not recover over the

entire 15.5 seconds (see Figure 8). The deceleratory part

of the curve was expected from the literature (e.g. Clifton,

1974). The brief acceleration apparently does not result

from an incomplete recovery of the response on the previous

trial: the response to stimulus offset trial 24, like all

the offset responses, was small, and already at 6.5 seconds

before the onset of trial 25 the heart rate had slightly

exceeded the final heart rate just prior to that offset.

























O
0



,
L_
om


/ /
z
0o















(uJdq) JONVHO 8H







Figure 8. Response of eight subjects to the final stimulus
omissions: onset omission on trial 25 (first omission,
solid line), and offset omission on trial 26 (second
omission, broken line).









39

Furthermore, the baseline value of trial 25 exceeds the

baseline value of trial 24 by .54 bpm and is only slightly

lower (.28 bpm) than that of trial 23. Even though there is

no immediate explanation of the acceleratory part of the

curve, accelerations smaller than the one on trial 25 had

been observed on several trials throughout the experiment to

precede the deceleratory response (see Figures 3, 6, 7, 9,

and 10). In an analysis with sex and frequency as between

subject factors and seconds as a within subjects factor, a

significant effect for overall seconds, F(16, 64) = 2.27, p

< .05 was found. Since there were no interactions with sex

or frequency, the analysis was repeated without these two

factors to provide some additional degrees of freedom. The

second analysis revealed a significant linear seconds

effect, F(l, 7) = 6.62, a < .05, in addition to the main

effect of seconds.

When the analyses were repeated with the additional

four subjects who had become fussy on previous trials,

there were no significant effects. The analysis was again

repeated with only seconds as a factor which again resulted

in a linear seconds effect, F(l, 11) = 6.28, p < .05, and an

effect for overall seconds, F(16, 176) = 2.75, p < .025.

The mostly accelerative response to the final stimulus

offset omission (trial 26) appears rather irregular and does

not indicate a definite response. The analysis of eight

good subjects with sex, frequency, and seconds as factors


























.0
w
0
z
3<


ME
x


-males
---females


.5 2.5 5.5 8.5 11.5 14.5


--1000 hz
-4600 hz


.5 2.5 5.5 8.5 115 14.5

SECONDS




Figure 9. Response of eight subjects to offset omission
trial 26 (second omission). Subfigure I: male subjects
(solid line), and female subjects (broken line). Subfigure
II: subjects who started the session at 1600 Hz (solid
line), and subjects who started the session at 1000 Hz
(broken line).






































/
N
N
N

/
/


/
/
/


(uwdq) 30NVHO UH










Figure 10. Response of eight subjects to the last
habituation trial (trial 19, broken line), and the first
dishabituation trial (trial 21, solid line).









42

showed a main effect for seconds, F(16, 64) = 4.1, p <

.00025, a linear seconds by sex by frequency interaction,

F(l, 4) = 15.87, p < .025, and a cubic seconds by sex by

frequency interaction, F(l, 4) = 20.90, p < .025. When

analyses were made for each frequency separately with sex as

a between subjects factor, there was only an interaction of

overall seconds by sex, F(16,32) = 3.02, p< .005 for those 4

subjects who started at 1000 Hz. For the subjects who

started at 1600 Hz there was a linear seconds effect, F(1,2)

= 72.15, g < .025 and a linear seconds by sex interaction,

F(l, 2) = 54.04, p < .025. Since there are so few subjects,

the sample will not be broken down further to examine

interaction effects. While the heart rate curves for the

subgroup of male subjects, female subjects, and subjects

who started the session at 1000 Hz are mostly accelerative,

the heart rate curve for those subjects who started at 1600

Hz consists of an acceleration with a subsequent

deceleration (see Figure 9), similar to the overall response

to trial 25. Next, an analysis of all 12 subjects who were

in a good state on trial 26 was performed in order to

confirm the above results. A linear seconds effect, F(l,

8) = 8.10, p < .025, an interaction of linear seconds by

frequency, F(l, 8) = 19.08, p < .0025, p < .005, and a

linear interaction of seconds by sex by frequency, F(l, 8)

= 17.23, a < .005, was found. However, when subjects were

examined for the two different frequencies separately with









43

only seconds included as a factor, there were no significant

effects.

The slight heart rate acceleration over all subjects on

trial 26 may in part be due to a recovery to baseline level

after the deceleration of trial 25; at the onset of trial

26, the heart rate level is still 1.29 bpm below what it was

at the beginning of trial 25. Only a subgroup, those

subjects who started the session at 1600 Hz, showed a heart

rate deceleration. It is perhaps relevant that this is the

same subgroup who shows dishabituation in the trials just

prior to these omission trials (described below).

To provide for a comparison with equivalent power and

subject selection between the early and late tests of

stimulus omission, trials 6 and 7 were analyzed with only

those eight subjects who were used in the analyses of trials

25 and 26. There were no significant results for either

trial, indicating that the deceleratory heart rate response

to the first of two stimulus omissions is not a response

specific to the subgroup of those eight subjects. The sex

effect may not have become significant because of the small

number of subjects.

In conclusion, the response to the test trials consists

of a deceleration to stimulus absence for the first omission

only, regardless of whether that omission is one of onset or

offset. However, there are some important qualifications

evident. In the case of the initial test trials, the









44

decelerative response occurs for only a subgroup of the

subjects and for the final test trials, this deceleration

starts after a 3 second delay. The linear accelerative

response that became significant on trial 26 only if two

grouping factors were included, seems to be in part a late

recovery of the decelerative response on trial 25 (see

Figure 8). Only a subgroup, those subjects who had started

the session with 1600 Hz, show a heart rate deceleration

to trial 26. Thus, while there is a response to omission of

a pulsing stimulus, it is not as clear as it was with

simpler stimuli (Davies & Berg, 1983).

Dishabituation

Response on trials 19, the stimulus onset trial

preceding the change in frequency, and 21, where the

frequency changed, were analyzed for those 8 subjects who

remained in a satisfactory state throughout the experiment.

Aside from a big initial deceleration, there is little

evidence of a coherent heart rate curve on the final

habituation trial (trial 19), consistent with the

expectation of a habituated response, whereas the new

stimulus frequency (trial 21) appears to produce an extended

deceleration following a brief initial acceleration (see

Figure 10). In an analysis of trials 19 and 21 including

sex, frequency, trials, and seconds a significant trials by

quadratic seconds by frequency interaction, F(1, 4) = 19.74,

p < .025, was found, indicating that the occurrence of









45

dishabituation depends on frequency. Further analyses of

the dishabituation trials are only summarized here and a

detailed account is given in Appendix E.

Only those subjects who changed from 1600 to 1000 Hz

showed statistical evidence of dishabituation. The

frequency specificity of the dishabituation effect was found

with eight subjects only, with all subjects who were in a

satisfactory state on these trials, and in analyses

including two stimulus onset trials from before as well as

after the stimulus change (trials 17, 19, 21, and 23). An

analysis of the stimulus onset trial preceding the frequency

change (trial 19) indicated that the response to regular

stimulus trials had habituated. The response to the

frequency change trial (trial 21) again was frequency

specific for eight as well as for all available subjects.

However, only for those eight subjects who remained in a

satisfactory state throughout the entire experiment,

significant results were obtained on the dishabituation

trial (trial 21), when the responses to both frequencies

were analyzed separately. For both frequencies, the heart

rate change is deceleratory. The response of those subjects

who switched from 1600 to 1000 Hz strongly resembles a

typical orienting response: after a brief delay there is a

deceleration followed by a return to baseline. The response

to a change from 1000 to 1600 Hz, on the other hand, is a









46

more irregular deceleration which has not started to recover

15.5 seconds after stimulus onset (see Figure 11).

Anticipation

Since there is no positive evidence in the literature

on the development of an anticipatory heart rate response in

one to two-month-olds, decisions on what trials to analyze

had to be based on logical inferences and on the appearance

of the heart rate curves. Based on inspection of the data,

there is generally very little evidence for anticipatory

responding, though the last two prestimulus onset intervals

before a frequency change (trials 19 and 21) appear most

promising for a significant decelerative heart rate response

(see Figure 12). These trials might have the greatest

expectation of anticipatory responses since they are onset

trials, where responding was previously greatest, and since

these come at the end of the most extended series of

uninterrupted regular stimulus trials.

An analysis of those eight subjects who remained in a

satisfactory state throughout the experiment with sex,

frequency, trials, and seconds included as factors yielded a

cubic seconds effect, F(l, 4) = 18.84, p < .025. Since

there was no interaction with sex or frequency, the analysis

was repeated under omission of these factors, resulting in a

quadratic seconds effect, F(l, 7) = 7.28, p < .05.

The next step was to assess the reliability of this

anticipatory response by repeating the analysis with all 12






















S0



\ 0







N /
('-












N 0
N N












0


/in
^-1-







\ I)







t cV 0

(wdq) 39NVHO 8H






Figure 1l. Response of eight subjects to the first
dishabituation trial (trial 21). Four subjects who
switched from 1000 to 1600 Hz (solid line) and four
subjects who switched from 1600 Hz to 1000 Hz (broken
line).



















- Trials 19,21
- Trials 23,25


-6.5 -4.5 -2.5 -.5
SECONDS
















Figure 12. Anticipatory response of eight subjects to
combined stimulus onset trials before the frequency change
(combined trials 19 and 21, solid line) and after the
frequency change (combined trials 23 and 25, broken line).









49

subjects who were in a satisfactory state on these

particular trials. One of the 12 subjects had briefly

become fussy during previous trials and three changed to an

unsatisfactory state after these trials but before the end

of the experiment. The analysis, including sex, frequency,

trials, and seconds as factors yielded no effect for seconds

per se, but only some interaction effects: a trials by

linear seconds interaction, F(l, 9) = 7.78, p < .025, and a

trials by linear seconds by sex by frequency interaction,

F(1, 9) = 7.07, p < .05. Since these results were very

different from the results obtained with only eight subjects

and since they were also rather complex, they were not

followed up on. Nevertheless, they indicated that the

anticipatory response is rather weak and unreliable.

The Anticipatory Response after Stimulus Change

In order to test whether the anticipatory response is

disturbed through the tone change, as figure 12 suggests,

an analysis was done comparing the anticipatory response to

trials 19 and 21 (block 1) with the anticipatory response

to trials 23 and 25 (block 2). Sex, frequency, blocks,

trials, and seconds were factors in the analysis with eight

subjects. A significant blocks by cubic seconds

interaction, F(l, 4) = 10.84, D < .05 indicated that the

anticipatory response is indeed influenced by the change in

frequency. An analysis of trials 23 and 25 with sex,

frequency, trials, and seconds as factors yielded a









50

significant cubic seconds effect only in interaction with

sex, F(l, 4) = 12.67, p < .025. When trials 23 and 25 were

examined for both sexes separately with frequency, trials,

and seconds as factors, no significant seconds effect

emerged. Thus, the anticipatory response seems to be

disturbed by the change in tone frequency. The analysis

comparing the anticipatory response before and after

stimulus change was not repeated with all available

subjects, since they did not show a significant response on

trials 19 and 21.

In conclusion, the anticipatory response seems to be

weak under the present experimental conditions, and what

little there is of an anticipatory response is disturbed by

the dishabituation trials. This makes the significant

time-locked response on trial 25 to stimulus omission even

more impressive. It confirms previous research that at this

early age a time-locked response is easier to elicit than an

anticipatory response (Haith, Hazan, & Goodman, 1984, Davies

& Berg, 1983).














CHAPTER IV

DISCUSSION

To summarize the results, analyses indicated that there

was a large orienting response to the onset of the regular

pulse trains but no consistent response to their offset.

This is different from the outcome for continuous stimuli,

where there is a small orienting response to stimulus onset

as well as to stimulus offset. The response to stimulus

omission was less clear than it had been with continuous

tones in Davies and Berg (1983). With the exception of a

subgroup on the very last stimulus omission trial,

significant responses were found only for the first of each

pair of stimulus omission trials; for the initial pair this

was true only for the female subjects. Thus, for response

to regular trials the critical factor determining responding

is whether it is stimulus onset or stimulus offset, whereas

for the omission response it seems more important whether it

is a first or a second omission. With regard to

dishabituation, even though the changed stimulus elicited a

significant response for both frequencies, this response was

significantly larger than that occurring prior to the change

only for those subjects who had switched from 1600 to 1000







52

Hz. Anticipatory responding was weak and it appeared to be

disrupted by the change in frequency.

Analyses of the regular stimulus onset trials (trials

1, 3 and 5) showed a deceleratory quadratic orienting

response which was larger than would be expected for this

age group based on research with continuous stimuli. A

comparison with data from a similar previous experiment

(Davies & Berg, 1983) yielded results consistent with Graham

et al. (in press), namely that in young infants pulsed

stimuli elicit larger orienting than continuous stimuli.

Graham et al. hypothesized that the reduced response to the

continuous stimulus, with its single onset and offset, may

relate to immature processing of transients by young

infants. The transient system is hypothesized to initiate

and/or facitate the attentional processing necessary for an

orienting response to occur. Since pulsed stimuli contain

a series of onsets, the infant's deficient transient system

becomes activated over and over and therefore an orienting

response is facilitated.

In contrast to the large onset response, there was no

reliable response to stimulus offset to the regular 20

second stimulus trials or even after the 60 second pulse

train (trial 8). The question arises as to why a stimulus

that elicits a comparatively large orienting response and

thus by implication elicits a large degree of attention does

not attract interest at its offset. Based on Sokolov's









53

(1960) concepts, an orienting response is expected as a

result of change of any sort and therefore should occur to

stimulus offset as well as to stimulus onset. Work by Berg

and Chan (1972) with adults supported the notion that with

continuous stimuli at least, offsets elicit responses

equivalent to those at onset.4 For a similar previous

experiment (Davies & Berg, 1983) using continuous tones

responses to onset and offset were statistically the same

for one to two month olds, and for seven-month-olds offset

responses were only slightly smaller than onset responses.

For pulsed tones, however, a statistically unequivocal

orienting response has been reported only by Leavitt et

al. (1976), while there have been several reports of either

no response (e g. Clifton & Meyers, 1969) or only a weak

response (Berg, 1974).

There are several possibilities to explain the reduced

chance of offset responding for pulsed stimuli. The

prediction of Bohlin et al. (1981) that an offset response

would occur if the pulse train was only long enough was not

supported by the present experiment. According to the

present experiment, the lack of an offset response is not


4 There have been instances in the literature
(e. g. Bohlin et al., 1981) when offset responses were
smaller than onset responses or even nonexistent. These
have occurred when conditions preceding offset were less
favorable than those preceding onset. Few studies have
tested offset and onset responses under truly comparable
conditions.









54

due to the fact that pulsed stimuli provide less total

stimulation than continuous stimuli, since there was still

no offset response after a 60 second pulse train. With the

50 percent duty cycle employed, this stimulus included a

total of 30 seconds of tone stimulation, more than

sufficient for an offset response to a continuous tone.

Another possible hypothesis is that differential

habituation may take place for onset and offset responses.

During a pulse train, the subject is exposed to a stimulus

offset after every pulse and therefore may have habituated

to offsets before the train stopped. Based on this

hypothesis, the longer the pulse train, the smaller would be

the probability of an offset response. At pulse train

onset, on the other hand, the subject hears the first onset

after 20 seconds of no stimulus. This view implies that a

pulse train is not (or only in a very limited way) perceived

as a gestalt. Developmental research indicates that in fact

gestalt perception is deficient in young infants

(Bertenthal, Campos, & Haith, 1980). An implication of this

view is that offset responses to pulsed stimulation might be

easier to elicit in older infants or adults, when gestalt

perception has matured.

Finally, there is the possibility that the lack of

offset response may be related to the pulse rate. Pulse

rates that closely match the frequency of rhythms that are

exhibited naturally by the infant may be processed












differently from other stimuli. In Berg's (1972) study

where pulses came on for 400 msecs and stayed off for 600

msecs and thus produced two slightly different time

intervals, none of which was a simple divisor of one second,

the offset response was not statistically different for

pulsed and continuous tones within the same analysis.

However, there was not a substantial overall response to the

offset of the 10 second stimuli employed.

The question remains why Leavitt et al. (1976) found an

offset response with their 60 second train of speech

stimuli. There is the possibility that the decelerative

response constituted a recovery from the extended

accelerative response to the speech stimuli. On the other

hand, it could be that due to the high complexity level of

speech stimuli the information difference between stimulus

train and no stimulus train was heightened. Such an

interpretation would be consistent with Sokolov (1960) since

a larger change in information should heighten the chance

for a significant orienting response.

The response to stimulus omission was less clear than

it had been in a previous experiment using continuous

stimuli (Davies & Berg, 1983). The predominance of a

response to the first omission is similar to Clifton's

(1974) results, where there was a large decelerative

response only to the initial stimulus omission. In

contrast, Davies and Berg (1983) found a decelerative









56

response to both consecutive omission trials. Not only does

the time-locked response (the response that occurs at the

omission of the event instead of in its anticipation) seem

less robust with pulsed than with continuous tones, but the

response curve itself also appears more irregular. The

explanation for this difference may lie in the

attention-getting power of the stimuli employed. Clifton

(1974) as well as the present study used stimuli that

strongly attract attention and likely demand more

information processing than the continuous tones used by

Davies and Berg (1983): glucose presentations and pulsed

pure tones, respectively. This is consistent with Thomas

and Weaver's (1975) theory that with more complex stimuli

timing becomes worse since attention is drawn away from the

timing mechanism by the complex stimulus. Their theory

assumes that attention is shared between an f processor and

a g processor. The f processor is the "timer" counting

subjective pulses and the g processor is an (in their

experiment visual) information processor. As the g

processor captures more attention, the output of the f

processor becomes less reliable, resulting in higher

variability of the subject's timing performance. Thomas and

Weaver formulated and tested their theory only with brief

visual stimuli (in the 1 second range), so this study

extends their finding to a larger set of stimuli. The

findings of the present study are also in agreement with









57

Brackbill et al. (1967), where infants rapidly acquired

simple temporal intervals but were impossible to condition

to a more complex temporal sequence. The Thomas and Weaver

hypothesis could also explain why the precise timing in

infants found by Brackbill et al. (1967) and Davies and Berg

(1983) is not more apparent in everyday life where

situations usually are rather complex.

In the present experiment, the females exhibited better

timing than the males, because only the females showed a

time-locked heart rate deceleration to the first stimulus

omission. They also showed evidence of habituation over the

first three stimulus onset trials. The orienting response

to the first stimulus was the same for both sexes, while for

the two following stimulus onset trials, the response became

significantly smaller for the females, which may indicate a

faster learning of the temporal interval as compared to the

males. With auditory stimuli sex effects on the orienting

response per se are unusual, while sex effects for timing

have been reported previously for anticipatory heart rate

responding (Stamps & Porges, 1975), though this was not

unequivocally replicated in a later study (Stamps, 1977).

In summary, there is weak evidence that females

may be somewhat better at timing than males: there is an

unreplicated advantage of female newborns at anticipatory

responding (Stamps & Porges, 1975), and an advantage of

female two-month-olds for the time-locked but not the









58

anticipatory response (present experiment) that was not

replicated in the latter part of the study. The combined

evidence may point to a somewhat better general timing

ability in females than in males; however, to date evidence

is too sparse to make conclusions on the parameters

necessary to reliably demonstrate the effect.

Only those subjects who switched from 1600 to 1000 Hz,

showed clear statistical evidence of dishabituation. The

differential dishabituation for the two frequency groups was

unexpected, since the two frequencies employed span a rather

narrow range. The frequency specific response could be due

to a peculiarity of those subjects tested at that frequency,

such as that they remained somewhat more alert in the latter

part of the experiment.

Another possiblility is that the frequency effect is

real. In a study with newborns by Brown, LeVita, Kush, and

Rothstein (cited by Graham et al., in press) using frequency

modulated tones centered at 1200 Hz and 2000 Hz the

significant response to frequency change was mainly the

result of the shift from the high to the low frequencies.

Along the same lines, lower frequencies have been shown to

be preferred by young infants (Berg & Berg, 1979).

A search of the literature revealed only one other

study that found heart rate dishabituation to auditory

stimuli with infants comparable in age to those in the

present study: Leavitt et al. (1976) used with 6-week-olds










59

a change of tone frequency from 1100 to 1900 Hz in a

no-delay paradigm; i.e., there were no intertrial intervals

between the pulse trains in order to minimize the memory

demands and thereby maximize the possibility of heart rate

dishabituation to a change in tone frequency. The

dishabituation effect of Leavitt et al. was not frequency

specific. The reason for this may have been that the

infants of Leavitt et al. had an easier task than the

infants in the present experiment: they had a larger

frequency change and a no-delay paradigm. The infants in

the present study may have been tested with a frequency

change that was closer to the threshold value for heart rate

dishabituation, since the dishabituation effect over trials

became significant only for the frequency change from high

to low. Thus a frequency preference that may have been

eliminated by a ceiling effect in Leavitt et al. could have

become apparent in the present study.

Statistical evidence for an anticipatory response was

very weak. Only through averaging trials could a

significant change in heart rate be obtained. The shape of

the anticipatory activity that developed, even though

deceleratory, was dissimilar to the shape typical for adult

subjects (see, e.g., Johnson & May, 1969) in that it

reached its lowest point before the onset of the event in

question. This may indicate temporal inaccuracy. The fact

that in Brooks and Berg's (1979) and in Davies and Berg's









60

(1983) experiments anticipatory responses developed in

infants four-month-old and older under rather demanding

temporal conditions while the two-month olds in the present

experiment give only a weak response to a more optimal

condition suggests a strong developmental change early in

infancy.

The weak anticipatory effect in the present study was

also subject to disruption by the change in tone frequency,

a change that was perhaps distracting. Unlike the

dishabituation effect, the disruption of the anticipatory

response was not frequency specific. This disruption is

consistent with the hypothesis that heightened complexity

interferes with timing capability: The change in frequency

makes the cognitive input more complex and therefore

attention may be taken away from the "timer" (Thomas &

Weaver, 1975), resulting in a disruption of the timing

response. Perhaps the integrity of timing responses

could provide a sensitive indication of stimulus

discrimination. Disruption of a timing response following a

shift from one stimulus to another would indicate that the

two stimuli were distinguished. Of course, the exact

parameters necessary to strengthen the anticipatory response

sufficiently to make it statistically unequivocal and thus

useful for stimulus discrimination experiments--possibly a

series of continuous tones without stimulus omission

trials--need to be ascertained through future experiments.









61

The present study confirmed, as previous studies had

shown, that in young infants the anticipatory response is

harder to elicit than the time-locked response (Davies &

Berg, 1983; Haith, Hazan, & Goodman, 1984; Clifton, 1974):

While on the stimulus omission trials ending the series

(trials 25 and 26) the infants were responding to stimulus

omission, the anticipatory response to trials 23 and 25 was

nonsignificant. Statistical significance in this case,

however, is no absolute measure of the prevalence of a

certain response type, since the two responses are

qualitatively different.

In conclusion, the present study found that pulsed

stimuli elicit a large orienting response in one-to-

two-month olds. Possibly, the multiple onsets inherent

in a pulse train facilitate the orienting response or

perhaps a stimulus that matches rhythms naturally occurring

in the infant is easier to process. The absence of a

reliable offset response to pulsed tones cannot be easily

explained by an energy summation concept, since even after a

60 second pulse train there was no significant offset

response. In contrast to the large orienting response that

the pulsed stimuli elicited, the heart rate responses to

stimulus omission were small and of irregular shape. This

outcome is clearly different from what was reported for

continuous tones (Davies & Berg, 1983), where the response

to stimulus omission was not statistically different from









62

the response to the regular stimulus trials. These findings

are in agreement with Thomas and Weaver's (1975) theory that

when attention is shifted to information processing, timing

becomes more variable. Furthermore, the present study

confirmed that in young infants a significant time-locked

response is more likely to occur than a significant

anticipatory response (Haith, Hazan, & Goodman, 1984, Davies

& Berg, 1983). The small and irregular anticipatory

response that occurred in the present experiment was

disrupted by a change in tone frequency. Unlike the

dishabituation effect during stimulus on time that became

significant only for the switch from a higher to a lower

frequency, the disruption of the anticipatory response was

not frequency specific.











APPENDIX A


INFORMED CONSENT











The purpose of this experiment is to investigate what sounds infants
can hear and how well they pay attention to the sounds. Your child will
be listening to a series of tones and speech sounds presented at various
intervals. Your child's responses will help us to better understand what
young infants listen to and like to hear.

We will measure heart rate by placing three sensors on your child, two
on his chest and one at the bottom of his rib cage. Respiration will also
be measured by taping to the infant's stomach a devise similar to a rubber
band. The experiment usually will be completed in less than an hour.

Neither the sounds nor the response measures will cause any harm or
discomfort to your child. We will carry out the experiment only after
you have given us your written permission. Even after you sign this
informed consent, you may withdraw your child from the experiment at any
time. We will be happy to discuss your infant's responses with you. All
information gathered here is confidential within legal limits. If you
have any questions at any time during the session, we will be happy to
answer them. You will be paid $5.00 for your participation, whether your
baby completes the session or not.

I have read and I understand the procedure described above. I agree
to allow my child to participate in the procedure,
and I have received a copy of this description.


Parent's name


Date



Date


Witness


Margarete Davies
303-2 Diamond Village
Gainesville, FL 32603


W. Keith Berg
2902 S. W. 1st Way
Cainesville, FL 32601









APPENDIX B


MEDICAL QUESTIONNAIRE








The data which we collect from your baby can be affected by a number

of factors. The more information we have about your baby, the easier

it will be to interpret these data. The following questions are for

our use in better understanding your baby's responses, and no individual

information will be reported. If you do not wish to answer any question,

just leave it blank. Please feel free to ask us to clarify any of the

questions.


Mother's name

Infant's name


Today's date

Mother's Age

Infant's Birthdate


1. Has the baby had any recent colds, ear infections, respiratory

difficulties, or any major health problems? If so, please specify, and

provide the approximate dates and durations.






2. Is the baby now taking any medication? If so, what is it, how often

and in what dosage is it taken, and what problem is it treating?






3. How many brothers and sisters does the baby have, and what are their

ages?


















4. If known, was the baby born early, late, or on timc?



5. If known, what sort of medication did the mother receive during labor

and during birth? If you do not know exactly, please indicate if there

were any pills, injections, or gas received.





6. Were there any special problems during pregnancy, such as illnesses,

infections, or injuries?





7. Were there any special conditions during birth, such as breech birth,

Cesarean section, or use of forceps?



8. Was the Lamaze method used?

9. Where was the baby born?



10. Did the baby have to stay in the hospital for any extra period of

time? If so, why?



11. Do you breast-feed your baby?

12. If so, are you taking any medication? If so, what is it, how often and

in what dosage is it taken, and what problem is it treating?


THANK YOU FOR YOUR HELP!









APPENDIX C

RESPONSE DEVELOPMENT OVER THE BEGINNING
STIMULUS ONSET TRIALS (1, 3, AND 5),
MALES COMPARED TO FEMALES

In order to more accurately describe the females'

change in response over trials and based on the significant

trials by seconds by sex by frequency interaction reported

in the previous section, trials 1, 3, and 5 were analyzed

separately with only sex, frequency, and seconds included as

factors. There was no difference in response between the

groups for the first trial. A significant quadratic seconds

effect, F(1, 16) = 12.74, p < .005, and a significant cubic

seconds effect, F(l, 18) = 6.91, p < .025 indicated the

expected orienting response to the first stimulus onset.

For trial 3, there was again evidence for an orienting

response in form of a quadratic seconds effect, F(l, 16) =

20.00, p < .0005 and a cubic seconds effect, F(l, 16) =

10.83, p < .005. In addition, there was a significant

interaction of quadratic seconds by sex, F(l, 16) = 5.42, p

< .05, suggesting a different response in males and

females. An analysis of trial 5 again yielded evidence of

an orienting response with a significant quadratic seconds

effect, F(l, 16) = 11.94, p < .005. Again, there was an

interaction with sex, this time by cubic seconds, F(l, 16) =

7.79, p < .025. Thus the orienting response to stimulus

onset does not differ for the first trial but merely

develops differently over trials.

66









APPENDIX D

PULSED VERSUS CONTINUOUS STIMULI

There were some minor methodological differences for

the subjects that were tested with continuous stimuli. The

mean age of the subjects was 67 days, compared to 68 days

for the pulsed group. The subjects in the continuous group

were not tested in a sound attenuating chamber but in an

ordinary laboratory room with a background noise level of 35

dB (as compared to 22 dB in the sound attenuated chamber).

While for the pulsed group a 60 second baseline interval

preceded the onset of the first stimulus, the baseline

interval for the continuous group was only about 10 to 20

seconds. Therefore trial number one was not used in any of

the analyses. For the continuous subjects, the mother and

the experimenter stayed with the infant throughout the

experiment, and neither one listened to noise to mask the

tones. The adults were kept out of the direct line of

vision of the infant. The parents were instructed not to

interact with the infant in any way. Fluorescent ceiling

lights and occasional presentation of a quiet toy by the

experimenter if the infant became restless helped the

infants to maintain a quiet alert state. As with the

pulsed group, data from infants in an unsatisfactory state

were not used.

Since for the group that was tested with continuous

stimuli data were available only up to 8.5 poststimulus

67









68

seconds, the same interval had to be used for the group

that was tested with pulsed stimuli. Since for the pulsed

group the response shows hardly any recovery over this time

period (see Figure 5), a linear trend is expected to be

characteristic for this group's orienting response. Because

the 10 subjects that were tested with continuous stimuli

all were tested with 1000 Hz, only those 10 subjects from

the pulsed group were used who had also been tested at 1000

Hz.

Onsets Compared to Offsets,
Trials 3 and 5 versus Trial 2 and 4

In the first analysis, age, sex, and pulse rate were

included as between subjects factors and blocks (onset

versus offset), trials, and seconds were included as as

within subject factors. A significant on/off by linear

seconds by pulse/continuous interaction, F(l, 12) = 8.83, p

< .025 indicated that onsets and offset do not relate to

each other in the same way for pulsed and continuous stimuli

(see Figure 5). This result confirmed the visual impression

from the heart rate curves that for continuous tones the

response to onset and offset is very similar while for

pulsed tones there is a large orienting response to stimulus

onset and no response to speak of for stimulus offset. Thus

a separate analysis of onset and offset trials became

necessary. There was also a significant on/off by quadratic

seconds by sex interaction, F(l, 12) = 5.37, p < .05, which









69

is impossible to interpret before more analyses have been

made.

Stimulus onset trials 3 and 5. For the continuous

group, the first trial was not included in the analyses,

since the response may have been influenced by an

insufficient baseline period. An analysis including age,

sex, pulsing, trials, and seconds as factors yielded a

significant linear seconds by pulsed/continuous interaction,

F(l, 12) = 19.12, p < .001, indicating that the response to

stimulus onset is different for pulsed and continuous

stimuli for this age group (see Figures 2 and 5).

Furthermore, the analysis yielded a main effect for seconds,

F(9, 108) = 2.8, p < .025, which generally signifies a

response to the stimulus, and a linear seconds by sex

interaction, F(l, 12) = 5.73, p < .05, which needed further

analyses to be interpretable.

Stimulus offset trials 2 and 4. An analysis of trials

2 and 4 including age, sex, pulsing, trials, and seconds as

factors was conducted to assess whether the response to

stimulus offset was different for pulsed and continuous

stimuli as well. A significant quadratic seconds by

pulse/continuous interaction, the only significant result in

this analysis, F(l, 12) = 6.29, p < .05, indicated that this

was the case: The orienting response to the offset of

continuous tones stands in contrast to essentially no

response to the offset to a pulse train (see Figure 5).









70

Since pulsed and continuous stimulation produce

different onset as well as different offset responses, a

comparison of onset and offset response of the continuous

group only could be valuable.

Onset versus offset response to continuous stimuli.

Trials 3 and 5 were combined into a block of onset trials

and trials 2 and 4 were combined into a block of offset

trials. Since sex was the only between subject factor that

had yielded significant results in the previous analyses, it

was the only between subjects factor maintained in this one.

Within subjects factors were blocks, trials, and seconds.

There was no significant blocks by seconds interaction with

F < 1 for the blocks by quadratic seconds interaction as

well as for the main interaction of blocks by seconds,

indicating that onset and offset responses do not

significantly differ for continuous stimuli. A significant

quadratic seconds effect, F(1, 8) = 15.19, p < .025

signified the expected orienting response (see Figure 5).

Thus, for continuous tones, the response to stimulus onset

is the same as the response to stimulus offset, namely a

heart rate deceleration of quadratic order, the typical

orienting response.

To follow up on the sex effect found for stimulus onset

trials 3 and 5 for the pulsed and continuous group combined,

trials 3 and 5 were analyzed for the continuous group only

with sex, trials, and seconds as factors. No interaction of








71

sex with trials or seconds was found. Thus, the sex effect

seems to be limited to the pulsed group. Therefore, the

analysis comparing stimulus onsets to stimulus offsets for

the continuous group only was repeated under omission of

sex as a factor in order to maximize the degrees of freedom

and thus the chance to find a blocks by seconds

interaction. Again, no blocks by seconds interaction was

found, reaffirming the finding that for continuous tones the

response is the same for stimulus onset and stimulus offset.









APPENDIX E

DISHABITUATION

A significant trials by quadratic seconds by frequency

interaction had been found for the trial of frequency change

(trial 21) and the previous stimulus onset trial (trial

19).Therefore, trials 19 and 21 were analyzed for each

frequency separately with only trials and seconds as

factors. Only those subjects who had changed from 1600 Hz

to 1000 Hz showed evidence of dishabituation in form of a

significant trials by quadratic seconds interaction, F(l, 3)

= 19.29, p < .025. The response of these subjects to trial

21 strongly resembles an orienting response, whereas the

response of those subjects who switched from 1600 to 1000 Hz

is more irregular even though it is still deceleratory, as

would be expected for an orienting response (see Figure

11). Since the first analysis had yielded an interaction

effect with trials, single trial analyses were done to

better describe the change in response. An analysis of

trial 19 with sex, frequency, and seconds as factors yielded

no significant results. This can be interpreted as

habituation of the heart rate response to the pulsed

stimuli, as had already been shown with the analysis on

trials 9, 11, and 13. The analysis of trial 21 with sex,

frequency, and seconds as factors showed a linear seconds

effect, F(1, 4) = 16.16, p < .025, a linear seconds by

frequency interaction, F(l, 4) = 18.32, p < .025, and a

72









73

cubic seconds effect, F(l, 4) = 7.92, p < .05. Thus, while

there was no significant response to trial 19, there was a

significant response to trial 21, which differed somewhat

depending on frequency. To follow up on the seconds by

frequency interaction, trial 21 was analyzed for each

frequency separately (see Figure 11) with only seconds

included as a within subjects factor. Those subjects who

had switched from 1000 Hz to 1600 Hz showed a linear seconds

effect, F(1, 3) = 15.94, whereas for those subjects who had

switched from 1600 to 1000 Hz there was a significant effect

for overall seconds, F(16, 48) = 4.66, p < .025. Thus,

there is a significant response to trial 21 with both

frequencies, even though the response shape differs

somewhat. While subjects who started with 1000 Hz show a

rather gradual and irregular deceleration that does not even

start to recover over the 15.5 second interval, subjects who

started with 1600 Hz show a response curve that except for a

brief initial acceleration closely resembles a typical

orienting response: a deceleration and a subsequent recovery

(see Figure 11).

In summary, for the eight subjects analyzed, the

dishabituation effect is frequency dependent: in a

comparison over trials it becomes significant only for those

subjects who changed from 1600 Hz to 1000 Hz. While there

is no significant response to trial 19, there is a

significant response to trial 21 for both frequencies, even









74

though the shape of the response differs somewhat depending

on frequency.

In order to confirm these results, the analyses

were repeated with all 11 subjects who were in a

satisfactory state on trials 19 and 21. Each of the

subjects was in a satisfactory state up to and including

trial 21, but three had a state change before the end of the

session. The analysis of trials 19 and 21 with sex,

frequency, trials, and seconds included as factors yielded a

significant trials by linear seconds by frequency

interaction, F(1, 7) = 6.21, a < .05, a significant trials

by quadratic seconds by frequency interaction, F(l, 7) =

8.27, p < .025, and a significant trials by quadratic

seconds by sex by frequency interaction, F(l, 7) = 6.85, p <

.05. Thus, it was reaffirmed that the phenomenon of

dishabituation was frequency dependent. However, when

trials 19 and 21 were compared for each frequency

separately, there were no significant results. Next, trials

19 and 21 were analyzed separately with sex, frequency, and

seconds included as factors. Similar to the responses of

the eight subjects only, there was no significant response

on trial 19 and a significant interaction of overall seconds

by frequency, F(16, 112) = 3.08, p < .05, on trial 21.

However, when trial 21 was examined for each frequency

separately, no significant results for either subgroup

emerged. Thus, even though the results with all subjects









75

included pointed in the same direction as results with only

eight subjects, they were not as clear cut. The reason for

this may be that the additional subjects were briefly before

a state change, which may have made heart rate responses

more variable.

In order to gain additional evidence on the change

in response and its frequency specificity, the two trials

preceding the stimulus change, trials 17 and 19, and the

two trials following the change, trials 21 and 23, were

combined into two trial blocks. An analysis including

sex, frequency, blocks, trials, and seconds as factors

yielded a blocks by cubic seconds by frequency interaction,

F(l, 4) = 20.65, p <.025, when only eight subjects were

examined, and an effect for blocks by overall seconds by

frequency, F(16, 96) = 2.44, p < .025, when all 11 subjects

were included in the analysis. Thus it was confirmed with a

larger number of trials, that the occurrence of

dishabituation in this experiment depends on frequency.

In conclusion, the results of both subgroups

indicated that dishabituation is frequency specific. In

addition, the data of subjects who remained in a

satisfactory state throughout the experiment showed that the

switch from 1600 Hz to 1000 Hz is more likely to lead to

dishabituation than the reverse.




























APPENDIX F

TABLES












































( X C
0 0




0 CO
>,


0




c W


C c 0) 0
S0 M C u c

0 0 U U 0 -
S 0 a)




44 O tU
S* U m C


o o O c C
UU 0- f 0
1 m 6' N N 1







o o 0 0 0 0)



c c ID .1 6'
q. lrI P4 V. U* -


x X
0
OJ




a V
C C .4

U U 0
eu



0 0 C
s E
m M
a 0
-o -




0 0 .0
u u

S0 C
No o

> >. -
0 0 .4



U 0 0
0 0 0 C

0 -

N N






0 0 '-


o N m-
o o 0
o a a





CD
I0 I) U>
tfl(N
1-
u m f


E
0
0


) I
0

.0 I


o a0

SU I
0 0
0 0 I


1




T I





0 0



O W0






N N I
o I








0 0
* I I





- I




1l*






* 0

0 0 O
-4 C





x U

mo0


0 0





0)

0 U
N 01
- .0
- 3










-4


4u 0
U u
a 0

.0 U
0 0





U .4
O 0
v 0
0 U
U 0
0 U



u

o t

0 4

.0 0
x -
U -


0 V
- C
0c 0
- u















0
- 0)
10 44
O U
0 0
E *n

.0
0 a
.1 0





in

r(










78

































>1 I t c C
4 I I C



4 I I
n 1 I mI
SI I W II.



I I V I m 0 0

0 1 0 1 I 1 0
u I I u I O I 1
4 I I I I V *
4 I I I 0 I C >. c
I I I .0 I 4 0
I I I a o u
S 1 0 1 0 0 0 0
n I I a I 41 .4

Og >W >* W '





SI V I I I V I 0 0
C C I IB C I V V U 41 41 41 C 4



0 0 I 0 V I C C 41 V 0 >




1 n o > >
0 0 0 0 -1 U U U 0 0 0 O
a4 C t4 C- I ki 4i m c 41 4'




m- I InI w o m m
0 4 > I I > 0 I- 3 3 0 0 0 > -
41 4 0 0 U 0 i-0 0' U U U 0. 0 .0


1 6 U I I W
1 0 4 'C uI 4 I I I N I c









w 0 0 v 0 0 0 0 N N 0 N 0 0





m I I


0* w* I x






SC 0 0 0 0 0





4 a I I I




C A1 44 U 0 0 0 I 0 U 4 0


U I I I 3I






*r-i I3 01 u I 4
0 0 m I ) I I 1o

4um I- IT I* Tu
4 I I I* 41 .0 V


(4 444 I 44 0 0









79




















w m

0 o c c u
o u I a
I m I m I
I 0 I C V V 0 I

wc I I 0 0 U)
a U 3 ET (a U U a a
I .0 I IL 0
II cl c Q uQ
a X aI a. -aa 9

a 1 Q CC 0
a a U a a a0 a > 0

SI > > > > c
at 1 11 I .0 0 a at


C C Ut ( I a... C ma C C I .' U -
0 0 .a V I O .-a O I U) a a a a
aO U 0 0 U 0I I U
S t1 .-a U >. I U) a 3. A a ) U >I > > >
m m IL x I to I x W W
3 I c I c I u c

0 ~ > > -< 3 a > 0) > a U a-4 .-. .. U i
a 0 0 -t U a o 0 a o E a r a' .0 .0 a .
0 a
a a a
) N I N 0 >N = N I
M D O I M c c I W O C


00 O I 10I
(U aI a a Y -




S0 0 0 0 0 0 0 n0 0

to .l -1 w C4 as ..




> toCI I a a N
t a a





C W 0
V a + I +

a1 c a a 4 a
a 0 S I

I U I I I






3 a 'a.. t .a 0 Ut I .0 4. Ut
C a t a






) U a. ILI C a
_; 3 ttt 4. I I 0)
a a m I a a
m a





oU u IT



I 1aa 0 0 0
a a a f




I
rI
ra na








































































0

o











V
VI
C
VI
VI












VI


VII a I -
I I
.0 LI LI
u VI 01 VI

VI u. 41 V
3n~ a o
B u j


0)

u ki
V) 0

.0 LI
C o




VI


VI 4
b ~4

a
0 VI


K






0
C
C


0 0
o 0


LI 4



m,


n o


c c
0 0
u u
0 0I
U w
o o
o 3


0 0
U U

>. >.
O C

m m



u u


U u

0 0
U U


to

m




0

01
Vi












t


u u
SL
aV
0 -

0 4
u 3


4 x
S-

1 1

a a



0 .0
u 01
W I

k VI



a a
0 0
4o



i 0


a 0





- M

o a
r-l 6























ne 1-
r- ct


, Ch 0







4 -4 o
o o




M ID 0


N ON -4


N 0 N N
O 0 0 0




Un












0
mD DOI
t"


O, OB LN
so


m

-n
+

o
u

m






* c

V 0
- u
u 0
- m



u

01

01

01 u






U
VI VI
VI
LI
VI
.0
U,


X -I
u ct
0 -
O WI
a -u


m
3
0
3
C
- I

C C
0 0
o 4
m
U,

fN



(Y


c~~rcr ~I
clrl~ ~


~clcl c~





































I NJ
in
*o >.
te .0
I C .0
t O
0 W
I U o
I o) C
0
I 0 U

*u n




1
u u
I -N


I NJ

I $4 U
0" U

m >
'0 -
S I P



I 0 0
I u
I .0 0
C C
S 4 $4 N $4
u I .4 t


0
Lj
U


N)
J3


NJ

0
7, >
u
8 P


U .
C


SN
0 N


A
N -
$ 4
U. U

.0 -2


OJ u

. a
.0 0'


m
C
0
U





01






N)
S 4






0 U





0 0















0




0


>1


o




u a




,0
c
1 >
0 0
a x
u N



0 0
u u

NJ NJ







a N NJ-


4')


E
0

o
0
























a 0
a c

0 E
U -
HU x
N N)


w M
1..4 .4 I





*W *V





O 0
m *
N N I




m ..4
o a~
* r I
o 4 I

N N I


B N O

0 -


w4 11 u 0
u II m W



N U


0 1
41 I 0)
4 N *
Ua N N
unJ N NJ 4.


N)
Ur
NJ

.0r
N-
.. N


>1
U
C

0'
S X) NJ
0' NJ $
NJ N) '4


41
A
.0
0 0
0 N)










82























a I


0

I I I II u
i I I C I
N 41 I 6i oI
33 I V0 I Iv
SC I p I
0 I 1 0 I 0 > I 0
SI u Io U 1 u I 0
1 1
4 I N I 4 I I


S> I > 0 > > 1 I > > I
N I O I Io I -
C I C C I C C C I 1 I C C I C I 4
S0 I 0 0 0 > 0 0 I 0 I >
SU U 0 U 0
It 0) I 41 Nf 4 a4
S U 1 I I U
N I I 4 I C I ..4 I C

U I I4 4 I 4 4 4 I 4( 4 I 4 I 4
U M I U U U I U U I U I 4
3 I 41 41 I 4 01 I -M 4 & 41 Ia Q 41
O > > > I> > > I M U > > I >
4 o I 0 0 I O 0 0 I 4 I 0 0 I o I 4
Mo MM
I 1 I W W I, I I N
4 O 4 I N N I N N N I I N I I -

m 411lI 1 I I I 1 I w I 0 w0 Ic
SI I I I I I I
0 M I I- I I I I I -


IuI
V 0 I I 0 I

I I I I I I I
N I O I O O O I O I I I
I I I I I I
N 0 I 01 I 01 W I I I N
N I 0 0 0I I I I O I
S0 0 0 0 0 N I0 N I 0

N I 0 0 I 0 0*I 0 I 0 I 0 I


II I


I I I I I I I I I


C 0 0 1 0 0 I0 I N N
0) C U 1 1 O y




C 0 0t O I
N n *0 10 I N 1 I N I N W I






4n I I I I
Sf I I I *
0 Nr ( I0
't N I N I NI I I I NI I






II I I I I) = I I I
o m 4 N0 I ( I j I Ia I 0



am) 0 2 cI I o I Ic L


N I I I I I
N I I I I I I
SO 01 (I I g

S 41 0 N C I C I C I C I
A NU I U B U I U I U I U I U U

3 O ) 41 N N I 41 I 41 oI 4 1 I 41 I 41 a
N I f I I I0



41 N I I U I U U I

) 3 1 N 01 I 01 It I *









I U N 4 I< t> It II (s rN N I 4
A2A











83


















54 54 >4
~ e >, x,
0 I I0 I > 0
44E I .01 I 0 4 N


r 0 x r 00
44 GD I GD I GD .0 G
M ~ B a0 au o
44I '0 V 3 4

4' I U I G U I> .0
>4 GD 0 G I >. U P, 0
0 .0 0D u c I u 0
>4 >4 11 G >4 >4 GD 'O
.0 d U .0 w .0 0 C
N n '0 I U' C O
mm c aP m I O a u x u U
a V 0 0 0 0 u G O
I O U C O U I 0' "
n 0 0 0 c u I D 0 w
N U G >4 o >4 4 G >4 >4 GD ). -. >4
I G a a a m .
44 GD u GD4 4 GD GD J3 ^ >
n U GD '4 U0 0 i. A
4 .0 GD I .0 4< GD I GD GD GD GD
N v > 3 'V 'V '
I U O 0 1 4 C C C0 0WG
N 4I0 0 0 -


u C C C C I B 1 0 u



0, 0 ~ O~-
u o oo u us ooo- B~ O- U~!~!'XO x
0 N U U u 0 U 0' U U 0 G 3 U -
0) N a .0 .0 '4 .0 4. JI .0 .0 I 4' 44 0 O U U .-I 0D .3



GD 00 IV
a on *r_
GD I Bi I



01

(Yn
0 It








a,0
m (D
a GD N Iw 0; II o


0 M c a> -*
GD 0 v4 N 00 0
'V 0f 44 N -4 -4 -4 r4 -4 o4 -4 -4 -4





N0 u 0 u 0 u u 0 u
GD N" 4


n *" *







0 w 0



A u
0 I I I
.1 N C I 4= 4 I N I -4 N a 0 4 -



N 01 .- o o -' I 0
GD N I ~ N (~~~



c 40 4 u 4 40 .






3 Nt 4 0 D" 0 4 4^ 4 U 0 04 U 0






















,Q E iu- U Is M m
44 4 >4 .









N> .4 o 4
G 0 I I 1 1







u TI a I+ I a








roc
0 0 2 2 2
8 U, GD 44 414 4 .1 4
aIw i I ^1 a

"" n0 ^
N1 4. 4 IN4( 0


*^L a (1 a- s ^I IM rf N
N 4 4 I

4 N G 0 U GD 0 1 4 GD 0

4' .0 U 4 0 U4 U 4 0


44u >. 4 >4 ~ 4 4 4 >


GD 4 I U
0 GD 4 I 01 4 0I 4
GD u u a 44 N GD 4 GD 44 G
GD O GD 0~ 44 04 0
(J) 3 I' 4' 44 -, 0' 4) 0


440 1 4I I
44nc 44 4e 4 4




.0 GD 4' N .0N 4N .0 4 .0 4


N 4 ~ 4t (r
4- 4
o 4 *4










84

























II 0
n 3
I C
n 0
u .0


4 0 0
( U U 0



N *4 4

N 1 0 0 4 4
N 0 0 It
N 0 U U 4
IT t I>










Y "OI I
N 0 0 01




C4 V 1- 0


0 4 M 1 m 1
S0 0 1 N 4
u 4 )) 1 14
N 0 0 0 C I






0 I I >

) 0 01















AI m 0
AN It ILI I u












a wI 0







a) rut 1t 0 0
N 0) .4 4






N 0 v 4 0














31
N 4J
NO W O N 4 C I







0 0 0 0 C
n 01 N I O













0 .- 40 0
01 a 3 0

N N O 0 N
0 a e o00
4 O N 0* 00

c a I a 4 -
0N )4 4 4 N




01 n I

4 N 0 N 0 C






a 4 ,
o NO 01 N 0 4







0 U 0 01 u U
0 0 0 44 N 4 U 3
30 3 03









.0 U


































































Vi u



C1 t I

S0 0 w I
1 44 u 4
n u a*


4 W4













0
-4 I



4n I







o
IA
;1,









0










I
1 0
u
44 41


-4 I I

I




rrl
o I
a a I
0 0
0 0

















0




41
in


>1
U
c
41
21
0'


N N
- -



ID N




w( 44
<"* rsi





P" N

o I"
o o





- M




inr
1 40


44




0
rJ






m

u f




Ua 4
0 -





41 U
Of C
m or
41 1

3
0*
o Pr

4T1 41
fl w

0' )4



u
0


0 2
N 41





IA


I
44 I
U I
tJ
u



0 I
N 41 I




I
-
Ii i


41
U


0 3
O 2
N 41


e
0


o*


0 a 4
'O C
41 c -<
04 0 6
C U 41
41 41i 4
u a
po


1 U U
111 C I C

0. I 0

0' 1 4 I 1
41 .4 I 1 .4


> 1
U




0)
2 I
0 I


n M1 I
W
-i p-
4 i


0 I



M4 I




0 40 I
ro
ou o





IS I






* W I
I
- 0
I I


u,



r4.
01
JJ

o

*r I






41




a
0 I





a
I




o
41 4






0) u-



u
0
41








XI I
0 3
41
2 I









GO







M
41 I



U I


.0


41
X




a







0

0
0







4r,

Ys




-4

+

o


In







'o
C
c





0
U


u
0
R
41
U
c




4 0



tO *



41

.0
0 2
N 4




Y 0


c
0

u
0
U
41
41


I













41






0
N






- I
- I

N






0











N I
* Io
B 0
In



-^ I








U W
4
41 V I






-j m




>1
U, I
C

u



41 4

2r I






41 I



44


O
N 01 I
(0 N
m


00 3
PI B




(S































































c C

3 3

M 0 M 0
0 4 0 u


>, I





S 0 I
M
0 U
W I

In nI








0 I

N NI








0 r1



















0
w I
o
In







aa

o












0 'u4


ID in
o e













44%




0 I




I I
in


m 0 0 0m

4 0) U 01




uu
c c 1




o) iu .u


m


01


A
jj


u 0
c^ a


m I
ju

0)


A I
I




o


* N
00


E
o




0 4u


-






O 0
O O-4
O o







o











0
- C


M -


m
C)
(U
-n
0,
A
0% 0
.1 0


C I
>1
C I
Ou








on
0 Iu


W u l
a I















O0
N N0



\D -I
-4 '.






000



I4 O
40 0
0. 0
















O 1
o o







m
o






44 I




-4 I
-4




u 0 I







0c





OJ
0 0
44 01
0) I

0) M





0 I




m

0 I
(A






r-4 C

-4 C

0% -


-~ 0




0
a


o

C


0)
*u








'a
0 0

0.


N


























































































































U I Ur
0 I 0

0 I 0
0 I 0'
41 4 41 I 41 4
0' 4 Sd I 4 S
41 41 '5 I 14IU


u
u

.0
.a

0 0
N 41


41
Ut




41


u,
U,

1-
m

o
4.
0




a
Un







*o

41
0
0






0)










ct
0
1
0
0'
41 4
41 Sd





4-1
U

0)
*r41


41 '0 I 0
o4 0 0 0
O O I 0
-x U U I U
w Q) 41




U U u

cc c 41
0)
tr



0Du 01 O 11

o
0 I 0
41 :' I ) X 4
41 d 4 Sd I 4 S
41 1 S I 1 '


n
41
u

.Q
0 0 w


41
U



01



*n
w 4




*-4


41

U



0
41




-Iv




f)


41
0
*0
41
4)
0 'S


0









r- r-

- -4
OO












N4 N
N W
0 0%
or a







0 0




N ^

N N
N -



U,

Un
i

0


U,






0
m


41




4o





0
u




a






41
a,




.0"
0'





41

U
41 u

a


.0
.4 0
-4 a

















88























It














0 4 0 1 f IS













I t


If l
40 m 4 1



















0 0
4
If










w 0






















I I ILI
41 u If 1 4) 4
0 4 w I u













IIM I I u I
0 n 1j fe Ir Io o
HI G I I W
















NgC 1 0 C
u ai u- N N
Z 4

E-4 II

N 0 0 4 9 0 4 0 0

NI Ny w I I I 4 M 4 X g 1





4 4 3 4 ( I I I





C a o M ( ( o m I 1 W i
0 I I 4 I 4 4 I I
4 4





0 0









& *M & 4 -. 4 41 U Ol
U &U UW ) 0 01 I
0 = I 3




















U m U U U ( U 0 U 0 B B U B


N 4 1 4 4 4 4 N M N 4 fM




0 I E M 0 4 0 4 0 4 0 0
N 4 4 a a
N I 4 4 4,


4 4 4 4 4 4
N 4 4 4C
N 4 4 4



C u 4. I N I N I I I '
-~a 0 M .. K 4 C 4 -
.C ~ 4 .44 4 U I 0 U NI 0 44 0 N 0 4 I





C 4. I C 4 C 4 C 4
0 0 U 4.4 0 4 4 0 4 0







444 44









REFERENCES


Ashton, R. (1976). Aspects of timing in child
development. Child Development, 47, 622-626.

Berg, W. K. (1972). Habituation and dishabituation of
cardiac responses in four-month-old, alert infants.
Journal of Experimental Child Psychology, 14, 92-107.

Berg, W. K. (1974). Cardiac orienting responses of 6- and
16-week-old infants. Journal of Experimental Child
Psychology, 17, 303-312.

Berg, W. K., & Berg, K. M. (1979). Psychophysiological
development in infancy: State, sensory function and
attention. In J. Osofsky (Ed.), Handbook of infant
development, (pp.283-343). New York: Wiley.

Berg, W. K., & Chan, S. (1972, November). Heart rate
responses to equivalent stimulus onsets and offsets.
Paper presented at the meeting of the Society for
Psychophysiological Research, Boston, MASS.

Bertenthal, B., Campos, J., & Haith, M. (1980).
Development visual organization: Perception of
subjective contours. Child Development, 51, 1072-1080.

Bohlin, G., Lindhagen, K., & Hagekull, B. (1981). Cardiac
orienting to pulsed and continuous auditory stimulation:
A developmental study. Psychophysiology, 18, 440-446.

Brackbill, Y., Fitzgerald, H. E., & Lintz, L. M. (1967). A
developmental study of classical conditioning. Monograph
of the Society for Research in Child Development, 32 (8,
Serial No. 116).

Brooks, P. R., & Berg, W. K. (1979). Do 16-week-old infants
anticipate stimulus offsets? Developmental
Psychobiology, 12, 329-334.

Clarkson, M. G., & Berg, W. K. (1983). Cardiac orienting
and vowel discrimination in newborns: Crucial stimulus
parameters. Child Development, 54, 162-171.

Clifton, R. K. (1974). Heart rate conditioning in the
newborn infant. Journal of Experimental Child
Psychology, 18, 9-21.

Clifton, R. K., & Meyers, W. J. (1969). The heart rate
response of four-month-old infants to auditory stimuli.
Journal of Experimental Child Psychology, 7, 122-135.












Conroy, R. T. W. L., & Mills, J. N. (1970). Human circadian
rhythms. London: J. & A. Churchill.

Davies, M. B., & Berg, W. K. (1983, April). Change in
mechanisms infants use to mark time. Paper presented at
the meeting of the Society for Research in Child
Development, Detroit, MICH.

Graham, F. K., Anthony, B. J., & Zeigler, B. (in press).
The orienting response and developmental processes.
In D. Siddle (Ed), The orienting response. Sussex,
England: John Wiley.

Haith, M. M. (1980). Rules that babies look by: the
organization of newborn visual activity. Hillsdale,
NJ: Lawrence Erlbaum.

Haith, M. M., Hazan, C., & Goodman, G. (1984, April).
Infants' expectation and anticipation of future visual
events. Paper presented at the International Conference
on Infant Studies, New York, NY.

Hoffman, H. S., Cohen, M. E., & DeVido, C. J. (1979). A
comparison of classical eyelid conditioning in adults and
infants. Unpublished manuscript.

Huyhn, H., & Feldt, L. S. (1970). Conditions under which
mean square ratios in repeated measurements designs have
exact F-distributions. Journal of the American
Statistical Association, 65, 1582-1589.

Huyhn, H., & Feldt, L. S. (1976). Estimation of the Box
correction for degrees of freedom from sample data in the
randomized block and split plot design. Journal of
Educational Statistics, 1, 69-82.

Johnson, H. J., & May, J. R. (1969). Phasic heart rate
changes in reaction time and time estimation.
Psychophysiology, 6(3), 351-357.

Lavie, P. (1982). Ultradian rhythms in human sleep and
wakefulness. In W. B. Webb (Ed.), Biological rhythms,
sleep, and performance (pp. 239-271). New York: J. Wiley
& Sons Ltd.

Leavitt, L. A., Brown, J. W., Morse, P. A., & Graham, F. K.
(1976). Cardiac orienting and auditory discrimination in
6-week-old infants. Developmental Psychology, 12,
514-523.












Miller, C. L., & Byrne, J. M. (1984). The role of temporal
cues in the development of language and communication.
In L. Feagans, C. Garvey, & R. Golinkoff (Eds.), The
origins and growth of communication (pp. 77-101).
Norwood, NJ: Ablex, 1984.

Palmer, J. D. (1978). An introduction to biological
rhythms. New York: Academic Press.

Rewey, H. H. (1973). Developmental change in infant heart
rate response during sleeping and waking states.
Developmental Psychology, 8, 35-41.

Richter, C. P. (1965). Biological clocks in medicine and
psychiatry. Springfield, IL: Charles C. Thomas.

Robertson, S. S. (1982). Intrinsic temporal patterning in
the spontaneous movement of awake neonates. Child
Development, 53, 1016-1021.

Shaffer, L. H. (1982). Rhythm and timing in skill.
Psychological Review, 89(2), 109-122.

Sokolov, E. N. (1960). Neuronal models and the orienting
reflex. In M. A. B. Brizier (Ed.), The central nervous
system and behavior: III. New York: Macy Foundation.

Stamps, L. E. (1977). Temporal conditioning of heart rate
responses in newborn infants. Developmental Psychology,
13,(6), 624-629.

Stamps, L. E., & Porges, S. W. (1975). Heart rate
conditioning in newborn infants: Relationships among
conditionability, heart rate variability, and sex.
Developmental Psychology, 11(4), 424-431.

Ter Vrugt, D., & Pederson, D. R. (1973). The effects of
vertical rocking frequencies on the arousal level in
two-month-old infants. Child Development, 44, 205-209.

Thomae, H., Erftmann, E., Lehr, U., & Schapitz, L. (1972).
Entwicklungspsychologie. In H. Thomae (Ed.), Handbuch
der Psychologie (Vol. III, pp. 46-324).
Goettingen: Hogrefe.

Thomas, E. A. C., & Weaver, W. B. (1975). Cognitive
processing and time perception. Perception and
Psychophysics, 17, 363-367.