MSU

LIBRARIES
M

5

RETURNING MATERIALS:

 

 

 

Place in book drop to
remove this checkout from
your record. FINES W111
be charged if book is
returned after the date
stamped below.

 

 

 

 

A PSYCHOPHYSIOLOGICAL STUDY or THE

NATURE OF HABITUATION IN INFANCY

BY

Steven R. Gitterman

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

I982

ABSTRACT

A PSYCHOPHYSIOLOGICAL STUDY OF THE

NATURE OF HABITUATION IN INFANCY

By

Steven R. Gitterman

A psychophysiological measure of sustained attention (decrement in
heart rate variability) was used to test two competing models of
habituation. The all-or-none model postulates that habituation occurs
in a single trial following an increase in behavioral response during
the preceding trial. The progressive decrement model describes
habituation as a successive decline in response on each trial.
Four-month-old infant subjects were repeatedly shown a checkerboard
stimulus and then separated into habituation and nonhabituation groups
on the basis of their total fixation times to the stimulus. These two
groups were then compared to see whether the physiological measure of
attention differentiated the two groups over all experimental trials.
An additional comparison of the physiological response on the trials
before and after the criterion for habituation was met was made for the
habituation group. Results for both tests were negative. A further
test of differences on the physiological response between the last

habituation and dishabituation trials was also nonsignificant.

Steven R. Gitterman

Taken together, these findings question the validity of use of a
decrease in heart rate variability as a measure of sustained attention.
However, certain methodological problems in the present study may have
introduced biases against positive findings. These problems are

discussed along with more recent psychOphysiological approaches for

studying attention and cognition.

ACKNOWLEDGEMENTS

I would like formemost to thank my committee chairman, Dr. Hiram E.
Fitzgerald, without whom this dissertation would not have been
completed. His support and encouragement has enabled me to complete
this work and face an exciting future using the skills I have acquired
in graduate school. For this I am truly grateful. I also wish to thank
the other members of my committee, Drs. Mark Rilling. Lawrence Messe',
William Crano, and especially Lauren Harris, for whom I have a great
personal fondness as well as professional respect. Many other students
and faculty in the Department of Psychology have also been exceptionally
helpful, though none more so than Dr. Jack Condon. I can only say that
Jack's help has been indispensible. and his friendship a source of much
enjoyment and stimulation. Brain Coyle and Esther Dienstag also have
been extremely helpful as consultants; I am glad to have known each of
them as friends in graduate school. To everyone else who has helped me

in my stay at Michigan State, thanks.

TABLE OF CONTENTS

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

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

Chapter

I. INTRODUCTION .. .........................................

ll. LITERATURE REVIEW ......................................
Theories of habituation ......- ..... . ....................
Psychophysiological studies of attention ...... .........
Studies of habituation and orienting in infancy ... .....

III. METHOD ............................... . ....... ..... .....
Subjects ...................... ......... ..... . ..........
Procedure ......... ................... . .......... . ......

Dependent Variables

Heart rate . ...... . ....... .... ....................
Visual fixation ........ ..... .. . ..... . .........
'v. RESULTS 00....................0....0......0.0. . 000000

Results for cardiac measures alone

Analysis of behavioral data with

cardiac measures .......... ...... . ..... . ....... . ......
Results for dishabituation testing . ....................
V. D'SCUSS'ON 00.0........ ..... 0.00.....0. . .0....... ..

APPEND'CES ...............................

35
37

no
bl
#5
#7

AB

59
6h

78

Appendix A - Forms ......

Appendix B - Subject characteristics .........

Appendix C - Analysis of variance tables .......... .....
REFERENCE NOTES .............................................
REFERENCES 0.0.0....000OO ...... 0. .... ..... .... .. ......

78
83
85
102

103

Table

10

11

LIST OF TABLES

Page

Analysis of covariance of heart rate response to
stimulus onset over trials for all 28 subjects. ... ..... 85

Analysis of covariance of variability response to
stimulus onset for all 28 subjects over trials. ........ 86

T-test comparing habituation and nonhabituation
groups on average of two longest fixations during
first five trials. .....0.....0.0...................00.. 87

T-test comparing habituation and nonhabituation
groups on duration of first fixation time. ............. 88

Analysis of variance comparing variability response
for prestimulus and poststimulus seconds on trial I
by habituation group. ...........00.0........00........0 89

Analysis of variance comparing heart rate response
for prestimulus and poststimulus seconds on trial I by
habituation group. .......0...0....0.......0............ 90

Analysis of covariance for heart rate response to
stimulus onset by habituation groups over trials
and over seconds. ................................ ...... 91

Analysis of covariance for heart rate responses on
trial preceding and trial of habituation for
habituation group. ....................................0 92

Analysis of covariance of variability response for
habituation group to stimulus onset over trials
and over peri0d50 ....0.........0.......0.0......000.0.. 93

Analysis of covariance for nonhabituation group
variability response to stimulus onset over trials
and over periOds. .........0.........00..0...0..0..0.0.0 9h

T-test comparing habituators fixation response on
last habituation trial and dishabituation trial. ....... 95

12

I3

1h

15

16

17

T-test comparing nonhabituators fixation response on
last habituation trial and dishabituation trial. .......

Analysis of covariance comparing habituators heart
rate response on last habituation and
dishabituation trials. .................................

Analysis of covariance comparing habituators
variability response on last habituation and
dEShabituation trials. ..............0....... ......... ..

Analysis of covariance comparing nonhabituators heart
rate response on last habituation and
dishabituation trials. ....... .. ...... .... ..............

Analysis of covariance comparing nonhabituator
variability response on last habituation and
dEShabituation trials. ........0................. ..... .0

Analysis of variance for groups over trials with the

variability of the entire 18 second stimulus presentation
as dependent variable. ........I........................

vi

96

97

98

99

100

101

 

LIST OF FIGURES

Figure

10

II

12

Hypothetical habituation curve (Thompson
et a1.. 1973). .........................................

Backward habituation curve (Cohen 8 Gelber, 1975). .....

Heart rate response for all 28 subjects
to stimulus onset. ............ . ...... . ........ ... ......

Variability response for all 28 subjects
to stimulus onset. ......C...................OCCOCOOOOCO

Total fixation response over trials by group. ........ ..

Total fixation time plotted backwards from

criterion. ......O...........0.........................0

Prestimulus period versus poststimulus period
for trial 1. ...............................0.... .......

Heart rate response for habituators. ...................
Heart rate response for nonhabituators. ................
Variability response of habituators over trials. .......
Variability response of nonhabituators over trials. ....

Comparison of final habituation trial
and dishabituation trial. ..................... ...... ...

vii

Page

10

19

“9

50
53

5A

56
57
58
6O
6l

62

CHAPTER 1

INTRODUCTION

The past two decades have seen a resurgence of interest in the study
of both child development and infancy. Despite the introduction of new
experimental methodologies that have been developed for use with verbal
children, experiments with infants are still restricted by many factors.
Especially limiting are difficulties associated with the experimenter's
inability to adequately quantify many aspects of the infant's response
repertoire. and the necessity of maintaining an alert biobehavioral state
throughout the duration of an experimental session. Since relatively
little progress has been made toward solving such methodological problems
of infant research, the basic paradigms of habituation, operant
conditioning, and classical conditioning continue to be prominent methods

for investigating infant behavioral development.

Both habituation and conditioning paradigms are well defined
experimental procedures and have been invaluable in the study of infant
perceptual development (see Bower (l97h) and Cohen 5 Gelber (1975) for
reviews). However, the conclusions of studies in which these procedures
were used have been discussed far more than the methods of study
themselves. For habituation (the focus of this study), the relative lack
of inquiry into the process of habituation no doubt stems from the common
interpretation and acceptance of habituation as a progressive decrement
of response to the repeated presentation of a stimulus. Support for this
belief depends on the analysis and interpretation of group data plotted
over trials of presentations [e.g., see Field, Dempsey. Hatch. Ting, 8
Clifton (l979)]. The progressive decrement in response (when the group
curves are plotted) was thought to represent the gradual attainment by
the infant of a mental "image“ of the stimulus. This idea was explicitly
incorporated by Sokolov (l963) into his theory of the orienting reflex.
The operational definitions of orienting theory by Sokolov have since
been incorporated into much of the research on habituation and cognitive
development and have been quite influential in developmental psychology
(Graham 5 Clifton, l966; Lewis, l97l).

Indeed, the practice of plotting group habituation curves over
trials also brings into question the exact nature of the habituation
process from the first trial onward. If curves are plotted backwards
from criterion [similar to the backward plotting of learning curves by

Zeaman and House (l963)], the curves describe an all-or-none model of

habituation; i.e., the curves appear level until the habituation trial;
then responses decline dramatically. This was first noted by Cohen and
Gelber (1975) as an alternative to the decremental (or negative
exponential) model of habituation. Yet, as Cohen later pointed out
(I976), the apparent rapid decline in response could be a statistical
artifact in the determination of a habituation criterion which forces
randomly short trials to be included in the habituation trials. This
effect acts to make the backwardly plotted data responses to decline more
precipitously when the habituation criterion is reached than is possibly
true. Results of a mathematical simulation by Cohen and Menten (198])
further confound the issue of the true nature of habituation. Their
results indicated that whether or not their simulation used parameters to
describe either an all-or-none habituation or a progressive decrement in
response, the data obtained from such a simulation, when plotted, appear
as an all-or-none habituation (also see McCall, l979).

The question as to which model is the more nearly correct with
respect to the period of infancy is a theoretically significant issue. as
certain learning theories (e.g., Jeffrey, 1968) as well as biologically
oriented theories of habituation (Thompson 8 Spencer, l966) assume the
progressive decrement to be true. Conversely, certain
information-processing models using Harkovian models (Kintsch. l967)
offer some support for viewing learning and habituation as all-or-none
phenomena. A third possibility is that neither model is always correct,
but that either stimulus or subject variables account for the observed

patterns of behavior. This suggests that both patterns can occur, but

that different stimulus properties (or methodological variables, e.g.,
interstimulus intervals or stimulus length) interact to yield the
observed pattern. Similarly, biobehavioral differences between infants
may yield intersubject variability in response pattern.

Psychophysiologic measures (e.g. heart rate, galvanic skin
response. evoked potentials) have been particularly useful in infant
studies, both as dependent measures for an experimental manipulation (see
Berg 5 Berg. l979, for a review) and as measures intrinsically important
by themselves as indices of psychological development (Graham 5 Jackson.
l970). The meaning of these autonomic measures (specifically heart rate)
as a causal influence on behavior in adults (e.g.. Lacey & Lacey, l97h)
currently is a controversial issue (Obrist, l98l). However, the use of
autonomic responses solely as a dependent measure that covaries with
certain psychological processes is well established, especially with
infants (Berg 6 Berg, l979). Heart rate is a very widely used autonomic
measure. and its prevalent use has led to several refinements in the
measurement and analysis of EKG records. Investigators originally
handscored polygraph outputs for simple measures of HR acceleration or
deceleration; more recently the change in heart rate variability (Forges,
l97h) has been proposed as a meaningful indicator of certain attentional
states (described below), as has the interrelationship between
respiration and heart rate assessed by spectral analytic techniques
(Forges, l976). These two examples (like many others) represent attempts
by psychophysiologists to acquire additional insight and meaning by

different analytic approaches to the same physiological responses.

Although the exact nature of habituation remains an unsettled issue. many
investigators have sought to relate the group curves obtained from
habituation procedures to heart rate measures. Most often, this involves
using parallel changes in heart beat responsitivity (acceleration or
deceleration) to infer that habituation has occured (Graham, l973). This
experimental strategy usually has presented with problems in two areas --
assumptions about the nature of habituation itself, and the appropriate
method for analysis of the heart rate response data. Previous studies
have implicitly (and often explicitly) assumed a progressive decrement in
response as the model for habituation as opposed to the all-or-none
phenomenon. An entirely different course of physiological change would
be expected if heart rate does parallel an all-or-none model of
habituation rather than a progressive decrement one. The other problem
in these studies is the use of an appropriate psychophysiologic response
measure. As will be discussed below, certain theoretical interpretations
of heart rate patterns predict different courses of habituation for
different components of an EKG record. Analysis of only
acceleration-deceleration (the most common method of analyzing EKG data)
ignores potentially more interesting aspects of the data.

The present study was designed to analyze the relationship between
habituation and heart rate; specifically (a) multiple components of the
heart rate response and (b) progressive decrement vs. all-or-none models
of habituation. The experimental procedure involved repeated
presentations of visual stimuli to four-month-old infants. A review of

the relevant literature followed by a statement of the specific

hypotheses of this study is presented below.

 

CHAPTER 2

LITERATURE REVIEW

The literature review is divided into three parts. The first part
discusses theories of habituation relevant to the present study. In it,
Sokolov's work is introduced and is further discussed in the second
section describing psychophysiological measures of attention. The last
section discusses studies of habituation in infancy (both with heart rate
and visual fixation measures) and some of the important methodological

issues in this area of research.

Theories 91 Habituation

Popular awareness of habituation no doubt preceded both its
"discovery" and operational definition in experimental psychology.
Sherrington (l906; l96l) and James (1890; 1950) each discussed it early
in this century. although from two different perspectives that somewhat
mirror debate and discussion in theories of habituation today.

Sherrington's view could be characterized as a more inductive approach.

Noting "reflex fatique" in the peripheral musculature of a spinal dog
preparation, Sherrington inferred from the observed decline in leg
flexion-extension reflexes in response to electrical stimulation, a
centrally occuring inhibitory process that similarly served as the
mechanism of habituation in more complex, common stimulus situations.
Conversely, James discussed habituation deductively from the perspective
of cortical inhibition, assuming that higher cortical functions existed
to mediate the behavioral change in response systems. James' ideas were
derived by observing the decline in response by man to repeated, common
complex stimuli. He assumed that cortical mechanisms were the focal
point for ignoring stimuli such as the "din of the machinery in the
mill", and for other changes (e.g., motor inhibition) that we associate
with the colloquial use of the term habituation.

In the early l960's, two theories with equivalent predictions of the
course of habituation, although from different theoretical and empirical
approaches, emerged in experimental psychology. Primarily through work
with the cat spinal flexion response, Thompson and Spencer (l966)
proposed the dual-process theory of habituation. In this conception,
habituation is viewed as two simultaneously occurring processes: response
decrement in the particular system of the animal under investigation and
a general "sensitization" occuring in all systems of the animal.
Dishabituation was defined in the dual-process framework as the increase
in behavioral response to a stimulus different from that used during
habituation training (Thompson 6 Spencer, l966). (To avoid possible

confusion, note the difference between this and the classic Pavlovian

definition of dishabituation as the recovery of the original
unconditioned response to a stimulus after a period during which the
stimulus was not presented). Dishabituation is described as the
behavioral response to the increased sensitization that has occured over
the habituation trials (see Figure l ). The theoretical sensitization
curve is superimposed on the theoretical habituation response such that
the observed behavioral responses over time are the product of the shape
of the two curves at that particular instant; increasing sensitization
therefore acts to mask somewhat the theoretical rate of decline, since
its effect is to raise the probability of responding. In other words, on
the early trials of a habituation paradigm sensitization is high and
increasing more than true habituation response is declining; thus the
observed behavioral response to a stimulus actually increases on the
first few stimulus presentations since the balance of sensitization and
true habituation favors sensitization. As stimulus presentation is
repeated sensitization decreases and true habituation continues. The
observed response also declines in tandem, ultimately to the level of the
true habituation curve since sensitization returns to the prestimulus
level and no longer influences response amplitude. Recent reviews by
Groves 8 Thompson (1973) and Thompson 8 Glanzman (l976) cite evidence
supporting this view of habituation in species ranging from invertebrates
to man to indicate its postulated role as a general mechanism of learning

in all species.

[igggg l. Hypothetical habituation curve (Thompson
et al., l973).

IO

mu

 

mucumh mo wxuoum

cu
_

 

 

 

onpcahumcz wzmp.lll
zouhczhumcz ow>mwwmo .....
zomh¢~uhumzww.lll

 

 

“mum“ ..uc pm zommzozh. w>¢=o zomhcabumcz Acoupwzbom>r

cm

03

QB

25

UOZHKOJ cszHHDDm

N

II

To further describe the anatomical basis for dual-process theory,
Thompson [Thompson 8 Glanzman (I976); Thompson, Rinaldi, Berry 5 Berger
(l978)] has developed a hypothetical neuronal model to explain the
mechanisms mediating reflex decrement in a research preparation, the cat
spinal reflex. The model proposes that sensitization occurs in a
separate pathway that receives afferents from the interneurons in the
habituation pathway, and sends its efferents into a state system that
simultaneously enables the increased response to stimuli different from
the habituation stimulus. The pattern of habituation which follows from
this concept is one of exponential decrement in response (see fijgggg l ).
Its physiological analogue is decrement of response or of activity in the
interneurons of the habituation pathway. This hypothesis has much
support from single cell recordings in a variety of species (see Thompson
et al., l978) and is conceptually reasonable in the context of recent
developments in neuroscience that identify specific loci for pre- and
post- synaptic inhibition (e.g., see Shepard, I979). The relevant aspect
of the dual-process hypothesis for the present study is that it
specifically predicts a progressive (or exponential) course of response
decrement during habituation. This refers to a pattern of habituation
where there is a decline in response on each repeated trial of stimulus
presentation, ultimately reaching an asymptotically low level of
response. There has been much debate of this model (e.g., Hinde, l970).
and little consensus on the proposed universality of the neuronal

mechanisms postulated by Thompson et al.

12

The second prominent model of habituation stems from the work of
Sokolov (I963, I969) and is cited far more often in the child development
literature. A major factor contributing to its popularity has been its
perceived simplicity in describing habituation, particularly of what is
often referred to as the orienting response (0R). Many investigators
have adopted Sokolov's early model of habituation, frequently referred to
as the "match-mismatch“ hypothesis. Simply put, this model postulates
that a neuronal model of a stimulus is built up as the stimulus is
presented repeatedly, with concurrently occuring cortical inhibitory
processes acting to dampen an orienting reaction that would normally
occur in response to the stimulus. When a novel, or dishabituating
stimulus is presented, the "mismatch" to the neuronal model causes the
reoccurence of the orienting response, and a dishabituation of the
response. An orienting response, as described by Sokolov, is a general
activation of the body's physiological state to facilitate the "taking
in" or attentional response to an environmental stimulus (Graham 8
Jackson, l970). This will be discussed more fully below; however, the
concept that orienting responses facilitate learning by mediating
attentional responses and, very importantly, are measurable as part of a
general physiological response is critical to many of the studies of
infant perception and learning (for reviews, see Berg 5 Berg, I979;

Jeffrey, l968; Sameroff, l97l).

13

However. as discussed by Velden (I978), the earlier version of
Sokolov's work no longer adequately reflects further revisions in
orienting theory by Sokolov. The most recent formulation by Sokolov
(1969) is an information theory, or "entropy” model, of orienting, which
stresses the informational value of a stimulus. In a habituation
paradigm, it is proposed that the information value of the stimulus on
repeated presentation diminishes, and thus the 0R habituates.
Consequently, the OR functions as an "information regulator" and is most
applicable to stimuli having some form of signal value; it also accounts
for the continued presence of the orienting response in classical
conditioning paradigms, since the conditional stimulus continues to have
informational value as a predictor of future events. In addition, since
reaction to a stimulus demands its being "taken in", simultaneous
habituation of the behavioral and orienting response to a stimulus must
occur.

Unfortunately, the new form of Sokolov's theory fails to address two
additional aspects of orienting that have been discussed by other
researchers. One is the subject's motivation for the task, which can be
defined as the relevance of the stimulus for the subject (e.g., see
O'Gorman, I973). As Velden points out, there most likely exists a
multiplicative relationship between the signal value of a stimulus and
the elicitation of an OR (i.e., all factors being equal, a subject will
give a greater OR to a relevant stimulus). A second aspect of current

orienting theory neglected by Sokolov is orienting on the basis of

IA

different non-signal physical properties, or stimulus novelty, which
Velden calls "orienting incentive" to distinguish it from "informational
novelty" in Sokolov's theory. The stimulus novelty aspect of orienting
has been the one most widely studied in investigations of infant
orienting, as it follows directly from the earlier match-mismatch
hypothesis. Similarly, in infant habituation paradigms the distinction
between the habituation and dishabituation stimulus most often also
reflects a change in stimulus properties. A recent example which
illustrates the informational perspective is the set of habituation
studies of Schwartz and Day (I979). Their approach was to use the
magnitude of a dishabituation response (or in context of the present
discussion, the return of an orienting reflex) to a set of stimuli
different from the habituation stimuli to index the learning of a
property of the stimuli. The test stimuli varied from the repetitive
stimuli along a number of dimensions; from the test responses could be
discerned the degree to which the infants perceived the relevant
dimension, i.e. the informational value in the new stimuli. (For a
review of older studies, see Cohen 6 Gelber, I975).

From Sokolovian theory, again a decremental model of habituation is
postulated. The informational value of a repetitive stimulus should
decline on each presentation of the stimulus as centrally mediated
inhibitory pathways from the brain decrease behavioral response to the
declining stimulus informational value. In this form, Sokolov's model
can be readily integrated into the information-processing models

currently prominent in cognitive psychology (Ohman, I979). Other

 

IS

theorists, notably Piaget (Ginsburg 8 Opper, 1979), predict a similar
course of habituation due to the buildup of a memory engram.

Other prominent approaches to habituation theory have been from
learning theory (Lubow, Heiner, 8 Schnur I98l) and developmental
psychology (Cohen, 1973; Cohen 8 Gelber, 1975). Both also predict a
decremental pattern of habituation, although Cohen's model is cited by
the authors of the latter paper as also "consistent" with all-or-none
habituation response (see below). The Lubow et al. hypothesis is
particularly interesting for its interpretation of habituation as
”conditioned inattention” of an attentional response (defined as a
unconditioned response (UCR) to the stimulus). Habituation (or
conditioned inattention) is also postulated to be the pivotal factor in
explaining latent inhibition effects. A definitive set of rules derived
from classical conditioning principles is used by the authors to predict
the behavioral effects including habituation under the rubric of
conditioned attention theory. The focus of habituation in conditioned
attention theory is not on response decrement per se but on the effects
of habituation to a stimulus on the later associability of the stimulus
when it is acting as the CS (or UCS) in a traditional classical
conditioning paradigm. The role of habituation as affecting the salience
of the stimulus for later associability parallels Jeffrey's (1976)
discussion of the importance of habituation as a mechanism for perceptual
development in the infant. Lubow et al. also cite evidence that the OR
elicitation (and also habituation of the OR) does not correlate with

later associability of the stimulus as a CS, contrary to certain

 

 

l6

interpretations of OR theory. Jeffrey (1976) makes a similar point in
reference to infant learning.

The Cohen model is also interesting and, although described as a
model of infant visual attention, owes much to Sokolovian and
information-processing theories from which it has borrowed heavily. An
important feature of the model is the postulate of two separate processes
in infant visual attention: an attention getting process and an attention
holding process. The attention getting process is responsive to the
physical and sensory properties of a stimulus; it is hypothesized as the
site of integration where the decision whether an orienting response will
occur to a stimulus is made. The attention holding process is mediated
by components of long and short term memory; it is described as
responding to the familiarity or complexity of an environmental stimulus.
The attention-holding process is viewed as interacting with the former
process to either inhibit or to sensitize the orienting response. Both
systems are thought to be able to habituate separately as well as
interact to yield the observed behavior. Cohen (Cohen 5 Gelber, 1975,
pg. 396) uses this model to account for all-or-none habituation patterns
by postulating that the attention getting process acts within a threshold
of recognition, above which no response occurs. Wagner's (1976) model of
habituation also relies heavily on the distinction between two components
in habituation, but it stresses even more heavily than Cohen the separate
involvement of short and long term memory in each process. Wagner's work
centers around the importance of the "priming" of the stimulus in

short-term memory, and how this stimulus representation can account for

17

two empirically difficult issues in habituation studies: the distinction
between short- and long- term habituation, and the reason for greater
retention of habituation (or less classical (Pavlovian) dishabituation)
in paradigms using longer interstimulus intervals. None of the infant
literature appears to cite the Wagner model, which has been generated
solely on animal experiments, specifically eyeblink reflex studies in
rabbits (although Ohman (1979) uses many of Wagner's ideas in his
information-processing model of orienting). Lubow et a1. (1981) briefly
criticize some of the Wagner model predictions and its inappropriate
generalizability. Jeffrey (I976) similarly criticizes Cohen's model as
being based primarily on adult verbal memory studies using artificial
conditions, and then being inappropriately applied to infancy.

Two additional developmental theorists deserve mention: Lewis (1971)
and McCall (I971: McCall 5 McGhee, 1977). Lewis' theory of the
development of a cognitive schema during habituation differs only
slightly from the original Sokolovian proposal and adds little to
habituation theory. McCall's research has proposed that the attentional
response of infants to varying stimuli is an inverted-U shaped function
of the degree of discrepancy of a test stimulus from a previously exposed
standard (see below). However, this theory does not account for certain
experimental findings (e.g., OeLouche, 1976), and is difficult to test
empirically. (For a critical review of these latter two proposals, see

Cohen 5 Gelber, 1975, pp. 293-295.)

18

The alternative hypothesis to a decremental model of habituation is
an all-or-none model ( Figure ; ), first discussed by Gelber and Cohen
(1975). As Figure g demonstrates, when behavioral data from a
habituation paradigm are plotted backwards, habituation appears to occur
in a single trial following an increase in attention on the preceding
trial. The continuous decline in behavioral response that characterizes
the progressive decrement model is inapparent. In contrast to the
well-developed models of Sokolov and Thompson, this model has had little
theoretical discussion in the literature. Following Cohen and Menten
(1981), it could be predicted from discrepancy theory (e.g., McCall,
1971) that the continued presentation of a stimulus that is sufficiently
discrepant could lead to increases in looking time which serve to
successively reduce the discrepant nature of the stimulus. At the point
of maximal interest, an immediate decline in stimulus interest would
occur, corresponding to an all-or-none decrement in responsitivity. If a
stimulus is unfamiliar, perhaps the infant increases looking time until
the discrepancy function is optimal, and then attention declines
immediately. McCall (1979) offers an essentially similar explanation for

this phenomenon.

ngchophysiological Studies 2: Attention

Early in this century, attempts were made to use non-invasive
physiological processes to index psychological functioning. An important

example of this was the attempt to examine the "fight or flight

Figure ;. Backward habituation curve (Cohen 5 Gelber, 1975).

19

zouhcahumcr roam mummmh

T N- m... T m- m- T o- m- 8.. 3..
_ _ _ P _ _ _ _ _ — _

Nun

 

 

 

 

ambmu

.mwmumo a zwroou m>m=u zouhcnhumcz omczxocm

[LHXOII—HDZ I—HZIJJ HZ (DUUOZDCD

20

responses” and their relationship to sympathetic arousal and subjective
perception (Cannon, 1929). Similarly, James' (l890; 1950) discussion of
emotion (later known as the James-Lange theory) sought to describe
emotion solely as the response to the physiological (often autonomic)
changes. This early work stressed the idea that psychological function
is inseparable from concurrent metabolic processes, a concept implicit in
much of the research done in both psychosomatic medicine and
psychophysiology (Obrist, 1981). Psychophysiologists have long tried to
identify physiological concomitants of personality and intellectual
characteristics (e.g., Orlebeke 5 Feij, I979: Zeiner, 1979). In
psychosomatic medicine research, the established relationship between
Type A typologies and coronary heart disease underscores the complex
interaction of physical and mental states (Blumenthal et al., 1978).
Though there is little question that psychological and physiological
processes do interact, the specification of this interaction and the
degree to which one process affects the other is an area of ongoing
research and debate in psychophysiology. An example of this is the study
of evoked potentials, which are the EEG potentials in response to a
stimulus averaged over many trial presentations. Although a
frequently-used dependent measure in psychophysiological research, there
is still much disagreement on the psychological meaning and
interpretation of the various waveform peaks as mirrors of cognitive

processes (e.g., Walker 5 Sandman, I979).

 

__m____
u '3

21

Cardiac electrophysiology, another frequently used
psychophysiological variable, is far easier to measure and quantify than
the complex waveforms derived from EEG studies that confound analysis:
however, it similarly has generated much theoretical controversy. The
ease of noninvasive measurement and the large body of physiological
knowledge about cardiac function have made measures derived from EKG
records widely used dependent variables. Despite its wide use, there has
been no lull in the controversy over whether HR changes are an
independent measure of certain cognitive processes. The Laceys (I970,
I97A) have long argued for the utility of heart rate as a measure of
”psychological set". Specifically, heart rate deceleration is thought to
accompany the facilitation of stimulus intake (e.g., attention to the
signal in a reaction time task), whereas heart rate acceleration is
thought to underlie rejection of the sensory environment (e.g., doing
mental arithmetic). In this formulation, the change in HR is viewed as
potentially either a response to the occurence of an external event
(stimulus generated), or as internally generated. timed to precede an
anticipated event. A more controversial aspect of the Laceys' work is
their belief that the change in heart rate is the cause of the difference
in attention to the environment and that the change in heart rate
mediates cortical changes via decreased inhibitory feedback on the
brainstem from aortic and carotid baroreceptors. The Laceys have also
extended this to the level of analyzing behavioral changes as a function

of the time within an R-R cycle that a stimulus occurs (Lacey 5 Lacey,

22

I978).

The Lacey hypothesis has generated much criticism (see Carroll 5
Anastasiades, 1978; Elliott. l97A: Green, 1980), primarily discussions of
experimental findings that conflict with their theoretical framework.

The most prominent alternative to the Laceys' work has been the
cardiac-somatic hypothesis developed by Obrist and co-workers (Obrist et
al., l97h: Obrist, 1981, for reviews). The cardiac-somatic hypothesis
views the change in heart rate directionality (described by the Laceys as
"directional fractionation") as solely secondary to the metabolic demands
of the body at that instant. In this framework, deceleration is viewed
as secondary to the motoric inhibition that occurs when the need for
attentional responding arises. A recent review (Obrist, I981) cites a
large set of studies showing results often contradictory to those
predicted by the Lacey hypothesis. Especially supportive of the somatic
interpretation of cardiac events are studies using pharmacological
blockade of sympathetic responses (and thus attenuating the ability of
the heart to respond to sympathetic stimulation) but showing no change in
behavioral performance as would be predicted by the Lacey hypothesis.

Despite the currently unresolved conflict over the meaning of HR
changes, the earlier publications by the Laceys were a critical influence
on the theories of psychophysiological attention developed by Graham and
her co-workers (Graham 5 Clifton, 1966: Graham 5 Jackson, 1970). Graham
observed that the heart rate fractionation observed by Lacey could serve
to distinguish between the orienting and defensive reflexes described by

Sokolov (1963). As previously mentioned, the Sokolovian orienting

23

responses are a series of physiological changes that occur concurrently
with sensory stimulation to increase the "taking in" of the stimulus.
Conversely, the defensive reflex serves to mediate stimulus rejection,
e.g., as response to an overly intense noise. Many predictions were
derived from this basic assumption: most important for the present
experiment is that the orienting response habituates over trials whereas
the defensive reflex (DR) usually does not, or does so with great
difficulty. The utility of HR is that it is potentially one of the few
measures that distinguish the OR and DR of Sokolvian theory (in addition
to cephalic vasodilation, obviously far more difficult to use
experimentally). The majority of psychophysiological measures that occur
with OR elicitation are cited by Sokolov as also components of the DR,
and thereby can be viewed only as general responses to stimulation
without any psychological interpretation.

This hypothesis of a relationship between heart rate directionality
and psychological set was an immediate source of theoretical discussion
and empirical testing. This was especially true in developmental
psychology where few useful measures exist to index subjective
psychological states: acceleratory vs. deceleratory HR response appeared
to be well suited for this task. Furthermore, Graham 5 Jackson (1970)
also noted a "deceleratory shift", or a change in the newborn HR response
from acceleration to simple stimuli at birth, to deceleration to the same
stimuli at approximately three months. The change in response was
hypothesized to be related to a change in the infant's ability to show

orienting or defensive reflexes, and thereby a source of the limitations

2h

on newborn learning (Jeffrey, 1968: Sameroff, 1971). Other studies used
the converse strategy, e.g., interpreting a change in HR response from
deceleration at three months when placed on a visual "cliff" to
acceleration at older ages when placed in the same situation as
reflecting the development of certain cognitive abilities (Campos,
Langer, 5 Krowitz, 1970; Campos, I976). The well documented deceleratory
shift stimulated a number of attempts, ultimately successful, in
demonstrating newborn deceleratory responses (Friedman, I972: Sameroff,
Cashmore, 5 Sykes 1973). This resulted in identification of a number of
factors important in studies of newborn responsitivity. For auditory
stimuli, control of stimulus parameters [e.g., rise time (Berg, l97h)]
proved necessary to elicit newborn HR deceleration. More importantly,
the need for controlling biobehavioral state (Clifton 5 Nelson, 1976) was
also a byproduct of these studies. In fact, the analyses presented by
the authors of the latter paper demonstrate that in context of the
expected alert periods for a newborn, many of the earlier attempts to
show newborn orienting (and similarly, newborn habituation) were severely
biased against obtaining positive results by paradigms lasting longer
than the infant's alert intervals.

Although the deceleratory shift has been interpreted as paralleling
a psychological shift in orienting ability, alternative explanations have
been offered in the literature. Porges (1976) discusses the shift solely
as an increase in vagal tone over the first three months of life: some
support for this proposal is provided by a recent cross-sectional study

by Ackles (Note 1). Berg 5 Berg (I979) also discuss this possibility in

25

their review. Demonstrations of conditionability to stimuli not usually
eliciting a cardiac deceleration early in infancy (e.g., auditory square
waves) also argue against the necessity of HR deceleration for learning

to occur (see Fitzgerald 5 Brackbill, 1976).

The theoretical disputes over the meaning of HR response are
currently unresolved and can charitably be described as a confusing body
of literature (Velden 5 Schumacher, I979). Volumes on these issues
(e.g., Kimmel et al., 1979, Obrist et al., l97h) offer chapter by chapter
disagreements on HR interpretation. Despite this, the reliability of the
response itself is unquestioned and has favored its popularity in
psychological studies. As a covariate of attentional reponse the measure
holds much utility and was used this way in the present study. Most
investigators in the field (e.g.. Graham, 1979) assume the presence of a
metabolic component, but question its relevance for using HR as a
dependent variable (although a recent study (Svebak, Oalen, 5 Storfjell,
I981) specifically manipulated both cognitive task and concurrent
physical demand, and found an effect for cognitive task above the
metabolic effects secondary to different physical tasks). Whether change
in HR is secondary to respiratory or motoric quieting, the fact that
these changes reliably occur allows this measure to be used in similar
studies. The importance of primary mechanisms for HR control arises when
results anomalous to those predicted by either the Lacey or Obrist models

occur: this is not of concern in the present study.

 

26

The heart rate response described above is the change in absolute
heart rate during a period of interest (e.g., the foreperiod of a
reaction time task or the time immediately pre- and post- stimulus
presentation). Another measure that has been derived from EKG records
has been the variability of the interbeat intervals. Porges (l97b;
Porges, Arnold, 5 Forbes, I973; Porges 5 Humphrey, 1977; Porges 5 Raskin,
1969; Walter 5 Porges, 1976) has hypothesized the change in variability
(specifically a decline in variability) to be a measure of sustained
attention. This is part of the older statement of Porges's
psychophysiological theory of attention. In this conception, the
immediate acceleration-deceleration response that occurs to stimulus
onset or offset is a measure of phasic, or involuntary, attention. The
sustained, or active, component is measured by the change in variability
of the cardiac response. Porges postulates these to paralllel the James'
similar concept (1890: 1950) of attention. Porges also cites evidence to
indicate that these two measures are statistically independent even
though obatined from the same cardiac record. This finding also has been
noted and replicated by others (e.g., Lacey, I967: Siddle, I980). The
variability measure has also been used to classify infants on the basis
of baseline (or tonic) variability to then predict performance in
conditioning studies; infants with higher baseline variabilities are
shown to learn conditioned discriminations easier than infants with lower
baseline variabilities (Stamps 5 Porges, I975; Stamps, I977). Since

heart rate variability can be considered synonomous with the amount of

27

vagal influence on the heart (Mulder 5 Mulder, I981), Porges (I976) later
attempted to quantify the degree of vagal tone on the heart as an
estimate of parasympathetic influence. The methodology he proposed was

to use the estimate of the coherence function when cross-spectral time

 

I.

series analysis is done between heart rate and respiration (since
respiratory cycling affects heart rate) as an estimate of the
parasympathetic effect. Unfortunately, after this one proposal, very
little additional supporting evidence has been published by the author or
has appeared in the literature, and these papers have been mostly
confined (as was the original research) to work with hyperactive children
(e.g., Porges, Bohrer, Keren, Franks, 5 Drasgow, 1981; Porges, Coles,
Cheung, Drasgow, 5 Bohrer (1979) with normal subjects). An unpublished
study by Ackles (Note I) finds little support for the Porges hypothesis.
Other studies have appeared, however, that use spectral estimates of
heart rate frequencies as dependent measures but avoid the theoretical
emphasis put forth by Porges. Mulder 5 Mulder (1981) have used the
strength of particular spectral components (or the power at that
frequency) to estimate the regularity in variability changes to stimuli
presented at that particular frequency. A major problem with this
technique, especially when applied to rapid, non-repetitous experimental
designs, is the requirement of a bare minimum of at least 100 equidistant
time points: even with .5 second intervals, obtaining this number in
infant studies would require a very long stimulus presentation. This is
probably not a difficulty with older subjects, and adult studies have

been successful in showing spectral changes independent of respiratory

28

effects in response to certain cognitive tasks, particularly situations
requiring prolonged attention.

Heselgrave, Ogilvie, 5 Furedy (1979) have replicated Porges's work

 

on variability changes in adults (Porges 5 Raskin, 1969: Walter 5 Porges,
1976) and has shown a decrease in variability during cognitive processing
(mental arithmetic in this case). As mentioned previously, Lacey (1970)
has also described a decrease in heart rate variability that occurs
during cognitive processing; Siddle (1980) also replicates this finding.
Heselgrave et al. have also proposed an alternative to variability
calculation of s2 by using delta2/2, a measure that removes the effects
of linear trends present in heart rate data that have highly correlated
successive measurements. In a comparison of two measures applied to
their data, Heselgrave et al. conclude that the "appropriate HR
variability analysis for the cardiac-activity variability hypothesis
should involve the delta2/2 variability statistic" (pg. 155). However,
Wastell (1980) discusses the falliability of using the delta2/2 measure,
and why its properties make it inappropriate as a substitute for the
traditional variability measure.

Porges (1976), in his discussion of the variability measure,
hypothesizes it to represent a prolonged attentional response, the latter
half of his two component model of attention (the first component being
the phasic change in heart rate to the physical properties of a
stimulus). This definition implies continuous "mental work" or
processing, as in the aforementioned mental arithmetic tasks: it is most

likely not true in a paradigm such as the one used in this experiment

29

where attention to the stimulus is not constant over the interval of
stimulus presentation. However, in habituation studies this confound can
be used to advantage, since one variable of interest is the degree to

which subjects manifest cardiac responses supposedly compatible with

r.

sustained attention over trials. This will be further discussed below.

Studies 21 habituation and orienting jg infancy ‘

Habituation has been one of the most widely used techniques for
studying perceptual development in infancy. Through use of the paired
comparison technique (Fantz, 1975) and habituation-dishabituation
paradigms (Fagan, Fantz, 5 Miranda, 1975), much of the empirical
knowledge about infant development has been acquired. Both of these
procedures involve visual habituation to a stimulus and then use the
response to a different stimulus as a reflection of the infant's ability
to remember the former stimulus and discriminate between it and the
latter. In these studies the variable of interest is the infant's
ability to make the discrimination. Often the occurence of visual
habituation to the standard stimulus is not even tested for (e.g., Welch,
197‘).

Despite the wide use of habituation experimentally, the exact
definition of when habituation occurs has been a persistent problem in
both the animal literature (Ratner, 1970: Thompson 5 Glanzman, 1973). and
the human literature (Cohen 5 Gelber, 1975). The two basic alternatives
are using either an absolute criterion (i.e., habituation occurs after a

response is at a certain absolute asymptotic level) or a relative

30

criterion (i.e., habituation has occured after a certain per cent
decrement in response is present). The use of an absolute criterion
leads to potentially absurd conclusions: if a subject doesn't respond on
the first trial (a not infrequent occurence in infancy studies), the
conclusion of single trial habituation is drawn. The relative criterion
approach also has problems, the most obvious being the question of what
the criterion should be. The studies by Cohen most often use a decrement
of 502 from the average of the first three habituation trials: if that
figure needs to be lowered to get an adequately "habituated" sample, as
in Olson (1976), at what point can it be conclusively said that the
"habituated" #02 decrement sample reliably differs from a slightly less
declining group ? Another obvious problem involves determining which
trials should be used as the criterion. As McCall (1979) points out, the
trial of peak response in infant studies is very infrequently the first:
a 50% decline from the first few trials is potentially misleading. As
discussed below, these problems were recognized in the design of the
present study but nevertheless were potentially serious problems.

Other than its use as a methodological technique for perceptual
studies, the study of habituation itself has also been of interest to
investigators for a variety of reasons. Habituation has often been
described as the most primitive of learning and has been documented in
almost all phyla (Ratner, 1970). For this reason alone, attempts have
been made to demonstrate habituation in newborns and infants to
illustrate the extent of cognitive development ontogenetically. As an

extreme example of this approach, attempts by Brackbill (1973) and

‘ I

r—vr— l

 

31

Graham, Leavitt, Strock, and Brown (I978) to obtain OR habituation in
anencephalic infants were made to examine the question of habituation at
birth and specifically the role of the cortex in this process.
Unfortunately, their results are conflicting (Brackbill reported no
evidence of habituation but Graham found habituation) and allows little
to be said definitively. The rate that an infant habituates has been
used as a measure of individual differences and then related to other
variables (e.g., Lewis, 1971). Similarly, groups differing on a variable
of interest have then been tested to see whether they differ in rate or
form of habituation pattern (e.g., nutritional status between groups,
Lester, 1975: or number of congenital defects, Schexnider, Bell,
Shebilske. 5 Quinn, 1981). Habituation items are also included on
standard scales of infant development (Bayley, I969; Coll, Spekoski, 5
Lester, 1981)

Until very recently, the course of habituation was assumed to be a
progressive decrement in response as some form of cognitive
representation was acquired. In agreement with this concept,
experimental reports were usually plotted as forward habituation curves
that demonstrated this trend. Clearly, if the nature of habituation is
not a progressive decrement, the conclusions drawn from these early
studies become questionable. Assumption of this model is implicit in the
standard analysis of variance techniques used in analyzing habituation
designs and can cause erroneously significant conclusions if this is not
the true course of habituation (Cohen 5 Gelber, 1975). An example of

this problem is the study by DeLoache [1973, cited in Cohen 5 Gelber

 

32

(1975)] that interpreted fast habituators as storing information more
rapidly than slow habituators. The reanalysis of her data when plotted
backwards by Cohen 5 Gelber (1975) superficially indicates that both
groups habituated in a single trial. However, an unpubished study by
Rissman (1976), cited by Cohen (1976), does offer support for DeLoache's
conclusion. Another serious problem in early studies of habituation
unrelated to the implicit or explicit model of habituation assumed is the
frequent lack of control of infant biobehavioral state: both Cohen 5
Gelber (1975) and Clifton 5 Nelson (1976) have discussed this problem as
a significant confound of early studies.

In the majority of infant studies using habituation paradigms, the
dependent variable has been visual fixation. However, as discussed by
Cohen (1973), visual fixation responses can be interpreted as reflecting
two separate processes of attention-getting and attention-holding.

Though Cohen's attempt to asssociate these experimentally with the number
of fixations and the duration of each fixation (respectively) was
unsuccessful (Cohen, 1976), the theme of two separate processes has been
present in much of the theorizing about attention, as in the Porges model
described before. Experimentally, Porges's model is especially
attractive since it offers the possibility of assessing the two aspects
of attentional responses psychophysiologically, separate from any
behavioral response, and it has been moderately supported by studies in

infancy (Porges, 197A).

:41
g,“ 'l

__.C ”...

 

 

33

As was mentioned previously, Sokolovian theory predicts habituation
of the orienting response. When defined as the deceleratory HR response
after stimulus onset, demonstrations of habituation of HR response in
infancy are numerous (although this component of HR response is most
often identified as tflg orienting response, there is disagreement over
this interpretation (Graham, 1979; Velden 5 Shumacher, 1979). HR has
been a preferred dependent variable response for habituation studies in
infancy, since it can be used in studies without visual stimuli (e.g.,
auditory or gustatory stimuli) where there are not well-defined
behavioral measures of habituation, and also because it is one of the
responses predicted to distinguish orienting and defensive reflexes (see
Clifton, l97A: Berg 5 Berg, 1979 for reviews). Although these studies
are frequent, none report analyses of changes in the variability patterns
of heart rate with the exception of Porges's work (e.g., Porges, I97h).
The relationship between variability and behavioral measures is an active
area of study in Europe (see Kitney 5 Rompelman, 1980) but little work in

this area has been done in the United States.

Up to this point I have tried to summarize briefly the three
overlapping fields of research relevant to the present study. The goal
of this study was to analyze concurrently a psychophysiological measure
of attention (the decline in the variability of heart rate) and a
behavioral measure of habituation (visual fixation time) to observe
whether parallel patterns of decline occur in both. It is uncertain

whether the peak and then decline pattern of all-or-none habituation is

3h

artifactual. If a parallel decline were to occur in a separately
measured index of attention, unrelated to the criterion judgment of
habituation, a far stronger argument could be made for the true nature of
habituation being rapid, few, or single trial learning, and not a gradual
decline in response.

Before concluding this section, the recent important conceptual and
methodological step by McCall (1979) in trying to unravel the
relationships between cardiac measures of attention, visual fixation
measures, and habituation must be mentioned. In this study, McCall
essentially ignored the issue of the true course of habituation by using
a multivariate technique to group infants by the pattern of visual
fixation response from trial I onward to a visual stimulus. After
clustering the response patterns into 3 groups, McCall then attempted to
differentiate these groups by the cardiac patterns shown by the infants
over habituation trials. The results of the behavioral clustering showed
no relationship to a similar clustering of cardiac pattern. However,
McCall reported no analysis for the variability pattern of the subjects:
from Porges's work, it would be predicted that this is the component of
cardiac function that would more likely show a relationship to behavioral
orienting. However, McCall's work appears to be the first application to
grouping subjects rather than variables via a multivariate approach, a
strategy strongly advocated at a recent meeting of the Society for
Psychophysiological Research (Kircher, I982: Porges, 1982). This
approach could not be used in the present experiment due to the

relatively small sample size.

CHAPTER 3

' METHOD

Subjects

The final sample analyzed consisted of 28 infants with a mean age of
125.2 days (SO 8 18.5 days). There were 16 males, l2 females (see
Appendix B). All infants were full term (birth within 3 weeks of
expected due date), with no history of pre-, peri-, or post- natal
complications as reported by the parents (see Appendix B). Four months
was selected as the target age for two major reasons: a) the relative
ease of using subjects at this age (at 2 months it is often difficult to
maintain infants in an alert state: after 6 months, babies are often too
fussy to use as subjects in a laboratory environment where movement
artifact can cause errors in the physiological monitoring of subjects),
and b)the numerous studies demonstrating habituation at this age to a
variety of stimuli (e.g., Cohen, 1973: Clifton 5 Nelson, 1976). In
addition, Harper, Hoppenbrouwers, Sterman, McGinty, 5 Hodgman (1976)

found little change in heart rate or baseline variability in the period

35

 

36

from three to six months (far less so than the preceding three-month
period).

Seventy-two infants were tested to yield the final sample of 28. Of
the AA infants not included in the results, 28 were excluded due to
inability to finish the experimental session (primarily due to state
changes, e.g. crying or sleeping), and data from the remaining 16 were
not analyzed due to equipment malfunctions or noise in the polygraph
record that made computer scoring inaccurate. (It should be noted that
this was the first study to be analyzed using the LSI-ll computer in the
psychophysiological laboratory: as such, certain aspects of the recording
of EKG records important for later computer analysis were unknown to the
author when the study began. It is anticipated that future studies in
the laboratory will not suffer such a large subject loss due to this
problem.) Of the 28 subjects constituting the final sample, 2 were
retested completely on a second occasion after the first experimental
session could not be completed. Four additional babies were tested at
the onset of the study to train observers and familiarize lab assistants
with the experimental procedures.

Subjects were recruited by contacting families through a mailing on
university letterhead (see Appendix A). The names of parents were
obtained by examining the birth records for Ingham County, Michigan.
which are open for public inspection. Addresses were found in the
Lansing telephone directory. After receiving a reply card that was
included with the letter, parents were contacted and asked to come to the

laboratory at a time of day they thought the infant would be in an alert

 

37

and cooperative state.

Procedure

On arrival at the laboratory, parents were greeted by the
experimenter who explained the experimental procedure. Parents were also
told that they had the option to withdraw their infant from the study at
any time during the session and that all records for their infant would
be destroyed; none of the parents exercised this option during the
experiment. Parents were then asked to sign an informed consent form
(included in Appendix A) indicating their approval of the experimental
procedures. If the infant subsequently completed the entire session,
parents were asked to fill out a short biographical questionnaire (see
Appendix A) requesting information about the baby's birth and present
status. After this introduction, the parents and infant were escorted to
the laboratory where two heart rate electrodes and an electrode ground
(standard Beckman Ag/AgCl biopotential miniature electrodes) were placed
on the infant to monitor the electrocardiogram. The two conducting
electrodes were attached via a non-irritating adhesive electrode collar
approximately 1 cm. below the infant's nipple. To increase conductivity,
Synapse brand electrode paste was used to fill the space between the
electrode and an adhesive collar. The paste is described as both
non-irritating and hypo-allergenic, and no parent subsequently called to
report a rash or other adverse reaction to the paste. The ground
electrode, prepared in the same manner as the other two electrodes, was

placed approximately 2.5 cm. above the infant's navel. After the

 

38

electrodes were securely attached by micropore surgical tape, the parents
were asked to place the subject in an infant seat chair in the
experimental booth. The booth was a large, sound attenuated chamber,
where the forward wall had a large section removed and replaced by a h8.5
cm. by Al.5 cm. rear projection screen. The infant seat was a standard
infant car seat adjusted so that the infant was strapped in securely and
reclining at approximately a 20-degree angle from vertical. Two infants
were tested in their own car seats at parental suggestion due to
fussiness in the laboratory seat. The seat was .75 meters from the
screen during stimulus presentations.

The HR electrodes were plugged into a junction box attached via
shielded cable to a Grass Model 7 Polygraph. A Phipps-Bird pneumograph
was placed around the infant's chest to monitor respiratory movements on
the polygraph, and the output from this transducer was also fed into the
polygraph. A parent was seated in the chamber with the infant at all
times directly behind the infant seat and out of the infant's sight. A
white noise signal was continuously played in the chamber to mask the
noise of the experimental equipment and any outside noise. Although the
ambient noise level was not measured for this study, a previous study in
the laboratory using the same white noise signal reported its value to be
55 db re .2000 dynes/cmz (Ackles, Note 1). After the infant was seated
comfortably, the experimenter monitored the EKG and respiratory movements
on the polygraph in an adjoining room, making appropriate adjustments on
the polygraph when necessary. All physiological measures, stimulus

markers, and signals from the assistants monitoring the infant's looking

 

39

behavior were simultaneously recorded on a Vetter Model 700 tape recorder
for later digitization and analysis by a DEC PDP 11/03 computer system.
The same computer system was used to control stimulus presentation and
timing during experimental sessions.

If the infant was judged to be in an alert state (State 5 using the
Brackbill and Fitzgerald (I969) scale for monitoring infant state), the
experimental procedure was begun. An assistant, viewing the infant
through a .65 cm. peephole in the front wall of the experimental chamber,
constantly observed the infant's state and looking behavior throughout
the experiment. A microswitch was mounted on the front wall in easy
reach of the assistant: by depressing the button, the assistant indicated
whether or not the infant was looking toward the screen. Reliability
training was done (see below) with all assistants before testing began.

Following a one minute prestimulus baseline period, each subject was
given 10 repetitive stimulus trials followed by a dishabituation trial to
a novel stimulus. For each subject, the repetitive stimulus was an 8x8
square red and white checkerboard: the dishabituation stimulus in each
case was a 16x16 square red and white checkerboard. During intertrial
intervals, a black slide with a center hole was used to present a 2 cm.
white spot on the screen. Stimulus images were projected from 2x2 slides
mounted in a Kodak Model 700 projector: when presented, the side of each
checkerboard was 28.2 cm., each square having a side of either 1.75 or
3.50 cm. The slide showed the checkerboard against a blue background so
that the entire stimulus image was 36 cm. by 36 cm. Both 8x8 and 16x16

checkerboards (with black and white checks) have been shown to elicit

hO

prolonged looking behavior by the infant at 3 months (Horowitz, 1975).
The duration of each stimulus presentation was 18.A seconds. For the
first 20 subjects, the ITI's varied with a mean of 12.1 seconds and a
standard deviation of 1.8 seconds. For the remainder, the ITI values
were fixed for each ITI with a mean of 11.8 seconds and a 1.7 second
standard deviation. The lTl values were varied in order to prevent any
temporal conditioning to stimulus onset. The entire experimental
procedure, including calibration and baseline, was usually completed in
under 10 minutes to minimize the possible confounding influence of state

changes.

Dependent Variables

Heart Rate

Heart rate was monitored continuously throughout the experiment.
The raw EKG signal was recorded on magnetic tape time locked to the other
physiological and experimental signals. The EKG was then fed into a peak
detector-Schmitt trigger circuit (Shimizu, 1978) which identified R waves
in the EKG signal. The digital pulse output was fed into a PDP Il/03
computer which timed the R-R interval (i.e., heart period) in
milliseconds (although the computer resolution is in milliseconds, a 12
flutter in the tape record playback introduced some error in the heart
rate measurements). Heart period values then were converted to the
weighted averages for each .5 second interval relative to stimulus onset

and offset. Although there is disagreement in the literature about

Al

whether interbeat interval (181) values or their conversion to heart rate
are the appropriate unit for later analysis, (Graham, 1978: Heselgrave,
Ogilvie, 5 Furedy, 1978: Richards, 1980; Sayers, 1980), the arguments
favoring the use of heart rate in beats per second (the reciprocal of the
181 multiplied by 60,000) seemed far more compelling. HR variability
estimates were derived by dividing the first 18 seconds of stimulus
presentation into A periods of h.5 seconds each. The weighted average
for the lBI's in each .5 second interval in the h.5 second unit were
transformed to heart rate per .5 second interval, and then used for the
variability calculations. Therefore, estimates for each period were

based on 9 data points.
Visual Fixation

Visual fixation measures were recorded by research assistants
viewing the infants through .65 cm. peepholes at the sides of the front
of the experimental booth. Infant fixation was assessed via the corneal
reflection technique (Fantz, 1975) where observers look for the image of
the stimulus on the infant's cornea. When this was observed, a silent
microswitch mounted on the booth wall was depressed. Reliability
training was done with all observers. For each .5 second interval after
stimulus onset the computer recorded whether each observer signaled that
the infant was looking. Two separate estimates of reliability were
obtained: the first measured all intervals, including intervals where
both observers agreed that the infant was not looking. This is the

traditional method of calculating reliability (e.g., Ackles, Note I), and

 

#2

yielded estimates from 76% agreement to 96% agreement, with a mean value
of 88%. An alternative method for calculating reliability when fixation
is the dependent variable is to examine only the intervals where one or
the other observer judged the infant to be looking at the stimulus. In
this case, the reliability is not truly per cent agreement but is
estimating the probability that if one of the two observers judged the
infant to be looking, then the other will also. For this measure, the
mean was 732, with a range of 6A2 to 81% agreement. Thus, although the
traditional reliability estimate is respectable, the measure of agreement
when the infant is looking is less so: the possible problems introduced
by this measure will be discussed later. When two assistants were
observing the infant during reliability training, one peephole on each
side of the booth was used: when only a single observer was present, the
side the observer preferred was used. Visual fixation responses were
recorded on the polygraph record and also time-locked to stimulus
presentation on tape. Computer analysis of these data yielded estimates
for the total fixation time, length of first fixation, and total number
of fixations, although only total fixation time was used for later
analysis.

A recent report (Maurer 5 Lewis, 1979) demonstrates substantial
peripheral vision in three-month-olds. This can be interpreted as
weakening the view that the infant is not fixating or scanning the
stimulus when the corneal reflection is not perfectly over the center of
the cornea. Although the traditional technique of scoring fixation when

the image is over the center of the cornea was used in this study, future

 

1‘3

experiments may require an estimate of the discrepancy of the reflection
from a central focus.

The procedure for this study used a fixed stimulus interval, with a
variable intertrial interval to prevent the possible occurence of
temporal conditioning. Although this was a standard paradigm for
habituation studies with infants in the 1960's, this methodology has
fallen into disrepute. The major reason for this is by not linking
stimulus onset to fixation of the screen by the infant, stimulus
presentation trials can occur in which the infant does not fixate the
stimulus for the entire presentation time. Primarily for this reason,
Cohen (1973) and Horowitz (1975) strongly advocate the use of "infant
control procedures". An example of this approach is described by Cohen
(1973) where first the infant's attention is attracted by a flashing
light: once the infant fixates the flashing light, the stimulus is
presented contiguous to the light which goes out. If the infant then
turns away, the flashing light reappears, and the cycle repeats itself.
This method, while highly applicable for visual fixation responses,
cannot be used for studies measuring physiological responses to stimulus
onset since there also will be a physiological reaction to the onset of
the flashing light to get the infant's attention. An alternative
procedure is that used by McCall (1979). In the McCall study, after a
fixed ITI, the stimulus was only presented when the infant was fixating
the screen after the ITI interval was completed. However, in this study
the stimulus was left on only while the infant fixated the stimulus (as

in Cohen's approach). This approach was similarly rejected for the

Ah

present study, since a fixed stimulus interval time was needed to yield
equal intervals for variability analysis. For the present study, the
decision was made to present the stimulus unlinked to any behavioral
response. The infant seat used in the study was arranged such that when
the infant was strapped in, it appeared almost certain he would be
fixating the screen. It was predicted a priori that the change in
ambient illumination in the experimental booth would be sufficient to
attract the infant's attention, therfore obviating the possibility of
excessively long intertrial intervals if the infant did not look at the
screen. In a sense, this was viewed as a combination of a fixed trial
procedure and the Cohen method: the change in illumination due to
stimulus presentation was acting both as the attractant toward the
stimulus, and the stimulus after orientation occurred. Personal
experience from a previous study by the experimenter (Gitterman, Note 2)
and in other studies in the laboratory (Ackles, Note I) lent support to
the decision to use this procedure. Although this proved correct for the
majority of stimulus trials, trials did occur when the infant did not
fixate the stimulus. The effects of this potential confound are

discussed with the conclusions of this study.

CHAPTER A . 1

RESULTS

 

All analysis of variance tables for the main effects and
interactions are contained in Appendix C. Calculations were done on the
MSU Cyber 750 computer using either BMD (Dixon 5 Brown, 1979) or SPSS
(Nie, Hull, Jenkins, Steinbrenner, 5 Bent, 1975). The analysis of
variance results obtained via these programs, although allowing multiple
independent factors, used a univariate dependent variable analysis of
variance approach to repeated measures designs (see Keppel, 1973). As
has been discussed by numerous authors (Bock, 1975: McCall 5 Applebaum,
1973: Richards, 1980), the univariate approach assumes homogeneity of the
variance-covariance matrix, a situation that is rarely obtained in
repeated measures designs. Heterogeneity of these matrices is especially
probable when change in the variable of interest is part of the
experimental treatment. Two alternatives to this problem have been

proposed in the literature. The prefered solution is to use a

‘05

A6

multivariate approach with a multivariate analysis of variance (MANOVA)
that makes no assumptions about the nature of the variance-covariance
matrix. Although this was the approach the author intended to use for
this analysis, statistical computer programs suited to this design were
unavailable. The other alternative that has been discussed in the
literature is the use of the Greenhouse-Geisser correction (Jennings 5
Wood, 1976; McCall 5 Applebaum, 1973) for the heterogeneity of the

variance-covariance matrix. Essentially, this is an ad hoc correction

that results in a reduced number of degrees of freedom for the fi-ratio
tests, thereby yielding a more conservative test of significance. The
reduction in degrees of freedom with this correction can be very
substantial, often causing significant tests to become clearly
nonsignificant. McCall and Applebaum (1973) suggest doing all repeated
measurse F-tests with I and k-l degrees of freedom (k- the number of
levels on the repeated factor) for the most conservative possible
testing: alternatively, Jennings and Wood (1976) advocate a three-stage
procedure wherein, if the normal test is significant, but the
conservative test nonsignificant. the formula for an exact correction of
the degrees of freedom based on the degree of heterogeneity of the
variance-covariance matrix should be applied. If the unadjusted test is
nonsignificant. there is no need for any further testing. The
three-stage procedure was not done in the present analysis since BMD does
not give the necessary figures for this calculation. However, given the
borderline significance for many of the results that will be described,

it is fairly certain that any, even small, correction of the fi-test would

“7

cause it to become nonsignificant. (Heterogeneity of the
variance-covariance matrices was tested by BMD program P2V for every
analysis: in almost every case the data in the present study were
heterogeneous (or nonsymmetric).) In those situations where the standard
test is significant and the conservative test nonsignificant, McCall and
Applebaum (I973) propose that the results be described as tentative (or
"ambiguous“), erring on the side of nonsignificance. Situations where
this occurred (almost every significant result) will be described below.
The level of significance conventionally used for F-tests in
psychological research is g<.05. However, given the small number of
subjects in this study, it was thought that a more lenient criterion for
significance would be appropriate, and a g<.10 was chosen for the F-tests

that will be described.
Resglts for cardiac measures alone:

Before analysis, the interbeat intervals were converted to heart
rate in units of beats per minute. An analysis of covariance was then
performed using the prestimulus second as the covariate for the
subsequent 18 second intervals. There was a significant main effect for
trials ( {-1.86, g1-9.2&2, 22.05) with the uncorrected F test for the
analysis of heart rate after stimulus onset for all 28 subjects. If the
most conservative correction possible is applied (reducing the degrees of
freedom to 1,17), the F test becomes nonsignificant ( g-.20). However,
as noted, this is the most conservative possible reduction in degrees of

freedom. This result will be discussed further below: it is plotted as

 

A8

ngggg 3. There also was a significant main effect for seconds
( {-6.73, gf-l7,h59, g<.OOI). This result also was significant after
testing with the conservative degrees of freedom ( gi-l,l7, g-.02). The
seconds main effect can be interpreted as evidence for an orienting
response to stimulus onset by the entire sample, and the trials effect,
although tentative, indicates evidence for habituation by the decline in
the HR response to stimulus onset over successive trials. The
nonsignificant interaction ( g-.15, uncorrected) indicates relatively
uniform change for the components of the orienting response over the
stimulus trials. The seconds effect and the uncertain trials effect were
both predicted and replicate previous studies.

There was no main effect for trials for the analysis of covariance
on the variability measure. The main effect for periods (the four h.5
sec. units after the prestimulus interval is covaried out) was
significant before correction ( {-3.8h, g1-3,81, g-.01). After
correction, this result was nonsignificant ( gf-l,3, g-.l6), but again
the extreme reduction in the degrees of freedom for correction must be

considered ( Figure 3 ).

Subjects were then grouped into habituator-nonhabituator groups
using the criterion of 3 consecutive trials each having a total fixation
at least 502 below the average of the two longest fixations of the first
5 trials. (However, due to the low number of habituators, one infant

with a decline between #02 and 502 of the criterion was included.) This

M 3. Heart rate response for all 28 subjects
to stimulus onset.

49

 

pmmzo pwom wozoumw

 

wetmamfiiflfizoemmpomemmH
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
a. \\null. w 1
\~ \ \\ .a
\/o \ |\t \\\\ a!
\\\ /I O \\I \ \\ \\o\\ /I/
\\s \\\t”olcllclsllt\\\ "I ssr.\\\\ul \\ V/ll
\ / \.>/.n. x/ \mxkixx /..\\
\sss /I(\\\\ /t\k|!lla IIII\\\\\
Xxx ..
\1) x
\I\\. / \
\ / \
\ // \\\ I
\ \II/uwxx 8 a m BEE ..I
/ \ \\ w a 2. SEE ......
/k/ \ m a m SSE ..i ...
\\ link s. a m SSE
w a e BEE .II

 

 

 

pmwzo msuazubm ow whom—dam 0N 13¢ mo... wmzommmm upcm .Ecu:

befi

9:

mv~

om~

am:

Nmfi

mma

«ma

QUEI—(D mum: ZHZDI—U

[igggg 5. Variability response for all 28 subjects
to stimulus onset.

50

 

...wwzo 94.55.; mupmm ooumwm

 

 

 

“a mm
warm
0 a m 95:31-1
... a. m wqmum» I...
N». H m

 

 

 

whom—63m ow ...—a mo... mwzomwwx >:.:mm~m¢>

b..—

m;

m;

TN

N.N

m.N

ZCEF—DMCEJ ..JDC) >CE01HCEZUUJ

51

yielded 6 habituators from the original sample. The low number of
habituators from the original sample of 28 is probably a result of a
number of factors, the most important being the methodological problem of
not making presentation of the stimulus contingent on the infant facing
toward the screen; this issue will be raised later in the discussion
section. The average total fixation time for the habituation group used
to calculate criterion levels (i.e., the average of the 2 longest total
fixations for the first 5 trials) was 11.3 seconds (range h.9-l7.l). The
minimum average fixation time of h.9 seconds implies that inclusion in
the habituation group was not an artifact of short initial fixations.
The remaining 18 subjects not meeting the criterion for habituation were
pared to a group of 6. 0f the 12 subjects eliminated, 7 were eliminated
either because they made no response on a majority of trials or showed
response patterns that came close to habituation without meeting the
criterion. This procedure should have discriminated maximally the
resulting group from the habituation group on the behavioral dimension.
Five additional subjects were also eliminated randomly from the
nonhabituation group to yield equal group sizes. For the nonhabituation
group, the average of the longest fixations was 15.5 secs. (range
12.5-18.2 secs.). A t-test comparing these two groups showed the two
groups significantly different on this criterion measure [ £31.93, g196,
g-.05: test corrected for heterogeneity of variance (Winer, 1971)].
However, since the trial of longest total fixation was rarely the first
trial, another t-test was done comparing the two groups on length of

total fixation on the first trial on which fixation occurred. For this

 

52

measure the means were 8.8 and 10.7 seconds for
habituators-nonhabituators respectively, with a nonsignificant t-test

( t-.65, gi-8, g>.20; test corrected for heterogeneiety of variance)
indicating that the two groups did not differ significantly at the onset
of fixating the screen. The fixation responses over trials are plotted
in filgggg 5 as a forward curve: [jgggg 6 represents the habituation group
data plotted backward from the trial when the criterion for habituation
was met. The importance of the difference in looking behavior will be
discussed in the following section.

After grouping was done, analysis of covariance (ANCOVA) was applied
to the heart rate and variability measures. In the analysis of
covariance, the regression of the poststimulus measures on the
prestimulus measurement (either heart rate or variability) is used to
decrease error variance by removing differences in the data correlated
with the covariate: in the present case the prestimulus level is used to
covary out differences in initial level. In a repeated measures design
with a nested factor, one of the assumptions of ANCOVA is that the
regression of the covariate on the dependent measures is homogeneous for
each level of the nested factor (Richards, 1980). Although there is a
test of this assumption in standard texts (e.g., Keppel, 1973), a simpler
approach taken here was to do a repeated measures ANOVA for the baseline
period (the prestimulus period for trial 1, before any testing began) and
the first post stimulus onset period. It was felt these would be most
highly correlated (hence most linearly related) and free of any influence

of the trials effect (in this case the experimental treatment). The 2x2

Figure 5. Total fixation response over trials by group.

53

 

mumume
8 m m s m m ... m N
. _ _ _ — _ — _ _
mmoEatquzoz .....

 

machcshwmmr.lll

 

 

macaw rm wuch» mw>o mmzomwwm zoupcxuu Jabs»

8

2

3

Hz (OLIJQOZDCD

papmg mwxmpuoz

fijgggg 6. Total fixation time plotted backwards from
criterion.

54

wumum» zouhczhumcr

N+ a+ zoumwhmmu an N1 m1
_ _ _ _ P _

 

 

 

 

zo~¢mwumu zomm wommxxucm cupboum wzuh zouhcxum 4¢h0h

ca

Na

I—OI—GZJ lLHXCEF-HOZ HZ (DUQOZDCD

55

ANOVA (group by period) was nonsignificant for group or trial. The
period by group interaction was very close to significance ( g-.12);
given the low power of the small sample size, this result was interpreted
as indicating heterogeneity of the regression slopes. The means for
these two periods are displayed in Eigggg 1. In addition, examination of
the beta weights after the analyses were done showed that the beta
weights for the two groups were substantially different. Therefore, for
the variability analysis. separate analyses were done for the habituation
and the nonhabituation groups. A similar test for the heart rate measure
was nonsignificant, so an ANCOVA containing all three factors (group by
trial by second) was done here.

For the heart rate results, there was no main effect for habituation
group or trials, although the uncorrected seconds effect remained
significant ( {-2.51, gj-l7,l70, g-.OOI). This result was also
nonsignificant when the more conservative test was done ( gjsl,l7,
2.,15), Although, the trials effect was nonsignificant, the trials by
seconds interaction was significant ( £-l,l8, gf-153,1530, g-.07)
uncorrected, but it was clearly nonsignificant with the conservative
test. The results for this analysis are plotted separately by group
( Figures 8 and 9 ). The three way interaction (group by trials by
seconds) was nonsignificant. After these tests for the entire 10 trials,
a test was made comparing HR responses on the first trial that was used
to classify habituators into the habituation group and the trial
preceding it (essentially comparing the trial before and after

habituation). The effect for trials in this comparison [a 2 (trials) by

fijgggg 1. Prestimulus period versus poststimulus period
for trial I.

56

msqszmpwhwom
_

mooummm

wzuazuhmmzm
_

 

 

machcnzmcxzoz .....
mucus...— “ mm: Ill

 

 

H .55: mo... ooumwm maqazahwpmom wammw> ooumwm msunzupmwmm

N.N

TN

9N

m.~

ZGI—DOCCEJ ..JDC) >CEOCHCZUUJ

f_igl_1r_e_ 8. Heart rate response for habituators.

57

[.1‘L.ni‘.l

hwwzo mzuatzm «EL: mozouwm
8339:2230" m o h o m ¢ m NH
_

 

 

 

 

_ — _ _ _ _ _ _ _ _ _ _ _ _ _
9;
o.— .n m mums: 11
x. N a 5 SEE ...i
x. z m N m BEE ..i
. / 5 a N BEE I
N x N a « 33E 1.. 02
0"
/o/ \X’
. N‘ss aoa .I Nwd
’ \I“ \ / col
to" so \\ .I/ ) ~s\s a.
Kiss oo isiiiiio’oo ssss was A o/\~ :\c 1.1.2.;
s. I”, l\|\ IIIGI |I\ v s» I I! \\ <0'Il\ /
. a \I \ .o s I: X o/
.s. \1/ w s\\ o/Illl ~\ .. \Ms C I/\ v.1 VWd
~-\\\ llfi/ls\ / s ks */ 1‘ .1
RN \./X\ // . Ne \. \ IIII\\\\
Ii”... \ I. z / ; \.\/ \
ss 1k //\\\/l \ o¢\ / \
N \. I s \ /\ I mm:
7 \ K)...
<
mm“

wmohcazmmz moh— wwzommmm upmm Hmcw:

GUJGI—(D LNG: zHZDF-IJJ

Ligggg 9. Heart rate response for nonhabituators.

58

pwmzo waqztahw «PE: wozouwm

mfibamamaeémflNHZHmebmw—c mwa
_____________

 

0¢~

om:

um:

 

 

 

 

«9
02
N a m m

c N N N N NE
\\ m a m BEE ii
\ s ... N BEE
N a s 325 II

2:

mmopmazmczzoz mom mmzommwm whim SEN:

(Dunn—w 0..qu IZHZDF-IJJ

59

18 (seconds)] analysis of covariance was nonsignificant.

For the variability measure, prestimulus variabilty was covaried out
in the analysis of covariance with the four A.5 second periods post
stimulus as the dependent variables. As described above, a separate
analysis was done for each group. For the habituators, the trials effect
was nonsignificant ( fi-l.3l, 81-9,Ah, g-.26). The uncorrected effect for
periods was marginally significant ( fi-Z.95, 8133,15, g-.07), but clearly
nonsignificant if the conservative test is applied ( 8131.3, g-.20). The
interaction was also nonsignificant ( 88.20), but in the proper
direction. The response of the habituators over trials is plotted (in
blocks of 2 trials) as Eigggg lg . For the nonhabituators the trial
effect was nonsignificant ( g>.25), as was the period effect ( g>.25).
There was an uncorrected significant interaction (see Ligggg 11) for
trials by periods ( {-1.57, 8:!27,135, g-.05). When tested
conservatively, this interaction was again nonsignificant ( 8:-3,9,

E>'25) .
Results jg; dishabituation testing:

The results for behavioral dishabituation were different for the two
groups. For the habituation group, a paired samples g-test between
trials 10 and 11 (the last habituation trial and the dishabituation
trial) was very close to significance, with an increase of looking time
at the dishabituation stimulus ( £3-I.88, 81r5, g-.08, one-tailed test:
Eigggg l; ). For the nonhabituators, the fixation response on the

dishabituation trial was nonsignificantly different from the last

M 12. Variability response of habituators over trials.

6O

.5sz magazmhm mukmc woouzum
e m N a

 

 

    
 

N a m mums: 11
m a b mug”: III
9 a m 95:: 1.1
v a m @455. .1...
N a a wqcum» III

 

 

9.5:: mm>o machcazmcz ...o mmzommwc >Cdm¢~m¢>

N.N

¢.N

m.N

ZCEPDMCEJ .JOCD >¢¢HCZUL|J

[jgggg ll. Variability response of nonhabituators over trials.

61

pwwzo angst—hm mm»u¢ oouzum

 

 

N N N N

_ _ _ _
\.

\ / S N m 35E ll

\ N N 2. SEE ...-..

\ / N N .o. :5: ..-...

. \ / N N N 325

/ \ / N N N SEE II

 

 

 

mucumb ¢w>o agapcspumczzoz no wwzommwc >h~4~mc~¢¢>

m.—

o.~

N.N

¢.N

0.N

Z¢F=KCJ .100 >C¢HIZUU

Ligggg 1;. Comparison of final habituation trial
and dishabituation trial.

-.-. ‘fw .~' sue-u.

l-
l .
I.

J

I
0“
‘

I
:1

 

 

62

umHmp onpmahummszo
b

ucum»

Jchp .pummr ham;
_

 

 

wmohmapwmmrzoz.li-
wro~¢2-m¢z.lll

 

 

ucump onpcapHmcrmHo 02¢ Acamp .hHmcz uczam no zomHmmmzoo

Dd

N~

HZ UDUJUOZCHD

I—DI—CEJ LLHXCEI—HOZ

63

habituation trial ( £§h.86, 8:85, g>.20, one-tailed test): however, this
nonsignificant result was due to the nonhabituators significantly
decreasing their looking time to the novel stimulus ( E‘-°°5 if a
two-tailed test is used). This result is puzzling. The fact that
looking times were higher on trial 10 for the nonhabituators would lead a
to the anticipation of a nonsignificant g-test due to ceiling effects,

but not a significant decline. N

 

For the physiological measures, separate repeated measures ANCOVAs 1
were done comparing the habituation and dishabituation trials for each
group on both the HR variable and the variablility measures. For
habituators, none of the main effects or interactions (for both heart
rate and variability) was significant. For the nonhabituators, the heart
rate analysis showed no significant main effects; nor was the interaction
significant (although the expected main effect for seconds was
significant ( £81.71, 81:17.85, g-.06), but again nonsignificant when
corrected). For the variability measure, none of the fi-tests was

significant.

CHAPTER 5

DISCUSSION

The findings of this study replicated previous results in the
literature that described infant orienting responses to stimulus onset.
Unfortunately, the nesults of the variability tests that were the primary
focus of analysis were inconclusive. Although these results support the
exponentially decaying model of infant habituation superficially, my
Opinion is that drawing this conclusion is unwarranted. The pattern of
habituation in infancy is probably still best considered an unresolved
issue. For both conceptual and methodological reasons, I do not believe
there is any unequivocal way to interpret these data so as to permit a
definitive statement on the true course of habituation. In addition, I
believe that the data collection was influenced by a number of
methodological problems that occurred during the experiment. I feel that
these problems, each of which will be discussed below, essentially

vitiate any conclusions to be drawn from this experiment.

6h

 

65

The results for heart rate orienting responses replicate many prior
studies that have elicited a deceleratory response in infants in response
to a relatively benign stimulus (see Berg 5 Berg, 1979). Similarly, the
decline of this deceleratory response over trials, intuitively
interpretable as habituation of the orienting reflex, also has often been
reported in the literature. The lack of a difference between the
habituation trial and the dishabituation trial for either group is
puzzling, especially in context of the significant difference in looking
behavior by both groups to the dishabituation stimulus. The small sample
size ( 8-6) in each ANCOVA comparing the two trials is most probably the
cause of this nonsignificant difference; nevertheless it weakens the
utility of simple heart rate as an indicator of cognitive process. The
results for orienting, showing a main effect for both seconds and trials,
offer little other than verification of previous studies and an
indication that, based on this one measure, the present sample did not
differ significantly from others that have been used in
psychophysiological paradigms.

The results for the variability response were disappointing and
somewhat unexpected. The nonsignificant trials effect, or trials by
periods interaction for the habituators was especially discouraging, as
both models of habituation predict a change in the response pattern over
trials for the habituators. If the 'tentative' period effect (i.e.,
nonsignificant after conservative L-testing) is assumed to be truly

significant, this would argue that the infants did differ reliably in

 

66

response after stimulus onset and that the variability response was
measuring a response to prolonged stimulus presentation. Therefore, the
lack of change in this response is doubly disappointing, as it would
indicate that when the variability does change it reflects a significant
reaction within trials that does not change as behavioral habituation
occurs. The plot of the mean responses over periods does seem to show a
general decline in variability over the first three periods with an
increase in the fourth period (as the Porges model would predict).
However, the small number of subjects makes any statement tenuous in the
absence of replication. (Although the curve appears to show a
significant habituation by comparison of trial block 9 5 10 with the
earlier trial blocks, it is more an artifact of combining trials 9 5 10
rather than a true difference. Examination of the separate trial means
bears this out.) The nonsignificant difference between the
dishabituation and preceding habituation trial also strongly argues
against the utility of measuring variability as an index of cognitive
processes. The same can be said for the nonsignificant result when the
two trials assumed to be pre- and post- habituation were compared. The
effect for trials was assumed a priori to be present only in the
habituation group, so the negative results for the nonhabituation group
add little to the discussion of habituation in infancy. The significant
interaction of trials by period for nonhabituators (also a 'tentative'
result) was unexpected, and in context of the previous results defies

easy interpretation or categorization.

 

 

‘l.

 

67

As mentioned earlier in this section, the issue of the nature of
habituation is still best considered unresolved: obviously the results
described offer little insight into the comparison of decremental vs.
all-or-none habituation models. The reasons this study has failed to
adequately test its major hypotheses can be broadly dichotomized as
either methodological or conceptual problems. The latter will be
discussed first.

The most disagreeable result for testing the hypothesis concerning
the nature of habituation was the nonsignificant trend over trials of
habituation for the habituation group (evidenced by the nonsignificant
periods by trials interaction). It is difficult to describe the-progress
of habituation using a dependent measure which does not show habituation.
A strong possibility is that the use of variability is a poor measure of
sustained or active attention. Porges et al. (1980), when discussing
spectral analysis, refer to variability as only a "crude measure of
attentional response”. If HR variability does have validity as a measure
of sustained attention, the strength of this effect is not sufficiently
great to yield significant results with the extremely small group sizes
obtained in this study. (Although no formal power calculations were done
post hoc, it is not unreasonable to assume that power with an 8 of is
undesirably low, especially with dependent variables measured with much
error). One possibility often cited to explain behavioral-physiological
dissociation (in the present case, behavioral habituation but no

variability response difference) is the degree of unreliability in the

 

68

measurement of the dependent variables. As reliability of either (or
both) measures decreases, there is less and less expectation that both
will covary appropriately enough to show coincident response patterns in
both systems. This does not appear to be true in the present study, as
no effect at all over trials was noted. A better alternative appears to
be the lack of construct validity in the use of variability as a measure
of sustained response. The support for variability measures as an index
for sustained attention is far from overwhelming, and (especially in
infants) stems almost entirely from Porges's laboratory. Even a recent
paper by Porges et al. (1981) describes variability responses in subjects
contradictory to the predictions of his theory. The choice of this
measure to examine the course of habituation thus appears to have been a
poor one (although the original expectation was to have used spectral
analytic measures as the dependent variable). It does, however, comment
on the possible lack of utility of this variable for monitoring cognitive
responses and should raise questions about its use in future studies. It
must be noted, however, that the description of variability used in this
study is not synonomous methodologically with other uses of this term in
the literature; a h.5 second interval was chosen to subdivide the post
stimulus period into intervals of reasonable size; some of the earlier
work with infants by Porges used longer intervals (approx. 30 seconds;
Porges, Stamps, 5 Walter, l97h), which might be more stable over time and
reflect true tonic changes in variability. (As a check, a post hoc
analysis of variability calculated for the entire 18-second interval over

trials was done for each group. These results here also showed no main

 

69

effect for trials or interactions for either group (see Appendix C). As
such, it appears this study ultimately evolved into more of a test of the
validity of the use of variability as a measure of sustained attention
rather than a true study of habituation. The single analysis that
probably bears most on the utility of the use of variability measures is
the nonsignificant comparison of the habituation and dishabituation
trials. As was noted previously, the nonhabituators and habituators
showed opposite responses to the dishabituation stimulus, with the
nonhabituators showing the odd result of significantly ( g<.01)
decreasing their fixation behavior to the novel stimulus. An appropriate
test of the validity of the variability measure is the comparison of the
last habituation trial and subsequent dishabituation trial. This test
was nonsignificant for the trial and second main effects (and the second
by trial interaction) for each group separately. Again, a nonsignificant
finding is not surprising, given the small sample size ( 8-6 for each
group) and concomitant low power. This is underscored by the F's being
below 1, most likely indicating significant heterogeneity of variance.
Although significant results would have important bearing on the question
of habituation, the nonsignificant results cannot be used to make any
definitive statements due to the many possible causes of this outcome.

From the preceding discussion, it is clear that deriving any
conclusions about the pattern of infant habituation from this study is
problematic. After examination of the important hypotheses, the problem
of interpreting the odd dishabituation result remained. The first

explanation that was considered was that this effect was secondary to

 

 

 

70

age; age-related changes in visual preference during infancy are well
documented (Fantz, Fagan, 5 Miranda, 1975). This was immediately
discounted because the two groups did not significantly differ in age at
time of testing. The second possibility entertained was that the groups
differed on some other individual difference characteristic. No obvious
descriptive characteristic differentiated the groups: this was expected,
since participation in the study was contingent on the infant being
healthy and having no history of pre- or post- natal insult. The next
area of inquiry was analysis of both the baseline and prestimulus
variability. Porges (l97h) has reported differential conditionability in
newborns correlated with baseline variability prior to stimulus onset.
Other studies have been reported describing behavioral differences
between groups differentiated on the basis of autonomic activity (e.g.,
Cousins (1976) and vanHover (l97h) for heart rate; Ingram 5 Fitzgerald
(1975) for skin conductance), but no study has been described replicating
the Porges work with newborns. A 8-test comparing the two groups on the
h.5 second period before the first stimulus period (i.e., during the end
of the baseline period) was suggestive of possible differences between
the two groups ( £Il.h8, 81210, 88.17). As this was too small an
interval to consider differences between the groups as real, a reanalysis
was done comparing the variability between the two groups for the final
10 seconds of the baseline period: the means for the two groups on this

measurement were essentially equivalent ( £-.38, 88-10, g>.25).

 

71

The foregoing has focused primarily on the conceptual problems in
interpreting results from this study. As was previously mentioned, I
feel strongly that methodological problems severely affected the
collection of data. It is necessary at this point to include a full
discussion of these problems in this experiment as I feel that they
influenced the outcome of this study. Moreover, recent ideas from other
branches of psychology that are now being applied in psychophysiology
call for a different way of conceptualizing psychophysiological variables
and change in physiological systems. Although some of these ideas were
incorporated in the proposal for this research, the same methodological
problems obviated any approach to analysis other than the analysis of
variance comparison between groups that was done.

The most obvious problem is a statistical one arising from the small
number of infants in the habituation group. The primary reason for this
low number was the use of a design that did not tie stimulus presentation
to having the infant look at the rear-projection screen where the slides
were displayed. Although various authors have discussed this problem for
over a decade (e.g., Cohen, 1973: Horowitz, 1975), for previously
discussed reasons I decided to present the slide without requiring the
infant to be looking at the screen. The raw data clearly indicated that
this is not a suitable alternative to the more commonly used linking
procedure. Of the 28 subjects whose physiological records were
sufficiently errbr-free to be usable, the literature suggests that 6 is

an underestimate of habituators. Although the six subjects who formed

72

the habituation group (obviously selected before analyses of the
physiological data were accomplished) were felt to be reliable examples
of habituators, the small n makes the power of detecting significant
differences between the pre- and post- habituation trials exceedingly
small. Another reason for the small number of habituators was the small
number of subjects with physiologically acceptable heart rate records.
This is mostly due to circumstances unique to the present study.
Although infant psychophysiological research usually suffers an
experimental loss of approximately A02 of subjects, the present study
suffered a loss of 672 of the subjects tested. In habituation research
the same stimulus is presented repeatedly and habituation results in the
increased likelihood of a state change (or movement artifact) occuring.
This is not as much a problem in studies where different stimuli are
presented that maintain an infant's interest. Also implicit in the
design of a habituation paradigm is that presentations cannot be stopped
and restarted as in other studies where the parameter of interest is the
physiological response to the stimulus.

A further cause of subject loss was the use of the laboratory
computer for timing the interbeat intervals from the electrocardiographic
record. As previously mentioned, this was the first use of the computer
for this task, an almost unbearably tedious job when done by hand. All
the data were collected and stored on audio tape before the programs were
written to perform the analysis. When the programs were used it was
discovered that baseline shifts are an important source of error during

computer analysis of heart rate, and little can be done if they are

 

 

73

already present in the EKG record. To solve this problem, the laboratory
program has been adjusted so that the experimenter has an auditory
feedback signal occuring with each heart beat during the experimental
procedure to indicate that computer analysis of the polygraph record is
acceptable, and if not, the experimenter can adjust the signal recorded
to the appropriate levels. The loss of over l/3 sample completing the
experimental procedure (16 of AA) is unacceptably high for this type of
research: future studies in this laboratory should not suffer as greatly
from this problem. Of the 72 infants in the original sample who came to
the laboratory, AA completed the experimental session. The 612
completion rate is in accordance with previously cited studies (e.g.,
Olson, 1976).

Though the methodological difficulties were severe in the present
experiment, the question of whether the approach taken in this study is
the best for investigating the relationship between psychophysiological
and behavioral events remains salient. In the time since the study was
first proposed, two significant developments in psychophysiology have
occurred. The first is the advocacy of multivariate techniques in
analyzing the relationships both between physiological measures and
between subjects (Kircher, 1981; Porges, 1981). This involves clustering
subjects into groups based either on behavioral or physiological
criteria, and then differentiating between the groups. This method was
applied successfully by McCall (1979) to the same problem that the
present experiment addressed. Although McCall was unsuccessful in

demonstrating a relationship between behavioral and physiological

 

 

7A

responses, conceptually the approach he used holds great promise for
future research in this area. Had there been a sufficient number of
subjects in the present experiment, this would have been the preferred

statistical technique. It allows the distinct advantage of forming

 

statistically post hoc groups of subjects that demonstrated either of the

two contrasting habituation patterns behaviorally (and in McCall's study,

a third group showing idiosyncratic patterns), and then allowing 1_
comparison with physiological measures. These groups are defined solely
by response pattern, thereby avoiding the necessity of forcing subjects
to meet an arbitrary criterion definition of habituation.

A second important advance in psychophysiology has been the use of
time series and spectral analytic procedures for analyzing cardiac data
(for a review, see Kitney 5 Rompelman, 1980). Porges and his associates
(e.g., I976: Porges et al., 1980) appear to be the most ardent American
supporters of this methodology, though they have not yet described any
experimental studies where this approach has been used with infants,
although in a recent paper, Porges (1979) described its potential use for
monitoring fetal distress in high risk deliveries. Richards (1980), in
his review of statistics for analyzing infant heart rate data, also
advocated the potential utility of this and other time series approaches
to infant heart rate analysis. However, the application of this
procedure absolutely requires a minimum number of equidistant points, and
although conventionally used time units can be subdivided to increase the
number of intervals (e.g., l-second units can be subdivided into four 250

msec. units [Porges et al. (1981); but see Wastell (1980)], there still

75

is the requirement of a minimum number of points for analysis, with
increasing reliability as the number of points increases. To apply this
to the present study would require 35 seconds of stimulus presentation to
obtain 150 points, which is much too long for this and most infant
paradigms. However, the application of spectral analysis offers the very
significant practical and conceptual advantage of being able to estimate
the amount of shared variance between the two processes from analysis of
the cross coherence between two rhythmic measures (in this instance,
heart rate and respiration). In Porges's work with hyperactive children
the coherence measure has been used as the dependent variable. This has
been used to test his hypothesis that one of the causes of hyperactivity
is an inappropriate balance of inhibitory neuronal forces that can be
measured as the degree of coherence between respiration and heart rate.
Other applications have used spectral analysis to partial out the
variability in heart rate records due to concurrently occurring
respiratory and blood pressure changes: what remains is a far cleaner
estimate of the variability due to the experimental manipulation,
although in the Porges example, use of coherence directly is appropriate,
since his manipulation (administration of Ritalin) affects both HR and
respiration and especially the relationship between them. For this
reason, Porges has described heart rate variability as only a crude
measure of sustained attention. This is underscored in the discussion of
the general application of spectral analysis by Porges et al. (1980)
where two variables have the same mean and variance but no relationship

in the time domain.

 

76

Given the preceding discussion of both multivariate grouping of
subjects and spectral analytic analysis of cardiac data, there are many
possible ways this approach could be applied profitably to questions
about habituation. One possible study would use a design exactly the
same as that used in this study, except that stimulus presentation would
be linked to viewing the slide image. Subjects would be monitored on
line until they reach a criterion of habituation, and at that point the
computer could be programmed to present the stimulus again. However, on
this trial it would be presented for a longer duration sufficient to
obtain enough points for spectral analysis. These subjects would then be
paired with nonhabituators who would receive the prolonged stimulus
presentation after the equivalent number of trials but without having
habituated. (Obviously a study of this nature would require many
controls and subtleties not mentioned here: this brief paragraph is meant
only to convey that designs using this approach are not inconceivable.
Clearly much research demonstrating the validity of spectral analytic
interpretation of physiological data is needed.)

In summary, while the overall results of this study are
inconclusive, they are entirely congruent with an unpublished thesis by
Ackles (Note I), conducted in the same laboratory which found little
relationship between variability and infant attention. The recent
developments in psychophysiological methodology are allowing more
sophisticated experimental designs in psychophysiological studies.

Perhaps they will shed more light on the nature of habituation. However,

 

 

 

77

more critical at the present juncture are intensive studies to establish
the validity and strength of psychophysiological measurements in a
variety of situations and paradigms: only then can they be used to answer

questions in other areas.

 

 

 

APPENDICES

APPENDIX A

Forms

 

 

 

78

Permission Form for Testing Infant

Date:

Dear Parent(s):

This form is to request permission for us or our staff to examine
your infant in tests of attention, and to observe your infant
during a short period of sleep.

You may withdraw permission at any time simply by informing us or
our staff members that you wish to do so. The information
collected is confidential; it will only be available to qualified
personnel, and information on any individual infant will be
identified by number only. If you have any questions about the
procedure to be used, please feel free to ask. The tests will not
disrupt the infant, or in any way be harmful; however,
participation in the study will not guarantee you or your infant
any beneficial results.

Your signature on this form verifies that the specific tests and
procedures to be used with your infant have been explained to your
satisfaction, and that you have voluntarily agreed to allow us to
test your infant. If at any time you wish to have the data for
your baby withdrawn from the study, simply advise us and we will
destroy all records relevant to your child.

None of the background information will be used in any form, other
than to ascertain that the infant has met the criteria for
participation in this study. Any reference to this information
will also be by number only, after which these records will be
destroyed.

Thank-you again for your help and participation.

Sincerely,

Steven R. Gitterman Hiram E. Fitzgerald
Experimenter Professor of Psychology

Parents signature

 

Experimenters signature

 

 

 

 

79

MICHIGAN STATE UNIVERSITY

Department of Psychology East Lansing, Michigan A882A
Snyder Hall
Developmental Psychobiology Laboratory .ai

Infant Learning Unit

Dear Parent(s):

 

For the past several years we have been studying the young
infant's ability to pay attention to and learn from his or her
environment. Over the next few months. we will be continuing our
work by examining the abilities of three-month old infants to pay
attention to certain picture presentations.

The procedure that will be used is quite simple. The infant

will be seated in an infant seat in front of a movie screen (either
parent. or anyone you wish, will be seated directly next to the
infant throughout this procedure). Your baby will be shown a
picture of a checkerboard for a minimum of 3 minutes, but in no
case longer than 12 minutes (in fact, the responses by your baby will
determine how long the checkerboard appears). While your baby is
watching, we will be monitoring his or her heart rate, breathing,
and whether he or she is looking at the screen. We are interested
primarily in how your infant changes his or her looking behavior
over time, a question that has been asked frequently in infant
research but will be analyzed differently in this particular study.

In order to measure heart rate, it will be necessary to place

three small surface electrodes on your baby's chest. To measure
breathing, we will place a small rubber bellows around the baby's
chest. These standard methods are completely safe and involve no
discomfort to the child. The entire session usually lasts no
longer than an hour. In order to have your baby feel as
comfortable as possible, we ask that one parent remain with the
infant throughout the study. if necessary, there are private areas
where you can change and feed your baby if you wish. If at any
time during the study you desire to stop, we will do so at once.

If you are interested in participating in this project, or learning
more about it before making your decision to participate, please
sign and mail the enclosed card. When we receive a signed
permission card, we will call you to make an appointment and to
answer any additional question you might have. Be assured

 

 

80

that your participation is wholly confidential, and we assign each
participant a number to preserve anonymity. As a result,

parents will be unable to receive information of their own infants
performance, but a summary of the group results will be sent to all
parents in the future. Please feel free to call us at the phone
numbers listed below if you wish any additional information.

Thank you for your time and interest. We look forward to

hearing from you in the near future.

Sincerely,

Steve Gitterman, M.A.
Project Coordinator (353-1651)

Dr. Hiram E. Fitzgerald
Professor of Psychology (353'3933)

SRG/hp

 

 

81

BACKGROUND INFORMATION SHEET

 

The information requested in this form will be used only to
report the general characteristics of the infants used in our
research. Only group results will be reported, and the identity
of individual infants will remain anonymous. All information you
provide on this form will remain strictly confidential.

 

Infant number Stimulus number
Date of testing Time of testing

Tape number and Ice

 

Parents:

Date of birth Sex
mo day year

 

Place of birth

 

 

 

 

 

City or town State
Weight at birth Weight now
lbs oz. lbs. oz.
Height at birth Height now
inches inches

Due date

 

mo day year

82

Any complications during pregnancy 7 If so, please
briefly describe them.

Was any anesthetic used during labor and/or delivery (for example,
local anesthetic, gas, saddle block)? If so, please
describe briefly.

Has your infant received any medication since birth? If so,
please briefly describe.

Is your infant, bottle fed, breast fed, or
some combination, with bottle feeding

275:.__so:._sz.

Is this your first child ? ___. If no, how many other children
do you have .

Any special problems with ( )colic ( )rashes ( )feeding ( )sleeping.

If so, please briefly describe.

Is there anything else special about your infant that you feel it
would be important for us to know about for this project 7
If so. please describe.

Thank you very much for your time.

 

 

 

 

APPENDIX 8

Subject Characteristics

 

Habituators

Subject number

aunt-mm“

Subject

O‘U’IL‘WN—a

age(days)

119
118

92
1111
157
131

Non-habituators

number

age(days)

107
167
1A6
113
122
102

Sex

Sex

3313“:

83

 

"I-

 

 

8A

Other subjects

 

Subject number age(days) Sex
1 128 M
2 137 M
3 1A8 M
A 109 F
5 11A M
6 1A3 F
7 1A9 M
8 11A F
9 118 M

10 119 M
11 113 F
12 11A F
13 1A8 F
1A 1A1 M
15 109 M
16 113 F

 

 

 

APPENDIX C

Analysis of variance tables

 

85

TABLE 1

 

Analysis of Covariance of heart rate response to stimulus onset
over trials for all 28 subjects.

 

---------------------------------------------------------- *------**---
Source df MS F p con. df p
Trials 9 622.088 1.86 .06 1,8 .20
error 2A2335.17l

Seconds 17 190.211 6.73 <.001 1,17 .02
error A59 28.262

T by S 153 26.2AA 1.12 .15

error A131 23.A63

 

*con. df - degrees of freedom for conservative test of effect
**probability of conservative test I

86

TABLE 2

Analysis of covariance of variability response to stimulus onset
for all 28 subjects over trials.

Source df MS F p con df p
Trials 9 1.197 .87 .55

error 2A2 1.371

Periods 3 3.101 3.8A .01 1,3 .17
error 81 .807

T by P 27 .969 1.16 .26

error 729 .835

37

TABLE 3

T-test comparing habituation and nonhabituation groups on average
of two longest fixations during first five trials.

 

T-TEST
Variable number mean Standard Standard error
deviation
habituators 6 9.8 A.Al 1.80
nonhabituators 6 15.5 2.23 .87

pooled variance estimate
T-value df 2-tail prob.

separate variance estimate
T-value df 2-tai1 prob.

 

88

TABLE A

 

T-test comparing habituation and nonhabituation groups
on duration of first fixation time.

 

T-TEST
Variable number mean Standard Standard error
deviation
habituators 6 8.82 6.62 2.51
nonhabituators 6 10.73 3.38 1.5A
pooled variance estimate separate variance estimate
T-value df 2-tail prob. T-value df 2-tail prob.

 

89

TABLE 5

Analysis of variance comparing variability response for prestimulus
and poststimulus seconds on trial 1 by habituation group.

Source df MS F p con. df p
Group 1 1.02A .53 .A8

error 10 1.918

Period 1 .007 .00 .96

P by G 1 .982 2.93 .12

error 10 .33

 

 

 

90

TABLE 6

Analysis of variance comparing heart rate response for
prestimulus and poststimulus seconds on trial I by
habituation group.

Source df MS F p con. df p
Group 1 51.0A1 .16 .70

error 10 326.229

Period 1 .327 .03 .87

P by G 1 15.360 1.2A .29

error 10 12.A28

 

 

 

91

TABLE 7

 

Analysis of covariance for heart rate responses to stimulus
onset by habituation groups over trials and over seconds.

 

 

Source df MS F p con df p
Group 1 I81 39A 31 59

error 9 583.851

Trials 9 730.0A0 l.A2 .19

T by G 9 12A.732 .2A .98

error 89 515.310

Seconds 17 79.272 ' 2.51 <.01 1,17 .15
S by G 17 A0.236 1.27 .22

T by S 153 3A.560 1.18 .07 9,17 >.25
T by S by G 153 19.508 .66 .99

error 1530 29.3A3

92

TABLE 8

Analysis of covariance of heart rate responses on trial preceding
and trial of habituation for habituation group.

Trials
error

Period
error

T by P
error

15

15

F p con. df p
.00 .95
.2A .86
.19 .90

"I.

 

 

93

TABLE 9

Analysis of covariance of variability response for habituation
group to stimulus onset over trials and over periods.

 

 

Source df MS F p con. df p
Trials 9 1.959 1.31 .26

error AA l.A92

.B l

Periods 3 1.579 2.95 .07 1,3 .21
error 15 .535

T by P 27 1.1A2 1.25 .20

error 135 .91

9A

TABLE 10

 

Analysis of covariance for nonhanbituation group variability
response to stimulus onset over trials and over periods

 

Source df MS F p con df p
Tr1als 9 .626 39 .93

error AA 1 622

Period 3 .321 62 .61

error 15 .517

T by P 27 1.013 1.57 .05 3,9 >.25

error 135 .6AA

 

95

TABLE 11

 

T-test comparing habituators fixation response on last
habituation trial and dishabituation trial.

 

T-TEST
Variable number mean Standard Standard error
deviation
last hab. trial 6 3.55 2.81 1.76
dishabituation 6 5.30 A.2A 1.73
Difference Standard Standard cor T-value df 2-tail prob.

deviation error

 

96

TABLE 12

T-test comparing nonhabituators fixation response on last
habituation trial and dishabituation trial

 

T-TEST
Variable number mean Standard Standard error
deviation
last hab. trial 6 11.33 A.055 1.656
dishabituation 6 8.25 A.AA6 1.815
Difference Standard Standard corr. T-value df 2-tail prob.

deviation error

97

TABLE 13

Analysis of covariance comparing habituators heart rate response
on last habituation and dishabituation trials.

Source df MS F p con. df p
Trials 1 99.A93 60 A8

error A I65.A20

Seconds 17 7.535 .25 .99

error 85 29.536

T by S 17 13.522 .30 .99

error 85 AA.866

 

 

98

TABLE 1A

Analysis of covariance comparing habituators variability response
on last habituation trial and dishabituation trial.

Source df MS F p con. df p
Trials 1 .011 00 .99

error A 2.019

.B 1

Periods 3 .303 .2A .86

error 5 2 218

T by P 3 .125 .19

error 15 .6A8

 

 

99

 

TABLE 15

Analysis of covariance comparing nonhabituators heart rate
response on last habituation and dishabituation trials.

 

Source df MS F p con. df p
Trials 1 203.819 .A2 .55

error A A88.89O

.B 1

Seconds 17 A2.930 1.71 .06 1,17 .22
error 85 25.090

T by S 17 16.32A .55 .91

error 85 29.660

100

TABLE 16

Analysis of covariance comparing nonhabituator variability
response on last habituation and dishabituation trials.

Source df MS F p con. df p
Trials 1 .059 .08 .79

error A .722

Periods 1 .199 .28 .8A

error 15 .721

T by P 3 .822 2.02 .15

error 15 .A06

101

TABLE 17

 

Analysis of variance for groups over trials with the variability
of the entire 18 second stimulus presentation as dependent

 

variable.
Source df MS F p con df p
Group 1 .A07 .IA .71
error 10 2.966
Trials 9 .A08 .85 .57
T by c 9 .1106 .85 .58

error 90 .A80

 

 

 

 

L I ST OF REFERENCES

 

REFERENCE NOTES

Ackles, P.K. A developmentgl study 2: visual attention and
respiratory sinus grrythmia'ig young human infants. Unpublished

doctoral dissertation, Michigan State University, 1980.

Gitterman, S.R. A psychophysiological study 21 attention in
early infancy. Unpublished M.A. thesis, Michigan State

University, 1979.

102

 

 

REFERENCES

Barry, R.J. A factor analytic examination of the unitary OR

concept. Developmental Psychology, 1979. 8. 161-178.

Bayley. N. Bayley scales 21 infant 8eyelopment. New York: The

Psychological Corporation, 1969.

Berg, W.K. Cardiac orienting responses of 6- and l6-week-old

infants. Journal 8: Exgerimental Child Psychology, 197A,.lz.
303-312.

Berg, W.K., 5 Berg. K.M. Psychophysiological development in
infancy: State, sensory function, and attention. In J.D.

Osofsky (Ed.), Handbook 21 infant 8evelopment. New York: Wiley,

1979-

Blumenthal, J.A., Williams, R.B., Kong, Y., Schanberg, S.M., 5
Thompson, L.W. Type A behavior pattern and atherosclerosis.

Circulation, 1978, 58, 63A-6AA.

Bock, R.D. Multivariate behavior analysis. New York: McGraw-Hill,

1975-

103

 

 

 

10A

Bower, T.G.R. Develggment 18 infancy. San Francisco: Freeman, l97A.

Brackbill, Y. The role of the cortex in orienting: OR in

anencephalic human infant. Develogmental Psychology, 1975. 5.

 

195-201.

Brackbill, Y., 5 Fitzgerald, H.E. Development of the sensory

analyzers during infancy. In L.P. Lipsitt 5 H.W. Reese (Eds.),

 

Advances lfl ch118 development and behavior (Vol. A). New York:

Academic Press, 1969.

Brooks, P.R., 5 Berg, W.K. Do l6-week-old infants anticipate

stimulus 7 Develgpmentgl Psychobiology, I979, 18,329-33A.

Campos, J.J. Heart rate: a sensitive tool for the study of

emotional development in the infant. In L. Lipsitt (Ed.),

Develgpmental Psychobiology: The significance 2E infancy.

Hillsdale, N.J. : Lawrence Erlbaum Associates, 1976.

Campos, J.J., Langer, A., 5 Krowitz, A. Cardiac responses on the

visual cliff in prelocomotor human infants. Science, 1970,

.1129 196-197-

Cannon, W.B. Organization for physiological homeostasis.

Physiological Review, 1929, 9, 399-A31.

 

105

Carroll, D., 5 Anastasiades, P. The behavioural significance of
heart rate: the Laceys' hypothesis. Biological Psychology,
1978. Z. 2A9-275.

Clifton, R.K. Cardiac conditioning and orienting in the infant. In
P.A. Obrist, A.H. Black, J. Brener, 5 L.V. DiCara (Eds.),

Cardiovascular psychoghysiology: Current issues 18 resgonse

 

mechanisms, biofeedback, and methodology. Chicago: Aldine,

197A.

Clifton, R.K., 5 Nelson, M.N. Developmental study of habituation in
infants: the importance of paradigm, response system, and
state. In H.V.S. Peeke 5 M.J. Herz (Eds.), Habituation:
Persgectives from child develogment, 881881 behavior, 888
neuroghysiology. Hillsdale, N.J.: Lawrence Erlbaum Associates,
1976.

Cohen, L.B. A two process model of infant attention.

Merrill-Palmer Quarterly, 1973, 19, 157-180.

Cohen, L.B. Habituation of infant visual attention. In T.J. Lighe
5 R.N. Leaton (Eds.), Habitggtion: Persgectives from child

develogmentI animal behaviorI 888 neuroghysiology. Hillsdale,

N.J.: Lawrence Erlbaum Associates, 1976.

 

 

 

 

106

Cohen, L.B., 5 Gelber. E.R. Infant visual memory. In L.B. Cohen 5
P. Salapatek (Eds.), Infant perception: From sensition £2

cognition (Vol. I). New York: Academic Press, 1975.

Cohen, L.B., 5 Menten, T.G. The rise and fall of infant
habituation. Infant Behavior and Development, 1981, 8,

269-280.

Coll, C.G., Spekowski, C., 5 Lester, B.M. Cultural and biomedical
correlates of neonatal behavior. Developmentgl Psychobiology,

1981, 18, lA7-15A.

Cousins, L.R. Individual differences in the orienting reflex and
children's discrimination learning. Psychophysiolggy, 1976,

13. “79-1187 0

Deloache, J.S. Rate of habituation and visual memory in infants.

Child Develogment, 1976, 81, lA5-15A.

Dixon, W.J., 5 Brown, M.B. Biomedical comguter grogramsI g-series.

Berkeley: University of California Press, 1979.

Elliot, R. The motivational significance of heart rate. In P.A.
Obrist, A.H. Black, J. Brener, 5 L.V. DiCara (Eds.),
Cardiovascular gsychoghysiology: Current 1&2222 jg resgonse
mechanisms, biofeedggck, 888 methodology. Chicago: Aldine,
197A.

 

 

107

Fantz, R.L., Fagan, J.F., III, 5 Miranda, S.B. Early visual
selectivity. In L.B. Cohen 5 P. Salapatek (Eds.), Infant
perception: From sensation £2 cognition (Vol. I). New York:

Academic Press, 1975.

Field, T.M., Dempsey, J.R., Hatch, J., Ting, G., 5 Clifton, R.K.
Cardiac and behavioral responses to repeated tactile and
auditory stimulation by preterm and term neonates.

Developmental Psychology, 1979, 15, A06-Al6.

Fitzgerald, H.E., 5 Brackbill, Y. Classical conditioning in

infancy. Psychological Bulletin, 1976, 8 , 353-376.

Friedman, S. Habituation and recovery of visual response in the
alert human newborn. Journal 21 Exgerimental Psychology, 1972,
ll. 339-3A9-

Ginsberg, H., 5 Opper, S. Piaget's theory 2: intellectual

develogment. Englewood Cliffs, N.J.: Prentice-Hall, 1979.

Graham, F.K. Habituation and dishabituation of responses innervated
by the autonomic nervous system. In H.V.S. Peeke 5 M.J. Herz
(Eds.), Habituation: Behavioral studies (Vol. I). New York:

Academic Press, 1973.

Graham, F.K. Constraints on measuring heart rate and period
sequentially through real and cardiac time. Psychoghysiology,
1978(3). 15. A92-A95.

 

 

108

Graham, F.K. Normality of distribution and homogeneity of variance
of heart rate and heart period samples. Psychophysiology,

1978(b), 15, A87-A91.

Graham, F.K. Distinguishing among orienting, startle, and defensive
cardiac responses. In H. Kimmel, E.H. van Olst. 5 J.F.
Orlebeke (Eds.), The orienting reflex 12 humans. Hillsdale,

N.J.: Lawrence Erlbaum Associates, 1979.

Graham, F.K., 5 Clifton, R.K. Heart rate change as a component of
the orienting response. Psychological Bulletin, 1966, _5,
305-320.

Graham, F.K., 5 Jackson, J.C. Arousal systems and infant heart rate
responses. In H.W. Reese 5 L.P. Lipsitt (Eds.), Advances 18
child development and behavior (Vol. 5). New York: Academic

Press, 1970.

Graham, F.K., Leavitt, L.A., Strock, B.D., 5 Brown, J.W. Precocious

cardiac orienting in a human, anencepahlic infant. Science,

1978, 199. 322-325.

Green, J. A review of the Laceys' physiological hypothesis of heart

rate change. Biological Psychology, 1980, 11, 63-80.

Greenhouse, S.W., and Geisser, S. On methods in the analysis of

profile data. Psychometrika, 1959, 88, 95-112.

 

 

109

Groves, P.M., 5 Thompson, R.F. Dual process theory of habituation:
Neural mechanisms. In H.V.S. Peeke 5 M.J. Herz (Eds.),

Habituation: Physiological substrates (Vol. II). New York:

 

Academic Press, 1973.

Hahn, W.W. The hypothesis of Lacey: a critical appraisal.

Psychological Bulletin, 1973, 19, 59-70.

 

Harper, R.M., Leake, B., Hoppenbrowers, T., Sterman, M.B., McGinty,
D.J., 5 Hodgman, J. Polygraphic studies of normal infants and
infants at risk for the sudden infant death syndrome: Heart
rate and variability as a function of state. Pediatric

Research, 1976, 1;, 778-785.

Heselgrave. R.J.. Ogilvie, J.C., 5 Furedy, J.J. Measuring
baseline-treatment differences in heart rate variability:
Variance versus succesive difference mean square and beat per
minute versus interbeat interval. Psychoghysiology, 1979, 18,
151-157.

Hinde, R.A. Behavioural habituation. In G. Horn 5 R.A. Hinde
(Eds.), Short-term changes 18 neural activity 888 behavior.

London: Cambridge University Press, 1970.

Horowitz, F.D. (Ed.), Visual attention, auditory stimulation, and

language discrimination in young infants. Monograghs 21 the

Society 18; Research jg_Child Develogment, I975, 39, Nos. 5-6.

110

Ingram, E. 5 Fitzgerald, H.E. Individual differences in infant
orienting and autonomic conditioning. Develogmental

Psychobiology, l97A. 1, 359-367.

James, W. The grincigles 21 gsychology (Vol. I). New York: Dover

Publications, 1950. (Originally published 1890.)

Jeffrey, W.E. The orienting reflex in cognitive developmnet.

Psychological Review, 1968, 19, 323-33A.

Jeffrey, W.E. Habituation as a mechanism for perceptual
development. In T.J. Lighe 5 R.N. Leaton (Eds.), Habituation:
Perspectives ILQE £2119 develogmentI 221221 behavior, 888
aggrophysiology. Hillsdale, N.J.: Lawrence Erlbaum Associates,

1976.

Jeffrey, W.E. 5 Cohen, L.B. Habituation in the human infant. In
H.W. Reese (Ed.), Advances jg child develoggent and behavior

(Vol. 6). New York: Academic Press, 1971.

Jennings, J.R., 5 Wood, C.C.‘ The e-adjustment procedure for
repeated measures analysis of variance. Psychophysiology.

1976. 11. 277-278.

Keppel, G. Design 888 analysis: A researchers handbook. Englewood

Cliffs, N.J.: Prentice Hall, 1973.

 

 

111

Kimmel, H.D., van Olst, E.H., 5 Orlebeke, J.F. (Eds.), 188
orienting reflex jg humans. Hillsdale, N.J.: Lawrence Erlbaum

Associates, 1979.

Kintsch, W. Memory and decision aspects of recognition learning.

Psycholggical Review, 1967, 18, A96-50A.

Kircher, J.C. Structured panel discussion: New trends of
quantification and analysis in psychophysiological research.

Psychophysiology, 1982, 19, 297. (Abstract)

Kitney, R.I., 5 Ropelman, O. The study f heart rate variability.

 

 

Oxford: Clarendon, I980.

Lacey. B.C., 5 Lacey, J.I. Studies of heart rate and other bodily
processes in sensorimotor behavior. In P.A. Obrist, A.H.
Black, J. Brener, 5 L.V. DiCara (Eds.), Cardiovascular
ggychophysiologyi Current issgg§,jg resgonse mechanisms,

biofeedbackI 888 methodology. Chicago: Aldine, 197A.

Lacey, B.C., 5 Lacey, J.I. Two way communication between the heart

and the brain: Significance of time within the cardiac cycle.

American Psychologis , I978, 33, 99-113.

Lacey, J.l. Somatic response patterning and stress: Some revisions
of activation theory. In M.N. Appley 5 R. Trumbull (Eds.),
Psychological stress: Issues 12 research. New York: Appleton

Century-Crofts, 1967.

112

Lacey, J.I., 5 Lacey, B.C. Some autonomic-central nervous system
relationships. In A. Black (Ed.), Physiological correlates 21

emotion. New York: Academic Press, 1970.

Lester, B.M. Cardiac habituation of the orienting response to an
auditory signal in infants of varying nutritional status.

Developmental Psychology, 1975, ll, A32-AA2.

Lewis, M. Individual differences in the measurement of early
cognitive growth. In J. Hellmuth (Ed.), The exceptional infant

;; Studies n abnormalities. New York: Brunner/Mazel, 1971.

Lubow,R.E., Weiner, I., 5 Schnur, P. Conditioned attention theory.
In G.H. Bower (Ed.), The gsychology 2: learning and motivation:
Advances in research and theory (Vol. 15). New York: Academic

Press, 1981.

Maurer, D., 5 Lewis, T.L. Peripheral discrimination by

three-month-old infants. Child Develoggent, 1979, 59, 276-279.

McCall, R.B. Attention in the infant: Avenue to the study of
cognitive development. In D.W. Walcher 5 D.L. Peters (Eds.),

Early childhood: IDS develoggent g: self-regulatory mechanisms.

New York: Academic Press, 1971.

McCall, R.B. Individual differences in the pattern of habituation

at 5- and 10- months of age. Develogmental Psychology, 1979.
15. 559-569-

 

 

113

McCall, R.B., 5 Applebaum M.I. Bias in the analysis of repeated
measure designs: Some alternative approaches. Child.

Develgpment, 1973, 89, ADI-A15.

McCall, R.B. 5 McGhee, P.E. The discrepancy hypothesis of
attention and affect in infants. In |.C. Uzgiris 5 F. Weizmann
(Eds.), 188 structuring 8: exgerience. New York: Plenum Press,
1977-

Mulder, G. 5 Mulder, L.J.M. Information processing and

cardiovascular control. Psychoghysiology, I981, 18, 392-A02.

Nie, N.H., Hull, C.H., Jenkins, J.C., Steinbrenner, K., 5 Bent, D.H.

Statistical package for the social sciences. New York:

 

McGraw-Hill, 1975.

Orlebeke, J.F., 5 Feij, J.A. The orienting reflex as a personality
correlate. In H. Kimmel, E.H. van Olst, 5 J.F. Orlebeke
(Eds.), IDS orienting reflex jg humans. Hillsdale, N.J.:

Lawrence Erlbaum Associates, 1979.

Obrist, P.A. Cardiovascular psychophysiology: A gersgective. New

York: Plenum, 1981.

Obrist, P.A., Black, A.H., Brener, J., 5 DiCara, L.V. (Eds.),
Cardiovascular psychophysiology; Current issues lg resgonse

mechanisms, biofeedback, 888 methodology. Chicago: Aldine,

197A.

 

11A

O'Gorman, J.C. Change in stimulus conditions and the orienting

response. Psychophysiology, 1973, 19, A65-A70.

Ohman, A. The orienting response, attention, and learning: An
information-processing perspective. In H. Kimmel, E.H. van
Olst, 5 J.F. Orlebeke (Eds.), The’orienting reflex 18 humans.

Hillsdale, N.J.: Lawrence Erlbaum Associates, 1979.

Olson, G.M. An information-processing analysis of visual memory and
habituation in infants. In T.J. Lighe 5 R.N. Leaton (Eds.),
flabituation: Perspectives ELEE chil8 develogment, animal
behavior, a88 neurophysiolggy. Hillsdale, N.J.: Lawrence

Erlbaum Associates, 1976.

O'Neill, M.T., 5 Hinton, J.W. Cardiac variability and attention:

the possible effects of test situational adaptation.

Biological Psychology, 1979, 9, 151-15A.

Porges, S.W. Heart rate indices of newborn attentional

responsitivity. Merrill-Palmer Quarterly, 197A, 89, 23l-25A.

Porges, S.W. Peripheral and neurochemical parallels of
psychopathology: A psychophysiological model relating autonomic
imbalance to hyperactivity, psychopathy, and autism. In H.W.
Reese (Ed.), Advances lg 88118 develogment ag8 behavior (Vol.

11). New York: Academic Press, 1976.

 

 

 

115

Porges, S.W. The application of spectral analysis for the detection
of fetal distress. In T.M. Field, A.M. Sostek, S. Goldberg. 5

H.H. Shuman (Eds.), Infants born _l risk. New York: Spectrum,

 

 

1979-

Porges, S.W. Structured panel discussion: New trends of
quantification and analysis in psychophysiological research.

Psychophysiology, 1982, l9, 297. (Abstract)

Porges, S.W., Arnold, W.R., 5 Forbes, E.J. Heart rate variability:
An index of attentional responsitivity in human newborns.

Developmental Psychology, 1973, 8, 85-92.

Porges, S.W., Bohrer, R.E., Cheung, M.N., Drasgow, F., McCabe, P.M.,
5 Keren, G. New time series statistic for detecting rhythmic
co-occurence in the frequency domain: The weighted coherence
and its application to psychophysiological research.

Psychological Bulletin, 1980, 88, 580-587.

Porges, S.W., Bohrer, R.E., Keren, G., Franks, G.J., 5 Drasgow, F.
The influence of methylphenidate on spontaneous autonomic

activity and behavior in children diagnosed as hyperactive.

Psychoghysiology, 1981, l8, A2-A8.

 

116

Porges, S.W., Coles, M.G.H., Cheung, M.N., Drasgow, F., 5 Bohrer, R.
Individual differences in cardiac-respiratory coupling and
heart responses during two attention demanding tasks.
Psychophysiolggy, 1979, l8, 179. (Abstract)

Porges, S.W., 5 Humphrey, M.M. Cardiac and respiratory responses

during visual search in nonretarded adolescents. American

Journal al Mental Deficiency, 1977, 8;, 162-169.

Porges, S.W., 5 Raskin, D.C. Respiratory and heart rate components

of attention. Journal 21 Exgerimental Psychology, 1969, 8l,
997'503-

Porges, S.W., Stamps, L.E., 5 Walter, G.F. Heart rate variability
and newborn heart rate responses to illumination changes.

Developmental Psychology, l97A. lg, 507-513.

Porges, S.W., Walter, G.F., Korb, R., 5 Sprague, R.L. The
influences of methylphenidate on heart rate and behavioral

measures of attention in hyperactive children. Child

Develogment, I975, 58, 727-733.

Pribram, K.H. The orienting reaction: Key to brain
re-representational mechanisms. In H. Kimmel, E.H. van Olst, 5
J.F. Orlebeke (Eds.), 188 orienting reflex lg humans.

Hillsdale, N.J.: Lawrence Erlbaum Associates, 1979.

117

Ratner, S.C. Habituation: Research and theory. In J.H. Reynierse,
Current issues lg animal learning: A colloguium. Lincoln:

University of Nebraska Press, 1970.

Richards, J.E. The statistical analysis of heart rate: A review

emphasizing infancy data. Psychophysiology, 1980, ll, 153-166.

Sameroff, A.J. Can conditioned responses be established in the

newborn infant: 1971? Develogmental Psychology, 1971, 5, 1-12.

Sameroff, A.J., Cashmore, T.F., 5 Sykes, A.C. Heart rate

deceleration during visual fixation in human newborns.

Develogmental Psychology, 1973, 8, 117-119.

Sayers, B.M. Signal analysis of heart rate variability. In R.I.

Kitney 5 O. Rompelman (Eds.), 122 study 2: heart rate

 

 

variability. Oxford: Clarendon Press, 1980.

Schexnider, V.Y., Bell, R.Q., Shebilske. W.L., 5 Quinn, P.
Habituation of visual attention in infants with minor physical

anomalies. Child Develogment, 1981, 5;. 812-818.

Schwartz, M. 5 Day, R.H. Visual shape perception in early infancy.

Monograghs EEC £22 Society IQL Research 10 Child Develoggent,
1979, 58(7, Serial no. 182).

 

118

Shepard, G.N. The sygaptic organization a: the brain. New York:

 

Oxford University Press, 1979.

Sherrington, C.S. The integrative action 21 the nervous system. New
Haven, Ct.: Yale University Press, 1961. (Originally published

1906.)

Shimizu, H. Reliable and precise identification of R-waves in the

EKG with a simple peak detector. Psychoghysiology, 1978, l5,
A99-501.

Siddle, D.A. The orienting response and stimulus significance: Some

comments. Biological Psychology, 1979, 8, 303-309.

Sokolov, E.N. Perception and the conditione8 reflex. New York:

Macmillan, 1963.

Sokolov, E.N. Modeling properties of the nervous system. In I.
Maltzman 5 M. Cole (Eds.), Handbook 21 contemgorary Soviet

gsychology. New York: Basic Books, 1969.

Stamps, L.E. Temporal conditioning of heart rate responses in

newborn infants. Develogmental Psychology, 1977, l5, 62A-629.

Stamps, L.E. The relationship between heart rate indices of the

orienting response and birth weight in normal full-term

newborns. Develogmental Egychobiolggy, 1980, l3, 33-35.

 

119

Stamps, L.E., 5 Porges, S.W. Heart rate conditioning in newborn
infants: Relationships among conditionability, heart rate
variability, and sex. Develgpmental Psychology, 1975, ll,
A2A-A31.

Svebak, S., Dalen, K., 5 Storfjell, 0. The psychological
significance of task-induced tonic changes in somatic and

autonomic activity. Psychophysiology, 1981, l8, AO3-A09.

Thompson, R.F., Berry, S.D., Rinaldi, P.C., 5 Berger, T.W.
Habituation and the orienting reflex: The dual process theory
revisited. In H. Kimmel, E.H. van Olst, 5 J.F. Orlebeke
(Eds.), IDS orienting galls; l8 hamaga. Hillsdale, N.J.:

Lawrence Erlbaum Associates, 1979.

Thompson, R.F., 5 Glanzman, D.L. Neural and behavioral mechanisms
of habituation and sensitization. In T.J. Lighe 5 R.N. Leaton
(Eds.), Habituation: Perspectives item 88ll8 development,
aglgal behavior, a88 neuroghysiology. Hillsdale, N.J.: Lawrence

Erlbaum Associates. 1976.

Thompson, R.F., 5 Spencer, W.A. Habituation: A model phenomenon for
the study of neuronal substrates of behavior. Psychological

Review, 1966, 13, 16-A3.

120

van Hover, K.I. A developmental study of the three components of

attention. Develgpmental Psychology, 197A, lg, 330-339.

Velden, M. Some necessary revisions of the neuronal model concept
of the orienting response. Psychophysiology, 1978, l5,

181-185.

Velden, M., 5 Schmacher, M. Orienting and defensive cardiac
responses. In H. Kimmel, E.H. van Olst, 5 J.F. Orlebeke
(Eds.), The orienting reflex l8 humans. Hillsdale, N.J.:

Lawrence Erlbaum Associates, 1979.

Wagner, A.R. Priming in STM: An information-processing mechanism
for self-generated or retrieval-generated depression in
performance. In T.J. Lighe 5 R.N. Leaton (Eds.), Habituation:
Perspectives from 88ll8 8gvelopment, aglmal behavior, a88
aaarophysiology. Hillsdale, N.J.: Lawrence Erlbaum Associates.

1976.

Walker, 8.8., 5 Sandman, C.A. Human visual evoked potentials are

related to heart rate. Journal 21 Comgarative and

Physiological Psychology, 1979, 95, 717-729.

Wastell, D.G. Measuring heart rate variability: Some comments on
the successive difference mean square statistic.

Psychophysiology, 1971, 8, 6A8-655.

 

121

Welch, M.M. Infants visual attention to varying degrees of novelty.

Child Development, l97A. 88, Al6-A2A.

Winer, B.J. Statistical principles l8 experimental 8esig8. New

York: McGraw-Hill, 1971.

Zeaman, D., 5 House, B.J. The role of attention in retardate
discrimination learning. In N.R. Ellis (Ed.), Handbook gl

mental deficiency. New York: McGraw-Hill, 1963.

Zeiner, A.R. Individual difference in orienting response magnitude
as related to academic performance. In H. Kimmel, E.H. van
Olst, 5 J.F. Orlebeke (Eds.), IDS orienting reflex l8 humans.

Hillsdale, N.J.: Lawrence Erlbaum Associates, 1979.