'--

PSYCHOPHYSIOLOGY

Copyright @ 1989 by The Society for Psychophysiological Research, Inc.

Vol. 26, No.4Printed in U.S.A.

Development and Stability in Visual Sustained Attentionin 14, 20, and 26 Week Old Infants

JOHN E. RICHARDS

Department of Psychology, Universityof South Carolina. Columbia. South Carolina

ABSTRACT

Infants were studied at 14, 20, and 26 weeks of age in a longitudinal design. They were presentedwith varying and complex patterns on a TV screen. Two-thirds of the presentations were accom-panied by a stimulus in the periphery delayed in time from the onset of fixation on the centralstimulus. As in previous research, the infants were not as easily distracted by the interruptingstimulus when the presentation occurred at the point of maximal heart rate deceleration as whenthe presentation occurred at the end of the heart rate response. Infants with large amounts ofrespiratory sinus arrhythmia (i.e., heart rate variability) in a baseline recording were less distractibleduring the deceleration-defined trials than were infants with low amounts of respiratory sinusarrhythmia. Intra-individual patterns of development in respiratory sinus arrhythmia over thetesting ages were closely paralleled by patterns of heart rate responding during sustained attention.Individual differences in baseline levels of heart rate and respiratory sinus arrhythmia were morestable than individual differences in sustained attention. The stability of attention responses overage may be mediated by the stability of the physiological system (e.g., heart rate, respiratory sinusarrhythmia, etc.), and by the within-age relation of attention to heart rate variability.

DESCRIPTORS: Attention, Heart rate, Infants, Respiratory sinus arrhythmia.

The stability over age of infant attention has re-cently become an important research issue. Onereason for this concern is that measures of atten-tion, habituation, and memory are correlated withcognitive or intellectual outcome in later childhood(e.g., Bornstein & Sigman, 1986;Cohen & Parme-lee, 1983;Fagan & McGrath, 1981;Fagan & Singer,1983; Lewis & Brooks-Gunn, 1981; Rose & Wal-lace, 1985). However, the predictive ability of ameasure is limited by its reliability. One measureof reliability is the test-retest correlation. Test-retestcorrelations of infant attention measures over bothshort (e.g.,days) and longer (1 to 3 months) periodsof time are low to moderate in level (e.g., Bornstein& Benasich, 1986;Colombo, Mitchell, O'Brien, &Horowitz, 1987;Rose, Feldman, & Wallace, 1988),or nonsignificant (Colombo et aI., 1987;Maitre, la-marre, & Pomerleau, 1987; McCall, 1979). Theselow test-retest correlations indicate that infant at-

This research was supported by a Social and Behav-ioral Sciences Research Grant from the March of DimesBirth Defects Foundation, #12-157, and a New Investi-gator Research Award from the National Institute of ChildHealth and Human Development, #R23 HD18942.

Address requests for reprints to: John E. Richards, De-partment of Psychology, University of South Carolina,Columbia, SC 29208.

tention measures have low reliability. The observedcorrelations between infant attention and child-hood outcome may be attenuated by the low levelof reliability (Bornstein & Sigman, 1986).

A second reason for the interest in the stabilityof attention in infancy is based on a theoreticalmodel of infant sustained attention (Porges, 1976,1980). Multiple components of infant attentionmay be distinguished, two of which are reactive andsustained attention. Reactive attention is an in-voluntary, phasic response to the onset of a stim-ulus, and lasts for 5 or 6 seconds. Sustained atten-tion is a voluntary, active response to stimulation,which begins after reactive attention and facilitatesinformation and stimulus processing. Porgesarguesthat the cholinergic nervous system activity is im-portant in the inhibition of peripheral body activity,which is a critical component of voluntary, sus-tained attention. An indirect index of cholinergicactivity is respiratory sinus arrhythmia, which isthe variability in heart rate that occurs at the samefrequency as breathing, and occurs due to vagal par-asympathetic activity, mediated via cholinergicmechanisms. Individual differencesin sustained at-tention should therefore be related to individualdifferences in level of respiratory sinus arrhythmia,because the latter is indexing a functional charac-

422

~~~....

July,1989 Development/Stability in Infant Sustained Attention 4~

teristic of the central nervous system that is in-volved in the control of sustained attention.

There are substantial individual differences ininfant sustained attention within a testing age(Richards, 1988). If these are stable characteristicsof the infant's attentional response, then thereshould be significant test-retest correlations be-tween sustained attention measures at two ages. Arecent report by Ruff, Lawson, and Parinello (1987)found evidence for the stability of sustained atten-tion. The latency to touch small, graspable objects,and the amount of time spent examining those ob-jects, were measured in infants at 12 and 24 monthsof age. There was a correlation between object ex-amination time at 12 and 24 months, but not. be-tween touch latency over these months of age. De-velopmental changes in object examination timeand touch latency differ. Latency to touch an objectdecreases with age, whereas object examinationtime does not (Ruff, 1986). Object examinationtime may be similar to sustained attention, becauseit involves the active encoding of stimulus infor-mation, whereas touch latency may be more similarto a reactive phase of attention, such as stimulusorienting. The results of these two studies implythat the developmental changes and individual dif-ferences affecting attention differ for the sustainedand reactive phases of attention.

However, stability in the measures of sustainedattention may depend on the level ofintra-individ-ual stability in respiratory sinus arrhythmia (RSA).Over the first 12 months of life, heart rate varia-bility and RSA levels increase (Harper, Hoppen-brouwers, Sterman, McGinty, & Hodgman, 1976;Harper et aI., 1978; Katona, Frasz, & Egbert, 1980;Richards, 1985b, 1987; Watanabe, Iwase, & Hara,1973). These changes may be due to the increasinginfluence ofthe parasympathetic nervous system onheart rate variability (Egbert & Katona, 1980).There is a close association between level of res-piratory sinus arrhythmia and infant attention re-sponses over this age range. RSA level has beenshown to be positively correlated with heart ratedeceleration in a variety of visual attention tasks,including habituation (Richards, 1985a), sustainedattention (Richards, 1985b, 1987), visual preference(Casey & Richards, 1988), and recognition memory(Linnemeyer & Porges, 1986). Respiratory sinus ar-rhythmia was related in those studies to behavioralmeasures of attention, such as visual fixation du-ration (e.g., Richards, 1987) and recognition mem-ory performance (Linnemeyer & Porges, 1986). At-tentional responses are coupled to the processescontrolling RSA level, so developmental changes inrespiratory sinus arrhythmia should affect the pat-tern of development of the attentional system. It is

not known whether the developmental changes ininfant RSA level are accompanied by intra-indi-vidual stability, i.e., reliable test-retest correlationsbetween respiratory sinus arrhythmia at differentages, or whether there are irregular patterns ofRSAchange within individuals. Irregular patterns of de-velopment of respiratory sinus arrhythmia couldresult in irregular patterns of attention develop-ment, and attenuate test-retest correlations of at-tention measures. On the other hand, stable RSAdevelopment patterns might be the basis for theobserved test-retest correlations of attention mea-sures.

The current study was an examination of bothdevelopment and stability in infant visual sustainedattention. The methods for measuring visual atten-tion replicated those of Richards (1987) and Caseyand Richards (1988). Visual attention and respi-ratory sinus arrhythmia were measured in infantswho were tested in a longitudinal design at 14, 20,and 26 weeks of age. These ages were the same asin previous studies (e.g., Richards, 1985b, 1987)that showed RSA-attention relations, and the de-velopment of both respiratory sinus arrhythmiaand sustained attention. These ages and the lon-gitudinal design allow for the examination of theparallels between intra-individual developmentalchange patterns in respiratory sinus arrhythmia andin the attention responses.

"

Methods

Subjects

Infants for this study were recruited from birth no-tices in a Columbia, South Carolina, newspaper. Theinfants were full-term, defined as having birthweightof greater than 2500 g and gestational age of 38 weeksor greater determined by the mother's report of herlast menstrual cycle. The parents reported no pre- orperinatal medical complications. Twenty-eight infantscompleted testing in a longitudinal design at ages 14,20, and 26 weeks. An additional 8 infants did not com-plete the longitudinal protocol and are not included inthis analysis. The mean testing ages of the infants were99.2 days (SD=2.23), 141.7 days (SD=3.69), and183.5 days (SD= 5.48), respectively. Testing was doneonly if the subjects maintained an alert, awake stateduring the entire procedure (eyes open, no fussing orcrying, responding to the protocol). Subjects that didnot maintain this state during testing were tested againin the same week. Four subjects were tested a secondtime at 14, two at 20, and three at 26 weeks of age.

Apparatus

The testing apparatus and stimuli were describedin detail in Richards (1987). The infant was held inits parent's lap approximately 51 cm from the centerof a black and white 49 cm TV monitor (19 in. TV).The primary stimuli were three patterns shown on the

-----

.L4

TV monitor. The stimuli were a recording of a SesameStreet TV program, a computer-generated checker-board pattern, and a series of computer-generated,concentric squares. All stimuli were presented in a 30cm square area on the monitor, subtending approxi-mately 32° visual angle. The stimuli for the inter-rupted stimulus trials consisted of two 17X II cmpanels with 20 LEDs which blinked on and off at 16Hz in a sequential pattern resembling a circle. Thepanels were located 42 cm (38°) to either side of thecenter of the screen.

Procedure

Respiration and the EKG were recorded for a 5-min period during which the infant was seated on theparent's lap away from the testing apparatus. The par-ent was instructed not to bounce or move the infant,but the infant's movements were not restrained. Theparent was then seated in the chair with the child onthe lap facing the TV screen. The LED panels werepresented for 4 trials in order to acquaint the infantwith their location. Each trial consisted of a 5-s periodwith no stimulus followed by the presentation of anLED panel. The panel remained on as long as the in-fant was looking at it, up to a maximum of 15 s.

The experimental trials consisted of 4 infant controltrials and 8 interrupted stimulus trials. The infant con-trol trials consisted of the presentation of the primarystimulus until the infant looked away from it. Theinterrupted stimulus trials consisted of the presenta-tion of the primary stimulus, and the presentation ofone of the secondary stimuli at a predetermined delayfrom the onset of visual fixation on the primary stim-ulus. The time of the secondary stimulus onset wasbased on one of the following criteria: (a) three-sec-ond-presented when 3 s had elapsed, (b) seven-sec-ond-presented when 7 s had elapsed, (c) heart ratedeceleration-presented when a significant decelera-tion of heart rate had occurred, and (d) heart rate ac-celeration-presented when heart rate began to returnto the prestimulus level following a heart rate decel-eration. A heart rate deceleration was defined as fivesuccessive beats with heart period longer than the me-dian period of the five heartbeats preceding the pre-sentation of the primary stimulus. The return of heartrate to the prestimulus level on the heart rate accel-eration trials was defined as five successive beats withheart period shorter than the median period ofthe fiveprestimulus heartbeats, and must have followed a de-celeration. Two 6-trial blocks were used, each with twoinfant control trials, and one trial from each of thefour types of interrupted stimulus criteria, with pro-cedure order being randomly chosen within each 6-trial block.

Measurement and Quantificationof Physiological Variables

The EKG was recorded with Ag-AgCIelectrodes onthe infant's chest. Beat-to-beat intervals were com-puted online with an IBM PC microcomputer by iden-tifying the R-wave of the EKG and measuring R-Rintervals with I-ms resolution. The evaluation of heart

Richards Vol. 26, No.4

rate during heart rate deceleration and heart rate ac-celeration trials was made with machine language pro-grams on the IBM PC which made the evaluation with-in 1.1 ms of the criterion beat. For quantitative anal-yses, the beat-to-beat heart period intervals were con-verted to rate (bpm) by assigning values to equal timeintervals based on the number of beats in the intervalweighted by the proportion of time that the beat oc-cupied the interval. The interval duration to whichheart rate values were assigned was 100 ms for thebaseline period (O.I-s by O.I-s heart rate intervals) and500 ms for the experimental trials (0.5-s by 0.5-s heartrate intervals) (see Graham, 1978). Rate was chosenas the cardiac function to coordinate the heart responsewith fixation patterns in the experimental trials (Gra-ham, 1978; Richards, 1980).

Respiration was measured during the baseline pe-riod with a pneumatic chest cuff, and a pneumaticrespiration transducer (Grass Instruments) quantifiedthoracic circumference changes due to respiration. Therespiration signal was digitized online at 50 Hz by anIBM PC microcomputer. Respiration frequency wasquantified for each baseline minute by detecting thenumber of breaths that occurred in each minute. Res-piration frequency was quantified only for the baselineperiod, in order to determine the modal respirationfrequency for the quantification of respiratory sinusarrhythmia.

A measure of respiratory sinus arrhythmia wascomputed from the baseline recording of the EKG withspectral analysis methods. The O.I-s by O.I-s heart rateseries was first detrended with a bandpass filter, de-signed to pass the frequencies associated with infantrespiration (Porges, 1986). The extent of respiratorysinus arrhythmia was defined as the power of heartrate summed over a frequency range of 0.1953 Hz(11.71 breaths per minute) and centered at the modalrespiration frequency for that baseline period (Rich-ards, 1985b, 1986, 1987; cf. Harper et aI., 1978, andPorges; McCabe, & Yongue, 1982). This power meas-ure was transformed by the natural logarithm functionfor the data analysis. The metric for extent of respi-ratory sinus arrhythmia is the natural logarithm of theroot-mean-squared variation of heart rate at the res-piration frequency ranges. The heart rate power spec-trum was computed from heart rate values assignedto O.I-s intervals, and the respiration signal sampledat O.I-s intervals. The first 512 O.I-s intervals of eachof the minutes were used, giving a frequency resolutionof 0.01953 Hz. The value was extracted separatelyfrom the data for each period, and was averaged overthese five baseline minutes.

Results

Baseline Heart Rate and

Respiratory Sinus Arrhythmia

The mean of the baseline O.I-s by O.I-s heartrate values, the standard deviation of those values,and the magnitude of respiratory sinus arrhythmiawere analyzed with a one-way repeated measures

July. 1989 Development/Stability in Infant Sustained Attention

Table 1

Mean baseline levelsand correlationsbetween the heart rate measures across the testing ages

41

BaselineMeasures 20 Weeks

Mean Baseline Levels (8Ds in Parentheses)

26 Weeks14 Weeks

Heart RateHeart Rate SOExtent of RSA

153.8 (11.4) 146.0 (12.6) 142.4 (11.5)13.8 (2.5) 12.9 (1.4) 10.3 (3.1)

1.24 (0.87) 1.52 (0.66) 1.96 (0.98)

.p<.05, ..p<.OI.

ANOV A with age as the factor. All three variablesshowed a significant age effect (P's<.OI). Across thethree testing ages there was a decline in mean heartrate level and the standard deviation of heart ratevalues, and an increase in extent of respiratory sin-us arrhythmia (Table I). Table I also contains thecorrelations of the variables across the three testingages. Mean heart rate level and extent of respiratorysinus arrhythmia were significantly positively cor-related across all three ages, whereas the standarddeviation of heart rate was not.

Heart Rate and Behavioral Responsesduring Attention

The heart rate responses during attention, andtime to distraction by the secondary stimulus, wereanalyzed as in Richards (1987). The results foundwith the experimental procedures (heart rate de-celeration, heart rate acceleration, three-second,seven-second) of that study were substantially rep-licated, and will be only briefly summarized here.The infants were less easily distracted by the sec-ondary stimulus when it was presented contingenton heart rate deceleration than when it was pre-sented contingent on heart rate returning to the pre-stimulus level. The level of respiratory sinus ar-rhythmia distinguished between large and smallheart rate responses, and distraction times, but onlyfor the heart rate deceleration trials. The level ofheart rate on the heart rate deceleration trials in the

period immediately following the secondary stim-ulus, which reflects the cognitive processing inten-sity, increased from 14 to 26 weeks of age.

Intra-Individual Development of Baselineand Attentional Measures

The longitudinal design of the study allowed theexamination of intra-individual patterns of changein the baseline physiological measures, and the re-lation of those patterns to the attentional responses.Subjects were divided into four groups based onchanges in respiratory sinus arrhythmia (RSA) oc-curring across the three testing ages: Group I: con-tinuously decreasing RSA (N = 3); Group 2: contin-

uously increasing RSA (N = 8); Group 3: increasein RSA from 14 to 20 weeks, followed by a decreasefrom 20 to 26 weeks (N = 8); and Group 4: decreasein RSA level from 14 to 20 weeks, followed by anincrease from 20 to 26 weeks (N = 9). Mean RSAlevel over this age range increased (Table I), andthis is reflected in the finding that Groups 3 and 4,which show decreases and increases in respiratorysinus arrhythmia over the testing ages, increaseoverall in RSA level.

Heart rate responses during attention were ex-amined for the four groups. The difference betweenthe mean heart rate level in the 2.5-s prestimulusperiod and the 2.5-s period following the secondarystimulus onset was visually examined. The patternof age changes for the heart rate response from theheart rate deceleration trials was nearly identical tothe pattern of the four baseline RSA groups. Thepattern of age changes in the heart rate responsefrom the three-second and seven-second trials wassimilar to the baseline pattern for some of the fourgroups, but not others. The heart rate response andbaseline patterns on the heart rate acceleration trialswere dissimilar in the four groups. Figure I presentsthe developmental changes in heart rate responsesduring the heart rate deceleration trials for the fourgroups. Group I, which decreased in RSA level overthe testing ages, also had a continuous decrease inthe attention-related heart rate change. Group 2showed a continuous increase in RSA level and acontinuous increase in the size of the heart ratechange. The two groups having mixed changes inbaseline respiratory sinus arrhythmia (Groups 3and 4) had parallel changes in the heart rate re-sponse at the same ages. Thus, for the heart ratedeceleration procedure, the patterns of intra-indi-vidual changes in respiratory sinus arrhythmia werematched by age changes in attention-linked heartrate responses.

An analysis of intra-individual change similar tothat done with baseline respiratory sinus arrhyth-mia was done for baseline heart rate. The subjectswere divided into four groups based on the patternof baseline heart rate change over the three testing

Correlations BetweenTesting Ages

14 & 20 20 & 26 14 & 26Weeks Weeks Weeks

.50.. .46.. .47..

.08 .05 .03

.29 .40.. .32.

,~6

'"

Richards Vol. 26, NO.4

Groups DsIIned by BueOne RSf\. C/"erge

1: Across age RSAdecre81192: Across age RSA IncI98&83: RSA Increese then decresse4: RSAdecreese 1hen lnaeese

14

~.~

"'..'..

"'..2-,~ ",

-~~~-~>'.. --~" '- ~>c-- ~~~ -'" s'.. -- '-.........-------'.. ~ >-'

.r/ ",~/-~><-~::~~~ ~

20 26

"RIsIIng Age (Weeks)

Figure 1. Mean heart rate change [ollowing the secondary stimulus onset, classified by baselinechange patterns in respiratory sinus arrhythmia (RSA) (see text).

Groups DsIIned by Baseline HR Age Change

1: Across age HR decreese

2: Across age HR lnaeese3: HR lnaease then decreese

4: HR decnI8se then Increese

;~;;~:=~:~-<~-~~ , ~,~.

~_.~~._~--~~=~.== ,-- --.,--~

-'~-~' S

~

20 26

"RI8IIng Age (W8eIc8)

Figure 2. Mean heart rate change [ollowing the secondary stimulus onset, classified by baselinechange patterns in heart rate (HR) (see text).

14

ages. Heart rate responses during attention were ex-amined for these groups. The pattern of age changesin the heart rate response and the pattern of thefour baseline heart rate groups were dissimilar forall four experimental procedures. For example, Fig-ure 2 presents the heart rate responses during theheart rate deceleration procedure for the four base-line heart rate groups. The pattern of develop-mental change in baseline heart rate was not par-alleled by the attention-linked heart rate responses,unlike the results found with baseline respiratorysinus arrhythmia.

The relation between the relative levels at thethree ages of baseline physiological measures andattentional responses was examined with correla-tion coefficients. Correlations were computed be-tween the baseline physiological variables, and theheart rate change and distraction time for the fourinterrupted stimulus procedures combined acrossthe three testing ages. Table 2 contains the resultingcorrelations. The heart rate response during theheart rate deceleration, three. second, and seven-second procedures was negatively correlated withthe level of respiratory sinus arrhythmia (RSA).

-6.0

-7.15

He8rtRateChange -10.0

dll1ngSustaInedAttentIon -12.15(bpm)

-16.0

-6.0

-7.15

HeartRateCtwIge -10.0

duTO&8bIInedAtI8nIIan -12.5(bp'n)

-15.0

- ---

Development/Stability in Infant Sustained Attention

Table 2

Correlations between baseline heart rate measures and the heart rate and

distraction time responses for the experimental procedures combined across ages

July, 1989 42,

BaselineMeasures

Correlations

Heart RateDeceleratiou

Seveu-Second

Heart RateAcceleration

Three-Second

Mean Heart Rate During 2.5 s Following Secondary Stimulus

Heart RateHeart Rate SDExtent of RSA

.09

.01- .40**

.19-.03- .68**

.07,02.06

.16-.11- .52**

Time to Distraction by Secondary Stimulus

Heart RateHeart Rate SDExtent of RSA

.23*

.44**

.08

.07

.21*-.10

.21*

.27*-.14

.17

.12-.04

*p<.05, **p<.OI.

Table 3

Correlations between successive testing ages for the heart rate and distractiontime responses across the experimental procedures

TestingAges

Correlations

Heart RateDeceleration

ProceduresCombined

Heart RateAcceleration

Three-Second

Seven-Second

14 & 20 weeks20 & 26 weeks

Mean Heart Rate During 2.5 s Following Secondary Stimulus

.26.02

.27*

.40**.36*.30

.23

.18-.01

.66**

.34*

.21

Time to Distraction by Secondary Stimulus

.34*

.2214 & 20 weeks20 & 26 weeks

.36*

.32*-.08-.28

.26

.23

*p<.05, **p<.OI.

The negative correlation indicates that large heartrate decelerations were associated with large levelsof respiratory sinus arrhythmia. The heart rate re-sponse during the heart rate acceleration procedurewas not significantly correlated with RSA level.Mean heart rate level was positively correlated withdistraction time on two of the procedures; and thestandard deviation of heart rate values was corre-lated with distraction time on three of the proce-dures. The pattern of these results corroborates therelations found between baseline RSA change pat-terns and heart rate responses during attention (e.g.,Figures I and 2).1

IThe inclusion of multiple observations from the samesubject violates the statistical assumption of independentobservations for correlation coefficients and may limit theinterpretation of the statistical significance of the corre-lations. However, this procedure should show the strengthof the baseline-attention association given the age-relatedchanges in both the baseline and attention measures.

Stability of Attention Responses

The stability of the heart rate responses duringfixation, as well as the distraction by the secondarystimulus, was examined by looking at the correla-tions between responses between successive testingages (14 and 20, and 20 and 26). Table 3 presentsthe first-order correlations between testing ages forthe mean heart rate level in the 2.5 s following thesecondary stimulus onset, and the time to distrac-tion by the secondary stimulus. Table 3 also pre-sents the between-age correlations of a combinedvariable which consists of the average Z-score ofthe four procedures at a testing age. Significant cor-relations across test ages occurred for the heart ratedeceleration and three-second trials. The combined

variable showed moderate and significant correla-tions across the two ages. The seven-second and theheart rate acceleration trials were not significantlycorrelated across the testing ages. The relationshipof the stability of the attention responses to the

~

428

baseline measures of heart rate was examined bypartialing out the variation of the within-age base-line measures from the across-age correlations inTable 3. Almost all correlations in Table 3 werenonsignificant when the covariation with the base-line physiological measures was partialed out of thecorrelation. The only statistically significant partialcorrelation was for the three-second procedure be-tween the 20 and 26 week testing ages (Pearson r=042, p<.05). The average absolute correlation sizewas .13. Thus, the variance shared at different agesby the attention variables also was shared by theheart rate variables.

Discussion

A consistent pattern of developmental changesin sustained attention has emerged from this andprevious studies (Casey & Richards, 1988; Rich-ards, 1985b, 1987). The most significant develop-mental change in heart rate occurred in the sus-tained phase of attention. The level of heart rateon the heart rate deceleration trials in the periodimmediately following the interrupted stimulus,which reflects the status of cognitive processing in-tensity, increased from 14 to 26 weeks of age. Thisfinding replicates data from previous studies (Casey& Richards, 1988;Richards, 1987).This agechangeparallels the finding of an increasingly sustainedheart rate response for the older age infants duringthe sustained period of attention (Richards, 1985b).Thus, with the increasing age levels used in thisstudy, the period in which the infant was not dis-tracted by the secondary stimulus wasaccompaniedby increased sustained heart rate responding. An-other developmental change was the time to dis-tract the infant from looking at the primary stim-ulus with the secondary stimulus, which decreasedwith age. However, the decline in distraction timesover this age range did not differ for the experi-mental procedures. The age changes in the heartrate responses were not exactly paralleled by agechanges in the behavioral responses.

The results of the current study support only alow stability of individual differences in sustainedattention. The correlations were moderately posi-tive, with an inconsistent pattern of statistical sig-nificance. The combination of the individual trialmeasures resulted in correlations from about .3 toA between successivetestingagesfor both heart rateand distraction time. The level of these correlationswas similar to that which has been reported in otherstudies (Colombo et aI., 1987; Rose et aI., 1988;Ruff et aI., 1987).Certain measures of infant atten-tion are more highly correlated across testing agesthan are other measures (Colombo et aI., 1987).Along this same line, in the present study the at-tention measures during the heart rate acceleration

Richards Vol.26, No.4

condition, and the seven-second condition, wereuncorrelated across age, whereas the heart rate re-sponses and distraction times on the heart rate de-celeration and three-second conditions were signif-icantly correlated. Although not entirely overlap-ping, the three-second and heart rate decelerationprocedures may produce similar results becausethey are more closely associated in time than theyare with the other two procedures. However, theheart rate deceleration condition is a better pro-cedure for assessing sustained attention, becausethe behavioral attention response and relation torespiratory sinus arrhythmia is related more closelyto level of the heart rate response than it is to timeof the secondary stimulus (Richards, 1987). Theseresults imply that sustained attention may be basedon a reliable individual difference system whereasreactive attention and attention termination arenot.

The stability of the baseline physiological vari-ables over the three testing ages was more impres-sive than was the stability of the attention re-sponses. The correlations between respiratory sinusarrhythmia (RSA) at the different ages were as ex-pected, and in the positive direction. The level ofthe correlations is low (.29 to 040), implying thatthere is as much instability as stability in the rel-ative placement of an individual's RSA level at thedifferent ages. The instability in RSA level is likelydue to the developmental changes in RSA levelacross this age range (e.g., Harper et aI., 1976, 1978;Katona et aI., 1980). The differing patterns of intra-individual change in respiratory sinus arrhythmia(e.g., Figure I) may account for the lack of intra-individual stability. The correlations between heartrate at the different ages were the largest in mag-nitude, and there was a decrease in heart rate levelacross this age. The correlation of heart rate acrossthis age range suggests that a stable physiologicalsystem may exist that is not exclusive to the mech-anisms responsible for respiratory sinus arrhyth-mia, but may include distinct (though related) pro-cesses of the central nervous system. These resultsare consistent with the existence of stable individ-

ual differences in a physiological system acrossthese ages in spite of the significant changes in heartrate level and heart rate variability.

The consistent relation between baseline heart

rate variability and attention responses is stronglysupported by this study. Respiratory sinus arrhyth-mia was related to the behavioral and physiologicalindices of attention on the trials in which attentionwas posited to be the most active (heart rate de-celeration trials), and for sustained attention ratherthan reactive attention. Intra-individual develop-

July. 1989 Development/Stability in Infant Sustained Attention 42'..

ment in baseline respiratory sinus arrhythmia wasparalleled by changes in heart rate during sustainedattention. Not only was the pattern preserved, butthe relative level of respiratory sinus arrhythmia(RSA) over the entire testing period was paralleledby the relative level of heart rate change duringattention over the same testing period. Also, thetest-retest correlations of the attention parameters(heart rate and behavioral) measured at the threetesting ages lost their statistical significance whenthe variance shared with the baseline heart rate var-iables was removed. Fluctuations in the RSA sys-tem that are local to a testing age show up as fluc-tuations in the attention system. The low levels of

stability in the attention system may be a result ofshort-term fluctuations in respiratory sinus arrhyth-mia, or in the physiologicalsystemsinvolved in thecontrol of heart rate, respiratory sinus arrhythmia,and respiration. An individual differences modelrelating heart rate variability and attention (e.g.,Porges, 1976, 1980; Richards, 1988) should em-phasize the stability of the underlying physiologicalsystem rather than the attentional parameters. Con-sistency between attention responses over ageseems to be mediated by the stability of the phys-iological system (e.g., heart rate, respiratory sinusarrhythmia, etc.) and the within-age relation of at-tention to heart rate variability.

REFERENCES

Bornstein, M.H., & Benasich, A.A. (1986). Infant habit-uation: Assessments of individual differences andshort-term reliability at five months. Child Develop-ment, 57, 87-99.

Bornstein. M.H.. & Sigman. M.D. (1986).Continuity inmental development from infancy. Child Develop-ment. 57. 251-274.

Casey. B.1.,& Richards, J.E. (1988).Sustained visual at-tention measured with an adapted version of the visualpreference paradigm. Child Development, 59, 1514-1521.

Cohen, S.E.,& Parmelee, A.H. (1983).Prediction of five-year Stanford-Binet scores in preterm infants. ChildDevelopment, 54, 1242-1253.

Colombo, J., Mitchell, D.W., O'Brien, M.. & Horowitz,ED. (1987).The stability of visual habituation duringthe first year of life. Child Development.58, 474-487.

Egbert, J.R., & Katona, P.G. (1980). Development ofautonomic heart rate control in the kitten during sleep.American Journal of Physiology,238, H829-H835.

Fagan, J.E, & McGrath, S.K. (1981). Infant recognitionmemory and later intelligence. Intelligence, 5, 351-356.

Fagan, J.F., & Singer, L. (1983). Infant recognition mem-ory as a measure of intelligence. In C. Rovee-Collier(Ed.),Advances in infancy research(Vol.2, pp. 31-78).Norwood, NJ: Ablex.

Graham, F.K. (1978).Constraints on measuringheart rateand period sequentially through real and cardiac time.Psychophysiology,15,492-495.

Harper, R.M., Hoppenbrouwers, T., Sterman, M.B.,McGinty, D.J., & Hodgman, J. (1976). Polygraphicstudies of normal infants during the first six monthsof life: 1. Heart rate and variability as a function ofstate. Pediatric Research. 10,945-951.

Harper. R.M., Walter, D.O., Leake, B., Hoffman, H.J.,Sieck, G.c., Sterman, M.B., Hoppenbrouwers, T., &Hodgman, J. (1978). Development of sinus arrhyth-mia during sleeping and waking states in normal in-fants. Sleep, 1, 33-48.

Katona, P.G., Frasz, A., & Egbert,J.R. (1980).Maturationof cardiac control in full-term and preterm infants dur-ing sleep. Early Human Development. 4, 145-159.

Lewis. M., & Brooks-Gunn, J. (1981). Visual attention atthree months as a predictor of cognitive functioningat two years of age. Intelligence, 5,141-147.

Linnemeyer, S.A., & Porges, S.W. (1986). Recognitionmemory and cardiac vagal tone in 6-month-old in-fants. Infant Behavior and Development. 9, 43-56.

Maitre, G., Lamarre, G., & Pomerleau, A. (1987, April).Infant visual habituation: Stability of response in four-month-aids. Paper presented at the meeting of the So-ciety for Research in Child Development, Baltimore,MD.

McCall, R.B. (1979). Individual differences in the patternof habituation at 5 and 10 months of age. Develop-mental Psychology, 15, 559-569.

Porges, S.W. (1976). Peripheral and neurochemical par-allels of psychopathology: A psychophysiological mod-el relating autonomic imbalance in hyperactivity, psy-chopathology, and autism. In H. Reese (Ed.), Advancesin child development and behavior (Vol. II, pp. 35-65). New York: Academic Press.

Porges, S.W. (1980). Individual differences in attention:A possible physiological substrate. In B. Keogh (Ed.),Advances in special education (Vol. 2, pp. 111-134).Greenwich, CT: JAI Press.

Porges, S.W. (1986). V: An accurate method of quantifyingRSA [Abstract]. Psychophysiology, 23, 414.

Porges, S.W., McCabe, P.M., & Yongue, B.G. (1982). Res-piratory-heart rate interactions: Psychophysiologicalimplications for pathophysiology and behavior. In J.Caccioppo & R. Petty (Eds.), Perspectives in cardio-vascular psychophysiology (pp. 223-264). New York:Guilford Press.

Richards, J .E. (1980). The statistical analysis of heart rate:A review emphasizing infancy data. Psychophysiology.17, 153-166.

Richards, J.E. (I 985a). Respiratory sinus arrhythmia pre-dicts heart rate and visual responses during visual at-tention in 14 to 20 week old infants. Psychophysiology,22, 101-109.

Richards, J .E. (1985b). The development of sustained vis-ual attention in infants from 14 to 26 weeks of age.Psychophysiology, 22, 409-416.

Richards, J.E. (1986). Power spectral analysis quantifi-

p~ 430

cation of respiratory sinus arrhythmia [Abstract]. Psy-chophysiology, 23, 414.

Richards, J.E. (1987). Infant visual sustained attentionand respiratory sinus arrhythmia. Child Development,58, 488-496.

Richards, J.E. (1988). Heart rate changes and heart raterhythms, and infant visual sustained attention. In.P.K Ackles, J.R. Jennings, & M.G.H. Coles (Eds.),Advances in psychophysiology (Vol. 3, pp. 189-221).Greenwich, CT: JAI Press.

Rose, SA, Feldman, J.F., & Wallace, I.F. (1988). Indi-vidual differences in infants' information processing:Reliability, stability, and prediction. Child Develop-ment, 59,1177-1197.

Richards Vol. 26, No.4

Rose, S.A., & Wallace, I. (1985). Visual recognition mem-ory: A predictor of later cognitive functioning in pre-term infants. Child Development, 56, 843-852.

Ruff, H.A. (1986). Components of attention during in-fants' manipulative exploration. Child Development,57, 105-114.

Ruff, H.A., Lawson, KR., & Parinello, R. (1987, April).Early individual differences in attention. Paper pre-sented at the meeting of the Society for Research inChild Development, Baltimore, MD.

Watanabe, K., Iwase, K, & Hara, K. (1973). Heart ratevariability during sleep and wakefulness in low-birth-weight infants. Biology of the Neonate, 22, 87-98.

(Manuscript received May 12, 1988; accepted for publication November 1,1988)