Infancy, 15(2), 107–124, 2010 Copyright  International Society on Infant Studies (ISIS) ISSN: 1525-0008 print / 1532-7078 online DOI: 10.1111/j.1532-7078.2009.00012.x What Habituates in Infant Visual Habituation? A Psychophysiological Analysis John Colombo Department of Psychology and Schiefelbusch Institute for Life Span Studies University of Kansas D. Jill Shaddy, Christa J. Anderson, Linzi J. Gibson, and Otilia M. Blaga Department of Psychology University of Kansas Kathleen N. Kannass Department of Psychology Loyola University of Chicago Despite the use of visual habituation over the past half century, relatively little is known about its underlying processes. We analyzed heart rate (HR) taken simultaneous with looking during infant-controlled habituation sessions collected longitudinally at 4, 6, and 8 months of age with the goal of examining how HR and HR-defined phases of attention change across habituation. There were four major findings. First, the depth and topography of decelerations and proportion of sustained attention (SA) did not vary across habituation at any age, which suggested (in contrast to the tenets of comparator theory) the persistence of substantial cognitive activity at the end of visual habituation. Second, attention termination (AT) robustly declined across trials, suggesting that, contrary to prior thinking, AT might be a sensitive indicant of visual learning. Third, infants at all ages showed an HR increase (startle) to stimulus onset on the first trial, the magnitude of which was associated with subsequent delayed HR deceleration and less SA; thus, stimulus events affect processing Correspondence should be sent to John Colombo, Department of Psychology, University of Kansas, 1415 Jayhawk Boulevard, Lawrence, KS 66045. E-mail: colombo@ku.edu 108 COLOMBO ET AL. during trials. Finally, mean overall HR reliably increased across trials for all ages. This last finding implies the need to distinguish between ‘‘phasic’’ HR changes (e.g., decelerations during looks) and longer term ‘‘tonic’’ HR changes (mean increases across trials) during habituation, and raises the question of what processes the tonic increases might reflect within the habituation paradigm. Since its introduction as a nonverbal technique for the study of visual learning in the 1960s (Berkson & Fitzgerald, 1963), visual habituation has been used as a technique for studying cognition and perception in infancy for over 50 years (Colombo & Mitchell, 2009). In this procedure, the infant’s visual responses to a repetitive series of stimulus presentations are monitored. Across these repetitive presentations, the duration of infants’ visual regard of the stimulus (i.e., ‘‘looking’’) declines. The fact that the duration of looking recovers to a novel stimulus presented at the end of the repetitive series (Pancratz & Cohen, 1970) rules out the possibility that the decline is attributable to sensory adaptation or fatigue. Thus, the decline in looking is generally considered to represent some form of learning; largely based on the topography of a response decline, the phenomenon has been characterized as a form of habituation (Rankin et al., 2009; Thompson & Spencer, 1966). Despite its widespread use for nearly half a century, the processes that contribute to visual habituation in infancy have never been well understood (Colombo & Mitchell, 1988). There is no dearth of models for what processes might govern, or contribute to, habituation (Groves & Thompson, 1970; Jeffrey, 1968; Kaplan & Werner, 1986; Malcuit, Pomerleau, & Lamarre, 1988; Miller, Ryan, Sinnott, & Wilson, 1976; Schoner & Thelen, 2006). However, the account of habituation that is most widely accepted for explaining the phenomenon in infancy is comparator theory (Sokolov, 1958, 1960). Comparator theory formally posits that the magnitude of the ‘‘orienting reflex’’ (most commonly operationalized as the duration of looking) declines as a function of the veracity of the match between the external stimulus and its internal representation. Thus, as repeated looks to an unchanging stimulus elicits looks of progressively shorter duration, the infant is thought to be encoding that stimulus until a reasonably accurate memory (engram) has been formed. Comparator theory generates a series of hypotheses about the processes that go on as habituation progresses. For example, it seems logical to expect that across habituation, as the internal representation becomes more accurately matched to the external stimulus, processing or cognitive demands should progressively decrease. While such a hypothesis would be relatively difficult to test behaviorally, a system for using heart rate (HR) as a means for parsing looking into specific and dissociable phases of attention WHAT HABITUATES IN INFANT VISUAL HABITUATION? 109 (Richards, 1985, 1987, 1989a, 1989b, 1997a) has been used for the last two decades. Indeed, we have used this approach (Maikranz, Colombo, Richman, & Frick, 2000) in studying some of the predictions concerning habituation made by dual-process theory (Groves & Thompson, 1970; Kaplan & Werner, 1986). The system is based on the widely replicated finding that HR decelerates during periods of infant attention (Graham & Clifton, 1966). However, this deceleration typically does not overlap entirely with periods of attention; for example, HR deceleration may occur some time after the initiation of a look. Similarly, the HR may return to baseline levels or above before the termination of a look. The decelerative phase has been characterized as sustained attention (SA) (Richards, 1985), and substantial evidence suggests that the period of deceleration reflects active cognitive processing (Richards, 1997a, 1997b) and probably reflects the encoding that is presumed to occur by the comparator model. Two other phases of attention have been delineated, based on the topography of the deceleration (Richards & Casey, 1992). The phase that precedes the onset of the deceleration has been defined as orienting (OR), and is thought to reflect basic processes involved in the initiation of engagement with the stimulus. The phase that follows deceleration has been defined as attention termination (AT), and this phase has been posited to reflect processes, such as disengagement and attention shifting (Colombo, Richman, Shaddy, Greenhoot, & Maikranz, 2001). Our recent conduct of a large-scale longitudinal study of visual habituation during the first year (Colombo, Shaddy, Richman, Maikranz, & Blaga, 2004) provides a unique opportunity to test such a prediction and to conduct a fine-grained examination of the processes that occur during infant visual habituation. Measurements of visual habituation were conducted using a traditional behavioral paradigm, but with the addition of simultaneous HR measurement that allowed the assessment of HR-defined phases of attention and other HR changes across the sessions. Based on the tenets of comparator theory, we began with two strong predictions about changes in psychophysiological markers in habituation. First, based on the presumption that cognitive load should decrease with habituation, we hypothesized that HR decelerations would diminish (i.e., become less strong and less sustained) as infants approached the attainment of behavioral habituation. Second, given that the measurement of SA has been standardized, and given that it has been determined that the phase reflects active and engaged stimulus processing, we expected that the proportion of time spent in SA would decrease as infants approached the attainment of visual habituation. In addition, the synchronization of HR with stimulus and looking events allowed us to examine whether there were other phenomena within the session protocol (e.g., 110 COLOMBO ET AL. reactions to stimulus onsets) that might bear on infants’ performance in the paradigm. METHOD Participants The sample for these analyses was drawn from the Kansas Early Cognition Project (Colombo et al., 2004). The entire longitudinal sample from that project included 226 infants tested every month from 3 to 9 months on an infant-controlled visual habituation protocol, augmented with simultaneous HR measurement. From this database, we chose sessions from three ages: 4, 6, and 8 months because we felt that these three ages would provide a fairly comprehensive representation of habituation across the first year (Colombo, 2001), while simultaneously allowing reducing the levels of the age factor to be included in analyses of the HR data. In addition, these three ages represented the best set of successfully completed habituation sessions within the database; the respective sample sizes were n = 208 for the 4-month-olds, n = 200 for the 6-month-olds, and n = 191 for the 8-month-olds. This provided us maximum power in examining data within a repeated measures (i.e., longitudinal) context. Behavioral protocol Testing set-up For the habituation session, infants sat in a car seat, 1.0 m from a rearprojection screen in a darkened room. Color slides of children’s faces (all faces showed positive affect, and were placed onto the same white background) were rear-projected (the slides subtended a visual angle of 25) and visual behavior was coded by a ‘‘live’’ observer blinded to stimulus identity from a video feed of the infant’s face. The video feeds were recorded for later analysis of the reliability of the live codes; as is typical with this technique, reliabilities for live coding were very high (+.95 > r > +.99). Habituation parameters Timing of looks, stimulus presentation, and calculation of habituation data were automated, with all aspects controlled by a microcomputer in real time. Valid looks were initiated by 1 sec of continuous looking at the stimulus, and were terminated by 1 sec of continuous looking away from the stimulus (Colombo & Horowitz, 1985). The habituation criterion was two WHAT HABITUATES IN INFANT VISUAL HABITUATION? 111 consecutive looks at 50% of the longest previous look; a ‘‘floating-point’’ habituation criterion was used, such that long looks occurring later in the sequence reset the criterion (Colombo & Mitchell, 1990). As expected, look duration declined with age but the number of looks to habituation did not vary significantly across ages (Colombo & Mitchell, 1990; Colombo et al., 2004). Heart rate measurement and data reduction HR was collected using a BioPac (Santa Barbara, CA) data acquisition system. Disposable electrodes were placed in a triangular configuration on the infant’s chest and abdomen, amplified, and digitized at 250 Hz. The BioPac software generates a graphical readout of the electrocardiogram (EKG); time stamps for individual R-waves were detected by commercial software that detected the r-wave using peak detection algorithms. This software detection was augmented by hand scoring of the EKG when necessary. In addition, stimulus and looking events were recorded at the time of testing and transmitted to the EKG record so that they could be combined with the R-wave time stamps to produce a sequential text file. This file was then analyzed with custom software to parse looking into HR-defined phases of attention based on previously published criteria (Richards, 1987). Briefly, a median was derived from each interstimulus interval; this served as the prestimulus level, and it varied from trial to trial (as we will report below, HR shows systemic changes across the session). Once looking began, the occurrence of five consecutive beats below the median triggered the classification of SA; beats prior to the onset of SA were placed into the OR category. When HR subsequently rose above the median during looking, this triggered the classification of beats into AT. Further details on procedures, data reduction, normative analyses, and prediction of later cognitive outcomes through later infancy are provided elsewhere (Colombo et al., 2004). ANALYSES AND RESULTS Behavioral data during habituation Habituation data from these three ages for the variables we are analyzing are reported in Table 1. As expected, look duration declined across the first three looks and to the habituation criterion. Look duration declined across ages, with the predominant decline occurring from 4 to 6 months. This pattern of outcome has been observed a number of times in previous reports (Colombo & Mitchell, 1990). 112 COLOMBO ET AL. TABLE 1 Habituation Data for the Three Ages Analyzed for this Study 4 months Habituation variable Trials to criterion Look 1 duration (sec) Look 2 duration (sec) Look 3 duration (sec) Criterion look 1 duration (sec) Criterion look 2 duration (sec) 6 months 8 months M SE M SE M SE 7.00 19.10 14.84 11.93 6.36 5.94 .24 1.67 1.51 1.17 .41 .47 6.73 16.47 9.85 9.06 4.56 4.23 .23 1.41 .81 .72 .26 .25 6.48 13.38 10.08 8.53 4.50 3.93 .20 .95 .77 .66 .31 .26 Analyses of stimulus onset We first examined the HR change of infants at all ages in response to the onset of the stimulus across the first three trials of the habituation session. This was done by comparing the average HR from the 2-sec period immediately preceding the onset of the stimulus (the ‘‘prestimulus’’ period, which constituted the intertrial interval) with the average of the interval after the presentation of the stimulus, but prior to the onset of looking (the ‘‘post-onset’’ period that preceded the onset of looking). These data were subjected to an Age (3) · Pre ⁄ Post Onset (2) · Trial (3) mixed model analysis, with all factors as repeated measures. We observed a significant Trial · Pre ⁄ Post Onset interaction, F(2, 3041.706) = 3.72, p = .024, as the direction of the HR change across the Pre ⁄ Post Onset periods varied significantly across the three trials tested. This interaction qualified significant main effects for Age, F(2, 3113.592) = 354.75, p < .001 (HR reliably decreased with increasing age) and Trial, F(2, 3041.706) = 22.060, p < .001 (HR reliably increased from Trial 1 to Trial 3). In probing the interaction, we observed that there was a significant increase in HR from the Pre to Post Onset periods on Trial 1 only (p = .021, d = .14). With the observation of this result, we examined the length of the postonset period as a possible contributor to this effect. While the ‘‘pre-onset’’ interval is fixed at 2 sec, the post-onset period can vary. For 4-month-olds, the post-onset length across the first three trials was 7.46, 4.86, and 3.47 sec, respectively; these values varied significantly, F(2, 199) = 4.67, p = .01, g2 = .045, due to the longer post-onset period on Trial 1. Post-onset values for 6 months (1.02, 1.17, and 1.93 sec) and 8 months (.99, 1.02, and 1.13 sec) did not vary significantly across trials. Even though the observed effect was constant across ages, we were concerned that the longer post-onsets observed on Trial 1 for the 4-month-olds might contribute to the effect. Thus, we analyzed the pre–post change in the 4-month-olds controlling for WHAT HABITUATES IN INFANT VISUAL HABITUATION? 113 post-onset length; with this factor controlled, the pre–post effect remained significant, F(1, 195) = 22.11, p < .001, g2 = .109. Thus, the increase in HR on Trial 1 was not an artifact of the longer post-onset intervals at 4 months. The absence of any other significant higher order interactions indicates that this effect was constant across ages. As the HR increase on Trial 1 does not appear to be artifactual, we are inclined to interpret this result as a mild ‘‘startle’’ response to the onset of the stimulus, perhaps attributable to the appearance of a relatively brightly lit stimulus in the darkened testing room. Given that this startle was present for only Trial 1, we also conclude that this startle habituated immediately. Figure 1 shows the means for the pre-onset and post-onset HRs across the first three trials, collapsed across all ages. Owing to the notion from dual process theory (Kaplan & Werner, 1986) that such a response might affect orienting or some other aspect of later behavior, we examined correlations between the degree of HR increase seen prior to looking on Trial 1 with the length of the first look and subsequent looks during habituation. At none of the ages did these associations attain statistical significance. However, we did find that, at all three ages, the presence and intensity of this HR increase affected HR-defined phases during the first look. The correlations between the difference in HR from the preto poststimulus onset periods are presented in Table 2 and show that, across all ages, the greater the HR increase to the stimulus onset, the higher the proportion of looking spent in OR, and the lesser the proportion of looking spent in SA. The nature of the association of this startle with AT varied with age; at 4 months, the startle marginally predicted more AT, at 6 months it was not significant either way, and at 8 months it significantly predicted less AT. Figure 1 Infants’ heart rate (HR) responses to stimulus onset on Trials 1–3 during habituation. Significant HR acceleration to stimulus onset occurs on Trial 1. 114 COLOMBO ET AL. TABLE 2 Correlations Between Size of Startle (HR Increase to Stimulus Onset) and Proportion of HRDefined Phases of Attention on First Trial Proportion of time in Age 4 months 6 months 8 months Overall Orienting Sustained attention Attention termination .284** .404*** .486*** .414*** ).404*** ).397*** ).414*** ).436*** .124a ).042 ).209** .010 Note. ap < .10, **p < .01, ***p < .001. These findings are consistent with the interpretation that the HR increase (startle) to stimulus onset interferes with subsequent cognitive activity during looking. It may also reflect the fact that the elevated HR as a result of the stimulus onset simply increases the latency of the infant’s HR deceleration during looking; this would also theoretically produce more OR and less AT. Heart rate during looking As look lengths varied across habituation, the analysis of the HR responses during looking was somewhat difficult to capture in a single approach. As a result, we addressed this in two ways. Analyses of modeled heart rate curves during looking The first approach was to use regression-based models of HR responses for each age and look. These models fit all of the available data, and provided a complete description of the HR profiles within looks, although they did not readily lend themselves to direct comparisons. Figure 2 shows the beat-by-beat HR within each of the five looks analyzed for the three age groups involved. We fit linear and quadratic terms to the raw (i.e., beat-by-beat) HRs for looks 1–3 and both criterion looks at each age. If, as we hypothesized, decelerations were to become less prominent with habituation, we would expect to see linear weights for the modeled curves start out strongly negative and then move toward zero (i.e., the HR profile move toward showing no change within looks later in the habituation sequence), and that the values of the quadratic weights for the modeled curves should decrease (i.e., the HR profile becoming increasingly ‘‘flatter’’ as decelerations decreased in depth WHAT HABITUATES IN INFANT VISUAL HABITUATION? 115 Figure 2 Beat-by-beat plots for infants’ heart rate within habituation looks, adjusted for the mean length of each look reported in Table 1. ‘‘L’’ signifies precriterion looks, ‘‘C’’ signifies criterion looks. The zero point is the poststimulus period prior to the onset of looking. and duration). The results of this modeling are presented in Table 3; linear and quadratic components that represent significant deviations from zero (i.e., which account for statistically significant amounts of variance in the HR change during the look) are asterisked. Also represented in Table 3 are the intercepts for HR across the trials; these values again show base HR increasing linearly across habituation by about 5 beats per minute on average from the first look to the last criterion look. TABLE 3 Intercepts (Constant) and Standardized Coefficients (b) for Linear and Quadratic Components for Heart Beats Within Looks During Habituation at 4, 6, and 8 Months 4 months 6 months 8 months Look Constant Linear Quadratic Constant Linear Quadratic Constant Linear Quadratic 1 2 3 Crit 1 Crit 2 145.36 147.45 148.89 150.20 150.74 ).349*** ).192** ).207** ).226** ).190** .434*** .203** .258** .215*** .183** 138.27 139.01 139.83 141.48 142.54 ).234** ).175** ).160* ).296*** ).250*** .179** .165** .183** .324*** .210* 135.23 135.96 137.74 140.64 140.82 ).272*** ).106 ).057 ).228*** ).159*** .313*** .151** .162** .198** .210** Notes. *p < .05, **p < .01, ***p < .001. ‘‘Crit’’ refers to criterion look. Linear and quadratic terms entered into the regression simultaneously; p-values are derived from tests of regression components in each model. 116 COLOMBO ET AL. Across all ages and across all looks analyzed, the modeled curves yielded significant, or nearly significant, linear and quadratic components. The linear components of looks 2 and 3 at 8 months approached (but did not achieve) conventional levels of statistical significance; on the other hand, the absolute value of curvilinear components actually increased across the sessions irrespective of age, and attained statistical significance at all points tested. Taken together, these results imply continued and strong decelerations across the habituation session, which is at odds with the predictions of comparator theory. Direct tests of this are featured in the next section. Analysis of studentized epochs Along with modeling all the available data as described above, we used a more constrained but also more manageable approach to the problem of analyzing HR sequences of different lengths. Here, to conduct direct lookby-look comparisons, we studentized HRs for each look into five epochs; that is, we standardized each look so that any epoch represented 20% of the entire look duration. The value for each epoch was the average HR for that one fifth of the look. This practice reduced the number of points per look to five, and in so doing, allowed for the application of more straightforward inferential analyses. We then sought to determine whether the nature of HR decelerations were from the first three looks in the habituation sequence from those observed in criterion looks at the attainment of habituation. We subjected the studentized HRs to another mixed model analysis, this time with factors of Age (3) · Trial (5) · Epoch (5), where the Trials factor included the first three two looks in the habituation sequence and the two criterion looks. The analysis yielded significant main effects for Age, F(2, 989.99) = 38.71, p < .001, g2 = .061, owing to the decrease in infants’ HRs across the ages tested, for Trial, F(4, 7475.05) = 17.90, p < .001, g2 = .017, reflecting the consistent finding here that infants’ HRs increased across looks, and for Epoch, F(4, 10,937.83) = 46.41, p < .001, g2 = .002, again attributable to infants’ HRs decelerating across epochs within looks. No interaction attained statistical significance. The main effects are evident in the plot of the studentized data shown in Figure 3. Given the predictions generated by the comparator model, we strongly expected the presence of a significant Epoch · Look interaction in this analysis. The absence of this (or any other) significant interactions was theoretically remarkable, as it suggested that, although the mean HR changed as a function of both age and habituation trial, the form and depth of the deceleration did not vary as a function of the look in the habituation sequence. That is, as infants presumably ‘‘learned’’ the stimulus across habituation, WHAT HABITUATES IN INFANT VISUAL HABITUATION? 117 Figure 3 Studentized heart rate profiles during habituation; each unit on the abscissa represents 20% of the infant’s look. Notations are as in Figure 2, and the first point is the poststimulus period prior to the onset of looking. decelerations did not appreciably change across trials and were, in fact, statistically equivalent. Heart rate–defined phases of attention The last set of analyses involved tests for changes in HR-defined phases of attention across habituation. We analyzed the proportion of time infants spent in OR, SA, and AT during looking as a function of the first three trials of habituation and the last two criterion looks. The results of analyses are shown in Table 4 below; the F-values and effect sizes (g2) reported at the bottom of the table represent the results of separate repeated measure multivariate analyses of variance (MANOVAs) conducted on each HR-defined phase across the first three looks and the two criterion looks. AT consistently declines across habituation trials; this is evident in significant and robust effects at all three ages. There are no changes in OR or SA up to 6 months of age, but at 8 months of age, there is a trend for the proportion of SA to decline (p = .055) across the habituation trials at 8 months. It is not clear whether this is artifactual or attributable to some underlying cognitive mechanism. The case for artifact rests on the very brief nature of looking at 8 months; look durations at that age are normatively short, and as infants proceed through the habituation sequence, they 118 COLOMBO ET AL. TABLE 4 Means for the Proportions of Time Looking Spent in OR, SA, and AT Across the First Three Habituation Looks and the Two Criterion Looks for the Three Age Groups Included in This Study 4 months Look 1 2 3 Crit 1 Crit 2 F-ratio p-value g2 6 months 8 months OR SA AT OR SA AT OR SA AT .213 .241 .303 .291 .313 1.45 ns .038 .574 .585 .542 .610 .602 .87 ns .005 .213 .175 .154 .099 .084 21.44 <.001 .352 .232 .291 .313 .360 .344 1.53 ns .011 .569 .567 .540 .543 .564 .64 ns .005 .199 .141 .146 .096 .093 15.29 <.001 .288 .237 .349 .310 .312 .378 1.54 ns .044 .602 .518 .541 .582 .517 2.37 .055 .064 .160 .133 .150 .107 .105 12.87 <.001 .256 Note. The F-ratios and p-values reported in the table are from within-subject univariate ANOVAs conducted on each of the phases across the habituation trials shown. become even shorter. SA may not be possible during the shortest of looks, given that it is defined by the attainment of five consecutive beats below the prestimulus median. As the criterion looks for 8-month-olds are quite short, this may contribute to the apparent decline in SA. DISCUSSION The simultaneous use of HR and behavioral indices of attention has been used for nearly two decades and has provided great advances to our understanding of the cognitive processes that occur in early development (Colombo, 2002; Richards, 1998). However, we are unaware of any other attempt to use this method to elucidate what goes on during visual habituation. It should be noted that these results were obtained with faces; although studies comparing face and nonface stimuli have generally observed constant developmental courses for looking (Courage, Reynolds, & Richards, 2006), it remains to be determined whether these results will hold for other types of stimuli. The analyses described here reveal four basic findings about visual habituation that are new to the field, and which have both theoretical and practical implications for future research. Heart rate decelerations across habituation The fundamental tenets of comparator theory hold that the magnitude of the response in habituation declines as an organism acquires an internal WHAT HABITUATES IN INFANT VISUAL HABITUATION? 119 representation of a novel stimulus. Given that past research has linked the processing of the stimulus to HR decelerations within looks in human infants, we expected that HR decelerations would decline in their depth and extent with looking across habituation. Three separate analyses conducted in this study failed to support this expectation. Modeling analyses of the raw HR changes within looks showed strong persistence of the quadratic component from the start of habituation through the attainment of the criterion looks. Analysis of data collapsed into equivalent studentized bins showed the same thing. Finally, analyses of the proportion of looking spent in HR deceleration (SA) show no appreciable change across habituation. This finding implies that either HR decelerations are not as indicative of cognitive engagement as the literature suggests, or substantial cognitive activity persists through the end of habituation. Given the preponderance of evidence linking HR decelerative phases with processing, the latter of these alternatives appears to be far more likely. That is, our analysis indicates that infants are engaged in active cognitive processing of the visual stimulus even at the attainment of visual habituation. If this is the case, it begs the question of whether the attainment of the habituation criterion accurately represents the infant’s complete processing of the familiar stimulus. This is a question with broad implications that can not be addressed adequately with the current data set and must be a topic for future consideration. Attention termination and habituation A second major finding was that, across all ages, the proportion of time spent in AT robustly decreased across habituation. Recall that AT is the maintenance of looking after the HR deceleration that has been long thought to reflect that cognitive activity during infant looking has dissipated (Richards, 1989b). To this point, AT has been characterized as a measure of the disengagement of attention (Colombo et al., 2001; Frick, Colombo, & Saxon, 1999), thought by at least some scientists to reflect a basic attentional function mediated by posterior (parietal) structures (Posner & Petersen, 1990). Within this framework, then, the release of attention is facilitated in habituation, presumably as a function of increasing familiarity with a stimulus. This finding is, to some degree, at odds with the contention that AT does not involve substantial cognitive activity (Richards & Casey, 1992). At the very least, this suggests that AT is affected by the cognitive processes involved in encoding or comparison. Given that the reduction in AT across habituation means that looking behavior and its concomitant HR response are becoming more synchronized, it does not seem unreasonable to 120 COLOMBO ET AL. characterize this change as increased somatic coupling or integration (Byrne & Miller, 1988; Byrne & Smith-Martel, 1987) across habituation. Any further discussion about the mechanisms through which this effect is mediated at this point would be speculative, but the source of this phenomenon is worth the study, given that a developmental course that features the decline in AT across ages is associated with better long-term outcomes (Colombo et al., 2004, 2009). Stimulus onset startle The third major finding here concerns fairly robust evidence for the existence of a startle response (in the form of an HR increase) to stimulus onsets on the first trial of the habituation protocol. This response is presumably attributable to the abrupt appearance of the stimulus in the visual field (if the infant is looking at the screen at stimulus onset), or an obvious change in ambient lighting in the darkened booth that was produced by the stimulus’ appearance (if the infant was looking elsewhere). This phenomenon was evident and statistically equivalent at each age tested, and independent of age and any the length of the stimulus onset period. Furthermore, the response habituated immediately; it was not statistically evident on any other trials but the first one. Our data suggest that although this response was short-lived, it did have effects on the distribution of HR-defined phases of attention on the trial on which it occurred. Based on the observation that larger magnitude responses were associated with more OR and less SA on the first trial (its effects on AT were equivocal and varied across ages), it appears that the presence of this response delayed and impaired HR deceleration on that trial, perhaps interfering with processing. It would be useful to know whether such responses occur within individual infants elsewhere in habituation, and whether they impede processing more broadly. The finding may also relate to observations of the contributions of sensitization to habituation and infant cognition (Colombo, Frick, & Gorman, 1997; Kaplan & Werner, 1986; Maikranz et al., 2000). Tonic heart rate changes across habituation Finally, HR was observed to increase across habituation trials. This was evident in the intercepts for the modeled curves, mixed model analyses of studentized HR intervals, and plots of raw HR data for all age groups. The same observation of increased HR within testing sessions has been reported in two other papers from our laboratory (Blaga & Colombo, 2006; Maikranz et al., 2000). WHAT HABITUATES IN INFANT VISUAL HABITUATION? 121 Clearly, phasic HR decreases (i.e., stimulus-based decelerations, presumably a concomitant of attention or orienting) are considered to reflect increased cognitive engagement and processing in human infants (Graham & Clifton, 1966). In the adult literature, phasic HR increases have been linked to increased cognitive load (Kennedy & Scholey, 2000). The source of these longer term tonic increases in infant HR over the course of habituation, however, is unclear. It can be argued that this increase in HR might be attributed to increased motor activity across in the sessions. HR increases with physical activity that makes metabolic demands, and some researchers have reported correlations between HR increases and motor activity (Byrne & Miller, 1988; Byrne & Smith-Martel, 1987; Gregg, Clifton, & Haith, 1976). However, other recent papers suggest that controlled or small movements that occur during attentional tasks may not produce appreciable increases in HR (O’Sullivan & Berthier, 2003; Porges et al., 2007). One study of motor activity and HR in habituation (Anderson, Rawlings, & Colombo, 2007) also reported equivocal findings with respect to the relation between movements and HR. Given that infants’ activity was restricted by placement in a car seat, the role of such movements in this change within sessions remains uncertain and should be the topic of future investigation. It may be that these longer term, tonic responses indicate some change related to learning and cognition. This, too, remains as a topic for future study. CLOSING COMMENTS For many years, visual habituation has been regarded as a tool through which other aspects of infant cognition might be accessed, and with a few exceptions, has not been the topic of study as an indicator of learning per se (Colombo & Mitchell, 2009). The current analyses address some gaps in our understanding of visual habituation in infants during the first year of life, and highlights the notion that the habituation paradigm is a rich source of information, and that critical questions still remain to be answered about the many and varied processes that it reflects. In answering the question of what habituates in infant habituation, the data provide some surprising answers: a stimulus onset–related HR increase and the HR-defined phase of AT. Overall HR also increased across trials. Also surprising is what does not habituate: the nature and depth of HR decelerations. The findings point up the importance of researching behavioral phenomena with concomitant measures at different levels of analysis. 122 COLOMBO ET AL. ACKNOWLEDGMENTS This research was supported by NIH grant R01-HD35903 to JC. Preparation of the manuscript was supported by the University of Kansas Intellectual and Developmental Disabilities Research Center (P30 HD002528), and the University of Kansas Center for Biobehavioral Neurosciences in Communication Disorders (P30 DC005803). Its contents are solely the responsibility of the authors and do not necessarily represent the views of these sponsoring agencies. These results were presented at the International Conference on Infant Studies in Vancouver, British Columbia, Canada, in April 2008. We are grateful to the 417 families that participated in this research. We also acknowledge the contributions of W. Allen Richman and Julie Maikranz to this project, and the suggestions of John Richards, Mary Courage, Greg Reynolds, and Les Cohen. REFERENCES Anderson, C. J., Rawlings, J., & Colombo, J. (2007). Spontaneous motor behavior and heart-rate defined phases of attention during visual habituation. Boston, MA: Society for Research in Child Development. Berkson, G., & Fitzgerald, F. L. (1963). Eye fixation aspect of attention to visual stimuli in infant chimpanzees. Science, 139, 586–587. Blaga, O. M., & Colombo, J. (2006). Visual processing and infant ocular latencies in the overlap paradigm. Developmental Psychology, 42, 1069–1076. Byrne, J. M., & Miller, C. L. (1988). Developmental course of early cardiac-somatic integration. Infant Behavior & Development, 11, 43–53. Byrne, J. M., & Smith-Martel, D. J. (1987). Cardiac-somatic integration: An index of visual attention. Infant Behavior & Development, 10, 493–500. Colombo, J. (2001). The development of visual attention in infancy. Annual Review of Psychology, 52, 337–367. Colombo, J. (2002). Infant attention grows up: The emergence of a developmental cognitive neuroscience perspective. Current Directions in Psychological Science, 11, 196–200. Colombo, J., Frick, J. E., & Gorman, S. A. (1997). Sensitization during visual habituation sequences: Procedural effects and individual differences. Journal of Experimental Child Psychology, 67, 223–235. Colombo, J., & Horowitz, F. D. (1985). A parametric study of the infant control procedure. Infant Behavior & Development, 8, 117–121. Colombo, J., & Mitchell, D. W. (1988). Infant visual habituation: In defense of an informationprocessing analysis. Cahiers de Psychologie Cognitive ⁄ Current Psychology of Cognition, 8, 455–461. Colombo, J., & Mitchell, D. W. (1990). Individual differences in early visual attention: Fixation time and information processing. In J. Colombo & J. W. Fagen (Eds.), Individual differences in infancy: Reliability, stability, prediction (pp. 193–227). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. WHAT HABITUATES IN INFANT VISUAL HABITUATION? 123 Colombo, J., & Mitchell, D. W. (2009). Infant visual habituation. Neurobiology of Learning and Memory, 92, 225–234. Colombo, J., Richman, W. A., Shaddy, D. J., Greenhoot, A. F., & Maikranz, J. M. (2001). Heart rate-defined phases of attention, look duration, and infant performance in the pairedcomparison paradigm. Child Development, 72, 1605–1616. Colombo, J., Shaddy, D. J., Blaga, O. M., Anderson, C. J., Kannass, K. N., & Richman, W. A. (2009). Early attentional predictors of vocabulary in childhood. In J. Colombo, P. McCardle, & L. Freund (Eds.), Infant pathways to language: Methods, models, and research directions (pp. 143–167). New York: Psychology Press. Colombo, J., Shaddy, D. J., Richman, W. A., Maikranz, J. M., & Blaga, O. M. (2004). The developmental course of habituation in infancy and preschool outcome. Infancy, 5, 1–38. Courage, M. L., Reynolds, G. D., & Richards, J. E. (2006). Infants’ attention to patterned stimuli: Developmental change from 3 to 12 months of age. Child Development, 77, 680–695. Frick, J. E., Colombo, J., & Saxon, T. F. (1999). Individual and developmental differences in disengagement of fixation in early infancy. Child Development, 70, 537–548. Graham, F. K., & Clifton, R. K. (1966). Heart-rate change as a component of orienting response. Psychological Bulletin, 65, 305–320.. Gregg, C., Clifton, R. K., & Haith, M. M. (1976). A possible explanation for the frequent failure to find cardiac orienting in the newborn infant. Developmental Psychology, 12, 75–76. Groves, P. M., & Thompson, R. F. (1970). Habituation: A dual process theory. Psychological Review, 77, 419–450. Jeffrey, W. E. (1968). Orienting reflex and attention in cognitive development. Psychological Review, 75, 323–334. Kaplan, P. S., & Werner, J. S. (1986). Habituation, response to novelty, and dishabituation in human infants: Tests of a dual-process theory of visual attention. Journal of Experimental Child Psychology, 42, 199–217. Kennedy, D. O., & Scholey, A. B. (2000). Glucose administration, heart rate and cognitive performance: Effects of increasing mental effort. Psychopharmacology, 149, 63–71. Maikranz, J. M., Colombo, J., Richman, W. A., & Frick, J. E. (2000). Autonomic correlates of individual differences in sensitization and look duration during infancy. Infant Behavior & Development, 23, 137–151. Malcuit, G., Pomerleau, A., & Lamarre, G. (1988). Habituation, visual fixation and cognitive activity in infants—A critical analysis and attempt at a new formulation. Cahiers de Psychologie Cognitive ⁄ Current Psychology of Cognition, 8, 415–440. Miller, D. J., Ryan, E. B., Sinnott, J. P., & Wilson, M. A. (1976). Serial habituation in 2-monthold, 3-month-old, and 4-month-old infants. Child Development, 47, 341–349. O’Sullivan, L. P., & Berthier, N. E. (2003). Attention during looking and reaching as assessed by heart rate in 7½-month-old infants. Developmental Psychobiology, 42, 292–300. Pancratz, C. N., & Cohen, L. B. (1970). Recovery of habituation in infants. Journal of Experimental Child Psychology, 9, 208–216. Porges, S. W., Heilman, K. J., Bazhenova, O. V., Bal, E., Doussard-Roosevelt, J. A., & Koledin, M. (2007). Does motor activity during psychophysiological paradigms confound the quantification and interpretation of heart rate and heart rate variability measures in young children? Developmental Psychobiology, 49, 485–494. Posner, M. I., & Petersen, S. E. (1990). The attention system of the human brain. Annual Review of Neuroscience, 13, 25–42. Rankin, C. H., Abrams, T., Barry, R. J., Bhatnagar, S., Clayton, D. F., Colombo, J., Thompson, R.F. (2009). Habituation revisited: An updated and revised description of the 124 COLOMBO ET AL. behavioral characteristics of habituation. Neurobiology of Learning and Memory, 92(2), 135– 138. Richards, J. E. (1985). The development of sustained visual-attention in infants from 14 to 26 weeks of age. Psychophysiology, 22, 409–416. Richards, J. E. (1987). Infant visual sustained attention and respiratory sinus arrhythmia. Child Development, 58, 488–496. Richards, J. E. (1989a). Development and stability in visual sustained attention in 14, 20, and 26 week old infants. Psychophysiology, 26, 422–430. Richards, J. E. (1989b). Sustained visual-attention in 8-week-old infants. Infant Behavior & Development, 12, 425–436. Richards, J. E. (1997a). Effects of attention on infants’ preference for briefly exposed visual stimuli in the paired-comparison recognition-memory paradigm. Developmental Psychology, 33, 22–31. Richards, J. E. (1997b). Peripheral stimulus localization by infants: Attention, age, and individual differences in heart rate variability. Journal of Experimental Psychology-Human Perception and Performance, 23, 667–680. Richards, J. E. (1998). Focusing on visual attention. Early Development & Parenting, 7, 153– 158. Richards, J. E., & Casey, B. J. (1992). Development of sustained visual attention in the human infant. In B. A. Campbell, H. Hayne, & R. Richardson (Eds.), Attention and information processing in infants and adults: Perspectives from human and animal research (pp. 30–60). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Schoner, G., & Thelen, E. (2006). Using dynamic field theory to rethink infant habituation. Psychological Review, 113, 273–299. Sokolov, E. N. (1958). Perception and the conditioned reflex. Oxford: Publishing House Moscow University. Sokolov, E. N. (1960). Neural model of the stimulus and the orienting reflex. Voprosy Psychologii, 4, 61–72. Thompson, R. F., & Spencer, W. A. (1966). Habituation: A model phenomenon for study of neuronal substrates of behavior. Psychological Review, 73, 16–43.