Symposium
Chair: Martha A. Bell
Discussant: Nora S. Newcombe
Both looking and reaching tasks have been used to assess innumerablecognitive abilities during infancy. Evidence that specific lookingmeasures examined during infancy are related to later cognitive functioninghas firmly established looking as a standard measure in many developmentalresearch labs. Evidence that specific reaching measures examined duringinfancy are related to higher order cognitive concepts in older childrenhas ensconced reaching as a standard measure in other developmentalresearch labs. The purpose of this symposium is to highlight contemporaryuses of looking and reaching tasks in the examination of infantdevelopment. The first participant will discuss use of a reaching task toexamine the development of perseverative reaching errors in a longitudinalsample where the objective is to reach for targets, not search for hiddenobjects. The second participant will relate looking to issues ofexpectancy in both cognitive and social tasks. The third participant willcompare looking and reaching versions of the classic A-not-B task in alongitudinal sample and provide evidence for comparable performance on theobject permanence tasks beginning at 8 months of age. The final speakerwill present evidence suggesting that object permanence explanations, notmeans-ends explanations, best describe the reaching and looking behaviorsthat young infants display during search tasks. Finally, the discussantwill integrate the four presentations and examine looking and reachingtasks in terms of current abilities and imminent development.
Details of individual items:
paper
From a dynamic systems perspective, perseverative reaching errors emergebecause the act of reaching creates a perceptual-motor habit which lingersafter each reach to the first location, influencing subsequent reaches.Based on this, we predicted that younger infants would perseverate lessthan older infants for 2 reasons. First, infants should reach lessconsistently on A trials because identical targets do not offer clearspecification of which side was just cued. Second, young infants' armmovements are too unstable for them to repeat the same movement and thusbuild up a strong motor memory. Based on these factors, new reachers (5months) should be less likely to perseverate, whereas stable reachers (8months) should be more likely to perseverate.We tested 14 infants monthly from 5 months through 8 months of age on 2perseverative reaching tasks, one with identical targets and one withdistinctive targets. Infants were presented with 6 A trials and 2 Btrials, all at a 3 second delay. We examined perseverative responses byobserving where infants reached on the first B trial, as well as theirpattern of reaching throughout the A trials.On the identical target task, of the 5-month-olds that reached, 2/3 reachedcorrectly on the A on the A and B trials. Performance steadily decreasedas infants got older. The majority of infants at 6 months of age wereincorrect on A trials, and random on B trials. By 8 months, 85% of theinfants perseverated, which was significantly above chance, p < .05. Inthe distinctive target task, infants of all ages perseverated close to 50%,and did not significantly differ from chance across all sessions.Interestingly, the 5- and 6-month-olds demonstrated the same reachingpattern when targets were distinctive as when the targets were identical,whereas 7- and 8-month-olds reached differently on the two tasks. We suggest two possible related mechanisms to explain these surprisingresults. The first is a general increase in motor skills. Infants'reaches transition from jerky, roundabout movements to relatively smoothand straight reaches between seven and eight months of age (Thelen et al.,1996). Thus, younger infants may have been unable to form a strong memoryfor each reach because their reaches were highly variable. The reaches ofthe older infants, however, were stable enough to converge on a singlepattern and form a strong memory. The second related mechanism is attentionto the visual layout of the task. If young infants do not attend to thetask layout, they should follow the transient cue of the experimenterwaving the lid. Indeed, the youngest infants reached correctly on the Aand B trials for both target tasks. As the infants got older, theirattention to the distinctive task layout increased, thus providinglandmarks that may help them avoid confusion. The important message fromthese results is the remarkable nonlinearity of development outcomes. Thecomponent processes that produce correct reaches under some circumstancesare the very same processes that produce reaching errors under othersituations.
paper
This paper addresses the issue of how infant attention operates acrosscognitive and social domains. Researchers studying both cognitive andsocial development have studied infants' abilities to detect rules forevents, and infants' responses when those rules are violated. Althoughcognitive and social developmentalists have often worked in separate camps,there are intriguing similarities in their findings about infants'expectations, and responses to violations of expectancies. The study of infant cognitive development has capitalized on infants'propensity to orient selectively to novel events in the environment, andinfants' ability to learn rules governing stimulus events. Infantsdemonstrate their understanding of these relations through their overtbehavior (for example, looking), as well as physiological responses (forexample, changes in heart rate). Research on early social development has also indicated that infants aresensitive to rules, and quickly detect changes in the expression of thoserules. In the laboratory, the 'still-face' procedure provides a usefulmethodology to examine infants' responses to changes in socialcontingencies. Infants show rapid behavioral and physiological reactionsto alterations of their social expectancies, and this behavior alsoindicates an understanding of lawful relations in the environment. In our laboratory, we have been exploring similarities and differences ininfants' behavioral (looking) and physiological (heart rate andrespiration) responses in both cognitive and social tasks that involvelearning and expectations about environmental events. I will present datafrom a cross-sectional study of 4- and 6-month-old infants who were testedin both cognitive 'expectation' and social 'expectation' tasks, and willinclude analyses of both their looking responses (such as habituation,reaction time and gaze aversion) and physiological reactions. Similaritiesin infant attentional responses across the two procedures will bepresented, as well as developmental and individual differences in infantattention. I will conclude with a discussion of how behavioral andphysiological measures provide both converging and independent information,and how attention operates in terms of expectancies across domains forinfants.
paper
The neuropsychological model of A-not-B performance focuses on thematuration of the frontal lobe and the development of skills associatedwith working memory, inhibition, and attention. These cognitive skills areessential for successful performance on the task itself and, therefore,should remain consistent across response systems. Obviously, the differentresponse systems (i.e., gross motor for reaching responses and oculomotorfor looking responses) mature at different rates. Thus, there areindications that infants may exhibit success on the A-not-B task vialooking response at an age when the reaching response is still maturing.It should be the case, however, that upon maturation of the gross motorsystem associated with the reaching response, the cognitive skills for bothlooking and reaching versions of the task would be the same and, thus,performance on the two versions of the task should be comparable. The purpose of this study was to examine the development of performance onthe A-not-B task from 5 to 10 months of age in both reaching and lookingmodalities. The hypothesis was that performance on the looking version ofthe task would exceed performance on the reaching version from 5 to 7months, but that from 8 to 10 months performance on the two versions wouldbe comparable. Additional hypotheses concerning brain electrical activityand heart rate guided examination of electrophysiological data collectedduring task performance.Fourteen infants (8 male) were participants in this longitudinal study.Infants were seen monthly from 5 to 10 months of age and were assessed onboth looking and reaching versions of the A-not-B task and were given anordinal scale score (e.g., Bell & Fox, 1997; Kermoian & Campos, 1988). EEGand ECG were recorded during baseline and looking task conditions. As a group, infants demonstrated increased performance on both looking(p.001) and reaching (p<.001) versions of the task across this 6-monthtime period. From 5 to 7 months of age, infants performed at a higher levelon the looking version of the task (all p's<.001). From 8 to 10 months ofage, performance on the looking and reaching versions did not differ (allp's >.13). Associations among looking and reach task performance, frontalbrain electrical activity, and heart rate changes will be discussed. These within-subjects data call into question data which demonstrate alooking advantage for object permanence tasks. They also lend support forthe hypothesis that similar cognitive skills are required for both lookingand reaching versions of the task.
paper
This paper addresses the important issue of whether or not young infantsare able to represent occluded objects. In the first part, the results of4 new experiments that address the means-ends explanation of why infants donot reach for the occluded object in the Piagetian cloth task arepresented, showing that the means-ends explanation will not account for theresults. In the first study by Shinskey, Bogartz, & Poirier (submitted)it is shown that infants retrieve and contact an object more, contact thecurtain more, and look away less when the object is occluded by atransparent curtain than when occluded by a semitransparent or opaquecurtain. In the second study by Poirier, Shinskey, & Bogartz (inpreparation), various results with graded curtains and an opaque curtainwith a small hole in it are obtained again infirming the means-endsexplanation. The third study is Jeanne Shinskey's dissertation. InExperiment 1, 6- and 10-month-old infants were presented with an objectvisible in water, partly visible in milk, hidden in milk, or hidden under acloth. 6-month-olds searched less, manually and visually, when the objectwas hidden than when it was visible or partly visible, as predicted by adeficit in object permanence, whereas 10-month-olds showed no differences.The means-ends hypothesis prediction that 6-month-olds would search lesswhen the object was hidden by a cloth than when it was hidden by milk wasnot confirmed. In Experiment 2, 6- and 10-month-old infants were presentedwith an object visible behind a transparent curtain, partly visible througha hole in an opaque curtain, partly visible ( a lit flashlight under acloth), and hidden behind a completely opaque curtain. 6-month-oldssearched less when the object was hidden than when it was visible or partlyvisible, but there were no differences at 10 months. The results areinterpreted as favoring a lack of object permanence rather than ameans-ends deficit.The remainder of the paper considers various research that bears on thequestion of whether or not young infants represent occluded objects by (1)discussing studies by Bogartz, Shinskey & Schilling (submitted), Cashon &Cohen (submitted) and Schilling (submitted) indicating that familiaritypreference may be responsible for the effects obtained in the Baillargeon,Spelke, & Wasserman (1985) drawbridge study, (2) considering Leslies objectindex minimal representation of location approach to young infantrepresentation of occluded objects, and (3) presenting a new approach(Bogartz, submitted) to analyzing infant arithmetic designs which showsthat the Wynn (1992) and the Simon, Hespos, & Rochat (1995) studies ofinfant arithmetic may not involve representation of the occluded object butinstead be based simply on familiarity preferences. This new approachprovides a method for analyzing infant arithmetic experimental designs inorder to determine whether and how much different possible stimulus aspectsof the design are confounded.