Our findings on disengagement in autism parallel those reported in normal 2-month-olds, in whom attention has been described as 'obligatory'. Discussion focuses on the potential role of general versus domain-specific processes in producing some of the core features of autism.

No abstract

We examined both the reliability of infant-controlled habituation and patterns of responding in a group of high-risk infants (mean age 16 mos). Good test-retest reliability was found for mean, total, and baseline looking time for one of the stimuli. Classification of infants' pattern of performance yielded two groups: linear (57%) and non-linear (29%) responders; 14% of cases could not be classified because of cross-session inconsistencies in performance. Linear responders had shorter total looking times, fewer trials to criterion, and showed more habituation than non-linear responders, thus validating the linear/non-linear distinction. Our findings indicate that the infant-controlled habituation task can be used reliably with infants who are at high risk for developmental disorders. Indeed, we provide evidence that this task is not only reliable, but may also provide meaningful distinctions between infants within the high-risk population. Discussion focuses on the role of attention in distinguishing between short (linear) and long (non-linear) lookers. Copyright # 2003 John Wiley & Sons, Ltd.Key words: habituation; high risk infants; reliability; attention; cognitive development Habituation is a process whereby an organism ceases to look at a repeated stimulus and, conversely, shows a preference for a novel stimulus (Malcuit et al.,