Objectives: This study was designed to investigate whether social or nonsocial stimuli elicit differential disengagement patterns in children with autism by comparing saccadic reaction times (SRTs) between a group of school-aged children with autism (n=13, M=136.5 months; sd=24) and a group of typically developing control children (n=14, M=166.5 months; sd=28). 

Methods: Data were recorded on a Tobii 1750 eye-tracker.  Diagnosis of the autism group was confirmed with an ADI-R and all children had IQs > 70.  The gap-overlap paradigm consists of a gap condition in which a central stimulus disappears before the onset of a peripheral target, and an overlap condition in which the central stimulus remains visible after the onset of the peripheral target.  Disengagement was operationalized as the latency to initiate an eye movement from the center of the array toward the peripheral target.

Results: Both groups demonstrated the gap effect, defined as increased SRTs in the overlap condition.  The autism group had significantly increased SRTs in both the overlap and the gap condition when compared to the control group.  In the overlap condition, the control group disengaged significantly faster from social stimuli, whereas the autism group disengaged equally slow from social and nonsocial central stimuli.  Interestingly, the groups did not differ in SRT in the nonsocial overlap condition. 

Conclusions: Typical children disengaged their attention faster from a social stimulus (e.g. a face), but children with autism disengaged slower than typically developing children regardless of the nature of the central stimulus.  Disengaging visual attention is a primary component of cognition that may contribute to the deficits in social behavior observed in autism.