Objectives: Here we used well-characterized electrophysiological assays of automatic contour completion to test whether a tendency toward the local, a characteristic of the autism phenotype, might originate in deficits in early automatic contour completion processes.
Methods: A common approach to understanding completion processes uses stimuli with incomplete contours that induce perception of complete contours –illusory contour (IC) stimuli (Kanizsa 1976). These are useful because rearrangement of elements of identical stimulus energy gives rise to different percepts. In the illusion-inducing configuration, continuous contours appear to form a two-dimensional object. In the non-inducing arrangement only the inducers are typically perceived. Robust modulation as a function of IC status of the visually-evoked potential (VEP) provides an index of the neural contributions underlying this perceptual change (Murray et al. 2002).
ERPs to IC and non-IC stimuli were compared between children with a diagnosis of ASD and their neurotypical counterparts, in three age cohorts: 6-9, 10-12, and 13-17. IC-inducing stimuli were presented at three spatial extents (4, 7 and 10 degrees). Participants performed an unrelated task. IC-effect amplitudes and latencies were measured for two well-characterized modulations of the VEP associated with perceptual completion: The IC-effect, occurring during the N1 timeframe (~100-200 ms) and associated with automatic filling-in of boundaries (Shpaner et al. 2009) and the Ncl, occurring between ~250-400 ms, reflecting more effortful visual object processing
Results: In typically developing controls, across all age-groups, modulation of the ERP as a function of IC status was observed in the N1-timeframe. This was followed by Nclmodulation that was greatest for the two younger groups. Spatial extent of the inducers did not influence contour integration processes in any age-group. In the ASD group, contour completion processes were less robust across age groups and across the manipulation of spatial extent, suggesting greater inter-subject variability in this group. Notably however, when IC-effects were present they were observed in both early and late timeframes.
Conclusions: In the ASD cohort we find greater variability in the automatic completion of visual information, as indexed by IC-effects. Data will be further analyzed to determine if there is a bimodal distribution within this group whereby some individuals show automatic contour completion and others do not, and whether this correlates with an index of ASD phenotype.