Date of Degree


Document Type


Degree Name





John J. Foxe

Committee Members

Sophie Molholm

Jennifer Mangels

Hilary Gomes

Lynne E. Bernstein

Subject Categories

Cognitive Neuroscience


Sensory input is inherently ambiguous and complex, so perception is believed to be achieved by combining incoming sensory information with prior knowledge. One model envisions the grouping of sensory features (the local dimensions of stimuli) to be the outcome of a predictive process relying on prior experience (the global dimension of stimuli) to disambiguate possible configurations those elements could take. Contour integration, the linking of aligned but separate visual elements, is one example of perceptual grouping. Kanizsa-type illusory contour (IC) stimuli have been widely used to explore contour integration processing. Consisting of two conditions which differ only in the alignment of their inducing elements, one induces the experience of a shape apparently defined by a contour and the second does not. This contour has no counterpart in actual visual space – it is the visual system that fills-in the gap between inducing elements. A well-tested electrophysiological index associated with this process (the IC-effect) provided us with a metric of the visual system’s contribution to contour integration. Using visually evoked potentials (VEP), we began by probing the limits of this metric to three manipulations of contour parameters previously shown to impact subjective experience of illusion strength. Next we detailed the developmental trajectory of contour integration processes over childhood and adolescence. Finally, because persons with autism spectrum disorders (ASDs) have demonstrated an altered balance of global and local processing, we hypothesized that contour integration may be atypical. We compared typical development to development in persons with ASDs to reveal possible mechanisms underlying this processing difference. Our manipulations resulted in no differences in the strength of the IC-effect in adults or children in either group. However, timing of the IC-effect was delayed in two instances: 1) peak latency was delayed by increasing the extent of contour to be filled-in relative to overall IC size and 2) onset latency was delayed in participants with ASDs relative to their neurotypical counterparts.