To recognize and interact with objects in an object-filled environment, an accurate perceptual representation of the surfaces that define the objects is essential. The surface representation process integrates local feature information from the early cortical processes to form surfaces. Complicating the process of representing surfaces, however, is the fact that most objects overlap one another, resulting in the occluded parts of the objects' surfaces not being encoded by the early cortical processes. Thus, to accurately represent surfaces, the surface representation process uses an arsenal of perceptual rules derived from environmental constants even as it relies on the early features of the non-occluded parts. But how these perceptual rules are implemented, at which stage of the surface representation process, and how top-down attention modulates the representation of both the occluding and occluded surfaces are not well understood. In light of these, our proposal addresses three fundamental issues in representing surfaces. The issues are: A. The boundary contour and surface property information in amodal surface interpolation B. The operational constraints in surface formation during binocular viewing C. Inhibitory mechanism and plasticity in binocular surface perception These issues will be investigated using psychophysical methods on human observers with normal vision, and those with significant sensory eye dominance. Observers in the latter group tend to suppress the image viewed by the non-dominant eye when the two eyes see different images; a situation almost resembling the extreme condition afflicting strabismic amblyopia patients. In this proposal, we will test and evaluate a new perceptual learning paradigm aimed at reducing sensory eye dominance and improving perception. Therefore, the outcomes of our investigations will not only further the knowledge of perceptual and attention mechanisms in perceiving surfaces, but also has a clinical significance. ? ?
Showing the most recent 10 out of 15 publications
