Our prior research utilizing random element displays that depict global stereoscopic and global kinetic forms has revealed that mildly mentally retarded adults have fundamental impairments in perceptual abilities that arise early in the visual processing of information. The specific nature of the impairments suggested that depth and motion are not accurately perceived by mildly mentally retarded adults and suggested, therefore, that tasks that require the subsequent use of this information would be impaired. In addition to the basic difficulties encountered in perceiving depth and motion we observed that reductions in the element density of the displays profoundly impaired the ability of the mildly mentally retarded adults to discriminate between objects of different shapes. This finding was quite surprising in that nonretarded adults are not impaired by element density reductions. Indeed, they report a phenomenal filling-in of boundaries such that the subjective contours enable vertical shape perception. This process of filling-in, termed interpolation by various computational theorists, is not restricted to random element displays. Instead it is theorized to be a general property of perceptual systems that affects performance ranging from visual segmentation skills (such as involved with figure-ground detections), shape perception, and perception-action skills (also called motor skills). The proposed research is designed to more specifically investigate the hypotheses that (1) deficits in perceiving depth and motion will impair performance on many types of tasks, and (2) interpolative abilities are impaired in mildly mentally retarded adults. These hypotheses will be evaluated by investigating four lines of research derived from a consideration of various theoretical approaches to perception including computational vision. First, sensitivity to elementary image features such as orientation are examined. Second, visual segmentation and filling-in abilities are explored. Third, the ability to utilize elementary features for object shape perception is considered. Lastly, performance on perception-action tasks as a function of perceptual status is examined.
