Abstract
The formation of topographic maps in sensory cortex and the reorganization of such maps following repetitive stimulation and deafferentation have been successfully sim ulated with neural models. The reorganization that occurs after acute cortical lesions, however, has been simulated only with the addition of unphysiological assumptions. We describe a computational model based on the hypothesis that inhibition in the cor tex is mediated through the competitive distribution of activation. This model exhibits spontaneous reorganization after focal cortical lesions. Reorganization occurs in two phases, a very rapid shift in map organization due to competitive activation dynam ics, and a second slower phase due to synaptic plasticity at the cortical level. This sec ond phase is shown to be accelerated by concentrating sensory input around the lesion and by appropriately selecting stimuli. These results demonstrate how neural models can be used to generate testable hypotheses regarding rehabilitation techniques for stroke and other injuries to the central nervous system. Key Words: Neural model— Rehabilitation—Stroke—Somatosensory—Computer model.
