Ocular Dominance and Activation Dynamics in a Uni ed Self-Organizing Model of the Visual Cortex

A neural network model for the self-organization of lateral connections and ocular dominance columns from uncorrelated binocular input is presented. The self-organizing process results in a network where (1) aaerent weights of each neuron organize into smooth hill-shaped receptive elds primarily on one of the retinas, (2) neurons with common eye preference form connected, intertwined patches, and (3) lateral connections primarily link regions of the same eye preference. Similar self-organization of cor-tical structures has been observed experimentally in strabismic kittens. The lateral connectivity in the model mediates activation dynamics that could explain, for example, why divergent squinters cannot form a single , binocular percept even for a completely binocular input. The model provides a basis for computational study of cortical self-organization and plasticity, as well as dynamic perceptual processes such as feature grouping and binding.