A self-organizing model for the development of ocular dominance and orientation columns in the visual cortex

Abstract

A binocular model describing the ontogenetic development in the visual nervous system is presented. It consists of a set of deterministic differential equations which have been derived from an statistical approach. The evolution of the solution is led by the spontaneous generation of input activity, characterized in this model by its spatial and temporal decorrelation. The development of a connection depends on the output activity of both connected neurons; for this purpose, Hebbian and anti-Hebbian learning have been used. The model can explain some properties observed in natural brains such as the appearance of ocular domains and orientation selectivity in the V1 visual cortex development.