DYNAMICAL COMPUTATIONAL PROPERTIES OF LOCAL CORTICAL NETWORKS FOR VISUAL AND MOTOR PROCESSING - A BAYESIAN FRAMEWORK

A major unsolved question concerns the interaction between the coding of information in the cortex and the collective neural operations (suc h as perceptual grouping, mental rotation) that can be performed on th is information. A key propel-ty of the local networks in the cerebral cortex is to combine thalamocortical or feedforward information with h orizontal cortico-cortical connections. Among different types of neura l networks compatible with the known functional and architectural prop erties of the cortex, we show that there exist interesting bayesian so lutions resulting in an optimal collective decision made by the neuron al population. We suggest that thalamo-cortical and corticocortical sy naptic plasticity can be differentially modulated to optimize this col lective bayesian decision process. We take two examples of cortical dy namics, one for perceptual grouping in MT, and the other one for menta l rotation in M1. We show that a neural implementation of the bayesian principle is both computationally efficient to perform these tasks an d consistent with the experimental data on the related neuronal activi ties. A major implication is that a similar collective decision mechan ism should exist in different cortical regions due to the similarity o f the cortical functional architecture.