STIMULUS-DEPENDENT CORRELATIONS IN STOCHASTIC NETWORKS

It has been observed that cortical neurons display synchronous firing for some stimuli and not for others. The resulting synchronous cell as semblies are thought to form the basis of object perception. In this p aper this ''dynamic linking'' phenomenon is demonstrated in networks o f binary neurons with stochastic dynamics. Analytical treatment within the mean field theory and linear response theory is possible and is c ompared with simulations. We establish that correlations are a sensiti ve function of the spatial coherence in the stimulus. We discuss the p ossibility to use these correlations as a mechanism for scene segmenta tion.