A coupled oscillatory neural network model for binding problem

A neural network model for the binding problem in visual system is proposed and analyzed in this paper. It contains clusters of neurons, and each of clusters behaves as a coupled oscillator, whose activity is represented by its phase. Two networks are composed of these clusters, and they are responsible for the different properties of an input object. Connections between clusters and between networks are defined. The numerical results show that the clusters in a network could extract distinct objects by the synchronization of clusters and the clusters corresponding to the same object between networks are coherent, there by these behaviors exhibit the binding of information.