Coherence and clustering in ensembles of neural networks

Abstract

Large ensembles of globally coupled chaotic neural networks undergo a transition to complete synchronization for high coupling intensities. The onset of this fully coherent behavior is preceded by a regime where clusters of networks with identical activity are spontaneously formed. In these regimes of coherent collective evolution the dynamics of each neural network is still chaotic. These results may be relevant to the study of systems where interaction between elements is able to give rise to coherent complex behavior, such as in cortex activity or in insect societies.