Chaotic dynamics and synchronization of multi-region neural network based on locally active memristor

The influence of neural synapses on the collaboration of different brain regions is an urgent need for current research. In this paper, a multi-region neural network (MRNN) is proposed using multistable locally-active memristor (MLAM). A new memristor is first designed with multistable, non-volatile, and locally-active. Then, the memristor is modeled as a neural synapse connecting two different regions to construct the MRNN, which is a multistable locally-active memristive Hopfield neural network. The neural network exhibits rich chaotic dynamics, and the dynamic coupling strength of the synapse is analyzed using bifurcation, phase diagrams, and two-parameter chaotic maps. The neural network also demonstrates self-boosting of attractors driven by the parameters of synapse. The effect of memristive parameters on the self-boosting of the attractor is revealed by describing the phase diagram and the basin of attraction. In order to explore the collective behavior of the proposed network, controllers are further designed to realize the state synchronization cross multiple brain regions.