Abstract:With the growing exploration of brain actions, memristive elements with biomimetic properties are urgently needed to estimate the activities of biological synapses. Based on this, a discrete memcapacitor is used as memristive synapses, which are applied in discrete neuron map to construct a memcapacitive-synapse neuron model in this paper. Firstly, the characteristics of the memcapacitor are studied, and its capability to perform memory behavior is demonstrated. Secondly, numerical methods are used to investigate the bionic behaviors and complex dynamical behaviors of the memcapacitive-synapse neuron model, including extreme multistability, and multiple firing patterns, which are all tightly related to the parameters of the memcapacitive-synapse. Finally, the chaotic attractor generated by this neuron model is also implemented based on a DSP hardware platform. It is justified from different perspectives that it is reasonable and feasible to adopt memcapacitor to estimated synapse behaviors.

A Novel Memcapacitive-Synapse Neuron: Bionic Modeling, Complex Dynamics Analysis and Circuit Implementation