文献类型：期刊文献

英文题名：Energy Aspects and Synchronizations Comparison of Memristive and Adaptive Neurons

作者：Wu, Fuqiang[1]; Wang, Rubin[1,2]

机构：[1] The Institute for Cognitive Neurodynamics, East China University of Science and Technology, Shanghai, 200237, China; [2] School of Computer Science, Hangzhou Dianzi University, Hangzhou, China

年份：2022

外文期刊名：SSRN

收录：EI(收录号：20220276687)

语种：英文

外文关键词：Dynamics - Electromagnetic induction - Hamiltonians - Memristors - Neural networks - Synchronization

摘要：Employing memristors to mimic biological synapses or to describe electromagnetic induction effects attracts much attention in current researches. Neurons behaved as nonlinear oscillators develop complex dynamics and interact to process information under the effect of electromagnetic induction. While the dynamics of single Hindmarsh-Rose neuron with respect to the adaptive current and the inductive current has been investigated, comparisons of both are less clear from the energy aspects and synchronization points of view. In this paper, a memristive Hindmarsh-Rose ( m HR) neuron model is proposed by using an electromagnetic inductive current involving quadratic memconductance to replace the adaptive current in the classical 3-D HR model ( a HR). The energy function of m HR is deduced by the generalized Hamiltonian formalism. The m HR neuron under inductive current maintains a lower level of energy than the a HR neuron under adaptive current. In particular, two m HR neurons with irregular firing reach phase synchronization under the electrical synapse, while achieve complete synchronization under the magnetic flux coupling. The coupled neuron operates with economy energy in the region of whether complete synchronization or phase synchronization. Corresponding results are further demonstrated by designing digital circuit. Comparing between the a HR neurons and the m HR neurons from synchronization and energy aspects, pursuant requirements can better choose appropriate model to explore dynamics of larger-scale neuronal network. ? 2022, The Authors. All rights reserved.