忆阻Morris-Lecar神经网络的同步行为研究  

Synchronization Behavior of Memristor Morris-Lecar Neural Networks

作　　者：黄微 韦笃取[1] HUANG Wei;WEI Duqu(College of Electronic and Information Engineering,Guangxi Normal University,Guilin Guangxi 541004,China)

机构地区：[1]广西师范大学电子与信息工程学院,广西桂林541004

出　　处：《广西师范大学学报（自然科学版）》2024年第1期67-78,共12页Journal of Guangxi Normal University:Natural Science Edition

基　　金：国家自然科学基金(62062014);广西自然科学基金(2021JJA170004)。

摘　　要：本文研究电磁感应系数和膜表面电荷尺寸对忆阻Morris-Lecar(ML)神经网络动力学行为的影响。以电荷尺寸为参数研究忆阻ML神经网络的放电行为,发现增大电荷尺寸会抑制神经元放电频率和幅度,神经元随着时间的推移进入静息状态。以电磁感应系数和电荷尺寸作为变量研究神经网络的同步行为。固定电荷尺寸,增大电磁感应系数,神经网络由混沌变得有序;电磁感应系数是决定神经网络同步的关键因素,当电磁感应系数K>3.2时,能更有效地诱导和增强同步。引入同步因子和标准差分别对网络的同步程度和神经元放电程度进行量化。实验结果表明,电磁感应系数可以有效调节神经网络的同步行为,具有控制神经网络的放电模式和网络同步的作用。The effects of electromagnetic induction coefficient and membrane surface charge size on the memristor Morris-Lecar(ML)neural network dynamics are investigated.Using charge size as a parameter to study the discharge behavior of memristor ML neural network,it is found that increasing charge size will inhibit the firing frequency and amplitude of neurons,and the neurons will enter the resting state over time.The synchronization behavior of neural network is studied with electromagnetic induction coefficient and charge size as variables.When the electric charge size is fixed and the electromagnetic induction coefficient is increased,the neural network becomes orderly from chaos.The electromagnetic induction coefficient is the key factor to determine neural network synchronization.When the electromagnetic induction coefficient K>3.2,the synchronization can be induced and enhanced more effectively.Synchronization factor and standard deviation are introduced to quantify the degree of network synchronization and neuron firing.The experimental results show that the electromagnetic induction coefficient can effectively regulate the synchronization behavior of neural network,and has the function of controlling the discharge mode and network synchronization.

关 键 词：Morris-Lecar 神经元 电磁感应 表面电荷 同步放电 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]