忆阻Chay神经网络系统同步转迁与时空动力学特性研究

Research on Synchronous Transition and Spatiotemporal Dynamics Characteristics of Memristor Chay Neural Network System

摘要: 为了探究电学突触与初始扰动对神经网络同步转迁的调控机制，本文基于磁控忆阻器模拟神经元电磁感应效应，构建含电耦合忆阻环状神经网络模型。利用双参同步云图、时空动力学图和时域图，分析不同参数与初始扰动下神经网络的集群动态演化规律。仿真结果表明，电耦合是决定网络同步能力的核心参数，增大电耦合强度可有效拓宽同步区间，正向磁耦合强度有利于促进神经元同步。全域梯度初值扰动下，网络仍可依靠耦合作用实现稳定全域簇放电；局部神经元初始扰动在弱耦合条件下会破坏同步，诱发无序放电与行波传播。随着耦合持续增强，神经元间的相互作用逐渐抵消初始相位偏差，系统重新收敛至全局簇同步，揭示电耦合与初始异常电位扰动协同调控神经集群动力学的内在机理。

Abstract: To explore the regulatory mechanism of electrical synapses and initial disturbances on the synchronous transition of neural networks, a magnetically controlled memristor is used to simulate the electromagnetic induction effect of neurons, and a ring neural network model with electrically coupled memristor is constructed. By using two-parameter synchronous cloud maps, spatiotemporal dynamics maps and time-domain maps, the dynamic evolution laws of neural network clusters under different parameters and initial disturbances are analyzed. The simulation results show that electrical coupling is the core parameter determining the synchronization ability of the network. Increasing the electrical coupling strength can effectively broaden the synchronization interval, and the positive magnetic coupling strength is conducive to promoting the synchronization of neurons. Under global gradient initial value perturbation, the network can still achieve stable global cluster discharge by relying on coupling effects. Under weak coupling conditions, the initial disturbances of local neurons will disrupt synchronization, induce disordered discharge and traveling wave propagation. As the coupling continues to strengthen, the interaction between neurons gradually cancels the initial phase deviation, and the system converges back to global cluster synchronization, revealing the intrinsic mechanism by which electrical coupling and initial abnormal potential disturbance jointly regulate the dynamics of neural clusters.