基于同构神经元簇的局域神经网络模型研究

Research of Local Neural Network Based on Homology Neural Cluster

该文在分析神经元离子通道及生物连接方式的基础上,提出同构神经元簇的概念,介绍了它的特点及采用同构神经元簇对外界特征表达的优势。在仔细分析单个神经元的非线性动力学特性之后,以时空整合的方法,构建了一个更符合同构神经元簇发放机制的非线形动力学方程,并结合离子通道放电特性提出阈值变化率的概念。最后对基于同构神经元簇局域神经网络的动力学模型仿真结果进行了分析,认为采用同构神经元簇构成的异构网络具有更符合人脑工作机制的发放特性。

Based on analyzing the structure ion channel and neural combination the paper gives the conception of homology neural cluster and introduces the characteristic of homology cluster and the advantage of representing external stimulus character . After analyzing the nonlinear dynamics of single neuron, the paper provides a nonlinear dynamics equation which is more appropriate to the firing processing of neural cluster, and gives the concept of threshold variety . At last the paper analyzes the simulation results of nonlinear dynamic model for local neural network based on homology neural cluster and argues that the model of isomerous network composed of homology neural is more advantageous than the previous models.