变权小世界生物神经网络的兴奋及优化特性

根据实际生物神经网络具有小世界连接和神经元之间的连接强度随时间变化的特点，首先构造了一个以Hodgkin-Huxley方程为节点动力学模型的动态变权小世界生物神经网络模型，然后研究了该模型神经元的兴奋特性、权值变化特点和不同的学习系数对神经元的兴奋统计特性的影响.最有意义的结果是，在同样的网络结构、网络参数及外部刺激信号的条件下，学习系数b存在一个最优值b^*，使生物神经网络的兴奋度在b=b^*时达到最大. 关键词： 动态变权生物神经网络 小世界网络 Hodgkin-Huxley方程

Excitement and optimality properties of small-world biological neural networks with updated weights

As the biological neural networks have small-world property and updating connection weights with time, we accordingly propose a new model of small-world biological neural networks based on biophysical Hodgkin-Huxley(H-H)neurons with updated weights. Then we study the statistical properties of excitement of this model and the updating of weights. The results show that, for networks with the same structure and parameters and external stimulation, there exists an optimal learning rate ralue b^* which makes the excitement strength of biological neural networks strongest.