联想记忆和模式分割是大脑的重要功能,大脑海马区是这一功能的重要物质基础。Hindmarsh-Rose神经元模型能够较好地模拟大脑海马区神经元的特性。本文以Hindmarsh-Rose神经元为节点,按Hopfield规则构造网络,同时,考虑到大脑中神经元连接的生理学实际情况,通过数值模拟,实现了在弱、中、强三种连接强度情况下的模式分割以及在随机概率连接情况下的模式分割,通过分析比较,提出了进一步改进的方向。
Associative memory and pattern segmentation are main functions of human's brain. Hippocampus of the brain is the important base of the two functions. Hindmarsh-Rose neuron model can simulate the neurons' activities in hippocampus well. Hindmarsh-Rose neuron was used as the node and the network was constructed according to the Hopfield rule. Considering the actual circumstance in the brain, the pattern segmentation was studied in the weak, medium and strong connection strength, and good results were obtained via numerical modeling. At the same time, the pattern segmentation was studied under random connection and the direction of improvement was proposed.