NEURAL NETWORK MODEL WITH DYNAMIC NEURONS I. COMPUTER SIMULATION
动态神经元的网络模型 Ⅱ.计算机仿真

On the basis of the dynamical model of neural networks proposed in a previous paper1], simulation was made with an IBM-PC computer. The simulation results showed; Our unit modelcould simulate the adaptation, inhibition after excitation, phase locking and position coding of receptors; A lateral inhibitory network composed of 50 such units could mimic the transient characteristics of the lateral inhibitory network in Limulus, and showed Mach band phenomenon when it reached the steady state. The simulation result also predicted that the network was especially sensitive to moving objects. The coincidence of the model and the biological neuron on the inherent mechanism of information processing and the coincidence of the characteristics of the composed lateral inhibitory network and the compound eye of Limulus sugested that the model would be promising to be a model more similar to the biological neural networks.