Dynamics of Hopfield Neural Networks with Synaptic Crosstalk Under Electromagnetic Induction
摘要
Electromagnetic induction and synaptic crosstalk are two different phenomena that exist in nervous system. Electromagnetic induction is mainly caused by the voltage difference between two neurons, while synaptic crosstalk is caused by the interaction of two synapses. In this study, we simulate the effects of electromagnetic induction and synaptic crosstalk on a Hopfield neural network using a hyperbolic-type memristor and coupled hyperbolic memristors. By using numerical measures including phase diagram, bifurcation diagram, basin of attraction and dynamic map, it is found that the new system model generates multiple coexisting behaviors depending on initial condition and various chaotic attractors are emerged relying on system parameters. That is, considering the effects of electromagnetic induction and synaptic crosstalk at the same time will lead to more complex dynamic behaviors of the system, which is expected to give new insights for the interpretation of brain neuroscience and the construction of artificial neurons. In addition, DSP implantation and Pspice simulation platform are used to complete the verification of MATLAB numerical simulation.