Izhikevich E M, Mikhailov A S, Sveshnikov N A
Department of Applied Mathematics and Cybernetics, Moscow State University, USSR.
Biosystems. 1991;25(4):219-29. doi: 10.1016/0303-2647(91)90020-l.
We consider a model of a neural network where the individual cells interact only by releasing and absorbing the molecules of a neuromediator. We show that such a system can realize the function of associative memory. A learning mechanism based on the chemotaxis is proposed and numerically investigated.