Identification of a Linearized Fitzhugh-Nagumo Neuron Model using a Combinational Multi Layer Perception Network

A.R. Kashaninia and S. Sadughi (Iran)


Multi Layer Perceptron (MLP), Linear Oscillatory Neuron (LON) model, Nonlinear System Identification (NSI), parametric and nonparametric SI, controllability of linear systems. 1-


: In this paper, the authors have focused their attention on developing a model, which is capable of being identified so that it mimics faithfully the behavior of an individual physiological neuron from input-output point of view. A hierarchical method is introduced to identify the parameterized nonlinear FN equation with respect to the achieved input-output data from a physiological trial on a neuron. In this hierarchy, four steps are outstanding, which are: linearization of nonlinear FN in the rest regime of Action Potential, identification of the parameters of this linear system with hard data, substitution of achieved parameters in original nonlinear system and finally, fine tuning the derived system via a combinational Artificial Neural Network. The output of achieved model is well consistent with the output of its corresponding physiological neuron. This idea can be followed for neuroanatomical bypassing purposes in living bodies.

Important Links:

Go Back

Rotating Call For Paper Image