THE DYNAMICS OF NONLINEAR OCILLATOR NEURON BY THE ACTION OF EXTERNAL NON-STATIONARY SIGNAL

Abstract

Context. Deals with the problem of time-frequency and time dependence of the morphology of the signal at the output of nonlinear oscillatory neuron with regard to its threshold effect. The object of study is non-linear modified model of Van-der-Pol, which describes the dynamics of nonlinear oscillatory neurons under the action of various shape, frequency and amplitude of an external non-stationary signals. Objective. Construction of nonlinear mathematical models of oscillatory dynamics of a neuron given its threshold effect under the action of a neuron to external non-stationary signals. Method. In the approximation of Krylov-Bogoliubov-Mitropolskii the proposed method of successive approximations method for solving nonlinear differential equations of second order with quadratic nonlinearity of the unknown function with the first derivative. The proposed solution method allowed us to obtain the frequency-temporal and temporal dependence of the morphology of the signal at the output of nonlinear oscillatory neuron given its threshold effect when the effect on the neuron structure of the various input signals. Proposed coding information oscillatory nonlinear neuron on the basis of frequency modulation and decoding using the inverse operator which acts on the vector of the output signal. Results. A non-linear model, the oscillatory neuron is implemented in the environment of computer algebra “Mathematica 7.0”. Investigation of frequency-temporal and temporal dependence of the morphology of the signal at the output of nonlinear oscillatory neuron given its threshold effect at various values of the parameters of the input non-stationary signal and different weight of synaptic connections. Conclusions. The existence of the resonance effect in nonlinear neuron with equal external frequency non-stationary signal and frequencies of the dynamics of the neuron. It is shown that nonlinear oscillatory neuron plays the role of the frequency modulator and significantly alters the structure of the input information signal is non-stationary (shape, frequency and amplitude). Proposed coding information oscillatory nonlinear neuron on the basis of frequency modulation and decoding by using the inverse operator.