Visualization of amplitude-frequency characteristics of EEG of pathological and cognitive functions of the brain from a position of nonlinear dynamics

Abstract

In recent years, medical imaging methods have been intensively developed to provide comprehensive and extensive data for studying the work of the brain, which makes the present study relevant. The purpose of this work is to develop a new approach for analyzing EEG signals based on nonlinear dynamics methods. To achieve this goal, it is necessary to adapt the methods of nonlinear dynamics to solving EEG analysis tasks. In this paper, the following methods are used: The principal component method for noise clearance, the wavelet transform, the Wolf, Kanz, Rosenstein, neural networks methods to calculate the spectrum of the Lyapunov exponent (LCEs). Studies have shown that the chaos of the schizophrenic's EEG signals is higher than for healthy people. It has been revealed that the best visualization of the amplitude-frequency characteristics of the brain's EEG pathological functions is provided by the Morle and Gauss 32 wavelets.