Motor Imagery-based Brain-Computer Interface: Neural Network Approach

Abstract

A neural network approach has been developed for detecting EEG patterns accompanying the implementation of motor imagery, which are mental equivalents of real movements. The method is based on Local Approximation of Spectral Power using Radial Basis Functions (LASP-RBF) and the original algorithm for interpreting the time sequence of neural network responses. An asynchronous neural interface has been created, the basic element of which is a committee of three neural networks providing the classification of target EEG patterns accompanying the execution of motor imagery by the upper and lower limbs. A comparative evaluation of the classification efficiency of EEG patterns of mental equivalents of real movements was carried out using the developed classifier and traditional classification methods in particular, Random Forest, Linear Discriminant Analysis and Linear Regression methods. It was shown that the classification accuracy using the developed approach is higher (up to 90%) than other classifiers.