Classification of motor imagery hands movement using levenberg-marquardt algorithm based on statistical features of EEG signal

Abstract

Motor Imagery (MI) is a highly supervised method nowadays for the disabled patients to give them hope. This paper proposes a differentiation method between imagery left and right hands movement using Daubechies wavelet of Discrete Wavelet Transform (DWT) and Levenberg-Marquardt back propagation training algorithm of Neural Network (NN). DWT decomposes the raw EEG data to extract significant features that provide feature vectors precisely. Levenberg-Marquardt Algorithm (LMA) based neural network uses feature vectors as input for classification of the two class data and outcomes overall classification accuracy of 92%. Previously various features and methods used but this recommended method exemplifies that statistical features provide better accuracy for EEG classification. Variation among features indicates differences between neural activities of two brain hemispheres due to two imagery hands movement. Results from the classifier are used to interface human brain with machine for better performance that requires high precision and accuracy scheme.