Enhancing the performance of motor imagery classification to design a robust brain computer interface using feed forward back-propagation neural network

This paper proposes a feed forward back-propagation neural network (FFBPNN) based method to enhance the performance of the motor imagery classification. The dataset consists of fifty nine channels of EEG signals which are first normalised using minmax method and then given as input to the FFBPNN network. Experimental outcomes of the FFBPNN are recorded in term of ‘0’s or ‘1’s for two classes of motor imagery signals. The accuracy of the proposed FFBPNN method has been measured using confusion matrix, mean square error and percentage accuracy. However, accuracy of the FFBPNN based method is recorded up to 99.8%. Hence the proposed method gives better accuracy of the classification which will ultimately help in designing robust BCI.