Detection of fast and slow hand movements from motor imagery EEG signals

Abstract

Classification of Electroencephalography (EEG) signal is an open area of re-search in Brain-computer interfacing (BCI). The classifiers detect the different mental states generated by a subject to control an external prosthesis. In this study, we aim to differentiate fast and slow execution of left or right hand move-ment using EEG signals. To detect the different mental states pertaining to motor movements, we aim to identify the event related desynchronization/synchronization (ERD/ERS) waveform from the incoming EEG signals. For this purpose, we have used Welch based power spectral density estimates to create the feature vector and tested it on multiple support vector machines, Nave Bayesian, Linear Discriminant Analysis and k-Nearest Neighbor classifiers. The classification accuracies produced by each of the classifiers are more than 75% with naïve Bayesian yielding the best result of 97.1%. © Springer International Publishing Switzerland 2014.