In this paper, a new classification method has been
developed utilizing a sparse representation (SR) of EEG
signals for the motor imagery based BCI application. We
use a common spatial pattern (CSP) for preprocessing of
EEG data and the SMRs as a feature of BCI system. To
evaluate the performance of the proposed classification
method, we use two classes, the left hand and the right hand,
of motor imagery data sets collected from four subjects. We
also evaluate the performance of the LDA method for
comparison with the performance of the proposed method.