Selection of relevant features for classification of movements from single movement-related potentials using a genetic algorithm

Classification of movement-related potentials recorded from the scalp to their corresponding limb is a crucial task in brain-computer interfaces based on such potentials. This paper demonstrates how the features for such a task can be selected from a large bank of features using a genetic algorithm. We show that it is possible to differentiate between the movements of contralateral fingers with a classification accuracy of 77% using a small number of features (10-20) selected from a bank containing roughly 1000 features.