Feature Analysis For Motor Imagery EEG Signals With Different Classification Schemes

A Brain-Computer Interface (BCI) is a communication system that decodes and transfers information directly from the brain to external devices. The electroencephalogram (EEG) technique used measure electrical signals corresponding commands occurring in control functions. for applications BCI are called Motor Imagery (MI) EEG signals. noisy, so it important use right methods recognize patterns correctly. This study examined performances of different classification schemes train networks using Ensemble Subspace Discriminant classifier. Also, most efficient feature space was found Neighborhood Component Analysis. maximum average accuracy classifying MI right-direction left-direction 80.4% with subject-specific scheme 250 features.