Hilbert-transform based predictions of hand movements from EEG measurements

New effective features, extracted from single trial EEG data measured on the human scalp capable of hand movement prediction are reported. The envelope and the instantaneous main frequency of the filtered (8–12´Hz) EEG signal, calculated by the Hilbert-Transform, are used as features. The EEG responses of right- and left-hand movements are studied. A Learning Vector Quantizer (LVQ) was used to predict the sides of the hand movements. The classification accuracy was about 90 percent.