Decoding Repetitive Finger Movements with Brain Signals Acquired Via Noninvasive Electroencephalography

Abstract

We investigated how well finger movements can be decoded from electroencephalography (EEG) signals. 18 hand joint angles were measured simultaneously with 64-channel EEG while subjects performed a repetitive finger tapping task. A linear decoder with memory was used to predict continuous index finger angular velocities from EEG signals. A genetic algorithm was used to select EEG channels across temporal lags between the EEG and kinematics recordings, which optimized decoding accuracies. To evaluate the accuracy of the decoder, the Pearson's correlation coefficient (r) between the observed and predicted trajectories was calculated in a 10-fold cross-validation scheme. Our results (median r = .403, maximum r = .704), compare favorably with previous studies that used electrocorticography (ECoG) to decode finger movements. The decoder used in this study can be used for future brain machine interfaces, where individuals can control peripheral devices through EEG signals.