A Study of Power Asymmetry Ratio of Single Trial Electroencephalogram During Finger Movement

This paper presents a study on power asymmetry ratio of the single trial electroencephalogram (EEG) during the preparation of left or right finger movement, or more precisely self-paced tapping. The objective is to predict which finger is going to be used to press a key in the computer keyboard. The change in EEG signal is more pronounced in the contralateral hemisphere. Therefore, the use of power band asymmetry ratio will enhance the identification of left or right finger tapping. Parametric spectral estimation, in this case autoregressive (AR) spectral analysis with Burg method is used to obtain the power spectral density (PSD) from 8 Hz to 30 Hz. The proposed method selects the most critical homologous channels in the classification of the task. It is discovered that the homologous pair of CP3 and CP4 is the most important channel-pair that gave a classification accuracy of seventy five percent.