A Hybrid Non-invasive Method for the Classification of Amputee’s Hand and Wrist Movements

The quest to develop dexterous artificial arm which supports multiple degrees of freedom for amputees has attracted a lot of study interest in the last few decades. The outcome of some of the studies had identified surface Electromyography (EMG) as the most commonly used biological signal in predicting the motion intention of an amputee. Different EMG based control methods for multifunctional prosthesis have been proposed and investigated in a number of previous studies. However, no any multifunctional prostheses are clinically available yet. One of the possible reasons would be that the residual muscles after amputations might not produce sufficient EMG signals for movement classifications. In this study, we proposed to use electroencephalography (EEG) signals recorded from the scalp of an amputee as additional signals for motion identifications. The performance of a hybrid scheme based on the combination of EMG and EEG signals in identifying different hand and wrist movements was evaluated in one transhumeral amputee. Our pilot results showed that the proposed hybrid method increased the classification accuracy in identifying different hand and wrist movements of the amputee. This suggests that the proposed method may have potential to improve the control of multifunctional prostheses.