Crosstalk Reduction in Epimysial EMG Recordings from Transhumeral Amputees with Principal Component Analysis
Paper in proceeding, 2018

Electromyographic (EMG) recordings of muscle activity using monopolar electrodes suffer from poor spatial resolution due to the crosstalk from neighbouring muscles. This effect has mainly been studied on surface EMG recordings. Here, we use Principal Component Analysis (PCA) to reduce the crosstalk in recordings from unipolar epimysial electrodes implanted in three transhumeral amputees. We show that the PCA-transformed signals have, on average, a better signal-tonoise ratio than the original unipolar recordings. Preliminary investigations show that this transformation is stable over long periods of time. If the latter is confirmed, our results show that the combination of PCA with unipolar electrodes allows for a higher number of muscles to be targeted in an implant (compared with bipolar electrodes), thus facilitating 1-to-1 proportional control of prosthetic hands.

Author

A. Matran-Fernandez

University of Essex

Enzo Mastinu

Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering

R. Poli

University of Essex

Max Jair Ortiz Catalan

Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering

L. Citi

University of Essex

Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS

1557170X (ISSN)

Vol. 2018-July 2124-2127 8512645

40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018
Honolulu, USA,

Subject Categories

Other Medical Engineering

Physiology

DOI

10.1109/EMBC.2018.8512645

PubMed

30440823

More information

Latest update

12/10/2018