The Effect of ICA and Non-negative Matrix Factorization Analysis for EMG Signals Recorded From Multi-Channel EMG Sensors.

Yeongdae Kim,Natsue Yoshimura,Sorawit Stapornchaisit,Yasuharu Koike,Makoto Miyakoshi

doi:10.3389/fnins.2020.600804

Abstract

Surface electromyography (EMG) measurements are affected by various noises such as power source and movement artifacts and adjacent muscle activities. Hardware solutions have been found that use multi-channel EMG signal to attenuate noise signals related to sensor positions. However, studies addressing the overcoming of crosstalk from EMG and the division of overlaid superficial and deep muscles are scarce. In this study, two signal decompositions—independent component analysis and non-negative matrix factorization—were used to create a low-dimensional input signal that divides noise, surface muscles, and deep muscles and utilizes them for movement classification based on direction. In the case of index finger movement, it was confirmed that the proposed decomposition method improved the classification performance with the least input dimensions. These results suggest a new method to analyze more dexterous movements of the hand by separating superficial and deep muscles in the future using multi-channel EMG signals.

Highlights

Electromyography (EMG) measures the electrical impulses from the muscle contraction induced by the central nervous system for voluntary body movement
We examined the following three parameters to compare EMG-based synergy (EMG-synergy) and independent components (ICs)-based synergy (ICA-synergy): (1) robustness of the synergy structure calculation in the two elbow posture, (2) preferred direction compared with anatomical basis, and (3) classification performance on the eight direction finger movements using convolutional neural network (CNN)
Modules, some EMG channel activation are similar with scalar product (SP) > 0.75, but their inclination for each finger movement is significantly different

Summary

Introduction

Electromyography (EMG) measures the electrical impulses from the muscle contraction induced by the central nervous system for voluntary body movement. The surface EMG signal contains different muscle signals and various noises such as baseline noise and movement artifacts (De Luca et al, 2010). Gazzoni et al, 2014 applied non-negative matrix factorization (NMF) to multi-channel EMG signals and distinguished the position on each forearm per movement of the wrist and single finger joint. Their study investigated deep muscle activities under singular joint movement and confirmed the feasibility of multi-channel EMG signals-based muscles synergy so as to identify deep muscle region. They did not dig deep into the dexterous finger movement and nor the structure of muscle synergy per joint movement.

Objectives

Methods

Results

Conclusion