A Comparison Study Of Emg Features For Force Prediction

Abstract

Myoelectric prosthetic devices can be controlled by use of surface electromyography (sEMG). However, intramuscular EMG (iEMG) has been proposed as an alternative, since it may provide more stable and selective recordings with several advantages. The purpose of this study was to assess the predictive capabilities of 14 features of iEMG and sEMG for force ranging from 0 to 100 % maximum voluntary contraction (MVC). Intramuscular EMG and surface EMG were recorded concurrently from the muscle flexor digitorum profundus from 11 subjects who exerted four force profiles during power grasping. The predictive capability of each feature was assessed using the mean R2-value with a 1st order polynomial (linear prediction). Wilson Amplitude showed the best results for both sEMG (R2 = 0.952 ± 0.007) and iEMG (R2 = 0.948 ± 0.008), with no significant difference (P = 0.658). Application of an advanced model based on artificial neural network did not improve the performance (P = 0.895). We have concluded that a linear model is sufficient for force prediction (0-100% MVC), and that iEMG is potentially suitable for proportional control in the same manner as when using a more global measure of intensity.