The use of surface EMG in knee extensor moment prediction

A systematic method of EMG quantification is developed to estimate the isometric muscle moment directly from quantified surface EMG. This method includes the EMG Signals acquired from an acupuncture point Fu-Tu located on the quadriceps muscle group, an EMG smoothing scheme, an electromechanical time lag estimation, and a mathematical model with the polynomial regression function to quantify the EMG. Three subjects were asked to be tested on the CYBEX II dynamometer with a knee joint angle of 90 degree flexion and hip joint angle of also 90 degrees. They were asked to perform "two" trials of maximal voluntary contraction and "three" trials of free voluntary contraction of the isometric exercise. The first two trials were used to build up the quantification model, and the latter three trials served as data for the validation of the method. A Medelec MS92 EMG system with surface EMG electrodes was used to acquire the EMG Signals. In the determination of the regression order of the polynomial equations, the threshold value 0.0001 of the difference of the coefficient of determination values was used. The results of the polynomial regression orders are all 6 for three subjects, which reflects a tendency of nonlinear behavior of the EMG/moment relationship. A validation scheme was proposed to calculate the root mean square difference (RMSD) of the measured knee extensor moments from the CYBEX II dynamometer and estimated moments from the EMG quantification. The mean values of the RMSD of the three subjects were 0.0597, 0.0679 and 0.1080. These results demonstrate that the present approach can estimate the isometric muscle moment exerted by the quadriceps muscle group.