Phasic behavior of EMG signals during gait: Use of multivariate statistics

Abstract

The present study represents an attempt to ascertain whether there are some underlying trends that, in some combination, can explain all the variations in the linear envelopes of 16 EMG signals of selected leg muscles. Two kinds of analyses were performed: (a) a factor analysis of EMG data corresponding to 16 muscles of the lower limb and (b) a "multi-dimensional scaling" (MDS) procedure. The latter technique involved mapping the Cartesian coordinates for 16 points subject to the constraint that the distance between any two points reflected the degree of coactivity for the corresponding muscles. The results of these analyses showed that four factors could account for 91.5% of the variance in the original data set. These factors could be clearly demarcated on the "muscle map," tending to support the notion that there are motor "programs" for groups of muscles that have to perform a given function during locomotion. Further analysis of the loading matrix (i.e., correlations between each EMG vector and each of the four factors) showed groups of muscles that acted in a similar manner. The muscle groups could be divided into those that act at the times of (a) heelstrike, (b) single limb loading response, (c) propulsion phase, or else (d) acted in a biphasic manner.