Repeatability of electrically-evoked myoelectric signals in the human tibialis anterior muscle.

The reproducibility of surface myoelectric signal measurements is of paramount importance for clinical applications of electromyography (EMG) techniques. The repeatability of electrically-evoked myoelectric signal shape (M-wave) as well as spectral and amplitude parameters, conduction velocity and elicited torque was tested, in isometric conditions, on the tibialis anterior muscle of 10 normal subjects. Contractions were elicited by stimulation of the main muscle motor point and repeated after removal and replacement of the stimulation and detection electrodes in the same carefully marked locations. This protocol was repeated five times on each subject on five different days. The test-retest Pearson correlation coefficient, the paired t test and analysis of variance (ANOVA) were used to quantify repeatability and estimate the fraction of variance due to repeated trials within experiments, repeated experiments within subjects and inter-subject variability. Results indicate that parameters of spectral variables are more repeatable than those of amplitude variables. Elicited torque and conduction velocity show the lowest repeatability. The intra-class correlation coefficient ranged from 87.9% for the initial value of median frequency to 11.5% for the initial value of conduction velocity. Fatigue indices based on the time course of the myoelectric signal variables showed even lower values of this coefficient. It is concluded that: (a) initial values and fatigue indices based on spectral variables are more repeatable than those based on amplitude variables; (b) the repeatability of conduction velocity and torque is very poor; (c) M-wave shape, rather than amplitude or width, seems to be a characteristic of individual muscles; and (d) electrode location is a critical issue in the study of M-waves elicited by stimulation of a muscle motor point. The methodology for estimation of muscle fibre conduction velocity must be refined and the characterization of evoked responses must be improved to allow widespread clinical applications.