Surface Electromyography for Noninvasive Characterization of Muscle :

The time course of muscle fiber conduction velocity and surface myoelectric signal spectral (mean and median frequency of the power spectrum) and amplitude (average rectified and root-mean-square value) parameters was studied in 20 experiments on the tibialis anterior muscle of 10 healthy human subjects during sustained isometric voluntary or electrically elicited contractions. Voluntary contractions at 20% maximal voluntary contraction (MVC) and at 80% MVC with duration of 20 s were performed at the beginning of each experiment. Tetanic electrical stimulation was then applied to the main muscle motor point for 20 s with surface electrodes at five stimulation frequencies (20, 25, 30, 35, and 40 Hz). All subjects showed myoelectric manifestations of muscle fatigue consisting of negative trends of spectral variables and conduction velocity and positive trends of amplitude variables. The main findings of this work are 1) myoelectric signal variables obtained from electrically elicited contractions show fluctuations smaller than those observed in voluntary contractions, 2) spectral variables are more sensitive to fatigue than conduction velocity and the average rectified value is more sensitive to fatigue than the root-mean-square value, 3) conduction velocity is not the only physiological factor affecting spectral variables, and 4) contractions elicited at supramaximal stimulation and frequencies greater than 30 Hz demonstrate myoelectric manifestations of muscle fatigue greater than those observed at 80% MVC sustained for the same time.

This paper describes an in vitro method for comparing surface-detected electromyographic median frequency (MF) and conduction velocity (CV) parameters with histochemical measurements of muscle fiber type composition and cross-sectional area (CSA). Electromyographic signals were recorded during electrically elicited tetanic contractions from rat soleus, extensor digitorum longus, and diaphragm muscles placed in an oxygenated Krebs bath. Fibers were typed as slow oxidative, fast oxidative glycolytic, and fast glycolytic based on histochemical enzyme stains. Muscles with a greater percentage of fast glycolytic and fast oxidative glycolytic fibers exhibited greater initial values of MF and CV as well as a greater reduction in these variables over the course of the contraction. Regression indicated that fiber type composition could be predicted based on two MF parameters. A weighted measure of muscle fiber CSA was found to be linearly related to both initial MF and CV. The results of this study suggest that MF and CV parameters recorded during a muscular contraction are related to muscle fiber type composition and muscle fiber CSA.

Surface EMG signals detected in dynamic conditions are affected by a number of artefacts. Among them geometrical factors play an important role. During movement the muscle slides with respect to the skin because of the variation of its length. Such a shift can considerably modify sEMG amplitude. The purpose of this work is to assess geometrical artefacts on sEMG during isometric contractions at different muscle lengths. The average rectified value (ARV) of 15 single differential signals was obtained by means of a linear array of 16 bar electrodes from the vastus medialis and lateralis muscles. The knee angle was changed from 75 degrees to 165 degrees in steps of 30 degrees and voluntary isometric contractions at a low, medium and high force level were performed for each angle. The ARV pattern was normalized with respect to the mean activity to compare signals from different joint angles. From the data collected it was possible to separate the geometrical changes from the changes due to different intensities of activation. In three out of five subjects, we found (within the resolution of our measures) a 1 cm shift for the vastus medialis muscle while no shift was observed for the other two subjects. For the vastus lateralis muscle a 1 cm shift was found in two out of four subjects. Such a shift produces the main contribution to geometrical artefacts. To avoid such artefacts the innervation zones should be located and the EMG electrodes should not be placed near them.