Frequency Parameters of the Myoelectric Signal as a Measure of Muscle Conduction Velocity

Abstract

Dunrng a sustained muscle contraction, the spectrum of the myoelectric signal is known to undergo compression as a function of time. Previous investigators have shown that the frequency compression is related to the decreasing conduction velocity of the muscle fibers. It is proposed that the frequency compression may be tracked by obtaining a continuous estimate of a characteristic frequency of the spectrum, such as the mean and median, or the ratio of low-frequency components to high-frequency components of the spectrum. A theoretical analysis was performed to investigate the restrictions in estimating the three parameters, as well as their sensitivity to the conduction velocity. The ratio parameter was found to be most sensitive to conduction velocity, but was the least reliable of the three. The median frequency was the least sensitive to noise. Therefore, from a theoretical point of view, the median frequency is the preferred parameter. A technique is described which determines an unbiased consistent estimate of the median frequency. The technique may be readily implemented in analog hardware.