Time- and frequency-domain monitoring of the myoelectric signal during a long-duration, cyclic, force-varying, fatiguing hand-grip task.
Three methods that can significantly reduce the variability of the EMG power density spectrum (PDS) variable by eliminating artifactual components are described. Two methods, one that allows the subtraction of power line noise in the time domain and one which allows the subtraction of system noise in the frequency domain from the EMG, were shown to be effective in helping to accurately estimate the median frequency (MF) of the PDS, and especially during low level contractions (0-25% maximal voluntary contraction, MVC) when the signal-to-noise ratio is unfavorable. The techniques eliminate the artifactual effects of system and power line noises from the EMG recordings throughout the force range (0-100% MVC) while preserving the native EMG power at all frequencies. It was also shown that if a technique to train subjects to produce their true MVC is employed, the absolute force/torque produced could be as much as 30% higher than in untrained MVC. The effect of true MVC production was also shown to be significant when interpretation of PDS variables are correlated to the processes which produce contraction.