A new approach for muscle fatigue analysis in young adults at different MVC levels

Spectral parameters such as the mean and the median frequencies have been documented to be reliable outcome variables for the assessment of muscle fatigue. However when recorded for long intervals, electromyography signals become non-stationary. Short-time Fourier transform (STFT) has been extensively used for computing the electromyography time-varying spectrum, but the joint time-frequency resolutions of the STFT is inherently limited. A new time varying auto regressive (TVAR) model is proposed to analyze EMG signals which does not have joint time frequency resolution limitations. The objective of this paper is to examine surface EMG signals of healthy young subjects at different levels of maximum voluntary contractions (MVC) and to analyze spectral shifts in mean frequencies (MNF) using STFT and TVAR models. Our results show, TVAR has a better accuracy in signal representation and high frequency resolution. Further, spectral estimation can be obtained even for shorter data sequence. In this study, continuous stream of EMG data sets from lower extremity muscles are used to characterize muscular fatigue at different MVC level. EMG data were recorded from the Rectus Femoris muscles during isometric contractions.