Estimation of the interplay between groups of fast and slow muscle fibers of the tibialis anterior and gastrocnemius muscle while running

Electromyograms recorded from the lower limbs of humans while running were submitted to a time/frequency analysis using wavelets. The results of the wavelet analysis yielded intensity spectra at every time point during the swing and the stance phase. It was previously shown that more or less high frequency components get activated during different periods of the movement. The purpose of this study was to test to what extent the spectra can be reconstructed by a linear superposition of two generating spectra that were associated to groups of fast and slow muscle fibers. The terms fast and slow do not only refer to the conduction velocity but also to the shape of the motor unit action potential and are used to characterize the groups in a broader sense. The principal component analysis of the spectra confirmed that a two dimensional spectral space was appropriate. A parametric spectral decomposition was used to extract the generating spectra within the two dimensional spectral space. The generating spectra were in turn used to compute the power with which the groups of muscle fibers contribute to the measured spectra and thus to the overall muscular activity. The power that was obtained for the different time points during the movement reflects the biomechanically important interplay between the groups of muscle fibers while running. (C) 2005 Elsevier Ltd. All rights reserved.