Spectral analysis of muscular sound at low and high contraction level.

The activated muscle generates a low frequency rumbling noise, which is known as the Sound-MyoGram (SMG). Spectral analysis of the SMG is carried out in this work, in order to: (i) check the adequacy of both the Fast Fourier Transform (FFT) and the Maximum Entropy Spectrum Estimation (MESE). Because it is a well known technique, the FFT method is only briefly described, while the philosophy of the MESE method is given in more detail and completed with a description of the recursive algorithm; (ii) select a frequency parameter suitable to describe the SMG. For this purpose two well-defined physiological conditions (20% and 80% Maximal Voluntary Contraction) have been adopted in order to provide a safe reference for the interpretation of the findings. The results show that: (a) both FFT and MESE are adequate to estimate the SMG Power Spectrum; (b) both the mean and the median frequency are suitable parameters, the mean frequency being the more favourable one; (c) the SMG Power Spectrum is a promising tool to study the muscle activation modalities.