Processing the surface mechanomyographic signal

Signal processing is an important issue that should be considered prior to recording mechanomyographic (MMG) signals. The first studies in this area found that the fast Fourier transform and maximum entropy spectrum estimation techniques provided MMG frequency spectra that were very similar in shape. The shape of the spectrum from the maximum entropy spectrum estimation technique, however, was highly dependent on the model order that was chosen. Recent investigations have also used joint time-frequency signal processing techniques, such as waveletbased methods and the short-time Fourier transform to analyze Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA. E-mail: [email protected]