Comparison of Fourier And Wavelet Transform Procedures For Examining The Mechanomyographic Frequency Versus Isokinetic Torque Relationship Of The Biceps Brachii

Abstract

PURPOSE The purpose of this study was to compare the isokinetic torque-related patterns for mechanomyographic (MMG) center frequency [mean power frequency (MPF), median frequency (MDF), and wavelet center frequency (CF)] derived from the fast Fourier transform (FFT) and discrete wavelet transform (DWT) algorithms. METHODS Ten adults [mean ± SD age = 22.8 ± 2.7 yrs] performed submaximal to maximal, concentric isokinetic muscle actions of the dominant forearm flexors at a velocity of 30°·sec−1. The surface MMG signal was detected by an accelerometer placed over the belly of the biceps brachii. RESULTS For both absolute and normalized values, MPF, MDF, and CF were intercorrelated at (r = 0.78–0.92). Quadratic models provided the best fit for the absolute and normalized MMG MPF, MDF, and CF versus isokinetic torque relationships (R2 = 0.47–0.78). CONCLUSION The primary finding of this study was the similarity in the MMG MPF, MDF, and CF versus isokinetic torque patterns. Thus, despite the concerns over signal nonstationarity, both Fourier and wavelet based methods can be used to examine the MMG center frequency versus isokinetic torque relationship.