WAVELET AND FOURIER TRANSFORAMTIONS OF EMG AND MMG SIGNALS DURING FATIGUING CYCLE ERGOMETRY
Dona J. Housh, Joel T. Cramer, Joseph P. Weir, Ian J. Kremenic, Malachy P. McHugh, Sharon R. Rana, Anthony J. Bull, G. O. Johnson, T. J. Housh · Medicine & Science in Sports & Exercise · 2003
PURPOSE The frequency domains of electromyographic (EMG) and mechanomyographic (MMG) signals have been used to examine fatigue-induced changes in motor control strategies during cycle ergometry. It has been suggested that continuous wavelet transformation (CWT) procedures, that do not assume signal stationarity, are more appropriate than Fourier transformations for determining the power density spectrum of EMG signals during dynamic muscle actions. The purpose of this study was to compare the patterns of frequency responses of EMG and MMG signals calculated from discrete Fourier transformation (DFT) and CWT procedures during fatiguing cycle ergometry. METHODS Five subjects (mean age ± SD = 22 ± 3 years) performed 195 s continuous cycle ergometer workbouts (70 rpm) at 95% of VO2max. EMG and MMG signals were recorded simultaneously from the vastus lateralis muscle for the descending phase of one pedal revolution every 15 s throughout the workbout. DFT analyses were used to calculate the mean power frequencies (MPF) and CWT analyses were used to determine the instantaneous mean frequencies (IMF) for the EMG and MMG signals. The average of the IMF (AIMF) across the descending phase of the pedal revolution was used to represent the center frequencies of the signals for the CWT analyses. RESULTS Polynomial regression analyses indicated that normalized EMG MPF decreased linearly (p < 0.001), while EMG AIMF decreased quadratically (p < 0.002) across time. There were no changes (p < 0.05) in normalized MMG MPF or MMG AIMF during the workbout. CONCLUSIONS The results indicated that both DFT and CWT identified the fatigue-induced decrease in center frequency of the EMG signal, as well as the lack of change in the MMG signal. The difference in the EMG patterns (linear vs. quadratic), however, suggested that DFT and CWT may provide unique information about frequency responses during fatiguing cycle ergometry.