Analysis of Some EMG Signals Using Multiresolution and Time-Frequency Techniques

In this paper several electro-myography signals are studied comparatively using multiresolution and time-frequency analysis. A 4-level wavelet decomposition is applied to several surface EMG signals, acquired from masseter, legs, abdominal or dorsal muscles, then a relative band energy distribution of each component is obtained. This analysis is performed comparatively using various wavelet families. Additional information can be obtained by plotting the variation in time of relative band energy distribution, which has a specific shape for a given EMG signal. A time-frequency analysis is performed as well, plotting the EMG signal spectrograms in some particular, relevant cases.