Cramer-Rao lower bounds for estimating the time-varying delay of surface EMG signals
It is known that the conduction velocity (CV) is a relevant estimator for fatigue and disease electromyographic (EMG) studies. CV estimation, which is linked to the time delay of an EMG signal propagation between two or more sensors, is particularly interesting in dynamic studies to detect local changes along the time. In this paper, we investigate three naturally time-frequency and time-scale methods to follow CV changes. In this work, the linear relationship between the phase information and the local time delay between two signals is used. Our results indicate that the three methods can be used to follow the conduction velocity evolution during a recording. Comparing the root mean square errors for each method highlight that the Fourier coherence method gives the best results compared to the two other methods (wavelet phase coherence and phase consistency).