The authors present a practical guide for studying nonstationary data on human motor behavior in a time-frequency representation. They explain the limits of classical methods founded exclusively on the time or frequency basis and then answer those limits with the windowed Fourier transform and the wavelet transform (WT) methods, both of which are founded on time-frequency bases. The authors stress an interest in the WT method because it permits access to the whole complexity of a signal (in terms of time, frequency, amplitude, and phase). They then show that the WT method is well suited for the analysis of the interaction between two signals, particularly in human movement studies. Finally, to demonstrate its practical applications, the authors apply the method to real data.