The surface electromyographic (SEMG) signal obtained during gait is often presented as the SEMG profile, the average SEMG activation pattern during one gait cycle. A disadvantage of this method is that it omits the step-to-step variability of the timing of the muscle activation patterns that might be relevant information as a performance measure of motor control and balance. In this paper, a method was used in which every step in the gait cycle could be analysed with respect to the timing of the muscle activation. For this purpose, the approximated generalised likelihood (AGLR) algorithm was implemented and tested. Results of the simulations show that the AGLR was much more accurate than a standard threshold criterion. Timing parameters could be calculated from a SEMG recording during gait and a measure for symmetry and coordination could be extracted. The amplitude distribution within and outside defined bursts is also presented to avoid the less precise classification into on and off patterns.