The amplitude of an EMG and the temporal pattern can be used when considering if an EMG profile is normal or not. In the method described in this paper a gain factor of the complete EMG profile was determined and then the profile normalised with this gain factor. This normalised individual profile was then compared with a standard profile. predicted on the basis of walking speed. Deviating profiles were identified when they fell outside the upper and lower 95% limits range for the average profiles of 14 leg muscles. The amount of deviation from the normal profile can be quantified with the normalised mean square difference D-2. Gain factors varied over a factor of 4 within a group of 10 normal subjects. For a normal population D-2 was below 1. Most muscles had consistent profiles but some patterns could be discerned which showed marked variability among muscles and subjects. (C) 2004 Elsevier B.V. All rights reserved.