The effect of manipulating root mean square window length and overlap on reliability, inter-individual variability, statistical significance and clinical relevance of electromyograms

Numerous ways exist to process raw electromyograms (EMGs). However, the effect of altering processing methods on peak and mean EMG has seldom been investigated. The aim of this study was to investigate the effect of using different root mean square (RMS) window lengths and overlaps on the amplitude, reliability and inter-individual variability of gluteus maximus EMGs recorded during the clam exercise, and on the statistical significance and clinical relevance of amplitude differences between two exercise conditions. Mean and peak RMS of 10 repetitions from 17 participants were obtained using processing window lengths of 0.01, 0.15, 0.2, 0.25 and 1 s, with no overlap and overlaps of 25, 50 and 75% of window length. The effect of manipulating window length on reliability and inter-individual variability was greater for peak EMG (coefficient of variation [CV] < 9%) than for mean EMG (CV < 3%), with the 1 s window generally displaying the lowest variability. As a consequence, neither statistical significance nor clinical relevance (effect size [ES]) of mean EMG was affected by manipulation of window length. Statistical significance of peak EMG was more sensitive to changes in window length, with lower p-values generally being recorded for the 1 s window. As use of different window lengths has a greater effect on variability and statistical significance of the peak EMG, then clinicians should use the mean EMG. They should also be aware that use of different numbers of exercise repetitions and participants can have a greater effect on EMG parameters than length of processing window. (C) 2014 Elsevier Ltd. All rights reserved.