Standard assumptions about the trial-by-trial distribution of averaged electromyography data could produce erroneous results.

The aim was to challenge the assumptions of standard statistical analyses of average surface electromyography (sEMG) data as a measurement of response magnitudes following the generation of a reflex. The ipsilateral tibial nerve was stimulated at three stimulation intensities and the response sEMG was measured in the contralateral soleus (cSOL) muscle. The magnitude of the cSOL response was measured at a set time window following ipsilateral tibial nerve stimulation. The averaged and trial-by-trial response magnitudes were assessed and compared. The analysis of the averaged and trial-by-trial response revealed significantly different results as the trial-by trial response magnitudes were log-normally distributed with between subject variance heterogeneity violating assumptions of standard statistical analyses. A statistical model has been suggested for the analysis of the responses. By ignoring trial-by-trial response variability and distribution, erroneous results may occur. This may change the interpretation of the results in some studies.