Objective To determine the intra- and inter-session reliability of motor evoked potential (MEP) size parameters in the lower limb of patients with stroke, focussing on the number of MEPs collected and the method of measuring MEP size.

Methods Transcranial magnetic stimulation was used to elicit MEPs in the soleus muscle of patients with stroke (n = 13) and age-matched healthy participants (n = 13) during low level muscle activation. Two sets of 10 responses were collected in the first session and a further 10 responses collected in a second session held 7 days later. Four MEP size measurements were made using 4, 6, 8, or all 10 of the MEPs collected. Intra- and inter-session reliability was examined using intraclass correlation coefficients (ICC) and typical percentage error.

Results Intrasession ICC statistics using 6 or more MEPs were >0.85 in the stroke group but intersession values were all <0.5. Reliability was best when measuring parameters from individual MEPs rather than averaged responses.

Conclusions Reliability of intrasession MEP size is excellent in the lower limb of patients with stroke using as few as 6 MEPs but intersession reliability is poor.

Significance Comparing MEP size measures across two or more sessions is questionable in the lower limb of patients with stroke.