Skill and selection bias has least influence on motor unit action potential firing rate/frequency.

INTRODUCTION: Motor unit action potential (MUAP) rise time in quantitative electromyography (QEMG) using multi-MUAP analysis may influence neuromuscular disease diagnosis.

OBJECTIVE: To demonstrate in QEMG that MUAP firing rate/frequency (FR) is the parameter least dependent on electrodiagnostician skill (EDXc) skill and selection bias.

METHODS AND MATERIALS: During minimal contraction (< 6 MUAPs/4.8 s analysis period) of tibialis anterior, 20 MUAPs were selected for QEMG at each of the 3 insertion sites. There were three EDXcs (professor & two fellows), each an independent MUAP selector + EMG machine itself (EMGM), for a total of four independent MUAP selectors. EDXc and selector order variation at each site involved the following. At insertion site1, during EMG by EDXc 1, the MUAP selector order was EDXc 1, EMGM, EDXc 2 and EDXc 3. At the following two sites, the previous 1st selector was sequentially rotated to become the last selector during EMG by the subsequent EDXc. The protocol was repeated in two subjects. MUAP inclusion criteria by EDXcs was MUAP rise time < 1 ms by visual inspection and, by the EMGM, any and all MUAPs during the 4.8 s analysis period.

RESULTS: The reproducibility of measurements between MUAP selectors for mean frequency was satisfactory (correlation coefficient for inter-EDXc was 0.57, and 0.66 between the most experienced EDXc and EMGM. Inter- EDXc correlation coefficient was poor for amplitude (0.10), duration (0.34), size index (0.35), phases (-0.13) and turns (-0.22). Of the MUAPs collected by EMGM, the mean amplitude was smaller and mean duration was longer than those collected by the other 3 selectors with no significant differences for the other parameters.

CONCLUSIONS: FR was the only satisfactorily reproduced parameter from QEMG of MUAPs between selectors, and least dependent on skill and/or selection bias. From this, FR appears a relatively objective parameter, while other parameters appear unduly influenced by EDXc skill and/or MUAP selection bias.