Reproducibility of motor unit number estimation in individual subjects

Although the reproducibility of motor unit number estimation (MUNE) for gro ups of subjects has been studied, there is little such data for individuals . Prediction intervals represent a tool to study individual MUNE reproducib ility and represent the range of values expected for a future MUNE if the t rue number of motor units remains unchanged. MUNE was performed using the s tatistical method on 48 normal individuals. The prediction interval was fou nd to be a function of the intrasubject coefficient of variation. Using a c ommercial manufacturer's recommended technique and software, prediction int ervals were found to be so broad as to be of uncertain value. We found that by averaging two MUNE observations for each determination, and using the m ethod of weighted averages for calculating the size of an average single mo tor unit potential, the intrasubject coefficient of variation was reduced f rom 16.48% to 8.77%, and the 90% prediction interval became sufficiently na rrow to be clinically useful. False-negative rates were also lowered substa ntially using these techniques. Thus, simple modifications of an existing M UNE program improved the clinical utility of this program for the longitudi nal study of patients in whom changes in motor unit number over time are of importance, such as those with motor neuron diseases. (C) 2001 John Wiley & Sons, Inc.