Comparison of parametric and nonparametric reference data in motor unit potential analysis.

For calculation of outlier reference intervals, by definition nonparametric statistics are applied, while for mean value reference intervals parametric or nonparametric statistics can be used. The aim of this study was to compare the mean value reference intervals and their sensitivity for diagnosis of myopathy. Quantitative concentric needle electromyography (EMG) of the biceps brachii muscle was performed using multi-motor unit potential (MUP) analysis. In 34 healthy subjects both parametric (mean+/-2SD) and nonparametric (2.5th-97.5th percentiles) reference intervals were calculated for mean values of MUP parameters, while for outliers nonparametric reference intervals (5th-95th percentiles) were calculated. Their sensitivity was tested in 29 patients with facioscapulohumeral muscular dystrophy. Nonparametric reference intervals were narrower than parametric intervals, which resulted in slightly higher sensitivities when combined with outlier limits (e.g., thickness=86% and 83%, respectively).