EMG mean power frequency: Obtaining a reference value.

Normalization of the electromyographic mean power frequency is generally performed on the basis of single estimates. Random variation of reference values based on such single estimates can seriously impair data reported in normalized form. In the present study we have examined five procedures for calculation of an initial reference value: calibration contractions at 0-kg (repeated) and 2-kg hand load, and regression routines at 0-kg and 2-kg hand load. There was a large between-subject variation if all different routines were considered, a moderate between-routine variation, and a quite small within-subject variation. The correlation between the different routines was high. At the same load level calibration and regression routines gave equivalent results, but the variability was less at the higher load level. The variations found are of such a magnitude that it is necessary to reduce the influence of random variation. Procedure proposed for reduction of this variability is repeated calibration measurements. A preferred test position for the trapezius muscle is straight arm, 90° of abduction in the scapular plane, 2-kg hand load.