Factors in the reproducibility of electromyographic power spectrum analysis of lumbar paraspinal muscle fatigue.

Abstract

An investigation of the reliability of various factors used to analyze electromyograms of paraspinal muscles of the lumbar spine. To determine the reproducibility of several electromyographic variables over time and to assess the influence of confounding factors. Previous reliability studies have included small numbers of subjects and often have not explored the factors that can influence the reliability of this technique. In the first study, 39 volunteers underwent electromyographic tests on three occasions at two loads. The time interval was 1 week between tests. Electromyographic variables measured included half-width (spectral width at half maximum amplitude), initial median frequency, median frequency slope, initial root mean square, root mean square, and root mean square slope. In the second study to determine the influence of posture, nine volunteers were tested on two occasions with a change in posture between tests. Reliability was better for tests at the higher loads. Initial median frequency, half-width, and root mean square all had acceptable reproducibility over time, with a coefficient of variation less than 10. Initial median frequency and half-width proved to be the most suitable variables; they yielded the largest between-subject variation and, of all variables, were affected the least by changes in load. The remaining variables were very load-dependent. A change in posture greatly influenced the reproducibility. These results demonstrate that only initial median frequency and half-width have acceptable reproducibility over time and at different loads, with the potential to be useful in discriminating between individuals. Changes in posture can greatly change an electromyographic result.