Analysis of multiple EMG and acceleration signals of various record lengths as a means to study pathological and physiological oscillations.

Quantitation of pathological and physiological oscillations are of interest to both the clinician and researcher. Power spectral analysis is a technique which allows for quantitation of both signal amplitude and frequency characteristics. For example, computer derived spectra of limb acceleration and demodulated electromyograms are used to define the properties of an oscillation. In addition, cross-correlation analysis (resultant cross-spectra, phase spectra, and coherence values) provides detailed information concerning the correlation between multiple signals: e.g., as to which muscle or muscles are driving an oscillation. We describe these techniques for analysis of pathological and physiological oscillation of the human motor system and suggest ways in which very short segments of data can be analyzed.