The spatial sampling problem in electromyographic studies of speech musculature

Abstract

The general practice of insertion of a single electrode pair per muscle for electromyographic recording raises the question of how representative are the signals obtained. Signals from three bipolar fine wire derivations high, middle, and low in the same palatoglossus muscle of speakers of American English were amplified, rectified, and integrated. The data were analyzed statistically to determine the patterns of similarity among simultaneous signals from different sites in the muscle for a given speaker, and the degree of independence of each signal. The middle signal was most often the best, and the low signal most often the poorest representative of the group of signals. The greatest similarity was found between the high and middle signals; next most similar were the middle and low signals. Similarity between signals varied as a function of the utterance. Nonlinear relations were observed between some signals. In some cases the relationship between two signals varied over time. This variation may be due to change in interelectrode distance during anisometric contraction. [V. Pollack, Med. Biol. Eng. 9, 657–664(1971)], or to recruitment patterns. [Work supported in part by NIDR.]