Distributed random electrical neuromuscular stimulation: effects of the inter-stimulus interval statistics on the EMG spectrum and frequency parameters.

An electrophysiological approach was used to study a distributed random electrical neuromuscular stimulation (ENMS) scheme in which a probability density is assigned to the inter-stimulus intervals (ISI) of the stimuli. One of the objectives of using ENMS techniques in the study of skeletal muscles is to obtain information about the electrical, physiological, and mechanical properties of muscles in a near-physiological situation under a well-controlled experimental design in which problems related to the uncertainty of firing patterns of the central nervous system and physiological interference are avoided. In particular, ISI with a Gaussian density were varied in mean rate, standard deviation (SD), and coefficient of variation. The influence of varying ISI, and the interaction of the ISI statistics with compound motor unit action potentials (CMUAP) on EMG power spectra and their frequency parameters, was assessed theoretically using a mathematical model which is similar to that of EMG signal generation in the electrophysiological case. In order to quantify the effects of ISI statistics on the EMG spectrum, the median frequency was calculated as a function of stimulation rate using analytical expressions for various values of the coefficients of a Gaussian ISI variation. The results obtained suggest that 1) the interaction between ISI statistics and the shape of the CMUAP plays a major role in determining the EMG spectrum; 2) the median frequencies (MF) determined from EMG spectra tend to increase with increasing mean rates of stimulation for a given CMUAP. The rate of increase of the MF depends on the coefficient of the ISI variation; 3) the EMG spectra of random electrically stimulated muscle show peaks at the mean rate of stimulation, and multiples of it, when the coefficient of variation of ISI is small. These peaks decrease in magnitude with increasing coefficients of variation of ISI; and, 4) a variation in the ISI should be introduced in the ENMS, when a reproduction of 'normal' EMG spectra is needed. These results are consistent with those reported for voluntary contraction of skeletal muscles.