SPATIO-TEMPORAL CHARACTERISTICS OF MULTICHANNEL ELCTROMYOGRAPHY

This paper discusses the spatio-temporal properties of multichannel electromyographic (EMG) signals and investigates muscle force estimates obtained by multichannel myoprocessor. The multichannel design is adopted to include the spatial information and thus enhances the sensitivity of the myoprocessor. In our experiment, fourteen pairs of electrodes with differential preamplifiers are applied to four major muscles about knee movement. Multichannel EMG signals and torque generated are measured during isometrical extension and flexion contraction in burst and ramp movements. To investigate the relationship between EMG signals recorded at different sites, the coherence function measuring the linear relationship in frequency between two EMG signals is utilized. To remove the spatial coupling effect between multichannel EMG signals, the spatial prewhitening processing is used to eliminate this effect. In this research, the power and shape of EMG spectra are found to be useful for determining the activation level and signature of a specific movement. It can be seen that the EMG signals measured from two adjacent electrodes have higher coherence values indicating better linear relationship in frequency. The frequency relationship of EMG signals measured at different electrode sites seems to be in a Gaussian distribution form. The results of muscle force estimation indicate that the more the electrodes are used, the less the estimate error is obtained. In the study of using model obtained in ramp force estimation to estimate constant force, the superiority of multichannel myoprocessor is observed. In addition, the result shows that the estimation error at lower contraction levels is less than that of higher contraction levels. The results of this research show the advantage of multichannel EMG processing. The extension of this study could be applied to functional neuromuscular stimulation and active prosthesis control. This study would be helpful in determining the applicability of multichannel surface EMG signals as a diagnostic and assessment tool for rehabilitation and neurology studies.