Noninvasive imaging of internal muscle activities from multi-channel surface EMG recordings

Surface Electromyogram (sEMG) technology provides a non-invasive way for rapid monitoring muscle activities, but its poor spatial resolution and specificity limit its application in clinic. To overcome these limitations, a noninvasive muscle activity imaging (MAI) approach has been developed and used to reconstruct internal muscle activities from multi-channel sEMG recordings. A realistic geometric hand model is developed from high-resolution MR images and a distributed bioelectric dipole source model is employed to describe the internal muscle activity space of the muscles. The finite element method and weighted minimum norm method are utilized solve the forward and inverse problems respectively involved in the proposed MAI technique. A series of computer simulations was conducted to test the performance of the proposed MAI approach. Results show that reconstruction results achieved by the MAI technique indeed provide us more detailed and dynamic information of internal muscle activities, which enhance our understanding of the mechanisms underlying the surface EMG recordings.