MODELISATION NUMERIQUE DU SIGNAL EMG DE SURFACE PAR LA METHODE DES ELEMENTS FINIS

Abstract

The modeling of the surface electromyography signal (SEMG) is basic for interpretation of EMG recordings, the conception and the test of algorithms for extraction information from the recorded signal. In this work, the finite element method is used to simulate the action potential of a muscle fiber. This potential is generated in a volume conductor of cylindrical shape, inhomogeneous, finite and composed of five layers (muscle, fat, skin, bone and air). All layers are isotropic except for the muscle layer which is anisotropic. The transmembrane current is modeled by a tripole source. The effect of additional layers on the signal detected with surface electrodes is studied. The effects of fiber depth, of adipose tissue thickness, of presence of bone and its location with respect to muscle fiber, of muscle anisotropy and of isotropic layers conductivities are also examined.