Modeling the spatial distribution of surface electromyogram amplitudes

Muscle excitations are distributed heterogeneously within the muscle, resulting in local differences in the amplitude of surface electromyograms. The functional significance of regionalized muscle excitation may be studied with high-density recordings, but these data are high-dimensional, making them difficult to analyze and interpret. Thus, we assessed the ability of two analytical approaches, Gaussian and sinusoidal functional forms, to represent uni- and bimodally (one or two peaks) distributed surface electromyography (sEMG) amplitudes. In the unimodal case, Gaussian functions fit the data well and had physically meaningful parameter estimates; sinusoidal fits were poor. In the bimodal case, neither Gaussian nor sinusoidal fits accurately represented the underlying amplitude distributions.