Simulation of Electromyographic Signals Based on Point Distribution Model

A central issue in EMG signal modeling is the simulation of its basic units designated as motor unit action potentials (MUAPs). In this paper we show how the Point Distribution Model (PDM) may be employed for generation of synthetic MUAPs based on experimental data. Our results show that the PDM allows for a simple way of controlling the distortion of synthetic MUAPs. This distortion takes into account underlying statistical properties of experimental data. Furthermore, we illustrate how these artificial MUAPs can be used for generation of electromyographic signals via a data-driven model. A comparison between the power spectra of synthetic and actual EMG signals shows that this model may generate signals that mimic the reality.