A method of EMG decomposition based on fuzzy logic

A new method for decomposing the EMG signal into their constituent motor unit action potentials (MUAPs) is presented. We propose a specific iterative algorithm with a classification method using fuzzy logic techniques. A statistical analysis of Inter-pulses Interval (IPI) and amplitude detected MUAPs is realised before its classification step. The algorithm was evaluated on simulated surface EMG signals (SEMG) before its application to real SEMG. On simulated signals, the rate of successfully classified MUAP was 88.4%. On real Laplacian SEMG, the algorithm identified correctly 21 MUAP trains (MUAPTs) on the 29 MUAPTs identified by an expert. The efficiency of the decomposition on Surface EMG makes the method very attractive for non invasive investigations.