Application of Independent Component Analysis and Expansion Matching Filter for Automatic Detection of Motor Unit Action Potential Trains

The purpose of this study is to strengthen feature of MU discharge for classification of motor unit action potentials (MUAPs) into each MUAP train by means of the combination between expansion matching filter and independent component analysis (ICA). From our results, it is found that, when a MUAP waveform is used as a template, expansion coefficients give higher responses at the time points when similarity between the template and candidate MUAP waveforms is higher by applying expansion matching filter to multi-channel sEMG signals. Furthermore, because it is expected that the distribution of an expansion coefficient sequence including the MUAP train similar with the template have higher kurtosis value, only a component including a small number of MUAP trains can be extracted by additional application of ICA to expansion coefficients. Therefore, the fact suggests that this method assist to simplify and decompose classification of MUAPs only with a computer without heuristic operations