Determination of feature relevance for the grouping of motor unit action potentials through a generative mixture model

The study of motor unit action potential (MUAP) activity from electromyographic signals is important for neurological investigations aiming to understand the state of the neuromuscular system. In this context, the identification and clustering of MUAPs that exhibit common characteristics, and the assessment of which data features are most relevant for the definition of such cluster structure, are central issues. In this paper, we propose the application of an unsupervised feature relevance determination (FRD) method to the analysis of experimental MUAPs. This method is embedded in a constrained mixture of distributions model that simultaneously performs data clustering and visualization. The experimental results of the analysis of a data set consisting of MUAPs measured from the First Dorsal Interosseous, a hand muscle, indicate that the features corresponding to the hyperpolarization period in the physiological process of generating muscle fibre action potentials are consistently estimated to be the most relevant. Moreover, the MUAP cluster structure of the data is shown to be only partially attributable to inter-subject differences, with the hyperpolarization period providing the best discrimination of the data by subject.