Decomposition of surface electrode-array electromyogram of biceps brachii muscle in voluntary isometric contraction

The study of the electric signals generated by contracting muscle fibers (EMG) is relevant to neurophysiological research and to the diagnosis of motoneuron diseases. Classical diagnosis methods make use of invasive, painful needle electrodes. We explored a non-invasive alternative to these methods by developing a specific algorithm that uses Independent Component Analysis (ICA) and template-matching techniques to decompose surface EMG (s-EMG). An experiment was carried out with two healthy subjects performing isometric contractions at different force levels. We measured s-EMGs with an electrode array and applied to them our decomposition algorithm. The obtained motor-unit firing pattern is in agreement with results obtained with needle electrodes.