Reconstruction of muscle motor unit structure from HD-sEMG signals using deep convolutional neural networks

In this paper, we propose a method for identifying the anatomic properties of skeletal muscle motor units from high-density surface electromyography (HD-sEMG) signals using deep convolutional neural networks. The method first applies decomposition of the HD-sEMG signal into sequences of motor unit action potentials (MUAPs). A deep convolutional neural network is then used to estimate the depth, location, size, and shape of each motor unit from the extracted MUAPs. The method is evaluated on synthetic datasets of the biceps brachii muscle.