Common drive detection for axial muscles cerebral control and coherence analysis of surface electromyography by neural networks

The present study investigates the presence of common modulation of motor unit discharge at very low frequency during voluntary contractions. The work is based on the surface electromyography recordings and on the coherence analysis between muscle activities related to different body side. In the experiments carried out, we aim to investigate on signals present in the brain without EEG acquisition, but by only processing the effect of these signals, i.e. the muscle activity. This mechanism of muscle control from the brain is not still well understood. In this work, the coherence analysis applied to the myoelectric signals, once the artifacts were removed by an independent component analysis neural network, reveals the existence of this common drive only in postural muscles, like the pectorals, whereas this low frequency component (<4 Hz) isn´t found in distal muscles like the dorsal first interosseous. The results fully agree with physiological studies, that assume the existence of this common drive only for axial muscles, in order to help the postural task. The distal muscles don´t need this control and effectively the coherence analysis shows a low value for all frequencies when applied on two first distal interosseus muscles.