Optimal multi-channel electrode array design for high-quality surface EMG acquisition and processing

Abstract

The acquisition of electromyography (EMG) signals has been a topic of interest for many decades, that aids introducing High-Density surface EMG (HD-sEMG) arrays. HD-sEMG found to be useful for neuro-rehabilitation applications as they provide a non-invasive window to the neural activities from the recorded muscle. This study investigates whether extremely dense electrode array can yield better performance in EMG acquisition and decomposition. The initial electrode parameters were selected as following: electrode radios = 0.1mm, inter-electrode distance (IED) = 0.3mm and grid size of 30x30. The IED and grid size were altered by disabling some electrodes in the grid. A synthetic EMG signal was generated using a biophysical simulation model. Additionally, two models were created which are thermal noise and crosstalk model in order to add noise and interference to the generated EMG signal. Lastly data were decomposed by convolutive blind separation model. Our result shows that 30x30 grid performed poorer compared to the other grids, where the best result was achieved by the 15x15 grid. We believe that dens array was subjected to higher crosstalk interference, due to short IED, which degraded its performance. Nevertheless, the inconsistency between increased IED and improved performance suggests that there is a relationship between IED and radius which we believe is worth investigating.