An Innovative Binary Quadratic Programming Approach for the Accurate Identification of Discharge Timings of Motor Units From High-Density Surface EMG Signals

Decoding motor unit (MU) activity in electromyo-graphic (EMG) signals traditionally relies on spike sorting techniques, which require time-consuming expert intervention at some stage, limiting its use in neural engineering applications. We introduce a novel method using binary quadratic programming to identify MU discharge timings. Our approach, validated on simulated signals, demonstrates improved accuracy compared to conventional methods, typically consisting of unsupervised learning algorithms. In its optimal setup, the new method achieves a superior accuracy <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(94\%\pm6$</tex>, considerably higher than traditional methods <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(74\%\pm 21$</tex>. Our approach offers a streamlined pathway for MU identification in high-density surface EMG (HDsEMG) recordings by automating the discharge identification process and reducing the need for expert oversight. This advancement holds significant promise for expanding the application of HDsEMG analysis in research and clinical settings within the field of neural engineering.