20.500.12556/DKUM-97209 Activity index outperforms cumulative spike train and amplitude envelopes in surface EMG coherence analysis BACKGROUND AND AIM: Coherence is frequently used to study functional coupling between two motor neuron pools. Decomposing surface electromyograms (sEMG) into spike trains of individual motor units (MUs) and summing them into cumulative spike train (CST) [1] may improve the coherence estimates provided by sEMG amplitude envelopes (AE), mainly due to removal of MU action potentials (MUAPs). However, the coherence estimation in CST depends on the number of MUs and reaches optimal values at several tens of detected MUs. Such a high number of MUs cannot always be guaranteed. Conversely, activity index (AI), which constitutes the first step of the CKC algorithm [2], compensates the MUAPs and combines contributions of all the MUs active in the detection volume of sEMG electrodes. In this study we compare coherence calculations using AE, AI and CST in synthetic and experimental sEMG. METHODS: 10 biceps brachii muscles were simulated with 500 MUs each. 20 s of 9×10 sEMG channels were simulated and sampled at 2048 Hz. Excitation levels of 10%, 30% and 50% of MVC were simulated with superimposed 10 Hz sinusoidal modulation with amplitude of 5% MVC. Experimental sEMG was acquired by 5×13 electrode array from gastrocnemius medialis (GM) and lateralis (GL) muscles of 5 healthy subjects during 20 s long 30% MVC contraction. All sEMG signals were decomposed by CKC algorithm [2], yielding MU spike trains and AI. MU spike trains with pulse-to-noise ratio > 25 dB were summed into CST and spatial average of rectified sEMG was used as AE. Coherence was calculated between all pairs of sEMG signals from two different simulated muscles and between simultaneous recordings of GL and GM. We evaluated the coherence values at 10 Hz (synthetic signals) and at the coherence peak on the 8-30 Hz interval (exp. signals). RESULTS: 14.2±6.1 and 26.1±16.7 MUs were identified from synthetic and exp. sEMG. In synthetic case, the AI yielded significantly higher (P<0.001) coherence values at 10 Hz (0.97±0.01) than the AE (0.83±0.07) and CST (0.78±0.23). In exp. signals coherence peaks were significantly (P<0.05) higher in the AI (0.48±0.20) than in the AE (0.35±0.26) or CST (0.24±0.07). CONCLUSION: The AI yielded higher coherence values than the AE and CST, likely due to the larger number of MUs (compared to the CST) and MUAPs compensation (compared to the AE). FUNDING: This research was funded by the European Union’s Horizon Europe Research and Innovation Program (HybridNeuro project, GA No. 101079392). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. high-density surface electromyography motor unit activity index cumulative spike train coherence nevroni elektromiogrami motorika true false true Angleški jezik Slovenski jezik Znanstveno delo 2026-02-23 15:59:13 2026-02-26 12:36:57 2026-02-27 03:17:41 0000-00-00 00:00:00 2024 0 0 1 0000-00-00 PreprintDelovna Neobjavljeno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.8:61 204427523 204432899 ISEK2024_Kutos.pdf ISEK2024_Kutos.pdf 1 6D9C8F0D40B5005C4AA535035F45FE3E 74888cfefb18cbe9c22855ce08bb5f456566e8ccbc9e545c9ae3834039e7f3c2 4341cf2d-10c8-11f1-95c9-00155d580e1a https://dk.um.si/Dokument.php?lang=slv&id=203699 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0