Resonate-and-fire neurons meet EMG : enhancing gesture classification with spiking neural networks

Tools

Manna, Davide Liberato and Bihl, Trevor and Di Caterina, Gaetano (2026) Resonate-and-fire neurons meet EMG : enhancing gesture classification with spiking neural networks. In: 2026 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2026-05-04 - 2026-05-08. (In Press)

[thumbnail of Manna-etal-2026-Resonate-and-fire-neurons-meet-EMG] Text. Filename: Manna-etal-2026-Resonate-and-fire-neurons-meet-EMG.pdf
Accepted Author Manuscript
Restricted to Repository staff only until 1 January 2099.
Download (188kB) | Request a copy

Abstract

Surface Electromyography (EMG) is widely used in rehabilitation, healthcare, and robotics, where low-power solutions are essential. However, EMG signal classification often demands significant computational resources. This paper introduces a novel, low-power approach using Resonate-and-Fire (RF) neurons combined with Spiking Neural Networks (SNNs). SNNs offer efficient, low-latency classification via neuromorphic (NM) hardware, and RF neurons serve as an encoding layer, enabling end-to-end NM processing. This encoding transforms EMG data into spike-frequency representations, emphasising signal relevance in the frequency domain. The proposed method is evaluated on the Ninapro DB5 dataset and demonstrates superior performance compared to both conventional and other NM-based approaches, while maintaining low latency. These results highlight the potential of fully NMsystems to outperform traditional methods, offering asynchronous, sparse, and energy-efficient computation for EMG classification.

ORCID iDs

Manna, Davide Liberato ORCID logoORCID: https://orcid.org/0000-0001-8963-5050, Bihl, Trevor and Di Caterina, Gaetano ORCID logoORCID: https://orcid.org/0000-0002-7256-0897;
CORE (COnnecting REpositories)