A topology-aware multiscale feature fusion network for EEG-based motor imagery decoding

Open Webpage

Chaowen Shen, Akio Namiki

Published: 01 Jan 2025, Last Modified: 06 Nov 2025Knowl. Based Syst. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Spectral coupling issues hinder MI-EEG decoding; topology analysis captures stable patterns to improve robustness.•A topology-aware multiscale fusion network integrates spatiotemporal, spectral, and topological EEG features.•Persistent homology extracts stable spectral-topological patterns, and recursive attention refines frequency modeling.•TA-MFF network achieves state-of-the-art accuracy on BCIC-IV-2a, BCIC-IV-2b, and BCIC-III-Iva MI-EEG datasets.