CFSPMNet: Cross-subject Fourier-guided Spatial-Patch Mamba Network for EEG Motor Imagery Decoding in Stroke Patients

2026-05-11

AI Summary

This work addresses the challenge of cross-subject decoding of motor imagery electroencephalography (MI-EEG) signals in stroke patients, whose pathological neural reorganization introduces significant inter-subject variability. To tackle this issue, the authors propose CFSPMNet, a novel framework that models MI-EEG as a sequence of latent neural states. The method innovatively reorganizes EEG token representations in the Fourier domain and integrates a state space model (Mamba) with joint spatial–spectral modeling. It further incorporates a shared–private prototype matching mechanism and a pseudo-label calibration strategy constrained by physiological consistency. Evaluated on the XW-Stroke and 2019-Stroke datasets, CFSPMNet achieves average accuracies of 68.23% and 73.33%, respectively, outperforming the strongest baseline by 5.63 and 8.25 percentage points, thereby substantially enhancing cross-subject transfer performance.

Abstract

Motor imagery electroencephalography (MI-EEG) decoding offers a non-invasive route for post-stroke rehabilitation, but cross-patient use remains difficult because pathological neural reorganization changes task-related EEG dynamics, aperiodic activity, local excitability, cross-regional coordination, and trial-level brain-state context. This makes source-learned MI representations unreliable for unseen patients. To address this problem, we propose CFSPMNet, a cross-patient adaptation framework that models post-stroke MI-EEG as latent neural-state organization. CFSPMNet combines a Fourier-Reorganized State Mamba Network (FRSM) with Shared-Private Prototype Matching (SPPM). FRSM represents each trial as a latent physiological token sequence, reorganizes token states in the Fourier domain, and uses Fourier-derived trial context to guide Mamba state-space propagation. SPPM improves target pseudo-label updating by combining semantic confidence with shared-private physiological consistency, filtering confident but physiologically inconsistent target predictions. Leave-one-subject-out experiments on two stroke MI-EEG datasets show that CFSPMNet outperforms representative CNN-, Transformer-, Mamba-, and adaptation-based baselines, achieving average accuracies of 68.23% on XW-Stroke and 73.33% on 2019-Stroke, with gains of 5.63 and 8.25 percentage points over the strongest competitors. Ablation, sensitivity, feature-alignment, pseudo-label selection, and neurophysiological visualization analyses further support the roles of Fourier-domain token-state reorganization and calibrated pseudo-label updating. These results suggest that latent neural-state modeling can improve rehabilitation-oriented cross-patient BCI decoding. Code is available at https://github.com/wxk1224/CFSPMNet.

Problem

• cross-subject
• MI-EEG
• stroke rehabilitation
• neural reorganization
• BCI decoding

Innovation

• Cross-subject adaptation
• Fourier-guided reorganization
• Mamba state-space model
• Prototype matching
• MI-EEG decoding