GCMCG: A Clustering-Aware Graph Attention and Expert Fusion Network for Multi-Paradigm, Multi-task, and Cross-Subject EEG Decoding

TL;DR

This paper introduces GCMCG, a novel EEG decoding framework combining graph attention, clustering, and expert fusion, significantly improving multi-paradigm, multi-task, and cross-subject BCI performance.

Contribution

The paper presents a unified GCMCG model that integrates graph-based electrode modeling, spectral clustering, and a mixture-of-experts architecture for robust EEG decoding.

Findings

Achieves high accuracy across multiple datasets (86.60%, 98.57%, 99.61%)

Enhances cross-subject generalization and robustness

Effectively models electrode relationships and brain regions

Abstract

Brain-Computer Interfaces (BCIs) based on Motor Execution (ME) and Motor Imagery (MI) electroencephalogram (EEG) signals offer a direct pathway for human-machine interaction. However, developing robust decoding models remains challenging due to the complex spatio-temporal dynamics of EEG, its low signal-to-noise ratio, and the limited generalizability of many existing approaches across subjects and paradigms. To address these issues, this paper proposes Graph-guided Clustering Mixture-of-Experts CNN-GRU (GCMCG), a novel unified framework for MI-ME EEG decoding. Our approach integrates a robust preprocessing stage using Independent Component Analysis and Wavelet Transform (ICA-WT) for effective denoising. We further introduce a pre-trainable graph tokenization module that dynamically models electrode relationships via a Graph Attention Network (GAT), followed by unsupervised spectral clustering to decompose signals into interpretable functional brain regions. Each region is processed by a dedicated CNN-GRU expert network, and a gated fusion mechanism with L1 regularization adaptively combines these local features with a global expert. This Mixture-of-Experts (MoE) design enables deep spatio-temporal fusion and enhances representational capacity. A three-stage training strategy incorporating focal loss and progressive sampling is employed to improve cross-subject generalization and handle class imbalance. Evaluated on three public datasets of varying complexity (EEGmmidb-BCI2000, BCI-IV 2a, and M3CV), GCMCG achieves overall accuracies of 86.60%, 98.57%, and 99.61%, respectively, which demonstrates its superior effectiveness and strong generalization capability for practical BCI applications.