Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fusion of Spatio-Temporal and Multi-Scale Frequency Features for Dry Electrodes MI-EEG Decoding

Tianyi Gong, Can Han, Junxi Wu, Dahong Qian

TL;DR

This work addresses the challenges of dry-electrode MI-EEG decoding, notably low SNR, baseline drift, and cross-subject/session variability, by introducing STGMFM, a tri-branch architecture that fuses spatio-temporal information and envelope dynamics. It combines dual-order graph processing (CCG->TSG and TSG->CCG) with a lightweight Multi-Scale Frequency Mixer, all guided by PLV-based connectivity priors and trained with a regularized, cosine-annealed schedule, followed by decision-level fusion. Experimental results on a 23-channel dry-EEG dataset show STGMFM outperforms CNN/Transformer/GCN baselines across cross-session, cross-subject, and fine-tuned scenarios, with notable gains in cross-subject generalization. The ablation study confirms that each module—dual graphs, MFM, PLV initialization, and simple fusion—contributes to robustness against noise and distribution shifts, supporting a practical path toward reliable dry-electrode BCIs.

Abstract

Dry-electrode Motor Imagery Electroencephalography (MI-EEG) enables fast, comfortable, real-world Brain Computer Interface by eliminating gels and shortening setup for at-home and wearable use.However, dry recordings pose three main issues: lower Signal-to-Noise Ratio with more baseline drift and sudden transients; weaker and noisier data with poor phase alignment across trials; and bigger variances between sessions. These drawbacks lead to larger data distribution shift, making features less stable for MI-EEG tasks.To address these problems, we introduce STGMFM, a tri-branch framework tailored for dry-electrode MI-EEG, which models complementary spatio-temporal dependencies via dual graph orders, and captures robust envelope dynamics with a multi-scale frequency mixing branch, motivated by the observation that amplitude envelopes are less sensitive to contact variability than instantaneous waveforms. Physiologically meaningful connectivity priors guide learning, and decision-level fusion consolidates a noise-tolerant consensus. On our collected dry-electrode MI-EEG, STGMFM consistently surpasses competitive CNN/Transformer/graph baselines. Codes are available at https://github.com/Tianyi-325/STGMFM.

Fusion of Spatio-Temporal and Multi-Scale Frequency Features for Dry Electrodes MI-EEG Decoding

TL;DR

This work addresses the challenges of dry-electrode MI-EEG decoding, notably low SNR, baseline drift, and cross-subject/session variability, by introducing STGMFM, a tri-branch architecture that fuses spatio-temporal information and envelope dynamics. It combines dual-order graph processing (CCG->TSG and TSG->CCG) with a lightweight Multi-Scale Frequency Mixer, all guided by PLV-based connectivity priors and trained with a regularized, cosine-annealed schedule, followed by decision-level fusion. Experimental results on a 23-channel dry-EEG dataset show STGMFM outperforms CNN/Transformer/GCN baselines across cross-session, cross-subject, and fine-tuned scenarios, with notable gains in cross-subject generalization. The ablation study confirms that each module—dual graphs, MFM, PLV initialization, and simple fusion—contributes to robustness against noise and distribution shifts, supporting a practical path toward reliable dry-electrode BCIs.

Abstract

Dry-electrode Motor Imagery Electroencephalography (MI-EEG) enables fast, comfortable, real-world Brain Computer Interface by eliminating gels and shortening setup for at-home and wearable use.However, dry recordings pose three main issues: lower Signal-to-Noise Ratio with more baseline drift and sudden transients; weaker and noisier data with poor phase alignment across trials; and bigger variances between sessions. These drawbacks lead to larger data distribution shift, making features less stable for MI-EEG tasks.To address these problems, we introduce STGMFM, a tri-branch framework tailored for dry-electrode MI-EEG, which models complementary spatio-temporal dependencies via dual graph orders, and captures robust envelope dynamics with a multi-scale frequency mixing branch, motivated by the observation that amplitude envelopes are less sensitive to contact variability than instantaneous waveforms. Physiologically meaningful connectivity priors guide learning, and decision-level fusion consolidates a noise-tolerant consensus. On our collected dry-electrode MI-EEG, STGMFM consistently surpasses competitive CNN/Transformer/graph baselines. Codes are available at https://github.com/Tianyi-325/STGMFM.
Paper Structure (15 sections, 6 equations, 2 figures, 2 tables)

This paper contains 15 sections, 6 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Overview
  4. Notation and Sliding Windowing
  5. PLV-driven Initial Channel Graph
  6. Branch A: CCG->TSG
  7. Branch B: TSG->CCG
  8. Branch C: Multi-Scale Frequency Mixer
  9. Decision-level Fusion
  10. Objective and Optimization
  11. Experiments and Results
  12. Dataset and Protocols
  13. Baselines and Overall Performance
  14. Ablation and Component Analysis (on Cross-Subject)
  15. Conclusion

Figures (2)

  • Figure 1: Pipeline of the proposed model. Raw EEG is windowed into overlapping segments and processed by three branches. A (CCG$\rightarrow$TSG) aligns functional connectivity via a PLV-initialized channel graph before aggregating on a time-slice graph. B (TSG$\rightarrow$CCG) first reconstructs intra-trial continuity and then fuses channels on the CCG. C (MFM) operates on the amplitude envelope (RMS) with multi-resolution imaging and cross-scale mixing. Linear head fuses branch logits at the decision level.
  • Figure 2: Before/after training comparison of adjacency under two priors: PLV initialization and the spatial-neighborhood prior