Research map / Label-efficient and robust learning / Self-supervised and contrastive
State Mamba: Spatiotemporal EEG State-Space Model with Dynamic Brain Alignment for Cross-Subject Representation
Weining Weng, Yang Gu, Yuan Ma, Yuchen Liu, Yingwei Zhang, Yiqiang Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add self-supervised pretraining (both alignment tasks) (ablation) | State Mamba · FACED (10-fold CV) | accuracy (%): 49.83 -> 54.69 | Improved |
| add self-supervised pretraining (ablation) | State Mamba · DEAP-Arousal (10-fold CV) | accuracy (%): 60.62 -> 64.38 | Improved |
| add representation alignment (ablation) | State Mamba · FACED (10-fold CV) | accuracy (%): 51.37 -> 54.69 | Improved |
| add pattern alignment (ablation) | State Mamba · FACED (10-fold CV) | accuracy (%): 53.82 -> 54.69 | Improved |
| add spatial channel-wise state transition (ablation) | State Mamba · DEAP-Valence (10-fold CV) | accuracy (%): 61.02 -> 63.97 | Improved |
| use LGANN instead of average pooling (ablation) | State Mamba · FACED (10-fold CV) | accuracy (%): 52.91 -> 54.69 | Improved |
| use LGANN instead of direct concatenation (ablation) | State Mamba · ISRUC-S1 (10-fold CV) | accuracy (%): 80.03 -> 79.32 | Hurt |
| use LGANN instead of direct concatenation (ablation) | State Mamba · FACED and DEAP (10-fold CV) | accuracy (%): 54.61 vs 54.69 (FACED) | No clear effect |
| use State Mamba instead of CBraMod | EEG encoder · DEAP arousal and valence (10-fold CV) | accuracy (%): 64.09 -> 64.38 (arousal) | No clear effect |
| use LOSO instead of 10-fold training | State Mamba · FACED | accuracy (%): 54.69 -> 58.33 | Improved |
| vary hidden dimension (ablation) | State Mamba · all four tasks | accuracy | Mixed |
| vary number of stacked layers (1 to 7) (ablation) | State Mamba · all four tasks | accuracy | No clear effect |
Closest papers by meaning
- CSBrain: A Cross-scale Spatiotemporal Brain Foundation Model for EEG Decoding (NeurIPS 2025 · similarity 0.89)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.88)
- Neural decoding from stereotactic EEG: accounting for electrode variability across subjects (NeurIPS 2024 · similarity 0.88)
- Multi-dataset Joint Pre-training of Emotional EEG Enables Generalizable Affective Computing (NeurIPS 2025 · similarity 0.87)
- CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding (ICLR 2025 · similarity 0.86)
- NeurIPT: Foundation Model for Neural Interfaces (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
