Research map / Graph learning / Graph self-supervised learning
NeuroCLUS: A Foundation Model with Functional Clustering for Intracranial Neural Decoding
Hui Zheng, Haiteng Wang
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 clustering loss L_clus to pre-training (ablation) | NeuroCLUS · Du-IN speech production | balanced accuracy (%): 44.72 -> 65.92 | Improved |
| add reconstruction loss L_mse to pre-training (ablation) | NeuroCLUS · Du-IN speech production | balanced accuracy (%): 32.70 -> 65.92 | Improved |
| use functional context prediction instead of masked signal prediction for the connectivity extractor (ablation) | NeuroCLUS · all benchmark tasks | performance | Improved |
| increase number of prototypes K from 1 to 8 (ablation) | NeuroCLUS · downstream tasks | performance | Improved |
| increase number of prototypes K beyond 8 (16, 32) (ablation) | NeuroCLUS · downstream tasks | performance | No clear effect |
| include downstream datasets (MAYO/FNUSA, Brain Treebank, Du-IN) in pre-training (ablation) | NeuroCLUS · seizure detection, Brain Treebank, Du-IN | performance | No clear effect |
| increase pre-training data volume (ablation) | NeuroCLUS · downstream tasks | performance | Improved |
| cluster channels into functional prototypes instead of channel-level or brain-level tokens | foundation model tokenization · Du-IN speech production | balanced accuracy (%): 42.57 (PopT) -> 65.92 | Improved |
| replace continuous embeddings with discrete code sequences (H2DiLR vs Du-IN) | region-level tokenization · Du-IN speech production | balanced accuracy (%): 62.70 -> 23.92 | Hurt |
Closest papers by meaning
- CSBrain: A Cross-scale Spatiotemporal Brain Foundation Model for EEG Decoding (NeurIPS 2025 · similarity 0.88)
- BaRISTA: Brain Scale Informed Spatiotemporal Representation of Human Intracranial Neural Activity (NeurIPS 2025 · similarity 0.87)
- BrainLM: A foundation model for brain activity recordings (ICLR 2024 · similarity 0.87)
- Leveraging unlabelled data for generalizable neural population decoding (NeurIPS 2026 · similarity 0.86)
- Du-IN: Discrete units-guided mask modeling for decoding speech from Intracranial Neural signals (NeurIPS 2024 · 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.
