Research map / Label-efficient and robust learning / Recognition and long-tail
Tokenizing Single-Channel EEG with Time-Frequency Motif Learning
Jathurshan Pradeepkumar, Xihao Piao, Zheng Chen, Jimeng Sun
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 |
|---|---|---|---|
| use joint raw-temporal and spectrogram input instead of raw only (ablation) | TFM-Tokenizer · TUEV | Cohen's Kappa: 0.5194 -> 0.5337 | Improved |
| use joint raw-temporal and spectrogram input instead of spectrogram only (ablation) | TFM-Tokenizer · TUEV | Cohen's Kappa: 0.5275 -> 0.5337 | Improved |
| remove position embedding from tokenizer (ablation) | TFM-Tokenizer · TUEV | Cohen's Kappa: 0.5119 -> 0.5337 | Improved |
| remove position embedding from tokenizer (ablation) | TFM-Tokenizer · TUEV | token utilization: 12.87% -> 9.78% | Mixed |
| reduce downstream transformer from 4 to 2 layers (ablation) | TFM-Tokenizer downstream model · TUEV | Cohen's Kappa: 0.5337 -> 0.5758 | Improved |
| increase downstream transformer layers beyond 4 (6, 12) (ablation) | TFM-Tokenizer downstream model · TUEV | Cohen's Kappa: 0.5337 -> 0.4996 (6 layers) | Hurt |
| add masked token prediction pretraining to downstream transformer (ablation) | TFM-Tokenizer downstream model · TUEV | balanced accuracy | Improved |
| add masked token prediction pretraining to downstream transformer (ablation) | TFM-Tokenizer downstream model · TUEV | Cohen's Kappa and Weighted F1 | No clear effect |
| increase vocabulary size from 256 to 8192 (ablation) | TFM-Tokenizer · TUEV | balanced accuracy | Improved |
| plug TFM-Tokenizer into BIOT | BIOT · TUEV (multi-dataset) | Cohen's Kappa: 0.5273 -> 0.5464 | Improved |
| plug TFM-Tokenizer into LaBraM | LaBraM · TUEV (multi-dataset) | Cohen's Kappa: 0.5175 -> 0.5367 | Improved |
| plug TFM-Tokenizer into BIOT | BIOT · TUEV (single-dataset) | balanced accuracy: 0.4679 -> 0.4228 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- EEGPT — EEGPT: Pretrained Transformer for Universal and Reliable Representation of EEG Signals (NeurIPS 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.87)
- EEGPT: Pretrained Transformer for Universal and Reliable Representation of EEG Signals (NeurIPS 2024 · similarity 0.87)
- What do EEG Foundation Models Capture from Human Brain Signals? (NeurIPS 2026 · similarity 0.87)
- LEAF: Language-EEG Aligned Foundation Model for Brain-Computer Interfaces (NeurIPS 2026 · similarity 0.87)
- NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces (NeurIPS 2026 · similarity 0.86)
- B[FM]$^2$: Brain Foundation Model via Flow Matching with SplitUNet (NeurIPS 2026 · 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.
