Research map / Label-efficient and robust learning / Recognition and long-tail
CAT-Net: A Cross-Attention Tone Network for Cross-Subject EEG-EMG Fusion Tone Decoding
Yifan Zhuang, Calvin Huang, Zepeng Yu, Yongjie Zou, Jiawei Ju
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 cross-attention fusion (ablation) | CATNet · Mandarin tone EEG-EMG (silent speech, 5-fold, 20 EEG channels) | precision: 77.63 -> 88.08 | Improved |
| add BiLSTM temporal encoder (ablation) | CATNet · Mandarin tone EEG-EMG (silent speech, 5-fold, 20 EEG channels) | precision: 78.42 -> 88.08 | Improved |
| add EMG-to-EEG directional attention (fusion EMG) (ablation) | CATNet · Mandarin tone EEG-EMG (silent speech, 5-fold, 20 EEG channels) | precision: 76.60 -> 88.08 | Improved |
| add EEG-to-EMG directional attention (fusion EEG) (ablation) | CATNet · Mandarin tone EEG-EMG (silent speech, 5-fold, 20 EEG channels) | precision: 76.46 -> 88.08 | Improved |
| add domain discriminator (gradient reversal) (ablation) | CATNet · Mandarin tone EEG-EMG (silent speech, cross-subject LOSO) | average accuracy: 84.69 -> 85.10 | Improved |
| add domain discriminator for subject S1 (ablation) | CATNet · Mandarin tone EEG-EMG (silent speech, cross-subject LOSO, subject S1) | accuracy: 48.96 -> 53.33 | Improved |
| use 20 EEG channels instead of 5 | CATNet · Mandarin tone EEG-EMG (silent speech, channel count) | average accuracy: 86.92 -> 88.08 | Improved |
| use full 64 channels instead of 20 | CATNet · Mandarin tone EEG-EMG (silent speech, channel count) | average accuracy: 88.08 -> 88.45 | No clear effect |
| use 10 EEG channels instead of 5 | CATNet · Mandarin tone EEG-EMG (silent speech, channel count) | average accuracy: 86.92 -> 87.02 | No clear effect |
| evaluate cross-subject instead of within-subject (silent) | CATNet · Mandarin tone EEG-EMG (silent speech) | accuracy: 88.08 -> 85.10 | Hurt |
| evaluate cross-subject instead of within-subject (audible) | CATNet · Mandarin tone EEG-EMG (audible speech) | accuracy: 87.83 -> 83.27 | Hurt |
Closest papers by meaning
- Towards Homogeneous Lexical Tone Decoding from Heterogeneous Intracranial Recordings (ICLR 2025 · similarity 0.88)
- CSBrain: A Cross-scale Spatiotemporal Brain Foundation Model for EEG Decoding (NeurIPS 2025 · similarity 0.87)
- Du-IN: Discrete units-guided mask modeling for decoding speech from Intracranial Neural signals (NeurIPS 2024 · similarity 0.86)
- CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding (ICLR 2025 · similarity 0.86)
- Neural decoding from stereotactic EEG: accounting for electrode variability across subjects (NeurIPS 2024 · similarity 0.86)
- SENSE: Semantic Neural Speech Synthesis from Brain Dynamics via Spatial Graph Encoding (NeurIPS 2026 · similarity 0.85)
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.
