Research map / Vision-language and multimodal / Human and embodied modeling
ROAMM: A Benchmark Dataset for Multimodal Human Attention Decoding and EEG-to-Text Modeling During Naturalistic Reading
Haorui Sun, Ardyn Vivienne Olszko, Niharika Singh, David C. Jangraw
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 EEG features to eye-tracking features (ablation) | Random Forest · ROAMM MW detection (LOSO) | AUROC: 0.565 -> 0.609 | Improved |
| add EEG features to eye-tracking features (ablation) | LogReg · ROAMM MW detection (LOSO) | AUROC: 0.592 -> 0.599 | Improved |
| add EEG features to eye-tracking features (ablation) | RBF SVC · ROAMM MW detection (LOSO) | AUROC: 0.586 -> 0.600 | Improved |
| add EEG features to eye-tracking features (ablation) | kNN · ROAMM MW detection (LOSO) | AUROC: 0.549 -> 0.557 | Improved |
| add EEG features to eye-tracking features (ablation) | Gradient Boosting · ROAMM MW detection (LOSO) | AUROC: 0.588 -> 0.587 | No clear effect |
| add EEG features to eye-tracking features (ablation) | MLP · ROAMM MW detection (LOSO) | AUROC: 0.582 -> 0.560 | Hurt |
| use EEG only instead of eye-tracking only (ablation) | Random Forest · ROAMM MW detection (LOSO) | AUROC: 0.565 -> 0.590 | Improved |
| use EEG only instead of eye-tracking only (ablation) | LogReg · ROAMM MW detection (LOSO) | AUROC: 0.592 -> 0.569 | Hurt |
| evaluate within-subject instead of leave-one-subject-out | MW classifiers · ROAMM MW detection | AUROC: 0.609 -> 0.72 | Improved |
| train and test on MW segments instead of all segments | EEG-to-text (Wang & Ji) · EEG-to-text ROAMM | BLEU-1: 0.168 -> 0.057 | Hurt |
| train and test on non-MW segments instead of all segments | EEG-to-text (Wang & Ji) · EEG-to-text ROAMM | BLEU-1: 0.168 -> 0.189 | Improved |
| train and test on MW segments instead of non-MW segments | EEG-to-text (Wang & Ji) · EEG-to-text ROAMM | ROUGE-1 F1: 0.258 -> 0.094 | Hurt |
Closest papers by meaning
- MindAlign: Bridging EEG, Vision, and Language for Zero-Shot Visual Decoding (NeurIPS 2026 · similarity 0.85)
- MindLLM: A Subject-Agnostic and Versatile Model for fMRI-to-text Decoding (ICML 2025 · similarity 0.85)
- EEG-Language Pretraining for Highly Label-Efficient Clinical Phenotyping (ICML 2025 · similarity 0.84)
- NeuroLM: A Universal Multi-task Foundation Model for Bridging the Gap between Language and EEG Signals (ICLR 2025 · similarity 0.84)
- CSBrain: A Cross-scale Spatiotemporal Brain Foundation Model for EEG Decoding (NeurIPS 2025 · similarity 0.83)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.83)
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.
