Research map / Image generation / Super-resolution
FlexiBrain: Resolution-Agnostic Voxel-Level Encoding for Native fMRI
mo wang, Wenhao Ye, Junfeng Xia, Minghao Xu, Hongkai Wen, Quanying Liu
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 MoE to Mamba-JEPA (ablation) | FlexiBrain · ABIDE | accuracy (%): 52.55 -> 65.45 | Improved |
| add MoE to Mamba-JEPA (ablation) | FlexiBrain · ADNI (MCI vs CN) | accuracy (%): 71.88 -> 79.16 | Improved |
| add pre-training before fine-tuning (ablation) | FlexiBrain · ADNI (MCI vs CN) | accuracy (%): 68.75 -> 79.16 | Improved |
| add pre-training before fine-tuning (ablation) | FlexiBrain · ABIDE | accuracy (%): 59.85 -> 65.45 | Improved |
| use Mamba backbone instead of ViT (ablation) | FlexiBrain · ADHD-200 | accuracy (%): 63.92 -> 69.76 | Improved |
| use Mamba backbone instead of ViT (ablation) | FlexiBrain · Backbone profiling (~50M params) | memory (GB): 8.29 GB -> 1.97 GB | Improved |
| use native-space fMRI instead of template-space (ablation) | FlexiBrain · ADNI (MCI vs CN) | accuracy: 5 hours -> 3 minutes preprocessing | Improved |
| patch size 6 s and 12 mm versus 24 mm spatial (ablation) | FlexiBrain · ADHD-200 (from scratch) | accuracy (%): 61.86 -> 65.98 | Improved |
| patch size 12mm/6s versus 12mm/12s (ablation) | FlexiBrain · ADHD-200 (from scratch) | accuracy (%): 64.95 -> 65.98 | Improved |
| downsample PPMI spatially and temporally by 2 (pretraining probe) (ablation) | FlexiBrain · PPMI downsampling | BACC (%): 67.05 -> 58.96 | Hurt |
| increase PPMI training data from 10% to 100% | FlexiBrain · PPMI few-shot | accuracy (%): 58.33 -> 76.04 | Improved |
| add MoE to reduce JEPA target collapse (ablation) | FlexiBrain · Target features | effective rank: 60.15 -> 144.27 | Improved |
Closest papers by meaning
- FlatClip: Reusing Image Foundation Models for fMRI Representation Learning via Cortical Flatmaps (NeurIPS 2026 · similarity 0.88)
- Omni-fMRI: A Universal Atlas-Free fMRI Foundation Model (ICML 2026 · similarity 0.88)
- MindEye2: Shared-Subject Models Enable fMRI-To-Image With 1 Hour of Data (ICML 2024 · similarity 0.87)
- MIRAGE: Adaptive Multimodal Gating for Whole-Brain fMRI Encoding (NeurIPS 2026 · similarity 0.86)
- A Compact Implicit Neural Representation for Efficient Storage of Massive 4D Functional Magnetic Resonance Imaging (AAAI 2025 · similarity 0.86)
- Meta-Learning an In-Context Transformer Model of Human Higher Visual Cortex (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.
