arXiv：cs.AI· Bui Van Hung, Dang Thanh Vu, Hoang-Son Vo, Mai Duc Minh Huy, Mai Tien Dung, Soo-Hyung Kim·· 20 小时前
StimFlow：用刺激条件化流匹配从图像合成脑活动
StimFlow: Synthesizing Brain Activity from Images via Stimulus-conditioned Flow Matching
AI 导读
StimFlow 通过条件化流匹配在原生体素空间直接合成 fMRI 响应，单样本即达到 0.371 的逐体素 Pearson 相关性，匹敌需要平均五次采样的基线。它结合 DINOv2 结构特征与 CLIP 语义特征以匹配视觉层级，并仅用一小时扫描数据即可恢复超过一半的全数据性能，支持高效的跨被试适配。
正文
Abstract:The inherent variability of fMRI responses across trials encourages a generative view of the stimulus-to-response mapping. Prior generative encoders often operate in compressed latent spaces and condition predominantly on semantic features, which provide limited low-level cues for early visual cortex. We propose StimFlow, which synthesizes fMRI directly in native voxel space using conditional flow matching with a linear path. This design aligns training and inference distributions, enabling competitive performance with a single sample and reducing reliance on averaging. To match the visual hierarchy, StimFlow combines structural DINOv2 features with semantic CLIP features, which prove complementary along the cortical processing stream. With a single sample, StimFlow achieves a per-voxel Pearson correlation of 0.371, matching baselines that average five draws. Generation in voxel space allows direct ROI analysis and concept localization without a decoder. Furthermore, it recovers over half of full-data performance from just one hour of scans, supporting efficient cross-subject adaptation.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.15341 [cs.CV] |
| (or arXiv:2610.15341v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.15341 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Van Hung Bui [view email]
[v1]
Sun, 11 Oct 2026 04:45:03 UTC (5,151 KB)
来源：arXiv：cs.AI · arxiv.org