BLEG: LLM Functions as Powerful fMRI Graph-Enhancer for Brain Network Analysis
Rui Dong, Zitong Wang, Jiaxing Li, Weihuang Zheng, Youyong Kong · 2026
Read BLEG: LLM Functions as Powerful fMRI Graph-Enhancer for Brain Network Analysis by Rui Dong, Zitong Wang, Jiaxing Li, Weihuang Zheng, Youyong Kong (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-07361.txt.
Description
Dong and four colleagues present BLEG, a way to boost graph neural nets for brain-network analysis with a large language model. Graph nets on fMRI data suffer from sparse features and thin domain knowledge in single-modality graphs. BLEG asks the language model for augmented texts, tunes a language model cheaply for richer text features, aligns those features with graph features in training, and tunes an adapter per task. Tests across datasets show BLEG beats prior methods, and a loss aligning language and graph outputs lifts results further.
Catalog record
| CREATOR | Rui Dong, Zitong Wang, Jiaxing Li, Weihuang Zheng, Youyong Kong |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47039 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.07361