Learning Structural-Functional Brain Representations through Multi-Scale Adaptive Graph Attention for Cognitive Insight
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915903290998784 |
|---|---|
| author | Mazumder, Badhan Wiafe, Sir-Lord Kotoski, Aline Calhoun, Vince D. Ye, Dong Hye |
| author_facet | Mazumder, Badhan Wiafe, Sir-Lord Kotoski, Aline Calhoun, Vince D. Ye, Dong Hye |
| contents | Understanding how brain structure and function interact is key to explaining intelligence yet modeling them jointly is challenging as the structural and functional connectome capture complementary aspects of organization. We introduced Multi-scale Adaptive Graph Network (MAGNet), a Transformer-style graph neural network framework that adaptively learns structure-function interactions. MAGNet leverages source-based morphometry from structural MRI to extract inter-regional morphological features and fuses them with functional network connectivity from resting-state fMRI. A hybrid graph integrates direct and indirect pathways, while local-global attention refines connectivity importance and a joint loss simultaneously enforces cross-modal coherence and optimizes the prediction objective end-to-end. On the ABCD dataset, MAGNet outperformed relevant baselines, demonstrating effective multimodal integration for advancing our understanding of cognitive function. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_29967 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Learning Structural-Functional Brain Representations through Multi-Scale Adaptive Graph Attention for Cognitive Insight Mazumder, Badhan Wiafe, Sir-Lord Kotoski, Aline Calhoun, Vince D. Ye, Dong Hye Computer Vision and Pattern Recognition Understanding how brain structure and function interact is key to explaining intelligence yet modeling them jointly is challenging as the structural and functional connectome capture complementary aspects of organization. We introduced Multi-scale Adaptive Graph Network (MAGNet), a Transformer-style graph neural network framework that adaptively learns structure-function interactions. MAGNet leverages source-based morphometry from structural MRI to extract inter-regional morphological features and fuses them with functional network connectivity from resting-state fMRI. A hybrid graph integrates direct and indirect pathways, while local-global attention refines connectivity importance and a joint loss simultaneously enforces cross-modal coherence and optimizes the prediction objective end-to-end. On the ABCD dataset, MAGNet outperformed relevant baselines, demonstrating effective multimodal integration for advancing our understanding of cognitive function. |
| title | Learning Structural-Functional Brain Representations through Multi-Scale Adaptive Graph Attention for Cognitive Insight |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.29967 |