Learning Structural-Functional Brain Representations through Multi-Scale Adaptive Graph Attention for Cognitive Insight

TL;DR

This paper introduces MAGNet, a novel Transformer-style graph neural network that adaptively models the interaction between brain structure and function for better cognitive insight.

Contribution

MAGNet is a new multi-scale adaptive graph neural network framework that effectively integrates structural MRI and resting-state fMRI data.

Findings

MAGNet outperforms relevant baselines on the ABCD dataset.

The model effectively captures structure-function interactions.

Joint cross-modal coherence improves prediction accuracy.

Abstract

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.