MPP-GNN: Subject-Adaptive Community Detection for fMRI-Based Alzheimer's Disease Classification

2026-07-29

AI Summary

This work addresses a key limitation in existing graph neural network (GNN) approaches for brain functional connectivity analysis: the assumption of a fixed number of functional modules across subjects, which overlooks individual variability and fails to leverage module discovery to inform connectivity learning. To overcome this, the authors propose MPP-GNN, the first model to integrate subject-adaptive hierarchical graph partitioning with GNNs in an end-to-end framework. Through a bilevel optimization scheme, MPP-GNN explicitly identifies subject-specific functional modules and uses them as priors to guide edge weight optimization and representation learning. This approach eliminates the rigid fixed-module assumption, achieves state-of-the-art AUC performance on two public Alzheimer’s disease datasets, yields modules highly consistent with the Yeo brain atlas, and reveals disease-related patterns of network dedifferentiation.

Abstract

Functional magnetic resonance imaging (fMRI) is a widely used technique for studying the brain. Recent methods that utilize graph neural networks (GNNs) for analysis of brain functional connectivity have shown great potential for the classification of brain disorders, such as Alzheimer's disease (AD). However, these methods often assume a preset number of functional modules across all subjects, which overlooks inter-subject variability. In addition, the discovered modules are rarely used to directly guide the learned connectivity patterns. Here, to address these issues, we propose a Meta Probabilistic Pooling GNN (MPP-GNN). We frame the model's task as a coupled, bilevel optimization that performs adaptive graph partitioning hierarchically to discover subject-specific modules and then uses the discovered brain modules as an explicit prior to guide edge refinement and representation learning. We validate MPP-GNN on two public datasets for AD classification, achieving the highest AUC in comparison to established baselines for both datasets. Furthermore, our analysis demonstrates that MPP-GNN shows significant alignment with the canonical functional-network organization defined by the Yeo brain atlas and reveals a network-level dedifferentiation pattern for AD.

Problem

• Alzheimer's disease
• functional MRI
• graph neural networks
• subject variability
• community detection

Innovation

• subject-adaptive community detection
• graph neural networks
• functional connectivity
• bilevel optimization
• Alzheimer's disease classification