Multi-Scale Adaptive Hypergraph Learning for High-Order Brain Connectivity Analysis

Jaeyoon Sim, Soojin Hwang, Seunghun Baek, Guorong Wu, Won Hwa Kim

Keywords: Neurodegenerative Disease, Brain Network Analysis, High-Order Graph Learning

Abstract: Understanding complex interactions between brain regions is critical for early neurodegenerative disease classification such as Alzheimer’s Disease (AD) and Parkinson’s Disease (PD). Graph-based models are typically employed to investigate brain networks with regional features and their interconnectivity. However, traditional approaches primarily focus on pairwise node interactions between directly connected nodes, limiting their ability to capture higher-order dependencies from multiple brain regions. Although hypergraph-based approaches have been proposed to capture higher-order relations beyond pair-wise connectivity, many existing methods rely on predefined hyperedges or restrict learning to hyperedge weights, limiting their flexibility and ability to capture multi-resolution structural patterns. In this regard, we introduce an adaptive multi-scale hypergraph learning framework, i.e., MASH, which constructs hierarchical node features and dynamically learns high-order interaction through continuous hyperedge construction over multi-resolution graph signals. Through extensive experiments on brain network benchmarks, we demonstrate the superiority of MASH by improving classification of different disease stages. Our model further identifies key regions of interest (ROIs) and their group-wise interactions from the learned hyperedges that are associated with disease progression, highlighting its potential as a powerful tool for brain network analysis with neurodegenerative disorders.

Primary Area: applications to neuroscience & cognitive science

Submission Number: 6625