Population graph learning improves brain disorder diagnosis accuracy
Beyond Phenotypic Similarity: Relation Refinement and Selective Population Graph Learning for Multimodal Brain Disorder Diagnosis
Artificial Intelligence
Summary
Diagnosing brain disorders by looking at brain data from many people at once can be tricky if the connections between people are not useful for diagnosis. The authors present PopDRP, a method that adjusts how people are connected based on actual diagnosis information instead of just similar traits. This helps the system learn better and improves accuracy when diagnosing conditions like autism or ADHD from brain data. Their tests show clear improvements compared to other methods.
What this means in practice
- •For medical imaging engineers: Enhance brain disorder diagnostic tools by integrating refined population graphs to improve accuracy using multimodal brain data.
- •For clinical data scientists: Develop better machine learning models that use diagnosis-informed relations to analyze brain scans and predict disorders more reliably.
Authors
Yinyu Lan
Abstract
Population graph learning supports brain disorder diagnosis by sharing information across subjects. Its effectiveness, however, depends on whether the relationships guiding this exchange are relevant to diagnosis. Similarity in demographic or acquisition attributes does not necessarily indicate diagnostic agreement, and heterogeneous subjects may require different representation updates within the same population graph. We propose PopDRP, a framework that combines diagnosis-guided relation refinement with hierarchical selective state-space learning. Using only training labels, Contrastive Affinity Refinement calibrates attribute-based affinities according to diagnostic agreement. Neural Graph Rectification then refines these relationships using fused functional-connectivity and non-imaging representations, guided by graph reconstruction and dual-view alignment objectives. On the resulting graph, a Selective Pyramid State-Space backbone integrates input-dependent state updates, graph diffusion, and hierarchical pooling and unpooling to learn population representations at multiple resolutions. Together, these components connect the refinement of inter-subject relationships with selective representation learning over the population graph. Experiments on ABIDE and ADHD-200 under stratified ten-fold cross-validation show improved accuracy and AUC over the compared baselines. Ablation studies further support the contributions of relation refinement and the selective hierarchical backbone to classification performance.