A Novel GNN Framework Integrating Neuroimaging and Behavioral Information to Understand Adolescent Psychiatric Disorders

Weifeng Yu, Gang Qu, Young-geun Kim, Lei Xu, Aiying Zhang

Keywords: adolescent physiatric disorder, neurobehavior, graph autoencoder

TL;DR: Graph learning framework integrates functional connectivity and behavioral characteristics

Abstract: Functional connectivity (FC) is widely used to study various psychiatric disorders, but its consistency is often undermined by significant inter-subject variability. While these differences can be reflected in behavioral characteristics, few studies have combined them with FC. To this end, we propose a novel graph learning framework that enhances the differentiation of psychiatric disorders by integrating FC and behavioral characteristics. Additionally, we apply Grad-CAM to enhance model interpretability by identifying key regions of interest involved in distinguishing individuals with psychiatric disorders from healthy controls. Experiments with the Adolescent Brain Cognitive Development dataset highlighted two critical insights: the thalamus and specific ROIs within the somatomotor and cingulo-opercular networks play a critical role for identifying psychiatric disorders. Additionally, visualization of latent representations demonstrated that individuals with externalizing disorders, specifically Attention Deficit Hyperactivity Disorder and Oppositional Defiant Disorder, can be distinguished from healthy controls. These findings underscore the utility of our graph learning framework for identifying psychiatric disorders and suggest its promise for improving diagnostic accuracy.

Primary Subject Area: Integration of Imaging and Clinical Data

Secondary Subject Area: Interpretability and Explainable AI

Paper Type: Both

Registration Requirement: Yes

Reproducibility: https://github.com/elleryyu/BEG-GAE

Visa & Travel: Yes

Midl Latex Submission Checklist: Ensure no LaTeX errors during compilation., Created a single midl25_NNN.zip file with midl25_NNN.tex, midl25_NNN.bib, all necessary figures and files., Includes \documentclass{midl}, \jmlryear{2025}, \jmlrworkshop, \jmlrvolume, \editors, and correct \bibliography command., Did not override options of the hyperref package, Did not use the times package., All authors and co-authors are correctly listed with proper spelling and avoid Unicode characters., Author and institution details are de-anonymized where needed. All author names, affiliations, and paper title are correctly spelled and capitalized in the biography section., References must use the .bib file. Did not override the bibliographystyle defined in midl.cls. Did not use \begin{thebibliography} directly to insert references., Tables and figures do not overflow margins; avoid using \scalebox; used \resizebox when needed., Included all necessary figures and removed *unused* files in the zip archive., Removed special formatting, visual annotations, and highlights used during rebuttal., All special characters in the paper and .bib file use LaTeX commands (e.g., \'e for é)., Appendices and supplementary material are included in the same PDF after references., Main paper does not exceed 9 pages; acknowledgements, references, and appendix start on page 10 or later.

Submission Number: 110