Brain Connectivity based Prediction of Trait anxiety using Graph Neural Network

Motivation: Trait anxiety is characterized as a personality disposition where individuals consistently respond to situations with elevated anxiety. Although numerous studies have revealed associations between trait anxiety and specific brain connectivity, methods for predicting trait anxiety using brain connectivity data require further enhancement. Goal(s): This study aims to develop a model that better predicts trait anxiety levels, measured by the STAI-T score, using MR images. Approach: We generate Connectivity Matrices from MR images, as input to Graph Neural Network (GNN) Model for prediction. Results: The proposed method more accurately predicted trait anxiety levels from Functional and Structural Connectivity compared to previously reported methods. Impact: This study suggests the possibility to complement traditional psychological assessments by predicting accurately trait anxiety levels using MR images. The approach could simplify mental health diagnostics, raise new questions about imaging-based predictions and improve access to timely interventions.