Interpretable Machine Learning Model for Characterizing Magnetic Susceptibility-based Biomarkers in First Episode Psychosis

Motivation: The research aims to improve diagnostic accuracy and treatment strategies for First-Episode Psychosis by identifying critical brain iron biomarkers. Goal(s): The present study aims to pinpoint potential predictive biomarkers derived from QSM and R2* for individuals experiencing first-episode-psychosis (FEP), along with their response to antipsychotic treatment. Approach: The study used MRI to assess brain iron levels in psychosis patients, employing machine learning for classification and treatment response prediction. Results: The study achieved 76.48% accuracy in classifying healthy individuals from First-Episode Psychosis patients and 76.43% accuracy in predicting treatment responses, identifying key biomarkers linked to dopamine pathways. Impact: This study identifies key brain iron concentration biomarkers using QSM and R2* in FEP, achieving 76% classification accuracy. It highlights the potential for improved early detection and treatment response prediction, paving the way for enhanced clinical outcomes in schizophrenia management.