Superficial White Matter Integrity in Early Multiple Sclerosis

This study investigates the sensitivity of superficial white matter (SWM) integrity in distinguishing early multiple sclerosis (MS) patients from healthy controls (HC). Using machine learning models, the best classification performance (AUC = 0.75) was achieved by combining SWM and deep white matter (DWM) features, with SWM integrity alone showing greater sensitivity than DWM, cortical thickness, and resting-state functional connectivity. The findings suggest that SWM abnormalities can be detected in early MS before more pronounced structural and functional changes appear.