Machine Learning Analysis of Pupillary Light Reflex for Assessing Autonomic Dysfunction in Multiple Sclerosis

This study investigated pupillography combined with machine learning to detect autonomic dysfunction in patients with relapsing-remitting MS, achieving 85.7% accuracy on test data but lower performance (75% accuracy) on independent validation. The research demonstrates feasibility of pupil-based biomarkers for MS assessment while acknowledging generalizability challenges.