Artificial intelligence in central nervous system autoimmune diseases: from clinical decision support to mechanistic discovery and translation

Abstract

Abstract Central nervous system (CNS) autoimmune diseases, including multiple sclerosis (MS), neuromyelitis optica spectrum disorder (NMOSD), autoimmune encephalitis (AE), and myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), remain difficult to diagnose, stratify, and treat because clinical phenotypes overlap, biomarkers are imperfect, and the disease course requires longitudinal observation. Artificial intelligence (AI) is increasingly being evaluated to integrate clinical, imaging, immunological, and molecular data for decision-making in neuroimmunology. Clinical applications include automated neuroimaging analysis, differential diagnosis, disease classification, relapse prediction, prognostic stratification, and risk-informed care planning. Evidence is unevenly distributed: MS has the largest body of published work, particularly in neuroimaging, whereas studies in NMOSD, AE, and MOGAD are smaller, more heterogeneous, and often lack external validation. Computational approaches are also being applied to multi-omics, single-cell and spatial data, molecular target prioritisation, drug repositioning, trial enrichment, digital biomarkers, federated learning, and longitudinal modelling; however, these applications generally provide association, prioritisation, or hypothesis-generating evidence rather than validated therapeutic benefit. This Review maps reported applications, evidence patterns, and methodological and implementation needs rather than ranking translational maturity or implying established clinical benefit.