AI Advances in CNS Autoimmune Care: Review Finds Early Evidence With Limits

The review shows AI has been evaluated in areas like automated neuroimaging analysis and disease classification, with multiple sclerosis receiving the most attention, especially for imaging tasks such as lesion assessment and disability prediction. While some methods showed promise, many studies were retrospective and lacked external validation, limiting how broadly results can be applied.

Researchers also noted a shift toward multimodal data and advanced concepts like federated learning and foundation models, but cautioned these approaches are currently hypothesis-generating rather than proven clinical benefits and require larger, multicentre validation to demonstrate real-world impact.