Authors: Martin Weygandt, Kerstin Hackmack, Caspar Pfüller, Judith Bellmann–Strobl, Friedemann Paul, Frauke Zipp, John–Dylan Haynes
DOI: 10.1371/journal.pone.0021138
Abstract Summary
Machine learning identified specific brain “hotspots” that distinguish multiple sclerosis patients from healthy controls with high accuracy. Beyond visible lesions, the technique detected MS-related changes in normal-appearing brain tissue, including cerebellar and posterior brain regions. This millimeter-scale precision surpasses current diagnostic guidelines and reveals that standard MRI captures more subtle MS-related tissue alterations than previously recognized.
Why Brain? 🧠
Machine learning identifies MS-specific brain changes in normal-appearing tissue on standard MRI, revealing diagnostic hotspots beyond visible lesions with up to 96% accuracy in millimeter-scale regions.
The image is AI-generated for illustrative purposes only. Courtesy of Midjourney.
