Open-Source AI Aims to Prevent Alzheimer’s Before Symptoms

MIT-led researchers unveil FINGERS-7B, the first AI foundation model designed to prevent Alzheimer's by integrating lifestyle, clinical, genomic, and proteomic data into a single biological fingerprint. On WW-FINGERS data it delivers four times more accurate preclinical diagnosis and 130% better responder stratification, predicting individual risk and timing of cognitive decline up to a decade ahead to guide personalized prevention, with open-source weights and training code deployed in the AD Workbench for global research use.