A global research consortium has released three open-source artificial intelligence tools built to speed up the search for Alzheimer's treatments. The Consortium for Biomedical Research and Artificial Intelligence in Neurodegeneration, known as C-BRAIN, announced the launch at the Alzheimer's Association International Conference in London. One tool synthesizes the sprawling body of Alzheimer's and neuroscience literature, a second pulls insights out of unpublished and so-called "dark" data that never made it into print, and a third gives researchers peer review-style feedback on their work.
C-BRAIN brings together academic researchers, pharmaceutical companies, and philanthropic organizations, with Washington University School of Medicine in St. Louis as a founding member. WashU Medicine led the formation of the 17-member group, whose stated goal is an "AI Biomedical Research Scientist" that works alongside human investigators rather than replacing them.
The problem the consortium is targeting is a brutal one: more than 99% of Alzheimer's drug candidates fail in clinical trials. Part of the reason, C-BRAIN argues, is that what scientists already know is scattered across millions of papers, enormous datasets, and results that were never published at all - including negative findings that could steer other labs away from dead ends. Tools that can read across that fragmented record may surface connections no individual researcher could hold in mind.
Randall J. Bateman, MD, the Knight Distinguished Professor of Neurology at WashU Medicine and C-BRAIN's founder and director, said the combination of AI with the volume of data scientists can now generate creates an opening, and that he expects discoveries in the next few years that would not have been possible otherwise. For families and patients, nothing changes today; the tools are aimed at researchers upstream of the clinic. Whether they shift the failure rate is a question only the trials that follow can answer.