Skip to content
NeuroTechGuide

Stanford launches five-year project to train AI models on human brain imaging data

The Stanford Digital Brain Project aims to collect about 14,000 hours of 7-tesla fMRI and wearable MEG data from roughly 100 people, and to share the data and models with researchers worldwide.

In briefResearch

Stanford’s Wu Tsai Neurosciences Institute announced the Stanford Digital Brain Project on November 2, a five-year effort to train foundation models on about 14,000 hours of imaging data from roughly 100 people. The first of 10 campaigns, due to start early next year, will scan 10 people about twice a week for a year using 7-tesla fMRI and wearable MEG. Mark and Mary Stevens are the lead funders, with added support from Clara Wu Tsai; the amount is undisclosed.

Stanford says the data and models will be freely available to researchers, and that a long-term goal is predicting individual responses to treatments such as TMS or focused ultrasound. That remains untested. Outside researchers told The Transmitter that no one yet knows how brain models scale with data.

Share:XLinkedInBlueskyEmail

More news