The Consortium for Biomedical Research and AI in Neurodegeneration is working with NVIDIA on artificial intelligence tools for Alzheimer's research, bringing together 18 member institutions.
The consortium, known as C-BRAIN, is led by Dr Randall Bateman. It was set up to help researchers work through large volumes of scientific and clinical data more quickly, including material that has often been difficult for both human researchers and AI systems to access. That includes unpublished findings, negative results and proprietary datasets held by member institutions.
Alzheimer's disease remains one of the world's biggest health challenges, with more than 10 million new dementia cases diagnosed each year and 57 million people living with dementia worldwide. Bateman said research is being slowed by fragmented data, a shortage of experts who can synthesise findings, and the pace at which new work is published.
The system under development is intended to let scientists use these data sources without moving sensitive information outside the institutions that control them. It combines centralised and federated methods so AI models can learn from local datasets while the underlying data stays in place.
This is one area where NVIDIA is involved. C-BRAIN currently runs 64 NVIDIA data centre graphics processing units on-premises alongside cloud-based nodes on Microsoft Azure, and expects to add more NVIDIA capacity, including Blackwell systems, as the work expands.
Two tools
Researchers in the consortium have already built two open-source tools, OpenScientist and InsightEngine. OpenScientist was developed with Lawrence Berkeley National Laboratory in an effort led by Research Scientist Justin Reese, while InsightEngine was built with 387 Labs.
OpenScientist is designed to take a research question and a dataset, generate hypotheses, run analyses and produce auditable results. InsightEngine tests hypotheses against published scientific literature, analyses user-supplied datasets, explores a biomedical knowledge graph and checks claims against evidence to reduce hallucinations.
Published case studies showed OpenScientist completed analyses in minutes that domain scientists estimated would otherwise take weeks or months. The cost was under USD $10 per run.
With the consortium now reaching more than 200 research groups, that level of automation could free up significant research time. C-BRAIN said this could exceed one million hours a year at scale.
Bateman described the work as a shift in how biomedical research can be carried out. "We have a window of opportunity with artificial intelligence to do research fundamentally different than we've ever done before, in a way that accelerates discoveries to help patients," said Dr Randall Bateman, Founder and Director of C-BRAIN and the Charles F. and Joanne Knight Distinguished Professor of Neurology at the Washington University School of Medicine in St. Louis.
He also put the project in simpler terms: "AI tools built by scientists, for scientists," Bateman said.
Dark data
A key part of the project is what the consortium calls dark data: research material that usually doesn't make its way into major AI models or broad scientific analysis. This can include negative results and unpublished work that often sit outside public databases and formal publication channels.
"A lot of life sciences data doesn't make its way into the frontier models," said Dr Ade Ojewole, Chief Technology Officer of C-BRAIN. "The consequence of that is we're leaving a lot of intelligence on the table."
Ojewole said handling those datasets is technically demanding, especially as C-BRAIN tries to combine secure data access with machine learning across institutions. "A great deal of optimization is going to be needed as we push the boundaries of what today's spatial omics machine learning architectures can model," he said. "NVIDIA software is definitely going to come into play there."
The collaboration is using NVIDIA BioNeMo and the MONAI framework in its research pipelines. The consortium is incorporating these tools as it builds infrastructure for data analysis across neuroimaging, multi-omics, fluid biomarkers and longitudinal clinical measures.
Bateman said the aim is not only to find patterns but also to improve understanding of disease mechanisms. "The C-BRAIN collaboration with NVIDIA aims to unlock those models in a way that they can deliver something quite impactful and important," he said.
The longer-term goal is a multimodal biological foundation model trained on the consortium's combined data resources. "AI that can think and reason has the potential to accelerate discoveries that will lead to powerful and implementable preventions and treatments for the entire population," Bateman said.