AI Accelerates Discoveries for Dementia Research

Currently, there are 57 million people worldwide living with dementia. For researchers dedicated to helping them, the main obstacle is not a lack of effort, but the enormous amount of data, the lack of experts to synthesize it, and the slow pace of discovery that does not keep up with the disease. The Consortium for Biomedical Research and AI in Neurodegeneration (C-BRAIN), in collaboration with NVIDIA, aims to improve this situation.

Alzheimer's disease is one of the greatest global health challenges, with over 10 million new cases each year. Despite a seven-figure increase in funding from the National Institutes of Health over the past 15 years, totaling $3.9 billion per year, no treatment has been found that can stop or reverse the disease. Bottlenecks such as insufficient key scientific findings, excessive data in institutional silos, and a slow pace of synthesis hinder progress.

C-BRAIN, led by Dr. Randall Bateman, includes 18 members actively engaged in academia, philanthropy, the pharmaceutical industry, and research organizations. Using AI, they aim to support scientists in achieving faster and better discoveries, with the goal of enabling new treatments. "Our goal is to fundamentally change how science is conducted by using AI to accelerate discoveries that benefit patients," said Bateman.

The consortium is working on a system that utilizes yet-to-be-available data in Alzheimer research, including unpublished findings and downloaded data from member institutions. C-BRAIN employs a hybrid centralized and federated approach to work with data without compromising patient privacy or violating institutional guidelines. Currently, 64 NVIDIA data center GPUs are running, combined with cloud nodes on Microsoft Azure, allowing them to scale their ambitious plans.

The consortium has already developed two tools: OpenScientist and InsightEngine. OpenScientist is an open-source AI co-scientist that generates hypotheses and conducts analyses. InsightEngine tests hypotheses against scientific literature to avoid pitfalls. Both tools have proven capable of performing analyses in minutes that would otherwise take weeks or months. This greatly increases the available research capacity, with the potential to free up over a million hours annually.

Looking ahead, C-BRAIN aims to build a multimodal biological foundational model that utilizes the full breadth of their data. This includes neuroimaging, multi-omics, and clinical data. This is expected to lead to new insights into the causes of brain aging and Alzheimer's disease, with the potential for implementing preventive and therapeutic measures.

Read the full article from Nvidia.