UVA Awarded $2 Million NIH Grant to Advance Brain Research Using AI

The University of Virginia’s College and Graduate School of Arts & Sciences has received a $2 million grant from the National Institutes of Health’s Office of Research Infrastructure Programs (ORIP) S10 Instrumentation Grant Program to acquire a state-of-the-art high-performance computing system dedicated to neuroscience research. 

The award will fund the purchase of 36 NVIDIA Grace Hopper computing nodes, a new generation of GPU-based technology that will dramatically expand the university’s capacity for applying artificial intelligence, machine learning and advanced data analysis to neuroscience research. The project is a partnership among Arts & Sciences, the School of Data Science and UVA’s Research Computing group. 

“These grants are highly competitive and provide essential tools for advancing cutting-edge research,” said John Van Horn, professor of psychology and data science and principal investigator on the grant. “This will give us one — if not the — largest neuroscience-dedicated computing systems in the country.” 

The new system will be installed behind the secure firewall of UVA’s Ivy computing environment, which is designed to handle sensitive human-subject data. It will augment the existing Ivy infrastructure — particularly its NEO cluster — to create a dedicated, high-capacity platform for brain sciences. Van Horn said the system will be roughly 400 times faster than current capabilities and will bring secure computing for neuroscience up to par with UVA’s more general-purpose high-performance system, Rivanna. 

“This investment supports UVA’s strategic priorities in the 2030 plan, particularly our focus on research preeminence and interdisciplinary collaboration,” said Christa Acampora, dean of Arts & Sciences. “The combination of talent with the most advanced tools available is powerful — accelerating the pace of discovery in areas that matter to society, including understanding the human brain.” 

The Grace Hopper architecture integrates CPUs and GPUs on the same board, allowing them to communicate directly and automatically assign workloads for optimal performance. This capability will help researchers process increasingly large and complex datasets more efficiently, leading to faster insights into brain form, function and connectivity. 

Van Horn expects about 30 “power users” from across UVA to take advantage of the system in its first year, with many of them working in human neuroscience. The new resource will support projects across Arts & Sciences, the School of Data Science and other units, while also enabling new collaborations with researchers at institutions nationwide. 

“This system will be a game-changer for data-intensive research at UVA,” said Phil Bourne, dean of the School of Data Science. “By combining the latest GPU-based computing power with secure, high-speed access to complex datasets, we’re giving our faculty and students the tools to tackle some of the most pressing questions in neuroscience and beyond. This project exemplifies how collaboration across schools and disciplines can push the boundaries of discovery.” 

The grant also aligns with the college’s ongoing Neuroscience Grand Challenges initiative, which seeks to position UVA as a leader in multidisciplinary brain research. “In neurosciences, we don’t collect less data — we keep collecting more,” Van Horn said. “This dedicated resource means we can process and analyze that data faster than ever before.” 

The technology, which Joshua Baller, UVA’s associate vice president for research computing described as an “ARM-based system with a unique CPU-GPU interconnect” is not typical of the kind of computational hardware currently being used for research today.

“As a result there is great potential for this system to accelerate certain workflows beyond the gradual rate of improvement we have seen in more standard designs,” Baller said.

Once operational, the new computing cluster will be used for studies aimed at unlocking the mysteries of how the brain works in both health and disease. “We’ll be able to answer important questions faster, develop new methods and models, and make discoveries that could lead to better treatments for neurological disorders,” Van Horn said. 

The NIH requires that the system be purchased, installed and serving researchers within one year of the award, Van Horn said, which means the system should be up and running this fall. 

“This grant is not just about buying hardware; it’s about building capacity for discovery,” Van Horn said. “We’re putting UVA at the forefront of computational brain science, and that’s something our entire research community can be proud of.”