Skip to main content

$2.3M NIH grant to advance AI-powered brain-imaging research tools

For Immediate Release Oct 31, 2023

BLOOMINGTON, Ind. — A $2.3 million grant from the National Institutes of Health will advance the development of AI-powered software tools to support breakthrough research on the brain, as well as help health care professionals who perform critical procedures, such as neurosurgery.

Eleftherios Garyfallidis. Photo by James Brosher, Indiana University Eleftherios Garyfallidis. Photo by James Brosher, Indiana University

The award, part of the prestigious NIH BRAIN 2.0 Initiative, will advance the development of community-supported open-source software for computational neuroanatomy led by principal investigator Eleftherios Garyfallidis, associate professor at the Indiana University Luddy School of Computing, Informatics and Engineering in Bloomington.

“This is an incredible opportunity to showcase some of the AI tools we have built at IU to support medical research worldwide,” Garyfallidis said.

Understanding brain networks and their relation to neural computation is crucial to developing medical breakthroughs, he added. The NIH BRAIN Initiative identifies new opportunities to apply emerging tools to revolutionize the understanding of brain circuits and designate valuable areas of continued technological development.

Networks that connect distinct brain regions are composed of large bundles known as tracts that contain the axons of millions of neurons. These brain networks are tied into neurological, cognitive and psychiatric health. Through Garyfallidis’ DIPY software project, an open-source platform collecting state-of-the-art medical imaging algorithms at the international level, researchers can study how these different brain regions are connected and function.

Garyfallidis is a pioneer in tractometry, which uses diffusion MRI data to create 3D models of the brain. The work supported under the award will apply computational methods and software tools using artificial intelligence to understand medical imaging data.

“A patient takes medicine, but how does a doctor know if it’s working?” he said. “How do we know if the targeted parts of the brain are getting better? This project aims to answer those questions.”

Garyfallidis and team created a series of state-of-the-art algorithms such as QuickBundles, the SLR, RecoBundles, BundleWarp and Bundle Analytics. These are freely available in the DIPY project.

The next phase of the project builds upon this work to further improve these tractometry tools, increase their performance and deploy end-to-end solutions. The project aims to enable the next generation of researchers to study the structure of the brain’s white matter, both in patients with neurological diseases as well as in healthy populations.

In addition, the project seeks to validate these methods and tools using human and non-human primate data.

Garyfallidis is also a leader in scientific open-source software. He is a core member of the Neuroimaging in Python team that revolutionized and democratized the way neuroscience research is performed.

“Open-source scientific software needs to be in the loop with medical practice,” he said. “Medical doctors want to learn how AI methods work; they want to understand. We need to stop pushing black-box solutions to the hospitals.”

The Garyfallidis Research Group will host the annual DIPY Workshop from March 11 to 15 online. The event provides instruction and skills needed to master the latest techniques and tools in structural and diffusion imaging. The event is available for researchers and clinicians of all levels.

Media Contact

Luddy School of Informatics, Computing and Engineering

Pete DiPrimio

More stories