The DIPY framework maps pathways in the brain
A team from the Luddy School of Informatics, Computing, and Engineering took home top honors in the Institute of Electrical and Electronics Engineers (IEEE) International Symposium of Biomedical Imaging MEMENTO Challenge, a scientific competition featuring more than 110 entries from around the world.
The MEMENTO Challenge focused on white matter microstructure using diffusion magnetic resonance imaging, a neuroimaging method used to study brain development. IU’s team, led by Shreyas Fadnavis, a Ph.D. student in the intelligent systems engineering program, used the free, open-source Diffusion Imaging in Python framework developed at Luddy to create a topologically separable inverse solver to super-resolve the MRI signal by leveraging the simplicial homology structure of the tissue microenvironment to gauge the intrinsic self-diffusivity.
“Currently, clinically used MRI has macroscale resolution of about one cubic millimeter,” said Eleftherios Garyfallidis, an assistant professor of intelligent systems engineering and an advisor with the team. “With this method, we are able to probe microstructural tissue parameters by sensing in microstructure. You can think of this as an extreme version of compressed sensing. The method will be used clinically for creating biomarkers.”
The solution can also be applied to other domains and problems, including dynamical systems, Hammerstein models, vibrations, control systems, and more. The competition motivated the team to develop an optimization for DIPY that is generic and works quickly.
“I’m really proud to see our team performing so well in international challenges,” Garyfallidis said. “It’s also great to see DIPY growing to more areas and solving harder and harder problems. We’re looking forward to the next challenge.”
The IEEE ISBI MEMENTO Challenge was run over two years, with the first year dedicated to ISBI designing the challenge, building appropriate datasets, and presenting it to competitors. The second year allowed researchers to develop, implement, and revise their submissions and algorithms.
“It’s exciting when our dedicated students showcase their skills on an international stage and gain recognition for their innovation,” said Martin Swany, chair of the ISE program at Luddy. “The DIPY software is known for its versatility, and the effort of this team really shows how useful it can be.”