In this project we aim to develop AI methods to improve the image quality of ultra-low-field MRI.

The accessibility of MRI is limited by logistical and economic constraints. Approximately 70 percent of the world’s population currently has little or no access to this imaging modality. Where access exists, MRI may be unavailable or of limited use to patients who have contraindications such as metal implants or claustrophobia. The same limitations also constrain research investigations that rely on MRI scans.

Now, new developments in ultra-low-field (ULF) MRI technology hold the promise to improve the mobility and accessibility of MRI. This promise extents to facilitating longitudinal and population-representative normative aging studies that can help researchers develop more accurate aging biomarkers and a deeper understanding of aging-related health disparities. Relative to conventional MRI, ULF MRI is currently challenged by low signal-to-noise ratio, long scan times, and low special resolution—challenges that must be overcome for ULF MRI to realize its potential.

In this collaborative project, we aim to address some of the key limitations of ULF MRI by developing AI methods that improve the image quality. Better images from ultra-low-field systems stand to make such systems more efficient in both clinical radiology and neuroscientific research.