Myelin water imaging from an 8 minute scan at 64mT via image reconstruction and multicomponent T2 fitting in a low-rank space

Motivation: Ultra-low field (ULF) MRI scanners are proliferating and their use to measure brain health through myelin content would benefit clinical applications and research. Goal(s): To leverage low-rank algorithms to achieve quantitative measurements of brain myelin content at ULF. Approach: Following low-rank reconstruction of 3D multi-spin-echo data, sets of 20 characteristic T2 decay curves were projected into the same low-rank space. A regularized non-negative least squares fit of these low-rank T2 vectors provided a T2 distribution; T2<40ms were assigned as myelin water. Results: Maps of short T2 components reflected historical myelin water fraction values and corresponded to a myelin water fraction atlas. Impact: Reconstruction and multi-component fitting in a low-rank space allows myelin water fraction mapping at 64mT, permitting quantitative ultra low-field measurements of brain development in infants and children and demyelination during the course of diseases like multiple sclerosis.