Correcting for b-value spatial non-uniformity arising from B0-inhomogeneity in an ultra-low field system

Motivation: Diffusion MRI relies on applying spatially uniform b-values across the field-of-view. However gradient and static field non-uniformity cause the b-value to become spatially dependent, biasing susequent analysis if not corrected for. Goal(s): To develop a method to correct for static-field induced diffusion weighting, and demonstrate its impact on an ultra-low-field MR system. Approach: Asymmetric fast spin-echo fieldmapping was used to compute spatially dependent b-value maps and these were used to correct simulated and in-vivo data Results: b-value variation upwards of 5% are observed in typical neuroimaging fields-of view in a commercial ULF-MRI system Impact: This work demonstrates that diffusion estimates in inhomogeneous static fields with weak encoding gradients are likely to be significantly skewed by that inhomogeneity. We demonstrate a simple fieldmap correction that does not require downstream retooling and induces minimal noise enhancement.