Multiband Diffusion Weighted Imaging for
Portable Low Field MRI Scanners

A DWI sequence using Patrick Le Roux's non-CPMG quadratic phase increment FSE readout that also builds on some of my previous work. It turns out the quadratic phase increment FSE is especially well suited for low field systems since it is spin-echo based, and the issues that quadratic phase increment FSE faces on superconducting systems (namely high SAR + desire for fast encoding) are not present in portable MRI.

This paper shows how to account for additional phase sources arising from system imperfections that are prominent on portable systems. I also improved the RF-encoding schemes used for self-navigated DWI which I think was a low-hanging fruit that had gone unnoticed.

Finally, there is a checklist that breaks the implementation into steps if someone attempts to reproduce this work. Programming sequences and acquiring data can be technically challenging since each sequence and system has its own pecularities. Hopefully this checklist helps with that.

I sincerely believe that if portable MRI takes off then this will be THE core approach for DWI. Compared to previous papers describing DWI acquisitions on portable MRI systems, I estimate multiband quadratic phase increment FSE is 5-10× more SNR efficient.

MATLAB code that implements the reconstruction and calibration steps, and some sample datasets are available here.

Lee PK, Qiu Y, Chen S, Wang C, Zhang Z. “Diffusion weighted Fast Spin Echo with RF-encoded slabs for portable low-field MRI systems,” Magnetic Resonance in Medicine. 2026. doi.org/10.1002/mrm.70292

Lee PK, Qiu Y, Wang C, Zhang Z. “Multiband Fast Spin Echo on portable low-field systems,” Magnetic Resonance in Medicine. 2025. doi.org/10.1002/mrm.70036

Lee PK, Hargreaves BA. “A joint linear reconstruction for multishot diffusion weighted non-Carr-Purcell-Meiboom-Gill fast spin echo with full signal,” Magnetic Resonance in Medicine. July 2022. doi.org/10.1002/mrm.29393