Recent Publication: Recovering high-quality FODs from a reduced number of diffusion-weighted images using a model-driven deep learning architecture

MBCIU team members  Joseph J. Bartlett, Catherine E. Davey, and Leigh A. Johnston with Jinming Duan from University of Birmingham have recently published  work  in diffusion MRI.

Diffusion MRI faces the challenge of long scan times to acquire high-quality microstructural information. In this work we investigate the potential of model-based deep learning and a fixel classification loss term, to improve the quality of fibre orientation distributions (FODs) reconstructed from a reduced number of diffusion weighted images (DWIs). In our experiments we show that the quality of FODs fit to 30 DWIs using our deep learning method appear similar in many regions to those fit to 288 DWIs using traditional methods.

Read the full article here.

Bartlett JJ, Davey CE, Johnston LA, Duan J. Recovering high-quality fiber orientation distributions from a reduced number of diffusion-weighted images using a model-driven deep learning architecture. Magn Reson Med. 2024; 92: 2193-2206. doi: 10.1002/mrm.30187