Q-space Conditioned Translation Networks for Directional Synthesis of Diffusion Weighted Images from Multi-modal Structural MRI

TL;DR

This paper introduces a novel generative adversarial network that synthesizes diffusion-weighted images conditioned on q-space information, enabling flexible sampling schemes and improved microstructural index estimation from sparse DWIs.

Contribution

It proposes a q-space conditioned translation network that removes fixed sampling constraints, allowing high-quality DWI synthesis from arbitrary structural images and sampling schemes.

Findings

Improved DWI synthesis accuracy over recent methods.

Enhanced downstream microstructural index estimation.

Supports arbitrary q-space sampling for flexible diffusion MRI analysis.

Abstract

Current deep learning approaches for diffusion MRI modeling circumvent the need for densely-sampled diffusion-weighted images (DWIs) by directly predicting microstructural indices from sparsely-sampled DWIs. However, they implicitly make unrealistic assumptions of static $q$-space sampling during training and reconstruction. Further, such approaches can restrict downstream usage of variably sampled DWIs for usages including the estimation of microstructural indices or tractography. We propose a generative adversarial translation framework for high-quality DWI synthesis with arbitrary $q$-space sampling given commonly acquired structural images (e.g., B0, T1, T2). Our translation network linearly modulates its internal representations conditioned on continuous $q$-space information, thus removing the need for fixed sampling schemes. Moreover, this approach enables downstream estimation of high-quality microstructural maps from arbitrarily subsampled DWIs, which may be particularly important in cases with sparsely sampled DWIs. Across several recent methodologies, the proposed approach yields improved DWI synthesis accuracy and fidelity with enhanced downstream utility as quantified by the accuracy of scalar microstructure indices estimated from the synthesized images. Code is available at https://github.com/mengweiren/q-space-conditioned-dwi-synthesis.