Self-Supervised Spatial And Zero-Shot Angular Super-Resolution by Spatial-Angular Implicit Representation For Rotating-View SNR-Efficient Diffusion MRI

TL;DR

This paper introduces a self-supervised neural representation that enables high-resolution diffusion MRI reconstruction from minimal views, significantly reducing scan time and allowing zero-shot synthesis of unseen directions.

Contribution

The proposed SA-INR model achieves accurate spatial and angular super-resolution in diffusion MRI from limited data, including unseen directions, breaking classical sampling limits.

Findings

High fidelity reconstruction for trained directions (34.82 dB)

Effective zero-shot synthesis of unseen directions (33.08 dB)

Improved downstream DTI model accuracy

Abstract

Rotating-view thick-slice acquisition is highly SNR-efficient for mesoscale diffusion MRI (dMRI) but requires numerous rotating views to satisfy Nyquist sampling, resulting in long scan time. We propose a self-supervised Spatial-Angular Implicit Neural Representation (SA-INR) that reconstructs high-resolution dMRI from a single view per diffusion direction, representing a massive acceleration. Our model, an MLP conditioned on a b=0 structural prior and the b-direction via FiLM, is trained end-to-end on the anisotropic input. The framework not only accurately reconstructs the trained b-directions (spatial SR) but also learns a continuous q-space representation, enabling high-fidelity "zero-shot" synthesis of unseen b-directions (angular SR). On simulated data, our method achieved high fidelity for both trained (34.82 dB) and unseen (33.08 dB) directions. Most importantly, the synthesized angular data also improved the quantitative accuracy of downstream DTI model fitting. Our SA-INR framework breaks the classical sampling limits, paving the way for fast, quantitative high-resolution dMRI.