SamRobNODDI: Q-Space Sampling-Augmented Continuous Representation Learning for Robust and Generalized NODDI

TL;DR

This paper proposes a q-space sampling augmentation-based continuous representation learning framework (SamRobNODDI) to achieve robust and generalized NODDI and introduces a sampling consistency loss to constrain the outputs of different sampling schemes, thereby further enhancing performance and robustness to varying q-space sampling schemes.

Abstract

Neurite Orientation Dispersion and Density Imaging (NODDI) microstructure estimation from diffusion magnetic resonance imaging (dMRI) is of great significance for the discovery and treatment of various neurological diseases. Current deep learning-based methods accelerate the speed of NODDI parameter estimation and improve the accuracy. However, most methods require the number and coordinates of gradient directions during testing and training to remain strictly consistent, significantly limiting the generalization and robustness of these models in NODDI parameter estimation. In this paper, we propose a q-space sampling augmentation-based continuous representation learning framework (SamRobNODDI) to achieve robust and generalized NODDI. Specifically, a continuous representation learning method based on q-space sampling augmentation is introduced to fully explore the information between different gradient directions in q-space. Furthermore, we design a sampling consistency loss to constrain the outputs of different sampling schemes, ensuring that the outputs remain as consistent as possible, thereby further enhancing performance and robustness to varying q-space sampling schemes. SamRobNODDI is also a flexible framework that can be applied to different backbone networks. To validate the effectiveness of the proposed method, we compared it with 7 state-of-the-art methods across 18 different q-space sampling schemes, demonstrating that the proposed SamRobNODDI has better performance, robustness, generalization, and flexibility.

Figures (8)

• Figure 1: Overview of SamRobNODDI. It contains both a training stage and a testing stage. During the training stage, the process consists of q-space sampling augmentation, continuous representation learning, and consistency loss. In the testing stage, only the continuous representation learning process is performed.
• Figure 2: Visualization results of NODDI parameter for SS testing using different methods with 30 diffusion directions per shell (1000, 2000 $s/{mm}^2$).
• Figure 3: Visualization results of NODDI parameter for RS testing using different methods with 30 diffusion directions per shell (1000, 2000 $s/{mm}^2$).
• Figure 4: Comparison of PSNR and SSIM for NODDI parameters estimated by SamRobNODDI and AMICO using different sampling rates.
• Figure 5: Visualization results of NODDI parameters estimated by SamRobNODDI using different sampling rates.