Rapid Whole Brain Motion-robust Mesoscale In-vivo MR Imaging using Multi-scale Implicit Neural Representation

TL;DR

ROVER-MRI introduces a multi-scale implicit neural representation framework to enable rapid, motion-robust mesoscale whole-brain MRI from rotating-view thick-slice acquisitions. By integrating coordinate mapping, multi-scale hash grid encoding, and end-to-end INR-based reconstruction, it achieves isotropic resolutions such as $180~\mu\mathrm{m}$ at 7T in 17 minutes with improved SNR and sharpness. The method demonstrates strong performance across ex-vivo and in-vivo datasets, reducing reliance on high-resolution ground truth, and remains robust to fewer views and motion. These capabilities promise substantial practical impact for fast, high-fidelity neuroimaging in both research and clinical settings.

Abstract

High-resolution whole-brain in vivo MR imaging at mesoscale resolutions remains challenging due to long scan durations, motion artifacts, and limited signal-to-noise ratio (SNR). This study proposes Rotating-view super-resolution (ROVER)-MRI, an unsupervised framework based on multi-scale implicit neural representations (INR), enabling efficient recovery of fine anatomical details from multi-view thick-slice acquisitions. ROVER-MRI employs coordinate-based neural networks to implicitly and continuously encode image structures at multiple spatial scales, simultaneously modeling anatomical continuity and correcting inter-view motion through an integrated registration mechanism. Validation on ex-vivo monkey brain data and multiple in-vivo human datasets demonstrates substantially improved reconstruction performance compared to bicubic interpolation and state-of-the-art regularized least-squares super-resolution reconstruction (LS-SRR) with 2-fold reduction in scan time. Notably, ROVER-MRI achieves an unprecedented whole-brain in-vivo T2-weighted imaging at 180 micron isotropic resolution in only 17 minutes of scan time on a 7T scanner with 22.4% lower relative error compared to LS-SRR. We also demonstrate improved SNR using ROVER-MRI compared to a time-matched 3D GRE acquisition. Quantitative results on several datasets demonstrate better sharpness of the reconstructed images with ROVER-MRI for different super-resolution factors (5 to 11). These findings highlight ROVER-MRI's potential as a rapid, accurate, and motion-resilient mesoscale imaging solution, promising substantial advantages for neuroimaging studies.

Figures (12)

• Figure 1: Overview of the ROVER-MRI framework. (A) Data Acquisition: Eight low-resolution images are acquired with varying slice orientations to ensure comprehensive spatial coverage. (B) SR Reconstruction: A neural network-based implicit representation maps RAS spatial coordinates to pixel values, enabling high-quality super-resolution reconstruction. (C) Coordinate Mapping: Matrix coordinates from low-resolution images are preprocessed to extract precise image values for reconstruction. (D) ROVER-MRI: Multi-resolution hash encoding integrates spatial mapping, voxel hashing, feature retrieval, and auxiliary input concatenation to achieve enhanced reconstruction performance.
• Figure 2: Reconstruction results on simulated LR data using Bicubic interpolation, LS-SRR, and ROVER-MRI. The first and third rows show reconstructed sagittal MRIs, while the second and fourth rows present the corresponding error maps calculated against the GT. The red and green boxes highlight zoomed-in regions, which allow a closer inspection of the reconstruction quality.
• Figure 3: Reconstruction results of our method and LS-SRR using fewer views. (A) LS-SRR reconstruction results, showing increased error as the view count decreases. (B) ROVER-MRI reconstruction results, demonstrating high-quality reconstruction even with fewer views. Rows 1 and 4 display typical reconstruction results, while Rows 2 and 5 show enlarged views of the regions within the red boxes. Rows 3 and 6 show the error maps for these regions, highlighting the superior reconstruction accuracy of ROVER-MRI compared to LS-SRR.
• Figure 4: Reconstruction results of our method and LS-SRR with varying noise levels. (A) LS-SRR results under added noise, where reconstructions at SNR 30 show reduced error compared to SNR 15. (B) ROVER-MRI reconstructs cleaner images with significantly reduced errors compared to LS-SRR at both SNR levels. Rows 1 and 4 display typical reconstruction results, while Rows 2 and 5 show enlarged views of the regions within the red boxes. Rows 3 and 6 show the error maps for these regions, emphasizing the robustness of ROVER-MRI against noise.
• Figure 5: SRR results at 180 µm isotropic resolution. LS-SRR introduces streaking artifacts, which degrade the image quality. In contrast, our ROVER-MRI method demonstrates superior performance, preserving sharper and more continuous anatomical structures while achieving improved SNR.