Meeting Banner
Abstract #4807

Zero-Shot Self-Supervised Distortion-Free Diffusion MRI Reconstruction

Mehmet Yigit Avci1, Jaejin Cho2, Yohan Jun3,4, and Berkin Bilgic3,4
1Department of Biomedical Engineering and Imaging Sciences, King's College London, London, United Kingdom, 2Department of Artificial Intelligence and Robotics, Sejong University, Seoul, Korea, Republic of, 3Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, MA, United States, 4Department of Radiology, Harvard Medical School, Boston, MA, United States

Synopsis

Keywords: Diffusion Reconstruction, AI/ML Image Reconstruction, Self-supervised learning, distortion correction

Motivation: Multi-shot Echo Planar Imaging (msEPI) for diffusion MRI captures detailed anatomy but suffers from phase inconsistencies and susceptibility-induced distortions.

Goal(s): To develop a zero-shot self-supervised reconstruction method that eliminates msEPI distortions using undersampled k-space data from single subject, without external datasets.

Approach: We present ZS-PRIME, the first zero-shot, high-fidelity framework for distortion-free msEPI reconstruction. ZS-PRIME leverages PRIME, a distortion-free multi-echo acquisition where the second echo, at lower resolution and acceleration, provides high-fidelity field maps. CNN-based k-space and image-space regularization ensure phase consistency and anatomical accuracy.

Results: ZS-PRIME outperforms existing methods, including PRIME with LORAKS regularization, delivering high-quality, distortion-free reconstructions, enhancing diffusion MRI fidelity.

Impact: ZS-PRIME combines advanced field map estimation (PRIME) with zero-shot self-supervised training, achieving distortion-free, high-resolution multi-shot diffusion MRI from undersampled data. This obviates the dependency on external training datasets, setting a new benchmark for efficient, high-fidelity diffusion MRI.

How to access this content:

For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.

After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.

After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.

Click here for more information on becoming a member.

Keywords