2023 / Medical Imaging 2023: Image Processing / Conference paper
SynBOLD-DisCo: synthetic BOLD images for distortion correction of fMRI without additional calibration scans
Summary
This conference paper introduces SynBOLD-DisCo, using a 3D U-Net to synthesize undistorted BOLD images for fMRI susceptibility correction. It is the proceedings contribution preceding the expanded 2023 journal study.
Study overview
A synthetic BOLD-contrast target allows FSL topup to correct single-phase-encoding fMRI when field maps or reverse-encoding images are absent. Evaluation shows improved alignment with structural anatomy and performance close to correction using reverse phase-encoded data. This separately published SPIE contribution precedes the expanded journal article.
Main findings
- A 3D U-Net supplied an undistorted target for topup correction.
- The proceedings paper and subsequent journal article are distinct publications.

Cite this work
Tian Yu, Leon Y. Cai, Victoria L. Morgan, Sarah E. Goodale, Dario J. Englot, Catherine E. Chang, Bennett A. Landman, Kurt G. Schilling (2023). SynBOLD-DisCo: synthetic BOLD images for distortion correction of fMRI without additional calibration scans. Medical Imaging 2023: Image Processing. https://doi.org/10.1117/12.2653647
This page summarizes the work; consult the original article for the complete methods, results, and qualifications.
