TY - GEN
T1 - Susceptibility DistortionSusceptibility distortion Correction of Diffusion MRIDiffusion MRI with a single Phase-EncodingPhase Encoding Direction
AU - Dargahi, Sedigheh
AU - Bouix, Sylvain
AU - Desrosiers, Christian
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
PY - 2026
Y1 - 2026
N2 - Diffusion MRI (dMRI)DMRIDiffusion MRI is a valuable tool to map brain microstructure and connectivity by analyzing water molecule diffusion in tissue. However, acquiring dMRIDMRI data requires to capture multiple 3D brain volumes in a short time, often leading to trade-offs in image quality. One challenging artifact is susceptibility-induced distortionSusceptibility distortion, which introduces significant geometric and intensity deformations. Traditional correction methods, such as topupTopup, rely on having access to blip-upBlip-up and blip-downBlip-down image pairs, limiting their applicability to retrospective data acquired with a single phase encodingPhase Encoding direction. In this work, we propose a deep learning-basedDeep learning approach to correct susceptibility distortionsSusceptibility distortion using only a single acquisition (either blip-upBlip-up or blip-downBlip-down), eliminating the need for paired acquisitions. Experimental results show that our method achieves performance comparable to topupTopup, demonstrating its potential as an efficient and practical alternative for susceptibility distortion correction in dMRIDMRI.
AB - Diffusion MRI (dMRI)DMRIDiffusion MRI is a valuable tool to map brain microstructure and connectivity by analyzing water molecule diffusion in tissue. However, acquiring dMRIDMRI data requires to capture multiple 3D brain volumes in a short time, often leading to trade-offs in image quality. One challenging artifact is susceptibility-induced distortionSusceptibility distortion, which introduces significant geometric and intensity deformations. Traditional correction methods, such as topupTopup, rely on having access to blip-upBlip-up and blip-downBlip-down image pairs, limiting their applicability to retrospective data acquired with a single phase encodingPhase Encoding direction. In this work, we propose a deep learning-basedDeep learning approach to correct susceptibility distortionsSusceptibility distortion using only a single acquisition (either blip-upBlip-up or blip-downBlip-down), eliminating the need for paired acquisitions. Experimental results show that our method achieves performance comparable to topupTopup, demonstrating its potential as an efficient and practical alternative for susceptibility distortion correction in dMRIDMRI.
KW - Deep learningDeep learning
KW - Diffusion MRIDiffusion MRI
KW - Susceptibility distortionsSusceptibility distortion
UR - https://www.scopus.com/pages/publications/105027635587
U2 - 10.1007/978-3-032-12837-9_7
DO - 10.1007/978-3-032-12837-9_7
M3 - Contribution to conference proceedings
AN - SCOPUS:105027635587
SN - 9783032128362
T3 - Lecture Notes in Computer Science
SP - 69
EP - 80
BT - Computational Diffusion MRI - 16th International Workshop, CDMRI 2025, Held in Conjunction with MICCAI 2025, Proceedings
A2 - Chamberland, Maxime
A2 - Chen, Yuqian
A2 - Filipiak, Patryk
A2 - Hendriks, Tom
A2 - Lv, Jinglei
A2 - Shailja, S.
A2 - Thompson, Elinor
PB - Springer Science and Business Media Deutschland GmbH
T2 - 16th International Workshop on Computational Diffusion MRI, CDMRI 2025, held in conjunction with 28th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2025
Y2 - 27 September 2025 through 27 September 2025
ER -