Skip to content
DraconiticResearch
Search Sources API
203,952 studies31 sourcesUpdated 10 min ago
Peer-reviewedCC BY-NC-ND 4.0Scientific Reports

On the accuracy of nonlinear image registration in portable low-field 3D brain MRI

  • J. Eugenio Iglesias1,2
  • Ian P. Johnson3
  • Jonathan Williams-Ramirez1
  • Dina Zemlyanker1
  • Lin Tian1
  • Karthik Gopinath1
  • Mark Olchanyi4,5,6
  • Ava D. Farnan7,8
  • Amelia Demopoulos7,8
  • Matthew S. Rosen1
  • Kevin N. Sheth3
  • Adam de Havenon3
2 more authors
  • W. Taylor Kimberly7,8
  • Annabel Sorby-Adams7,8
8 affiliations
  1. 1Martinos Center for Biomedical Imaging, Mass General Brigham and Harvard Medical School, Boston, MA 02114 USA
  2. 2Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139 USA
  3. 3Center for Brain and Mind Health, Departments of Neurology and Neurosurgery, Yale School of Medicine, New Haven, CT 06510 USA
  4. 4Picower Institute, Massachusetts Institute of Technology, Cambridge, MA 02139 USA
  5. 5Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, MA 02139 USA
  6. 6Center for Neurotechnology and Neurorecovery, Mass General Brigham and Harvard Medical School, Boston, MA 02114 USA
  7. 7Department of Neurology, Mass General Brigham and Harvard Medical School, Boston, MA 02114 USA
  8. 8Center for Genomic Medicine, Mass General Brigham and Harvard Medical School, Boston, MA 02114 USA
Read full textPDFDOICite

Abstract

Portable low-field MRI offers an affordable and mobile alternative to conventional high-field scanners, enabling imaging in point-of-care and resource-limited settings. However, its lower signal-to-noise ratio, reduced resolution, and acquisition artifacts raise concerns about the accuracy of standard image registration methods. Reliable nonlinear registration is critical for a wide range of emerging applications, including frequent brain monitoring, assessment of neurodegenerative disease progression, and evaluation of treatment effects such as those associated with Alzheimer’s therapeutics. In this work, we systematically evaluated state-of-the-art registration approaches on simulated low-field scans (obtained by downsampling high-field images) and on real low-field brain MRI data. We compared three approaches that are representative of current registration methods: classical optimization (NiftyReg), learning-based registration (SynthMorph), and synthesis-based registration (SynthSR+NiftyReg). Using downsampled high-field scans, all methods performed well, achieving high Dice scores and smooth deformation fields, indicating that reduced resolution alone does not hinder registration. In contrast, real low-field data exhibited lower accuracy, primarily due to geometric distortion and other acquisition-specific artifacts. Among the tested approaches, the synthesis-based pipeline achieved the most robust performance across subjects and modalities. Overall, existing algorithms can accommodate resolution limitations, however, future methods could further enhance coregistration by explicitly addressing the distortions present in low-field MRI scans.

Related

  1. Quantitative assessment of deformable image registration accuracy for cumulative dose estimation in multi-site stereotactic re-irradiation
    Ahamed Badusha Mohamed Yoosuf, Mohd Zahri Abdul Aziz, Salem Alshehri +1
    Radiation Oncology (London, England)Peer-reviewedCC BY-NC-ND 4.0Full textApplied physics & instrumentationPMC
  2. Distinguishing Strain Field Visualization Stability from Condition-Level Incremental Strain Accuracy in High-Temperature 2D Digital Image Correlation
    Yan Wang, Xiaoshan Ma, Xuan Zhang +3
    Sensors (Basel, Switzerland)Peer-reviewedCC BYFull textEurope PMC
  3. Research protocol for the diagnostic accuracy of intelligent portable impulse oscillometry across different respiratory diseases: a multicentre, cross-sectional study
    Bingxia Chen, Shubing Chen, Zhongping Wu +9
    BMJ OpenPeer-reviewedCC BY-NC 4.0Full textGeneral medicineCrossrefEurope PMC
  4. BundleMorphometry: Spatially Resolved White Matter Bundle Shape Analysis Using Nonlinear Streamline-based Registration
    D. Dasanayake, S. N. Amit, B. Q. Chandio
    bioRxivPreprintAll rights reservedFull textNeurosciencebioRxiv
  5. Optimal exposure time for satellite imaging? A trade-off between residual tip-tilt and registration accuracy
    Florian Cheyssial, Laurent M. Mugnier, Cyril Petit
    arXivPreprintarXiv licenceFull textInstrumentation & methodsOptics & photonicsarXiv

Cite

BibTeXRIS
BibTeX
@article{iglesias2026accuracy,
  title = {{On the accuracy of nonlinear image registration in portable low-field 3D brain MRI}},
  author = {J. Eugenio Iglesias and Ian P. Johnson and Jonathan Williams-Ramirez and Dina Zemlyanker and Lin Tian and Karthik Gopinath and Mark Olchanyi and Ava D. Farnan and Amelia Demopoulos and Matthew S. Rosen and Kevin N. Sheth and Adam de Havenon and W. Taylor Kimberly and Annabel Sorby-Adams},
  journal = {Scientific Reports},
  year = {2026},
  month = jul,
  doi = {10.1038/s41598-026-63646-1},
  url = {https://doi.org/10.1038/s41598-026-63646-1},
  note = {CC BY-NC-ND 4.0},
}
RIS
TY  - JOUR
TI  - On the accuracy of nonlinear image registration in portable low-field 3D brain MRI
AU  - J. Eugenio Iglesias
AU  - Ian P. Johnson
AU  - Jonathan Williams-Ramirez
AU  - Dina Zemlyanker
AU  - Lin Tian
AU  - Karthik Gopinath
AU  - Mark Olchanyi
AU  - Ava D. Farnan
AU  - Amelia Demopoulos
AU  - Matthew S. Rosen
AU  - Kevin N. Sheth
AU  - Adam de Havenon
AU  - W. Taylor Kimberly
AU  - Annabel Sorby-Adams
PY  - 2026
DA  - 2026/07/30
JO  - Scientific Reports
DO  - 10.1038/s41598-026-63646-1
UR  - https://doi.org/10.1038/s41598-026-63646-1
AB  - Portable low-field MRI offers an affordable and mobile alternative to conventional high-field scanners, enabling imaging in point-of-care and resource-limited settings. However, its lower signal-to-noise ratio, reduced resolution, and acquisition artifacts raise concerns about the accuracy of standard image registration methods. Reliable nonlinear registration is critical for a wide range of emerging applications, including frequent brain monitoring, assessment of neurodegenerative disease progression, and evaluation of treatment effects such as those associated with Alzheimer’s therapeutics. In this work, we systematically evaluated state-of-the-art registration approaches on simulated low-field scans (obtained by downsampling high-field images) and on real low-field brain MRI data. We compared three approaches that are representative of current registration methods: classical optimization (NiftyReg), learning-based registration (SynthMorph), and synthesis-based registration (SynthSR+NiftyReg). Using downsampled high-field scans, all methods performed well, achieving high Dice scores and smooth deformation fields, indicating that reduced resolution alone does not hinder registration. In contrast, real low-field data exhibited lower accuracy, primarily due to geometric distortion and other acquisition-specific artifacts. Among the tested approaches, the synthesis-based pipeline achieved the most robust performance across subjects and modalities. Overall, existing algorithms can accommodate resolution limitations, however, future methods could further enhance coregistration by explicitly addressing the distortions present in low-field MRI scans.
SN  - 2045-2322
ER  -