AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI
This study compares volumetric measurements of various brain regions using different magnetic resonance imaging (MRI) modalities and deep learning models, specifically 3T MRI, ultra-low field (ULF) MRI at 64mT, and AI-enhanced ULF MRI using SynthSR and HiLoResGAN. The aim is to evaluate the alignmen...
Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-06-01
|
| Series: | Frontiers in Neuroimaging |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnimg.2025.1588487/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849694192931962880 |
|---|---|
| author | Kh Tohidul Islam Shenjun Zhong Parisa Zakavi Helen Kavnoudias Helen Kavnoudias Shawna Farquharson Gail Durbridge Markus Barth Markus Barth Andrew Dwyer Andrew Dwyer Katie L. McMahon Paul M. Parizel Paul M. Parizel Richard McIntyre Richard McIntyre Gary F. Egan Meng Law Meng Law Zhaolin Chen Zhaolin Chen |
| author_facet | Kh Tohidul Islam Shenjun Zhong Parisa Zakavi Helen Kavnoudias Helen Kavnoudias Shawna Farquharson Gail Durbridge Markus Barth Markus Barth Andrew Dwyer Andrew Dwyer Katie L. McMahon Paul M. Parizel Paul M. Parizel Richard McIntyre Richard McIntyre Gary F. Egan Meng Law Meng Law Zhaolin Chen Zhaolin Chen |
| author_sort | Kh Tohidul Islam |
| collection | DOAJ |
| description | This study compares volumetric measurements of various brain regions using different magnetic resonance imaging (MRI) modalities and deep learning models, specifically 3T MRI, ultra-low field (ULF) MRI at 64mT, and AI-enhanced ULF MRI using SynthSR and HiLoResGAN. The aim is to evaluate the alignment and agreement among field strengths and ULF MRI with and without AI. Descriptive statistics, paired t-tests, effect size analyses, and regression analyses are employed to assess the relationships and differences between modalities. The results indicate that volumetric measurements derived from 64mT MRI deviate significantly from those obtained using 3T MRI. By leveraging SynthSR and LoHiResGAN models, these deviations are reduced, bringing the volumetric estimates closer to those obtained from 3T MRI, which serves as the reference standard for brain volume quantification. These findings highlight that deep learning models can reduce systematic differences in brain volume measurements across field strengths, providing potential solutions to minimize bias in imaging studies. |
| format | Article |
| id | doaj-art-35b65dbbf84f4d31ab2e936f8ab23a26 |
| institution | DOAJ |
| issn | 2813-1193 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroimaging |
| spelling | doaj-art-35b65dbbf84f4d31ab2e936f8ab23a262025-08-20T03:20:10ZengFrontiers Media S.A.Frontiers in Neuroimaging2813-11932025-06-01410.3389/fnimg.2025.15884871588487AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRIKh Tohidul Islam0Shenjun Zhong1Parisa Zakavi2Helen Kavnoudias3Helen Kavnoudias4Shawna Farquharson5Gail Durbridge6Markus Barth7Markus Barth8Andrew Dwyer9Andrew Dwyer10Katie L. McMahon11Paul M. Parizel12Paul M. Parizel13Richard McIntyre14Richard McIntyre15Gary F. Egan16Meng Law17Meng Law18Zhaolin Chen19Zhaolin Chen20Monash Biomedical Imaging, Monash University, Clayton, VIC, AustraliaMonash Biomedical Imaging, Monash University, Clayton, VIC, AustraliaMonash Biomedical Imaging, Monash University, Clayton, VIC, AustraliaDepartment of Radiology, The Alfred, Melbourne, VIC, AustraliaDepartment of Surgery, School of Translational Medicine, Monash University, Clayton, VIC, AustraliaAustralian National Imaging Facility, Brisbane, QLD, AustraliaHerston Imaging Research Facility, University of Queensland, Brisbane, QLD, AustraliaSchool of Electrical Engineering and Computer Science, University of Queensland, Brisbane, QLD, AustraliaCentre for Advanced Imaging, Australian Institute for Bioengineering and Nanotechnology, University of Queensland, Brisbane, QLD, AustraliaSouth Australian Health and Medical Research Institute, Adelaide, SA, AustraliaSA Medical Imaging, SA Health, Adelaide, SA, Australia0School of Clinical Sciences, Faculty of Health, Queensland University of Technology, Brisbane, QLD, Australia1David Hartley Chair of Radiology, Royal Perth Hospital, Perth, WA, Australia2Medical School, University of Western Australia, Perth, WA, AustraliaMonash Biomedical Imaging, Monash University, Clayton, VIC, Australia3Monash Healthcare Network, Melbourne, VIC, AustraliaMonash Biomedical Imaging, Monash University, Clayton, VIC, AustraliaDepartment of Radiology, The Alfred, Melbourne, VIC, Australia4Department of Neuroscience, School of Translational Medicine, Monash University, Clayton, VIC, AustraliaMonash Biomedical Imaging, Monash University, Clayton, VIC, Australia5Department of Data Science and AI, Monash University, Clayton, VIC, AustraliaThis study compares volumetric measurements of various brain regions using different magnetic resonance imaging (MRI) modalities and deep learning models, specifically 3T MRI, ultra-low field (ULF) MRI at 64mT, and AI-enhanced ULF MRI using SynthSR and HiLoResGAN. The aim is to evaluate the alignment and agreement among field strengths and ULF MRI with and without AI. Descriptive statistics, paired t-tests, effect size analyses, and regression analyses are employed to assess the relationships and differences between modalities. The results indicate that volumetric measurements derived from 64mT MRI deviate significantly from those obtained using 3T MRI. By leveraging SynthSR and LoHiResGAN models, these deviations are reduced, bringing the volumetric estimates closer to those obtained from 3T MRI, which serves as the reference standard for brain volume quantification. These findings highlight that deep learning models can reduce systematic differences in brain volume measurements across field strengths, providing potential solutions to minimize bias in imaging studies.https://www.frontiersin.org/articles/10.3389/fnimg.2025.1588487/fullaccessible MRIultra-low-field MRIdeep learning in neuroimagingbrain volume measurementquantitative MRI analysis |
| spellingShingle | Kh Tohidul Islam Shenjun Zhong Parisa Zakavi Helen Kavnoudias Helen Kavnoudias Shawna Farquharson Gail Durbridge Markus Barth Markus Barth Andrew Dwyer Andrew Dwyer Katie L. McMahon Paul M. Parizel Paul M. Parizel Richard McIntyre Richard McIntyre Gary F. Egan Meng Law Meng Law Zhaolin Chen Zhaolin Chen AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI Frontiers in Neuroimaging accessible MRI ultra-low-field MRI deep learning in neuroimaging brain volume measurement quantitative MRI analysis |
| title | AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI |
| title_full | AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI |
| title_fullStr | AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI |
| title_full_unstemmed | AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI |
| title_short | AI improves consistency in regional brain volumes measured in ultra-low-field MRI and 3T MRI |
| title_sort | ai improves consistency in regional brain volumes measured in ultra low field mri and 3t mri |
| topic | accessible MRI ultra-low-field MRI deep learning in neuroimaging brain volume measurement quantitative MRI analysis |
| url | https://www.frontiersin.org/articles/10.3389/fnimg.2025.1588487/full |
| work_keys_str_mv | AT khtohidulislam aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT shenjunzhong aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT parisazakavi aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT helenkavnoudias aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT helenkavnoudias aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT shawnafarquharson aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT gaildurbridge aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT markusbarth aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT markusbarth aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT andrewdwyer aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT andrewdwyer aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT katielmcmahon aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT paulmparizel aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT paulmparizel aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT richardmcintyre aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT richardmcintyre aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT garyfegan aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT menglaw aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT menglaw aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT zhaolinchen aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri AT zhaolinchen aiimprovesconsistencyinregionalbrainvolumesmeasuredinultralowfieldmriand3tmri |