From Low Field to High Value: Robust Cortical Mapping From Low-Field MRI.

Overview

1. Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
2. University of Washington, Seattle, Washington, USA
3. Washington University, St. Louis, Missouri, USA
4. Danish Research Centre for Magnetic Resonance, Copenhagen University Hospital, Hvidovre, Denmark
5. Massachusetts Institute of Technology, Cambridge, Massachusetts, USA
6. University College London, London, UK

Abstract

Three‐dimensional reconstruction of cortical surfaces from MRI for subsequent morphometric analysis is fundamental for understanding brain structure. While high‐field Magnetic Resonance Imaging (HF‐MRI) is the standard in research and clinical settings, its relatively limited availability hinders widespread use. Low‐field MRI (LF‐MRI), particularly portable systems, offers a cost‐effective and accessible alternative. However, existing cortical surface analysis tools, such as FreeSurfer, are optimized for high‐resolution HF‐MRI and struggle with the lower signal‐to‐noise ratio (SNR) and resolution of LF‐MRI. In this work, we present a machine learning method for 3D reconstruction and analysis of portable LF‐MRI scans over a range of contrasts and resolutions. Our method works “out of the box” and does not require retraining. It leverages a 3D U‐Net trained on synthetic LF‐MRI data to predict signed distance functions of the cortical surfaces, followed by geometric processing to ensure topologically accurate reconstructions. We evaluate our approach using paired HF‐/LF‐MRI scans of the same 15 subjects and 50 subjects from the ULF‐EnC dataset. The results show that our method robustly recovers surfaces across LF‐MRI acquisitions, with accuracy depending on MRI contrast mechanism (T1 vs. T2), slice anisotropy (axial vs. isotropic), and resolution. A 3 mm isotropic T2‐weighted scan acquired in under 4 min, which is comparable in duration to typical HF‐MRI acquisitions, yields strong agreement with HF‐derived surfaces: surface area correlates at r=0.96, cortical parcellations reach a Dice coefficient of 0.98, and gray matter volume achieves r=0.93. Cortical thickness remains more challenging but achieves correlations up to r=0.70, reflecting the difficulties of achieving sub‐mm precision with ~3 × 3 × 3 mm voxels. Our results also show that recon‐any performs robustly across other sequences and contrasts, though thickness estimates are particularly sensitive and degrade substantially with anisotropic or low‐resolution scans. We also validate our method on challenging postmortem LF‐MRI scans, further illustrating its robustness. Our method represents a significant step toward making cortical surface analysis feasible for portable LF‐MRI systems. The tool is publicly available at https://surfer.nmr.mgh.harvard.edu/fswiki/ReconAny.

Reproduced under the paper's license (CC BY-NC), from the paper cited above.

Code

No file of the authors' code could be read here: it is described below, and read at its source.

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

• 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
• 0 scripts, each with its path and the digest of its content;
• no match between paragraphs and code yet;
• neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

Datasets cited

• doi:10.7303/syn65485242, at the source; found in the references

Data Availability Statement

Data used in preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp‐content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf (http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf).

Reproduced under the paper's license (CC BY-NC), from the paper cited above.

Data Availability Statement

The datasets used in this study for training our recon‐any model include publicly available high‐field MRI scans from the Human Connectome Project (HCP) (Glasser et al. 2013) and the Alzheimer's Disease Neuroimaging Initiative (ADNI) Jack Jr. et al. (2008) datasets. For portable LF‐MRI analysis, the paired high‐field (3T) and low‐field (64 mT, Hyperfine) MRI scans used for evaluation in this work were collected under an institutional research protocol and are not publicly available due to participant privacy restrictions and ongoing regulatory constraints. Access to the HCP and ADNI datasets requires approval from their respective data use agreements. Individual patient data can be accessed by academic researchers under restricted conditions, requiring an institutional data use agreement.

The recon‐any pipeline is integrated into FreeSurfer and is freely available for research use. The source code, trained models, and documentation are provided at https://surfer.nmr.mgh.harvard.edu/fswiki/ReconAny. After sourcing FreeSurfer software, recon‐any can be executed using the following command:

where the command‐line options are defined as follows:

INPUT_SCAN: Path to the MRI image to be processed.

SUBJECT_ID: Identifier for the subject where a corresponding output directory is created.

SIDE: Specifies the hemisphere(s) to process: ○ left‐c: Left cerebrum (postmortem single hemisphere). ○ left‐ccb: Left cerebrum, cerebellum, and brainstem (if intact). ○ right‐c: Right cerebrum. ○ right‐ccb: Right cerebrum, cerebellum, and brainstem. ○ both: Both hemispheres (in vivo or full postmortem brains).

THREADS (optional): Number of CPU threads (default: 1); higher values speed up processing.

SUBJECT_DIR (optional): Alternative output directory, required if SUBJECTS_DIR is not set or needs overriding.

Data used in preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp‐content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf (http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf).

Reproduced under the paper's license (CC BY-NC), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 16 authors, 7 keywords, 9 MeSH terms, 59 funders, 26 references.

Cite

This paper

Gopinath, K., Sorby‐Adams, A., Williams‐Ramirez, J., Zemlyanker, D., Guo, J., Hunt, D., Mac Donald, C. L., Keene, C. D., Coalson, T., Glasser, M. F., Van Essen, D., Rosen, M. S., Puonti, O., Kimberly, W. T., Iglesias, J. E., & Alzheimer's Disease Neuroimaging Initiative. (2026). From Low Field to High Value: Robust Cortical Mapping From Low-Field MRI. Human brain mapping, 47(7), e70515. https://doi.org/10.1002/hbm.70515

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1038/s41598-026-55397-w [code]

Fast surface reconstruction of human brain MRI: benchmarking deep-learning based morphometry tools.

Journal: Scientific reports

In common: structural MRI / diffusion, 7 references

[2] doi:10.1002/mrm.70488 [code]

Advancing the Volumetric Analysis of Ultra-Low-Field Brain MRI Using Image-to-Image Translation.

Journal: Magnetic resonance in medicine

In common: structural MRI / diffusion, 5 references

[3] doi:10.3389/fncir.2026.1814667 [code]

A modular and flexible pipeline for intraoperative electrode reconstruction and localization in patients with brain lesions.

Journal: Frontiers in neural circuits

In common: structural MRI / diffusion, 5 references

[4] doi:10.1162/imag.a.1183 [code]

Learning-based segmentation of diffusion-weighted MR images with arbitrary <i>q</i>-space samplings.

Journal: Imaging neuroscience (Cambridge, Mass.)

In common: structural MRI / diffusion, 5 references

[5] doi:10.1002/nbm.70353 [code]

Automated Surface-Based Segmentation of Deep Gray Matter Regions Based on Diffusion Tensor Images Reveals Unique Age Trajectories Over the Healthy Lifespan.

Journal: NMR in biomedicine

In common: structural MRI / diffusion, 5 references

[6] doi:10.1002/alz.71649 [code]

Postmortem brain MRI reveals differential associations of subcortical and limbic volumes with cortical thinning and neurodegenerative pathologies.

Journal: Alzheimer's & dementia : the journal of the Alzheimer's Association

In common: histology / microscopy, structural MRI / diffusion, 4 references

[7] doi:10.1002/advs.76596 [code]

DDSurfer: A Weakly-Supervised Dual-Stream Deep Learning Framework for Cortical Surface Reconstruction From Diffusion MRI.

Journal: Advanced science (Weinheim, Baden-Wurttemberg, Germany)

In common: structural MRI / diffusion, 5 references

[8] doi:10.1162/imag.a.1352 [code]

Brain-age in ultra-low-field MRI: How well does it work?

Journal: Imaging neuroscience (Cambridge, Mass.)

In common: structural MRI / diffusion, 4 references

[9] doi:10.1162/nol.a.265 [code]

Structural Brain Correlates of Poor Reading Comprehension.

Journal: Neurobiology of language (Cambridge, Mass.)

In common: structural MRI / diffusion, 4 references

[10] doi:10.3389/frai.2026.1771088 [code]

Few-shot deployment of pretrained MRI transformers in brain imaging tasks.

Journal: Frontiers in artificial intelligence

In common: structural MRI / diffusion, 4 references

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.