Developmental brain age gap in prematurity and postnatally emerging delay in congenital heart disease

Overview

1. Center for MR-Research, University Children’s Hospital Zurich, Switzerland
2. Children’s Research Center, University Children’s Hospital Zurich, Switzerland
3. University of Zurich, Zurich, Switzerland
4. Department of Early Life Imaging, School of Biomedical Engineering and Imaging Sciences, King’s College London, London, UK
5. URPP Adaptive Brain Circuits in Development and Learning, University of Zurich, Switzerland
6. Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
7. Department of Neonatology, University Children’s Hospital Zurich, Switzerland
8. Pediatric Heart Center, Pediatric Cardiology, University Children’s Hospital Zurich, Switzerland
9. Department of Radiology, Shanghai Children’s Medical Center, Shanghai Jiaotong University School of Medicine, Shanghai, China

Abstract

Brain development follows a precisely regulated biological timetable, with defined periods of vulnerability increasingly recognized in congenital disorders affecting early brain development. This biological timing can be captured by the emerging concept of brain age, a measure of brain maturation, enabling the detection of deviation from normative developmental trajectories. Clinical conditions affect the degree of brain development during this critical period, including preterm birth and congenital heart disease (CHD).

We developed a deep learning-based brain age estimation framework across the fetal–neonatal period (21-44 gestational weeks) to quantify neurodevelopment from structural MRI. Using 1056 scans from six datasets acquired at three centers, Zurich, Shanghai, and the Developing Human Connectome Project, we trained models on normative fetal and neonatal MRI data. Both structural MRI-based and segmentation-derived cortical morphology-based models were implemented to assess representation effects and cross-center generalisability. The framework was applied to two clinically relevant conditions, preterm birth and CHD, to estimate the brain age gap (BAG), defined as the difference between predicted brain age and chronological age.

In preterm neonates scanned at term-equivalent age (n=90, 37-44 weeks), BAG was progressively more negative with lower gestational age at birth. Neonates born before 28 weeks showed delays of -0.7 to -0.8 weeks relative to term-born controls. In CHD (n=50, 22-34 weeks), fetal brain age did not differ from center-matched controls and no association with cardiac defect severity was observed. After birth, neonates with CHD (n=110, 37-44 weeks) showed significant (p<0.05) negative BAGs before surgery (-1.3 to -1.8 weeks) and BAGs increased significantly (p<0.05) after surgery (up to -3 weeks in center-specific analyses), indicating a delay in brain maturation from postnatal stage, but not in prenatal stage in CHD patients. These patterns were found across both structural MRI-based models and cortical morphology-based models, despite the need for cross-center calibration to minimize systematic bias. Voxel-based morphometry showed that a larger BAG was associated with regional contraction in deep frontal and peri-Rolandic white matter in preterm neonates, and perioperative spatial shifts in neonates with CHD. Saliency maps converged on deep white matter and periventricular regions, highlighting a potential link between BAG and delayed maturation of rapidly developing projection pathways.

These findings may indicate neurodevelopmental delays in preterm birth and a postnatally emerging maturational gap in CHD that increases following cardiac intervention. Despite limited generalisability of our methods, these results support a continuous fetal-neonatal brain age metric as a sensitive marker of global neurological maturational timing.

Reproduced under the paper's license (CC BY), 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

No dataset and no data link were found in the paper.

Data availability

The data that support the findings of this study are not publicly available due to ethical and data protection restrictions related to human subjects. Access may be granted upon reasonable request to the corresponding author, subject to institutional approval and the establishment of a data transfer and use agreement (DTUA). Data derived from the Developing Human Connectome Project (dHCP) are subject to the data access policies of that consortium and must be requested directly through their established application procedures.

Training and inference code, as well as the trained models necessary to reproduce the main analyses, are available in a public repository at: https://github.com/Mishakaan-dorp/Developmental-brain-age-estimation.

Reproduced under the paper's license (CC BY), 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 2, 28 September 2026

• Authors: added Kelly Payette (0000-0001-7041-0150); Anna Speckert (0009-0004-4728-6381); Hui Ji (0000-0002-9455-974X); Ruth Tuura (0000-0001-5932-7786); Su-Zhen Dong (0000-0002-9519-3057); Andras Jakab (0000-0001-6291-9889); removed Kelly Payette; Anna Speckert; Hui Ji; Ruth Tuura; Su-Zhen Dong; Andras Jakab
• Funding: added National Science Foundation; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung: 10003124; Universität Zürich; Vontobel-Stiftung

Version 1, 28 September 2026: the first record

Recorded: type, journal, dates, 14 authors, 70 references.

Cite

This paper

Kaandorp, M., Payette, K., Speckert, A., Steger, C., Ji, H., Ull, H. A., Tuura, R., Hagmann, C., Knirsch, W., Latal, B., Ren, J.-Y., Dong, S.-Z., Kim, H. G., & Jakab, A. (2026). Developmental brain age gap in prematurity and postnatally emerging delay in congenital heart disease. medRxiv (preprint). https://doi.org/10.64898/2026.04.01.26349523

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:10.3389/fnins.2026.1817732 [code]

Evaluating reliability of automated quantitative brain morphometry from fetal T2-weighted MRI.

Journal: Frontiers in neuroscience

In common: developmental, structural MRI / diffusion, 4 references

[2] doi:10.1016/j.dcn.2026.101775 [code]

Neonatal brain-age models in full- and preterm infants.

Journal: Developmental cognitive neuroscience

In common: developmental, structural MRI / diffusion, 4 references

[3] doi:10.1186/s41747-026-00702-5

MRI-based spatio-temporal atlas of ganglionic eminence.

Journal: European radiology experimental

In common: structural MRI / diffusion, 3 references

[4] 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, 3 references

[5] doi:10.1186/s40708-026-00316-y [code]

Generalizable and explainable deep learning for brain MRI: a multi-cohort evaluation of 3D architectures for age and sex prediction.

Journal: Brain informatics

In common: structural MRI / diffusion, 3 references

[6] doi:10.1016/j.ebiom.2026.106367 [code]

Linking maternal blood pressure with fetal cerebral haemodynamics and cortical growth in congenital heart disease.

Journal: EBioMedicine

In common: developmental, structural MRI / diffusion, 2 references

[7] doi:10.1038/s41398-026-04078-3 [code]

Brain age prediction in generalized anxiety disorder using a convolutional neural network.

Journal: Translational psychiatry

In common: structural MRI / diffusion, 3 references

[8] doi:10.1016/j.isci.2026.116072 [code]

Surface-based brain age index reflects hematological impacts on cognitive development in children with beta-thalassemia major.

Journal: iScience

In common: developmental, structural MRI / diffusion, 2 references

[9] doi:10.1162/imag.a.1293 [code]

PANDA: Patch-based unsupervised deep learning for brain anomaly detection via age prediction in fetal MRI.

Journal: Imaging neuroscience (Cambridge, Mass.)

In common: developmental, structural MRI / diffusion, 2 references

[10] doi:

Orientation-Aware Diffusion Super-Resolution for 3T-Like Fetal MRI from Routine 1.5T Scans.

Journal: Proceedings of machine learning research

In common: developmental, structural MRI / diffusion, 2 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.