Baby Open Brains: An open-source dataset of infant brain segmentations

Baby Open Brains: An open-source dataset of infant brain segmentations

Abstract Reproducibility of neuroimaging research on infant brain development remains limited due to highly variable processing approaches. Progress towards reproducible pipelines is limited by a lack of benchmarks such as gold-standard brain segmentations. These segmentations are limited by the dif...

Full description

Saved in:
Bibliographic Details
Main Authors: Eric Feczko, Sally M. Stoyell, Lucille A. Moore, Dimitrios Alexopoulos, Maria Bagonis, Kenneth Barrett, Brad Bower, Addison Cavender, Taylor A. Chamberlain, Greg Conan, Trevor K. M. Day, Dhruman Goradia, Alice Graham, Lucas Heisler-Roman, Timothy J. Hendrickson, Audrey Houghton, Omid Kardan, Elizabeth A. Kiffmeyer, Erik G. Lee, Jacob T. Lundquist, Carina Lucena, Tabitha Martin, Anurima Mummaneni, Mollie Myricks, Pranav Narnur, Anders J. Perrone, Paul Reiners, Amanda R. Rueter, Hteemoo Saw, Martin Styner, Sooyeon Sung, Barry Tiklasky, Jessica L. Wisnowski, Essa Yacoub, Brett Zimmermann, Christopher D. Smyser, Monica D. Rosenberg, Damien A. Fair, Jed T. Elison
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Scientific Data
Online Access:https://doi.org/10.1038/s41597-025-05404-y
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Reproducibility of neuroimaging research on infant brain development remains limited due to highly variable processing approaches. Progress towards reproducible pipelines is limited by a lack of benchmarks such as gold-standard brain segmentations. These segmentations are limited by the difficulty of infant brain segmentations, which require extensive neuroanatomical knowledge and are time-consuming in nature. Addressing this, we constructed the Baby Open Brains (BOBs) Dataset, an open source resource of manually curated and expert reviewed infant brain segmentations. Anatomical MRI data was segmented from 71 infant imaging visits across 51 participants, using both T1w and T2w images per visit. Images showed dramatic differences in myelination and intensities across 1–9 months, emphasizing the need for densely sampled gold-standard segmentations across early life. This dataset provides a benchmark for evaluating and improving pipelines dependent upon segmentations in the youngest populations. As such, this dataset provides a vitally needed foundation for early-life large-scale studies such as HBCD.
ISSN:2052-4463

Similar Items