Back

sf-pediatric: A robust and age-adaptable end-to-end pipeline for pediatric diffusion MRI

Gagnon, A.; Bore, A.; Valcourt Caron, A.; Edde, M.; Thoumyre, S.; Lepage, J.-F.; Talati, A.; Posner, J.; Ouellet, A.; Brunet, M. A.; Takser, L.; Rheault, F.; Descoteaux, M.

2026-01-20 neuroscience
10.64898/2026.01.19.700454 bioRxiv
Show abstract

Diffusion MRI (dMRI) provides a powerful, non-invasive window into white matter (WM) development. Yet, most existing processing pipelines are not well-suited to the rapidly evolving neurophysiology of the pediatric brain. Here, we introduce sf-pediatric, a scalable, end-to-end, age-adaptable dMRI pipeline that integrates normative models of brain diffusivities to enable optimal subject-specific analysis from birth through 18 years old. Leveraging normative trajectories derived from nearly 2,000 participants from six cohorts, sf-pediatric dynamically calibrates diffusion priors, template selection, segmentation, and WM atlases based on the subjects age. By incorporating automatic quality control into a portable, tested, containerized, open-access, and press-button framework across computing environments, sf-pediatric provides a robust pipeline for large-scale pediatric dMRI studies. We validated this approach by showing improved local modeling and cortical fanning while preserving reproducibility and the ability to derive brain-behavior relationships. Additionally, we demonstrated robust recovery of known developmental trajectories of WM microstructure and connectome-derived network organization.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.