Scaling global tractography with ExaTract: from large cohorts to ultra-high-resolution datasets
Abstract
Global tractography offers a theoretical framework for reconstructing white matter connectivity from diffusion MRI by jointly inferring all streamlines and their consistency with the diffusion signal and underlying anatomical priors. However, its sequential nature and log-linear complexity make it computationally prohibitive at the scales demanded by neuroimaging, from large population cohorts such as the Human Connectome Project (HCP) to ex vivo ultra-high-field acquisitions at mesoscopic resolution. In this work, we present ExaTract, a parallelized spin-glass tractography framework designed for high-performance computing (HPC) architectures. Implemented as an open-source MPI+Kokkos application, ExaTract partitions the tractography domain into spatially overlapping regions processed across distributed memory nodes, with regular synchronization of boundary modifications to preserve global convergence guarantees. We validate ExaTract on both in vivo and ex vivo diffusion MRI data: the complete HCP dataset demonstrates population-scale throughput, while the high-resolution Chenonceau dataset provides the first mesoscopic structural connectome derived from global tractography.
Related
- Soliton resolution conjecture for the Calogero--Moser derivative nonlinear Schrödinger equation in the scaling-critical space
- Anatomical Connectivity from Tractography-free Diffusion Tensor Imaging and Its Application to EEG Brain Imagin
- Diagonal Multi-omics Integration of Heterogeneous Datasets
- Measuring Structure in Graph Benchmark Datasets Using Graph Invariants
- Criticality in Dissimilar Decomposition and Undersampling of Random Datasets with Anomalies
BibTeX
@misc{legeay2026scaling,
title = {{Scaling global tractography with ExaTract: from large cohorts to ultra-high-resolution datasets}},
author = {Simon Legeay and Felix Matuschke and Bastien Herlin and Jan André Reuter and Christophe Destrieux and Markus Axer and Ivy Uszynski and Cyril Poupon},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.03419},
eprint = {2611.03419},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.03419},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Scaling global tractography with ExaTract: from large cohorts to ultra-high-resolution datasets
AU - Simon Legeay
AU - Felix Matuschke
AU - Bastien Herlin
AU - Jan André Reuter
AU - Christophe Destrieux
AU - Markus Axer
AU - Ivy Uszynski
AU - Cyril Poupon
PY - 2026
DA - 2026/11/02
PB - arXiv
DO - 10.48550/arxiv.2611.03419
UR - https://arxiv.org/abs/2611.03419
AB - Global tractography offers a theoretical framework for reconstructing white matter connectivity from diffusion MRI by jointly inferring all streamlines and their consistency with the diffusion signal and underlying anatomical priors. However, its sequential nature and log-linear complexity make it computationally prohibitive at the scales demanded by neuroimaging, from large population cohorts such as the Human Connectome Project (HCP) to ex vivo ultra-high-field acquisitions at mesoscopic resolution. In this work, we present ExaTract, a parallelized spin-glass tractography framework designed for high-performance computing (HPC) architectures. Implemented as an open-source MPI+Kokkos application, ExaTract partitions the tractography domain into spatially overlapping regions processed across distributed memory nodes, with regular synchronization of boundary modifications to preserve global convergence guarantees. We validate ExaTract on both in vivo and ex vivo diffusion MRI data: the complete HCP dataset demonstrates population-scale throughput, while the high-resolution Chenonceau dataset provides the first mesoscopic structural connectome derived from global tractography.
ER -