Inter-protocol reproducibility of multi-shell diffusion MRI metrics across major neuroimaging study protocols
Diffusion MRI is a core component of large-scale neuroimaging initiatives, yet the reproducibility of quantitative dMRI metrics across different acquisition protocols remains unclear. We acquired seven multi-shell diffusion protocols representative of major consortia and clinical settings on a single scanner in 41 healthy adults, isolating protocol effects from hardware and site variability. A linear mixed-effects model was used to estimate the inter-protocol coefficient of variation (CV) and the intraclass correlation coefficient (ICC) for diffusion tensor imaging (DTI), diffusion kurtosis imaging (DKI), and standard model imaging (SMI) metrics across atlas-based white matter regions. Rank agreement was assessed with Kendall's W, and repeated scans of one protocol provided a reference for test-retest variability. Overall, inter-protocol variability was low, with median CVs of 2.9-6.1% for DTI, 3.7-5.0% for DKI, and 3.0-4.7% for SMI metrics. Protocols ranked subjects consistently (W = 0.73-0.86) but agreed less in absolute value (ICC = 0.31-0.66), with fractional anisotropy (FA) being the most stable metric. Acquisition parameters, including TE, b-values, and signal-to-noise ratio jointly explained 53-93% of the inter-protocol variance. Differences in the b-values used for DTI and DKI estimation were particularly important for the diffusivities: matching b-values across protocols reduced the median inter-protocol CV for MD from 5.0% to 1.8%, with smaller reductions for mean kurtosis (MK), axial kurtosis (AK) and the axonal water fraction (f). These findings demonstrate good reproducibility of quantitative higher-order diffusion MRI metrics across commonly used multi-shell acquisition protocols, which can be further improved by aligning b-values, providing a foundation for data pooling, harmonization, and normative modeling in large-scale neuroimaging studies.
Authors
- Valentin N. Stepanov (ORCID: https://orcid.org/0000-0003-0007-8054)
- Benjamin Ades‐Aron (ORCID: https://orcid.org/0000-0001-5077-3341)
- Jamie Wren-Jarvis (ORCID: https://orcid.org/0000-0001-6918-8599)
- Ricardo Coronado‐Leija (ORCID: https://orcid.org/0000-0002-5846-2932)
- Dmitry S. Novikov (ORCID: https://orcid.org/0000-0002-4213-3050)
- Els Fieremans (ORCID: https://orcid.org/0000-0002-1384-8591)
- Santiago Coelho (ORCID: https://orcid.org/0000-0002-1039-4803)
- Jelle Veraart (ORCID: https://orcid.org/0000-0003-0781-0420)
- Subah Mehrin
- Nalini Jeet
- Jenny Chen (ORCID: https://orcid.org/0000-0002-8211-7346)
Institutions
- New York University (US)
Publication Details
- Journal
- bioRxiv (Cold Spring Harbor Laboratory)
- Published
- 2026-11-16
- DOI
- https://doi.org/10.64898/2026.11.13.766526
- Primary Topic
- Advanced Neuroimaging Techniques and Applications
- Type
- preprint
Funders
- Irma T. Hirschl Trust
- Center for Advanced Imaging Innovation and Research
- Centre d'Imagerie BioMédicale
- National Institutes of Health
- National Institute on Aging
- National Institute of Neurological Disorders and Stroke
- National Institute of Biomedical Imaging and Bioengineering