Multi-Site, In Vivo MRI Dataset of Brain Diffusivity Measures Before and after Harmonization, and Atrophy Measures Following Controlled Cortical Impact in Male and Female Adult Rats

Gregory Kislik¹Rachel Fox¹Alexandru Korotcov²Jinyuan Zhou³Marcelo Febo⁴Babak Moghadas³Adnan Bibic³Yunfan Zou³Jieru Wan³Ray Koehler³Temitope Adebayo²Mark P Burns⁵Joseph T McCabe²Kevin K W Wang⁴J. Russell Huie⁶Adam R Ferguson⁶Afshin Paydar¹Ina Beate Wanner¹Neil G. Harris^1*

• ¹Neurosurgery, University of California, Los Angeles, Los Angeles, United States
• ²Uniformed Services University of the Health Sciences, Bethesda, United States
• ³Johns Hopkins University, Baltimore, United States
• ⁴University of Florida, Gainesville, United States
• ⁵Georgetown University, Washington, United States
• ⁶University of California San Francisco, San Francisco, United States

Traumatic brain injury research faces persistent challenges in data comparability and reproducibility. Structured, interoperable datasets are essential to identify robust imaging biomarkers and validate cross-site findings. This dataset comprises 343 diffusion-weighted MRI scans from 186 male and female Sprague Dawley rats subjected to controlled cortical impact (CCI) or sham procedures at four research sites. Imaging was performed at 3 and 30 days post-injury using harmonized acquisition protocols and field strengths ranging from 7-11.7T. A standardized processing pipeline generated scalar maps (FA,MD,AD,RD), and z-score–based indices of injury. Harmonization was conducted using NeuroCombat on univariate data and at a voxel-based level. The dataset provides a foundation for testing image harmonization methods, developing quantification tools for automated assessment of injury, and training machine learning models. The dataset is published in the FAIR² framework, with machine-actionable metadata, responsible AI indicators, and structured documentation to support ethical, reproducible, and AI-ready reuse.