UNC Department of Radiology congratulates Assistant Professor of Radiology Gang Li, PhD, on his National Institute of Mental Health (NIMH) award funding a three-year project ($627K) entitled, “Harmonizing and Archiving of Large-scale Infant Neuroimaging Data.” Dr. Li’s study addresses the scarcity of widely available, large-scale infant MRI datasets that are comprehensively processed, harmonized, mined and archived within the National Institute of Mental Health Data Archive (NDA). This study aims to grow the number of centralized, large-scale infant MRI big datasets that boost statistical power and reproducibility in multi-site investigative studies of normative early brain development and neurodevelopmental disorders. 
Computational tools used to process many adult MR image datasets cannot easily be used to process and analyze challenging multi-site infant MR images that exhibit extremely low tissue contrast, large within-tissue intensity variations, and regionally heterogeneous dynamic changes. Dr. Li addresses these challenges through applying infant-dedicated, cortical surface-based computational tools and advanced machine learning techniques that his team developed and validated at UNC’s Biomedical Research Imaging Center Image Analysis Core Lab. This study seeks to fill a gap in early brain development through applying innovative tools and techniques to successfully process, harmonize and increase large-scale infant MRI datasets archived in the NDA that are accessible and available to benefit early brain development research.
The tools and techniques used for this study approach characterization of early brain development through producing quality-ensured, vertex-wise maps of multiple biologically distinct cortical properties in infant MR images (eg, cortical thickness, surface area, curvature, myelin). They preserve biological associations and remove image effects associated with varying contrasts and variations found with multi-site scanners and imaging protocols. They also leverage the informative growth patterns of the harmonized maps of multiple cortical properties to discover distinct regional partitions for infant brains.
Over three years (6/01/2021 – 5/31/2024), this study aims to advance computational characterization of early brain development through growing the 3,000+ publicly available, multi-site infant MRI datasets that have been acquired and released through the NDA. This study enables Dr. Li to further innovate and refine infant-dedicated, cortical surface-based computational tools and machine learning techniques within the Image Analysis Core Lab, toward a better understanding of normative early brain development and neurodevelopmental disorders.
Dr. Li noted: “This project builds upon the extensive experiences in infant neuroimaging analysis and unique image computation tools in the BRIC Image Analysis Core lab. The disseminated techniques and data to be developed in this project will make numerous MRI studies of early brain development more convenient and accurate.”