Developmental Score of the Infant Brain: Characterizing Diffusion MRI in Term- and Preterm-Born Infants

Wu Dan,Johns Hopkins University School of Medicine,Ernst Thomas M.,Caffo Brian S.,Oishi Kenichi
DOI: https://doi.org/10.1007/s00429-020-02132-4
2020-01-01
Abstract:Large-scale longitudinal neuroimaging studies of the infant brain allow us to map the spatiotemporal development of the brain in its early phase. While the postmenstrual age (PMA) is commonly used as a time index to analyze longitudinal MRI data, the nonlinear relationship between PMA and MRI data imposes challenges for downstream analyses. We propose a mathematical model that provides a Developmental Score (DevS) as a data-driven time index to characterize the brain development based on MRI features. 319 diffusion tensor imaging (DTI) datasets were collected from 87 term-born and 66 preterm-born infants at multiple visits, which were automatically segmented based on the JHU neonatal atlas. The mean diffusivity (MD) and fractional anisotropy (FA) in 126 brain parcels were used in the model to derive DevS. We demonstrate that transforming the time index from PMA to DevS improves the linearity of the longitudinal changes in MD and FA in both gray and white matter structures. More importantly, regional developmental differences in DTI metrics between preterm- and term-born infants were identified more clearly using DevS, e.g. 79 structures showed significantly different regression patterns in MD between preterm- and term-born infants, compared to only 27 structures that showed group differences using PMA as the index. Therefore, the DevS model facilitates linear analyses of DTI metrics in the infant brain, and provides a useful tool to characterize altered brain development due to preterm-birth.
What problem does this paper attempt to address?