Abstract #2949

Generating a Human Neonatal Brain Atlas for Superior Normalization Accuracy

Yajing Zhang¹, Linda Chang², Thomas Ernst², Jon Skranes³, Steven Buchthal², Daniel Alicata², Heather Johansen², Antonette Hernandez², Robyn Yamakawa², Lillian Fujimoto⁴, Michael Miller¹, Susumu Mori⁵, Kenichi Oishi⁵

¹Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, United States; ²University of Hawaii at Manoa, Honolulu, HI, United States; ³Department of Laboratory Medicine, Children's & Women's Health, Norwegian Univ. of Science and Technology, Trondheim, Norway; ⁴Straub Mililani Clinic, Mililani, HI, United States; ⁵Department of Radiology, School of Medicine, Johns Hopkins University, Baltimore, MD, United States

MRI is a sensitive method for detecting subtle anatomical abnormalities in neonatal brains. The tissue composition of the neonatal brain is substantially different from that of the adult, and therefore, the use of a neonate-specific atlas might be more appropriate. To optimize the normalization, we introduce a method to create a Bayesian neonatal brain atlas to represent the studied population. Anatomical parcellation can be obtained automatically, avoiding the labor-intensive manual drawings of 3D ROIs. This tool is expected to be applicable for whole-brain detection of subtle developmental abnormalities, and for identifying MRI-based markers of neurological disorders in neonatal brain development.