Neonatal Brain Normalization Based on Flattening Using Spring Model

Abstract

Image registration (IR) has been used in brain function analysis, voxel-based-morphometry, and so on. The conventional IR methods in MR images mainly use MR signal based likelihood. However, they cannot prevent miss-registration to different gyri because they do not evaluate correspondence of sulci. Also, we cannot directly apply methods for adult brain to neonatal brain because there are large differences in MR signal and sulcal width. This paper proposes a non-rigid 3D IR method using flattening with sulcal-distribution index (SDI) which is calculated from MR signal around the cerebral surface. And, control points are spatially smoothed by using a spring model. The likelihood used is mutual information of SDI. The method will be effective for neonatal brain in which the accurate delineation of cerebral surface is difficult. Results in 7 neonates (modified age; 3 weeks-10 months) showed that an angle difference between a landmark was improved about 20% in comparison with the conventional method.