Accurate and Consistent 4D Segmentation of Serial Infant Brain MR Images Li Wang 1 , Feng Shi 1 , Pew-Thian Yap 1 , John H.Gilmore 2 , Weili Lin 3 , and Dinggang Shen 1 ,⋆ 1 IDEA Lab, Department of Radiology and BRIC dgshen@med.unc.edu 2 Department of Psychiatry, 3 MRI Lab, Department of Radiology and BRIC University of North Carolina at Chapel Hill, USA Abstract. Accurate and consistent segmentation of infant brain MR im- ages plays an important role in quantifying the early brain development, especially in longitudinal studies. However, due to rapid maturation and myelination of brain tissues in the first year of life, white-gray matter con- trast undergoes dramatic changes. In fact, the contrast inverses around 6 months of age, where the white and gray matter tissues are isointense and hence exhibit the lowest contrast, posing significant challenges for segmen- tation algorithms. In this paper, we propose a novel longitudinally guided level set method for segmentation of serial infant brain MR images, ac- quired from 2 weeks up to 1.5 years of age. The proposed method makes optimal use of T1, T2 and the diffusion weighted images for complimen- tary tissue distribution information to address the difficulty caused by the low contrast. A longitudinally consistent term, which constrains the dis- tance across the serial images within a biologically reasonable range, is employed to obtain temporally consistent segmentation results. The pro- posed method has been applied on 22 longitudinal infant subjects with promising results. 1 Introduction The first year of life is the most dynamic phase of postnatal brain development. The brain undergoes rapid tissue growth and experiences development of a wide range of cognitive and motor functions. For precise quantification of structural growth, algorithms dedicated to accurate tissue segmentation of infant brains in the first year of life is indispensable. Current methods are able to segment neonates (less than 3 months) and infants (over 1-year-old) with great success [1]. However, works dealing with the segmentation of serial infant images in the first years of life have been few. During this period of growth, the contrast be- tween white matter (WM) and gray matter (GM) reverses owing to maturation and myelination. There are three distinct WM/GM contrast patterns in images of developmentally normal infants (in chronological order) [2]: infantile (birth), ⋆ Corresponding author. T. Liu et al. (Eds.): MBIA 2011, LNCS 7012, pp. 93–101, 2011. c  Springer-Verlag Berlin Heidelberg 2011