Brain MR Image Tissue Segmentation And Cortical Surface Analysis During Early Development

Brain magnetic resonance(MR)images are one kind of brain images obtained by using tomographic techniques.Accurate tissue segmentation could separate the gray matter,white matter and cerebrospinal fluid from the brain MR image,and provide foundation to further cortical surface reconstruction,which is vital to brain MR image processing.The cortical surface is a highly convoluted gray matter structure surrounding the white matter,which can be reconstructed from the gray matter in tissue segmentation results.Cortical surface is covered by sulcus and gyrus.These sulcus and gyrus appear during fetal stage and develop rapidly and asymmetrically during early postnatal years after birth.The structure of cortical surface was highly related with the establish and formation of functions during early development stage.Moreover,increasing studies also suggest that the abnormal cortical surface area expansion during infancy is related to many neurodevelopmental disorders.Therefore,the analysis of cortical surface structure during early development could provide important references for the normal development studies.Studies from brain MR image segmentation to early cortical surface structure analysis are longitudinal brain research from basic level to high level.During these years,researchers have put a lot of efforts into improving the accuracy of tissue segmentation,as well as establishing more real cortical surfaces,to aid the analysis of early cortical surface structure and fill the gaps in this field,such as orientation-expansion,parcellation and hemispheric asymmetries.This paper focuses on two aspects-tissue segmentation of brain MR image and analysis of early cortical surface structure.In the brain MRI,due to the partial volume effect and the intensity inhomogeneity,it is difficult to assign the affected pixels to one single tissue.Considering that the fuzzy clustering algorithm allows one pixel belong to multiple clusters,as well as the noise resistance can be adjusted by setting the smoothing term,it is suitable for the brain MR image segmentation.However,because they do not enhance the clarity of the original image,and lack the using of local boundary details,the tissue segmentation results still need to be improved.This paper proposed a segmentation algorithm mainly based on fuzzy clustering method to improve the segmentation accuracy.Based on tissue segmentation,this paper uses popular early cortical surface process pipeline to generate cortical surfaces,and analyzes the early cortical surface structure from 3 aspects,such as orientation-expansion,parcellation and hemispheric asymmetries.Specifically,1)Cortical surface expansion can be divided into the lengthening and widening of cortical sulcus during first postnatal year after birth.However,due to lack of dedicated computational methods,our knowledge on lengthening and widening of cortical sulcus during infancy remains scarce.2)Next,for the generation of early brain cortical surfaces atlas,the available cortical parcellations created for adults are inapplicable for fetuses,due to dramatic differences in brain appearances and dynamic growth of fetal brains.3)In primates,macaque monkeys are with close phylogenetic relationship to humans.Studying cortical hemispheric asymmetries during the dynamic early postnatal stages in macaque monkeys would increase our limited understanding on the possible origins,developmental trajectories,and evolutional mechanisms of brain asymmetries in non-human primates.However,due to the small size of the cerebral cortex of early macaques,the hemispheric asymmetries in macaque cortical surfaces during early development remains a blind spot to the community.In view of the above problems,this paper focuses on the following research work based on brain MR images and early cortical surfaces:(1)To smooth noise meanwhile improve image clarity by utilizing more local information,we propose a novel algorithm for brain MR image segmentation,combining super resolution,spatial constraint based on clustering and fine-tuning.With the introduction of these two parameters,the robustness of the clustering is enhanced,and the trade-off between noise smoothing and detail preservation can be controlled more accurately.Furthermore,the local regions around boundaries of different brain tissues are re-segmented in a fine-tuning process,and a weighted voting strategy is adopted for adjusting the incorrect pixels,which makes full use of the boundary details between different tissues(2)The knowledge on the oriented cortical surface expansions would help us better understand the cortical growth mechanisms and provide important insights into neurodevelopmental disorders,but still remains scarce,due to the lack of dedicated computational methods.To address this issue,we propose a novel method for longitudinal mapping of orientation-specific expansion of cortical surface area in these two orthogonal directions during early infancy.We apply our method to 30 healthy infants,and for the first time reveal the orientation-specific longitudinal cortical surface expansion maps during the first postnatal year.(3)Considering sulcal-gyral landmarks are not stably established in fetal brains,we propose to use the growth patterns of cortical properties to parcellate fetal cortical surfaces.Growth patterns,which reflect underlying changes of microstructures,enable an accurate definition of distinct regions in development and function,as microstructures determine the molecular organization and functional principles of the cortex.Finally,we generated biologically meaningful parcellations.(4)To fill the gap.of lacking hemispheric asymmetry of macaque cortical surface during early development,we first comprehensive analysis of cortical hemispheric asymmetries in typically developing macaques from birth to.20 months of age using multiple cortical properties.We achieve it by using an infant-specific cortical surfaces pipeline and a cortical surface statistical toolbox.Our results thus provide a valuable reference in studying early brain development and evolution.