Morphological Development Of Human Fetal Brain And Construction Of Structural Covariance Network During Second Trimester

During the prenatal period,the human brain undergoes immensely complicated and elaborate development.It is of great significance to study early developmental rules of fetal brain for understanding the connection of brain structure,the formation of functions,and revealing the origin of brain development diseases.The development and application of magnetic resonance imaging(MRI)technique has greatly promoted our understanding of the fetal brain.Compared with traditional ultrasound detection,MRI has the advantages of higher spatial resolution and better tissue contrast,which is helpful for the clinical diagnosis of developmental brain diseases.On the other hand,the development and application of image post-processing methods and software tools enable the study of early brain morphological development.The second trimester is a specific window of vulnerability for fetal brain development.During this period,the enormous proliferation and migration of neurons happen,and the volume of the cerebrum and cerebellum increase rapidly And,developmental related brain disease and malformations are more likely to occur.In clinical settings,in utero fetal MR scans can usually be acquired after 19 GW.The magnetic strength used in utero fetal MRI scan is commonly 1.5T.It is difficult to clearly distinguish the fetal structure with detailed local anatomy due to the sequence selection,frequent fetal movement,and the pulse of the maternal artery.Postmortem fetal specimens provide an opportunity to study earlier brain development,which have advantages in obtaining high spatial resolution and high signal-to-noise ratio(SNR)images by allowing high-field-strength magnets,increasing acquisition time,reducing slice thickness and using a smaller field of view.The study of specimen MR provides more and more reference value in the study of developing fetal brain.Using high field strength magnetic resonance data of fetal specimens during the early second trimester(14 to 22 gestational weeks(GW)),this study included three parts of work.In the first part,we explored the morphological development of the fetal cortical plate.In the second part,the structural covariance network was constructed based on the morphology of the cortical plate.In the third part,we explored the morphological development of fetal cerebellum.This results of this study can enrich radiological knowledge about human fetal brain morphological development and provide an anatomical reference for clinical diagnosis during early fetal brain development.Part 1:Morphological development of the fetal cortical plate during the second trimesterIntroduction:During the second trimester,the cerebral wall exhibits typical fetal laminar organization.The cortical plate,"the developing cortex",is composed of radially densely packed cortical columns,and undergoes immensely complex and rapid development.Starting around the seventh or eighth week of gestation,most neurons migrate to the cortical plate along the radially oriented glial fiber.The neuronal migration activity peaks between 12 and 20 GW.In clinical settings,in utero fetal MR scans are usually performed after 19 weeks of gestation.Most studies of fetal cortical plate development mainly focus on the late pregnancy.In this study,using high-field MRI data of fetal specimens during the second trimester,we aim to explore global and local morphological development,and hemispheric differences of the cortical plate,and to enrich the morphological study of early development of the cortical plate.Materials and methods:Twenty-seven human fetal specimens of 14 to 22 GW were included in the study and scanned by a 7.0 T Micro-MR scanner.Based on anatomy atlas of fetal brains and previous studies of MR images segmentation,we manually segmented the cortical plate using the Amira Software.The segmented cortical plate surfaces were processed with CIVET MRI analysis pipeline developed at Montreal Neurological Institute.We first constructed a surface template representing the age of our brain dataset using the surface registration algorithm SRFTRUACC and an unbiased template construction framework.The inner and outer surfaces of the cortical plate were converted to meshes,which were composed of vertices and triangular faces.The surfaces mesh of the cortical plate were sampled of 40962 vertices for each hemisphere.The morphology(thickness,surface area,and volume)of the cortical plate at each vertex was measured.The mean thickness of all vertices was the thickness of the whole cortical plate.The sum of the volume and surface area of all vertices were the volume and surface area of the whole cortical plate,respectively.The statistical analyses were performed using SurfStat.Linear regression analysis was performed to measure the correlation between average thickness,surface area and volume of the whole brain and increasing gestational age.A t-test was used to examine the differences of morphology between the left and right hemispheres.The vertex-based regression analysis was performed.The FDR procedure was performed to correct the multiple comparisons.Results:During the period of 14 to 22 GW,the average thickness?surface area and volume of the cortical plate increased significantly linearly.The growth rates of surface area and volume were faster than thickness.There were no significant differences in morphological characterizes between left and right hemispheres.The morphological characterizes of the cortical plate underwent uneven increases across different cortical regions.The regions around the Sylvian fissure and cingulate cortex developed faster.Conclusions:With high-filed MRI,this study provided a quantitative description of the morphological development of the fetal cortical plate from 14 to 22 weeks of gestation.The results of this study can provide a morphological reference for the further study of early development of the cerebral cortex.Part 2:Structural covariance network in developing brain of fetuses during the second trimester and neonatesIntroduction:During the fetal period,the proliferation and migration of neurons,together with axon growth and synaptogenesis,constantly reshaped the neural circuits in the brain.The complex network theory based on graph theory has been used to construct the structural and functional network of human brain,which provides a new way to study the development of human brain.Morphological characteristics(such as cortical thickness,surface area and volume,etc.)of the brain contains a large amount of brain connection information.Structural covariance networks(SCNs)have been used to characterize the inter-regional co-variation patterns of gray matter morphology across subjects.Because of difficulties associated with in utero imaging,the majority of existing early connectome studies have analyzed preterm infants and term neonates.The early configurations of developmental SCNs have not been fully explored.In this study,we constructed the structural covariance network for fetuses during the early second trimester and term neonates to explore early development of structural brain network.Materials and methods:Twenty-seven human fetal specimens of 14 to 22 GW were included in the study and scanned by a 7.0 T Micro-MR scanner.Thirty-nine healthy term-born neonates were also included,which was part of the Developing Human Connectome Project.The neonatal MRI were processed using the NEOCIVET pipeline developed based on CIVET.First,the native MR images were corrected for intensity non-uniformity artifacts,registered into the stereotaxic space,and classified as different types of brain tissue.The CLASP algorithm was used to extract the inner interface and outer interface of the cortical surface and was applied to generate a mesh model of the cortical surface composed of 40962 vertices for each hemisphere.The cortical thickness was measured using the t-link metric at each vertex.The thickness value of cortical plate in the fetal group used the results of the Part 1.We implemented a template-free approach to evenly sample the cortex into 1284 vertices,which were used to define the nodes of network.A general linear model(GLM)was used to account for covariance between vertices while accounting for the confounding variable of gestational age.The correlation matrices of the fetal and neonatal groups were binarized based on the same sparsity threshold.The characteristic network property(nodal degree,shortest path length,clustering coefficient,modularity and mean Euclidean distance between each edge)were calculated using the Brain Connectivity Toolbox.The nonparametric permutation test procedure was used to determine the statistical comparison in network properties between two groups.Results:The structure covariance network of fetal group exhibited a node degree distribution of an inverted-U shape,while the neonatal group basically followed an exponentially truncated power law distribution.The number of hubs significantly increased in the neonatal group.The number of nodes with moderate average path lengths were significantly increased in the fetal group,while the number of nodes with shorter and longer average path lengths were significantly increased in the neonatal group.From the early second trimester to birth,the normalized clustering coefficient and modularity of brain structural network significantly decreased,and the mean Euclidean distances between pairs of connected nodes significantly increased.The distributions of modular structure in the two group were different.The distribution of modules was concentrated on surrounding brain areas of closer proximity in the fetal group,while most modules in the neonatal group contained brain regions of physically long distancesConclusions:Though the fetal brain structural network exhibited modular structure during the early second trimester,this network was still considered to be in a primitive form.The brain connectome experienced rapid development and reorganization throughout the prenatal period to form a more efficient and well-integrated neonatal network.This study constructed,for the first time,structural covariance networks for fetuses during the second trimester and neonates,and provides a reference for early development of structural and functional networks.Part 3:Morphological development of the fetal cerebellum during the second trimesterIntroduction:The cerebellum,resided in the posterior cranial fossa,is a critical structure.The development of the cerebellum spans a long period(from approximately the fourth week of gestation into the first postnatal year)and undergoes a dramatic change in its morphological structure.The high complexity and protracted nature of development makes the cerebellum vulnerable to a broad spectrum of pathologic conditions,especially during the early fetal period.The cerebellum is connected to the brainstem via the cerebellar peduncles,and the impairment of fetal cerebellum could affect both structure and function of the developing brain.Clinically,the evolution of cerebellum is an important part of the prenatal examination,and could give information related to the developmental stage of the fetus.Previous research used intrauterine ultrasound and magnetic resonance image to measure main diameters and volume of the fetal cerebellum,which quantified the global development of cerebellum.However,these studies could not evaluate the local changes of cerebellum.In this study,we applied the volume and shape analysis to characterize the global and regional growth patterns of the fetal cerebellum during the early second trimester.Materials and methods:Thirty-five normal human fetal specimens with gestational ages ranging from 15 to 22 weeks were chosen and scanned by a 7.0 T Micro-MR scanner.Based on prior studies of cerebellar development and the atlas of human central nervous system development,we manually segmented the cerebellum using ITK-SNAP software.The absolute and relative volume of the cerebellum were computed using Matlab software.The relative volume of the cerebellum was defined as the ratio of cerebellar absolute volume to the sum of the cerebellar and cerebral volumes.To improve the accuracy of the registration step of the shape analysis,a template of the fetal cerebellum was constructed using Advanced Normalization Tools(ANTs).Shape analysis was using pipeline workflows developed by Laboratory of Neuro Imaging(LONI).The binary mask of each subject and the template were first converted to triangular meshes,which were then resampled to obtain 3000 uniformly distributed vertices.The thickness measured at each vertex of the mapped surfaces was defined as the distance from the vertex to the medial core of the cerebellum.Each individual triangulated mesh was then registered to our constructed cerebellar template mesh.As a result,the vertices of all surface were in one-to-one correspondence.To model the local development of the fetal cerebellum,the thicknesses associated with increasing gestational age were measured by linear regression analysis and the FDR procedure was performed to correct the multiple comparisons.Results:From 15 to 22 GW,the growth rate of the absolute volume of the fetal cerebellum steadily increased,and the second polynomial model better fit the growth trajectory.The pattern of relative volume changes revealed that the cerebellum had a faster growth rate than the supratentorial brain after 17 GW.The cerebellar template showed a dumbbell shape,which represented the average surface of the cerebellum during the early second trimester.Shape analysis found that from 15 to 22 GW,the extreme lateral portions of both cerebellar hemispheres showed the lowest rate of growth,and the anterior lobe grew faster than most of the posterior lobe.Conclusions:This study,for the first time,found that the greater growth pattern of the cerebellum compared to the cerebrum may begin as early as the 17 weeks of gestation.The sub-regions of cerebellum were characterized with different development rates,which may be linked to their phylogenetic and functional characteristics.There may be similar trajectories of growth within cerebro-cerebellar circuits.