We propose in this work a patch-based image labeling method relying on a label propagation framework. Based on image intensity similarities between the input image and an anatomy textbook, an original strategy which does not require any non-rigid registration is presented. Following recent developments in non-local image denoising, the similarity between images is represented by a weighted graph computed from an intensity-based distance between patches. Experiments on simulated and in-vivo MR images show that the proposed method is very successful in providing automated human brain labeling.

This study investigates the effect of class imbalance in training data when developing neural network classifiers for computer aided medical diagnosis. The investigation is performed in the presence of other characteristics that are typical among medical data, namely small training sample size, large number of features, and correlations between features. Two methods of neural network training are explored: classical backpropagation (BP) and particle swarm optimization (PSO) with clinically relevant training criteria. An experimental study is performed using simulated data and the conclusions are further validated on real clinical data for breast cancer diagnosis. The results show that classifier performance deteriorates with even modest class imbalance in the training data. Further, it is shown that BP is generally preferable over PSO for imbalanced training data especially with small data sample and large number of features. Finally, it is shown that there is no clear preference between oversampling and no compensation approach and some guidance is provided regarding a proper selection.

Early cortical folding and the emergence of structural brain asymmetries have been previously analyzed by neuropathology as well as qualitative analysis of magnetic resonance imaging (MRI) of fetuses and preterm neonates. In this study, we present a dedicated image analysis framework and its application for the detection of folding patterns during the critical period for the formation of many primary sulci (20-28 gestational weeks). Using structural information from in utero MRI, we perform morphometric analysis of cortical plate surface development and modeling of early folding in the normal fetal brain. First, we identify regions of the fetal brain surface that undergo significant folding changes during this developmental period and provide precise temporal staging of these changes for each region of interest. Then, we highlight the emergence of interhemispheric structural asymmetries that may be related to future functional specialization of cortical areas. Our findings complement previous descriptions of early sulcogenesis based on neuropathology and qualitative evaluation of 2D in utero MRI by accurate spatial and temporal mapping of the emergence of individual sulci as well as structural brain asymmetries. The study provides the missing starting point for their developmental trajectories and extends our understanding of normal cortical folding.

In recent years post-processing of fast multi-slice MR imaging to correct fetal motion has provided the first true 3D MR images of the developing human brain in utero. Early approaches have used reconstruction based algorithms, employing a two step iterative process, where slices from the acquired data are re-aligned to an approximate 3D reconstruction of the fetal brain, which is then refined further using the improved slice alignment. This two step slice-to-volume process, although powerful, is computationally expensive in needing a 3D reconstruction, and is limited in its ability to recover sub-voxel alignment. Here, we describe an alternative approach which we term slice intersection motion correction (SIMC), that seeks to directly co-align multiple slice stacks by considering the matching structure along all intersecting slice pairs in all orthogonally planned slices that are acquired in clinical imaging studies. A collective update scheme for all slices is then derived, to simultaneously drive slices into a consistent match along their lines of intersection. We then describe a 3D reconstruction algorithm that, using the final motion corrected slice locations, suppresses through-plane partial volume effects to provide a single high isotropic resolution 3D image. The method is tested on simulated data with known motions and is applied to retrospectively reconstruct 3D images from a range of clinically acquired imaging studies. The quantitative evaluation of the registration accuracy for the simulated data sets demonstrated a significant improvement over previous approaches. An initial application of the technique to studying clinical pathology is included, where the proposed method recovered up to 15 mm of translation and 30 degrees of rotation for individual slices, and produced full 3D reconstructions containing clinically useful additional information not visible in the original 2D slices.

Existing knowledge of growth patterns in the living fetal human brain is based upon in utero imaging studies by magnetic resonance imaging (MRI) and ultrasound, which describe overall growth and provide mainly qualitative findings. However, formation of the complex folded cortical structure of the adult brain requires, in part, differential rates of regional tissue growth. To better understand these local tissue growth patterns, we applied recent advances in fetal MRI motion correction and computational image analysis techniques to 40 normal fetal human brains covering a period of primary sulcal formation (20 -28 gestational weeks). Growth patterns were mapped by quantifying tissue locations that were expanding more or less quickly than the overall cerebral growth rate, which reveal increasing structural complexity. We detected increased local relative growth rates in the formation of the precentral and postcentral gyri, right superior temporal gyrus, and opercula, which differentiated between the constant growth rate in underlying cerebral mantle and the accelerating rate in the cortical plate undergoing folding. Analysis focused on the cortical plate revealed greater volume increases in parietal and occipital regions compared to the frontal lobe. Cortical plate growth patterns constrained to narrower age ranges showed that gyrification, reflected by greater growth rates, was more pronounced after 24 gestational weeks. Local hemispheric volume asymmetry was located in the posterior peri-Sylvian area associated with structural lateralization in the mature brain. These maps of fetal brain growth patterns construct a spatially specific baseline of developmental biomarkers with which to correlate abnormal development in the human.

Modeling and analysis of MR images of the developing human brain is a challenge due to rapid changes in brain morphology and morphometry. We present an approach to the construction of a spatiotemporal atlas of the fetal brain with temporal models of MR intensity, tissue probability and shape changes. This spatiotemporal model is created from a set of reconstructed MR images of fetal subjects with different gestational ages. Groupwise registration of manual segmentations and voxelwise nonlinear modeling allow us to capture the appearance, disappearance and spatial variation of brain structures over time. Applying this model to atlas-based segmentation, we generate agespecific MR templates and tissue probability maps and use them to initialize automatic tissue delineation in new MR images. The choice of model parameters and the final performance are evaluated using clinical MR scans of young fetuses with gestational ages ranging from 20.57 to 24.71 weeks. Experimental results indicate that quadratic temporal models can correctly capture growthrelated changes in the fetal brain anatomy and provide improvement in accuracy of atlas-based tissue segmentation.

Imaging of the human fetus using magnetic resonance (MR) is an essential tool for quantitative studies of normal as well as abnormal brain development in utero. However, because of fundamental differences in tissue types, tissue properties and tissue distribution between the fetal and adult brain, automated tissue segmentation techniques developed for adult brain anatomy are unsuitable for this data. In this paper, we describe methodology for automatic atlas-based segmentation of individual tissue types in motion-corrected 3D volumes reconstructed from clinical MR scans of the fetal brain. To generate anatomically correct automatic segmentations, we create a set of accurate manual delineations and build an in utero 3D statistical atlas of tissue distribution incorporating developing grey and white matter as well as transient tissue types such as the germinal matrix. The probabilistic atlas is associated with an unbiased average shape and intensity template for registration of new subject images to the space of the atlas. Quantitative whole brain 3D validation of tissue labeling performed on a set of 14 fetal MR scans (20.57–22.86 weeks gestational age) demonstrates that this atlas-based EM segmentation approach achieves consistently high DSC performance for the main tissue types in the fetal brain. This work indicates that reliable measures of brain development can be automatically derived from clinical MR imaging and opens up possibility of further 3D volumetric and morphometric studies with multiple fetal subjects.

In the latter half of gestation (20 to 40 gestational weeks), human brain growth accelerates in conjunction with cortical folding and the deceleration of ventricular zone progenitor cell proliferation. These processes are reflected in changes in the volume of respective fetal tissue zones. Thus far, growth trajectories of the fetal tissue zones have been extracted primarily from 2D measurements on histological sections and magnetic resonance imaging (MRI). In this study, the volumes of major fetal zones—cortical plate (CP), subplate and intermediate zone (SP+IZ), germinal matrix (GMAT), deep gray nuclei (DG), and ventricles (VENT)—are calculated from automatic segmentation of motion-corrected, 3D reconstructed MRI. We analyzed 48 T2-weighted MRI scans from 39 normally developing fetuses in utero between 20.57 and 31.14 gestational weeks (GW). The supratentorial volume (STV) increased linearly at a rate of 15.22% per week. The SP+IZ (14.75% per week) and DG (15.56% per week) volumes increased at similar rates. The CP increased at a greater relative rate (18.00% per week), while the VENT (9.18% per week) changed more slowly. Therefore, CP increased as a fraction of STV and the VENT fraction declined. The total GMAT volume slightly increased then decreased after 25 GW. We did not detect volumetric sexual dimorphisms or total hemispheric volume asymmetries, which may emerge later in gestation. Further application of the automated fetal brain segmentation to later gestational ages will bridge the gap between volumetric studies of premature brain development and normal brain development in utero.