3D U-Net Improves Automatic Brain Extraction for Isotropic Rat Brain Magnetic Resonance Imaging Data

Brain extraction is a critical pre-processing step in brain magnetic resonance imaging (MRI) analytical pipelines. In rodents, this is often achieved by manually editing brain masks slice-by-slice, a time-consuming task where workloads increase with higher spatial resolution datasets. We recently demonstrated successful automatic brain extraction via a deep-learning-based framework, U-Net, using 2D convolutions. However, such an approach cannot make use of the rich 3D spatial-context information from volumetric MRI data. In this study, we advanced our previously proposed U-Net architecture by replacing all 2D operations with their 3D counterparts and created a 3D U-Net framework. We trained and validated our model using a recently released CAMRI rat brain database acquired at isotropic spatial resolution, including T2-weighted turbo-spin-echo structural MRI and T2*-weighted echo-planar-imaging functional MRI. The performance of our 3D U-Net model was compared with existing rodent brain extraction tools, including Rapid Automatic Tissue Segmentation, Pulse-Coupled Neural Network, SHape descriptor selected External Regions after Morphologically filtering, and our previously proposed 2D U-Net model. 3D U-Net demonstrated superior performance in Dice, Jaccard, center-of-mass distance, Hausdorff distance, and sensitivity. Additionally, we demonstrated the reliability of 3D U-Net under various noise levels, evaluated the optimal training sample sizes, and disseminated all source codes publicly, with a hope that this approach will benefit rodent MRI research community.Significant Methodological Contribution: We proposed a deep-learning-based framework to automatically identify the rodent brain boundaries in MRI. With a fully 3D convolutional network model, 3D U-Net, our proposed method demonstrated improved performance compared to current automatic brain extraction methods, as shown in several qualitative metrics (Dice, Jaccard, PPV, SEN, and Hausdorff). We trust that this tool will avoid human bias and streamline pre-processing steps during 3D high resolution rodent brain MRI data analysis. The software developed herein has been disseminated freely to the community.

Peptidylarginine Deiminase and Alzheimer’s Disease

Peptidylarginine deiminases (PADs) are indispensable enzymes for post-translational modification of proteins, which can convert Arg residues on the surface of proteins to citrulline residues. The PAD family has five isozymes, PAD1, 2, 3, 4, and 6, which have been found in multiple tissues and organs. PAD2 and PAD4 were detected in cerebral cortex and hippocampus from human and rodent brain. In the central nervous system, abnormal expression and activation of PADs are involved in the pathological changes and pathogenesis of Alzheimer’s disease (AD). This article reviews the classification, distribution, and function of PADs, with an emphasis on the relationship between the abnormal activation of PADs and AD pathogenesis, diagnosis, and the therapeutic potential of PADs as drug targets for AD.

Rodent Brain Pathology, Audiogenic Epilepsy

The review presents data which provides evidence for the internal relationship between the stages of rodent audiogenic seizures and post-ictal catalepsy with the general pattern of animal reaction to the dangerous stimuli and/or situation. The wild run stage of audiogenic seizure fit could be regarded as an intense panic reaction, and this view found support in numerous experimental data. The phenomenon of audiogenic epilepsy probably attracted the attention of physiologists as rodents are extremely sensitive to dangerous sound stimuli. The seizure proneness in this group shares common physiological characteristics and depends on animal genotype. This concept could be the new platform for the study of epileptogenesis mechanisms.

Relationship between Blood Vessels and Migration of Neuroblasts in the Olfactory Neurogenic Region of the Rodent Brain

Neural precursors originating in the subventricular zone (SVZ), the largest neurogenic region of the adult brain, migrate several millimeters along a restricted migratory pathway, the rostral migratory stream (RMS), toward the olfactory bulb (OB), where they differentiate into interneurons and integrate into the local neuronal circuits. Migration of SVZ-derived neuroblasts in the adult brain differs in many aspects from that in the embryonic period. Unlike in that period, postnatally-generated neuroblasts in the SVZ are able to divide during migration along the RMS, as well as they migrate independently of radial glia. The homophilic mode of migration, i.e., using each other to move, is typical for neuroblast movement in the RMS. In addition, it has recently been demonstrated that specifically-arranged blood vessels navigate SVZ-derived neuroblasts to the OB and provide signals which promote migration. Here we review the development of vasculature in the presumptive neurogenic region of the rodent brain during the embryonic period as well as the development of the vascular scaffold guiding neuroblast migration in the postnatal period, and the significance of blood vessel reorganization during the early postnatal period for proper migration of RMS neuroblasts in adulthood.

Tracking the Migration of Injectable Microdevices in the Rodent Brain Using a 9.4T Magnetic Resonance Imaging Scanner

Wirelessly powered microdevices are being miniaturized to improve safety, longevity, and spatial resolution in a wide range of biomedical applications. Some wireless microdevices have reached a point where they can be injected whole into the central nervous system. However, the state-of-the-art floating microdevices have not yet been tested in chronic brain applications, and there is a growing concern that the implants might migrate through neural tissue over time. Using a 9.4T MRI scanner, we attempt to address the migration question by tracking ultra-small devices injected in different areas of the brain (cortico-subcortical) of rats over 5 months. We demonstrate that injectable microdevices smaller than 0.01 mm3 remain anchored in the brain at the targeted injection site over this time period. Based on CD68 (microglia) and GFAP (astrocytes) immunoreactivity to the microdevice, we hypothesize that glial scar formation is preventing the migration of chronically implanted microdevices in the brain over time.

3D U-Net improves automatic brain extraction for isotropic rat brain MRI data

Brain extraction is a critical pre-processing step in brain magnetic resonance imaging (MRI) analytical pipelines. In rodents, this is often achieved by manually editing brain masks slice-by-slice, a time-consuming task where workloads increase with higher spatial resolution datasets. We recently demonstrated successful automatic brain extraction via a deep-learning-based framework, U-Net, using 2D convolutions. However, such an approach cannot make use of the rich 3D spatial-context information from volumetric MRI data. In this study, we advanced our previously proposed U-Net architecture by replacing all 2D operations with their 3D counterparts and created a 3D U-Net framework. We trained and validated our model using a recently released CAMRI rat brain database acquired at isotropic spatial resolution, including T2-weighted turbo-spin-echo structural MRI and T2*-weighted echo-planar-imaging functional MRI. The performance of our 3D U-Net model was compared with existing rodent brain extraction tools, including Rapid Automatic Tissue Segmentation (RATS), Pulse-Coupled Neural Network (PCNN), SHape descriptor selected External Regions after Morphologically filtering (SHERM), and our previously proposed 2D U-Net model. 3D U-Net demonstrated superior performance in Dice, Jaccard, Hausdorff distance, and sensitivity. Additionally, we demonstrated the reliability of 3D U-Net under various noise levels, evaluated the optimal training sample sizes, and disseminated all source codes publicly, with a hope that this approach will benefit rodent MRI research community.Significant methodological contributionWe proposed a deep-learning-based framework to automatically identify the rodent brain boundaries in MRI. With a fully 3D convolutional network model, 3D U-Net, our proposed method demonstrated improved performance compared to current automatic brain extraction methods, as shown in several qualitative metrics (Dice, Jaccard, PPV, SEN, and Hausdorff). We trust that this tool will avoid human bias and streamline pre-processing steps during 3D high resolution rodent brain MRI data analysis. The software developed herein has been disseminated freely to the community.