The Experts below are selected from a list of 15564 Experts worldwide ranked by ideXlab platform
Hazel Sive - One of the best experts on this subject based on the ideXlab platform.
-
an assay for permeability of the zebrafish embryonic neuroepithelium
Journal of Visualized Experiments, 2012Co-Authors: Jessica T Chang, Hazel SiveAbstract:The Brain ventricular system is conserved among vertebrates and is composed of a series of interconnected cavities called Brain Ventricles, which form during the earliest stages of Brain development and are maintained throughout the animal's life. The Brain ventricular system is found in vertebrates, and the Ventricles develop after neural tube formation, when the central lumen fills with cerebrospinal fluid (CSF) 1,2. CSF is a protein rich fluid that is essential for normal Brain development and function3-6. In zebrafish, Brain Ventricle inflation begins at approximately 18 hr post fertilization (hpf), after the neural tube is closed. Multiple processes are associated with Brain Ventricle formation, including formation of a neuroepithelium, tight junction formation that regulates permeability and CSF production. We showed that the Na,K-ATPase is required for Brain Ventricle inflation, impacting all these processes 7,8, while claudin 5a is necessary for tight junction formation 9. Additionally, we showed that "relaxation" of the embryonic neuroepithelium, via inhibition of myosin, is associated with Brain Ventricle inflation. To investigate the regulation of permeability during zebrafish Brain Ventricle inflation, we developed a ventricular dye retention assay. This method uses Brain Ventricle injection in a living zebrafish embryo, a technique previously developed in our lab10, to fluorescently label the cerebrospinal fluid. Embryos are then imaged over time as the fluorescent dye moves through the Brain Ventricles and neuroepithelium. The distance the dye front moves away from the basal (non-luminal) side of the neuroepithelium over time is quantified and is a measure of neuroepithelial permeability (Figure 1). We observe that dyes 70 kDa and smaller will move through the neuroepithelium and can be detected outside the embryonic zebrafish Brain at 24 hpf (Figure 2). This dye retention assay can be used to analyze neuroepithelial permeability in a variety of different genetic backgrounds, at different times during development, and after environmental perturbations. It may also be useful in examining pathological accumulation of CSF. Overall, this technique allows investigators to analyze the role and regulation of permeability during development and disease.
-
multiple roles for the na k atpase subunits atp1a1 and fxyd1 during Brain Ventricle development
Developmental Biology, 2012Co-Authors: Jessica T Chang, Laura Anne Lowery, Hazel SiveAbstract:Formation of the vertebrate Brain Ventricles requires both production of cerebrospinal fluid (CSF), and its retention in the Ventricles. The Na,K-ATPase is required for Brain Ventricle development, and we show here that this protein complex impacts three associated processes. The first requires both the alpha subunit (Atp1a1) and the regulatory subunit, Fxyd1, and leads to formation of a cohesive neuroepithelium, with continuous apical junctions. The second process leads to modulation of neuroepithelial permeability, and requires Atp1a1, which increases permeability with partial loss of function and decreases it with overexpression. In contrast, fxyd1 overexpression does not alter neuroepithelial permeability, suggesting that its activity is limited to neuroepithelium formation. RhoA regulates both neuroepithelium formation and permeability, downstream of the Na,K-ATPase. A third process, likely to be CSF production, is RhoA-independent, requiring Atp1a1, but not Fxyd1. Consistent with a role for Na,K-ATPase pump function, the inhibitor ouabain prevents neuroepithelium formation, while intracellular Na þ increases after Atp1a1 and Fxyd1 loss of function. These data include the first reported role for Fxyd1 in the developing Brain, and indicate that the Na,K-ATPase regulates three aspects of Brain Ventricle development essential for normal function: formation of a cohesive neuroepithelium, restriction of neuroepithelial permeability, and production of CSF.
-
epithelial relaxation mediated by the myosin phosphatase regulator mypt1 is required for Brain Ventricle lumen expansion and hindBrain morphogenesis
Development, 2010Co-Authors: Jennifer H Gutzman, Hazel SiveAbstract:We demonstrate that in the zebrafish hindBrain, cell shape, rhombomere morphogenesis and, unexpectedly, Brain Ventricle lumen expansion depend on the contractile state of the neuroepithelium. The hindBrain neural tube opens in a specific sequence, with initial separation along the midline at rhombomere boundaries, subsequent openings within rhombomeres and eventual coalescence of openings into the hindBrain Ventricle lumen. A mutation in the myosin phosphatase regulator mypt1 results in a small Ventricle due to impaired stretching of the surrounding neuroepithelium. Although initial hindBrain opening remains normal, mypt1 mutant rhombomeres do not undergo normal morphological progression. Three-dimensional reconstruction demonstrates cell shapes within rhombomeres and at rhombomere boundaries are abnormal in mypt1 mutants. Wild-type cell shape requires that surrounding cells are also wild type, whereas mutant cell shape is autonomously regulated. Supporting the requirement for regulation of myosin function during hindBrain morphogenesis, wild-type embryos show dynamic levels of phosphorylated myosin regulatory light chain (pMRLC). By contrast, mutants show continuously high pMRLC levels, with concentration of pMRLC and myosin II at the apical side of the epithelium, and myosin II and actin concentration at rhombomere boundaries. Brain Ventricle lumen expansion, rhombomere morphology and cell shape are rescued by inhibition of myosin II function, indicating that each defect is a consequence of overactive myosin. We suggest that the epithelium must `relax', via activity of myosin phosphatase, to allow for normal hindBrain morphogenesis and expansion of the Brain ventricular lumen. Epithelial relaxation might be a widespread strategy to facilitate tube inflation in many organs.
-
Zebrafish Brain Ventricle injection.
Journal of visualized experiments : JoVE, 2009Co-Authors: Jennifer H Gutzman, Hazel SiveAbstract:Proper Brain Ventricle formation during embryonic Brain development is required for normal Brain function. Brain Ventricles are the highly conserved cavities within the Brain that are filled with cerebrospinal fluid. In zebrafish, after neural tube formation, the neuroepithelium undergoes a series of constrictions and folds while it fills with fluid resulting in Brain Ventricle formation. In order to understand the process of Ventricle formation, and the neuroepithelial shape changes that occur at the same time, we needed a way to visualize the Ventricle space in comparison to the Brain tissue. However, the nature of transparent zebrafish embryos makes it difficult to differentiate the tissue from the Ventricle space. Therefore, we developed a Brain Ventricle injection technique where the Ventricle space is filled with a fluorescent dye and imaged by brightfield and fluorescent microscopy. The brightfield and the fluorescent images are then processed and superimposed in Photoshop. This technique allows for visualization of the Ventricle space with the fluorescent dye, in comparison to the shape of the neuroepithelium in the brightfield image. Brain Ventricle injection in zebrafish can be employed from 18 hours post fertilization through early larval stages. We have used this technique extensively in our studies of Brain Ventricle formation and morphogenesis as well as in characterizing Brain morphogenesis mutants (1-3).
-
Characterization and Classification of Zebrafish Brain Morphology Mutants
Anatomical record (Hoboken N.J. : 2007), 2009Co-Authors: Laura Anne Lowery, Jennifer H Gutzman, Gianluca De Rienzo, Hazel SiveAbstract:The mechanisms by which the vertebrate Brain achieves its three-dimensional structure are clearly complex, requiring the functions of many genes. Using the zebrafish as a model, we have begun to define genes required for Brain morphogenesis, including Brain Ventricle formation, by studying 16 mutants previously identified as having embryonic Brain morphology defects. We report the phenotypic characterization of these mutants at several timepoints, using Brain Ventricle dye injection, imaging, and immunohistochemistry with neuronal markers. Most of these mutants display early phenotypes, affecting initial Brain shaping, whereas others show later phenotypes, affecting Brain Ventricle expansion. In the early phenotype group, we further define four phenotypic classes and corresponding functions required for Brain morphogenesis. Although we did not use known genotypes for this classification, basing it solely on phenotypes, many mutants with defects in functionally related genes clustered in a single class. In particular, Class 1 mutants show midline separation defects, corresponding to epithelial junction defects; Class 2 mutants show reduced Brain Ventricle size; Class 3 mutants show midBrain–hindBrain abnormalities, corresponding to basement membrane defects; and Class 4 mutants show absence of Ventricle lumen inflation, corresponding to defective ion pumping. Later Brain Ventricle expansion requires the extracellular matrix, cardiovascular circulation, and transcription/splicing-dependent events. We suggest that these mutants define processes likely to be used during Brain morphogenesis throughout the vertebrates. Anat Rec, 2009. © 2008 Wiley-Liss, Inc.
Robert Bartha - One of the best experts on this subject based on the ideXlab platform.
-
Association between gait variability and Brain Ventricle attributes: a Brain mapping study.
Experimental gerontology, 2014Co-Authors: Cédric Annweiler, Manuel Montero-odasso, Robert Bartha, John Drozd, Vladimir Hachinski, Olivier BeauchetAbstract:Abstract Background It remains unknown which Brain regions are involved in the maintenance of gait dynamic stability in older adults, as characterized by a low stride time variability. Expansion of lateral cerebral Ventricles is an indirect marker of adjacent Brain tissue volume. The purpose of this study was to examine the association between stride time variability and the volume of sub-regions of the lateral cerebral Ventricles among older community-dwellers. Methods One-hundred-fifteen participants free of hydrocephalus from the GAIT study (mean, 70.4 ± 4.4 years; 43.5% female) were included in this analysis. Stride time variability was measured at self-selected pace with a 10 m electronic portable walkway (GAITRite). Participants were separated into 3 groups based on tertiles of stride time variability (i.e., 2.8%). Brain Ventricle sub-volumes were quantified from three-dimensional T 1 -weighted MRI using semi-automated software. Age, gender, Cumulative Illness Rating Scale for Geriatrics, Mini-Mental State Examination, Go-NoGo, Brain vascular burden, 4-item Geriatric Depression Scale, psychoactive drugs, vision, proprioception, body mass index, muscular strength and gait velocity were used as covariates. Results Participants with the highest (i.e., worst) tertile of stride time variability exhibited larger temporal horns than those with the lowest (P = 0.030) and intermediate tertiles (P = 0.028). They also had larger middle portions of ventricular bodies than those with the intermediate tertile (P = 0.018). Larger temporal horns were associated with increase in stride time variability (adjusted β = 0.86, P = 0.005), specifically with the highest tertile of stride time variability (adjusted OR = 2.45, P = 0.044). Conclusions Higher stride time variability was associated with larger temporal horns in older community-dwellers. Addressing focal neuronal losses in temporal lobes may represent an important strategy to prevent gait instability.
-
a novel mri compatible Brain Ventricle phantom for validation of segmentation and volumetry methods
Journal of Magnetic Resonance Imaging, 2012Co-Authors: Amanda Khan, Robert Moreland, John Drozd, Michael Borrie, Robert BarthaAbstract:Purpose: To create a standardized, MRI-compatible, life-sized phantom of the Brain Ventricles to evaluate Ventricle segmentation methods using T1-weighted MRI. An objective phantom is needed to test the many different segmentation programs currently used to measure Ventricle volumes in patients with Alzheimer's disease. Materials and Methods: A Ventricle model was constructed from polycarbonate using a digital mesh of the Ventricles created from the 3 Tesla (T) MRI of a subject with Alzheimer's disease. The Ventricle was placed in a Brain mold and surrounded with material composed of 2% agar in water, 0.01% NaCl and 0.0375 mM gadopentetate dimeglumine to match the signal intensity properties of Brain tissue in 3T T1-weighted MRI. The 3T T1-weighted images of the phantom were acquired and Ventricle segmentation software was used to measure Ventricle volume. Results: The images acquired of the phantom successfully replicated in vivo signal intensity differences between the Ventricle and surrounding tissue in T1-weighted images and were robust to segmentation. The Ventricle volume was quantified to 99% accuracy at 1-mm voxel size. Conclusion: The phantom represents a simple, realistic and objective method to test the accuracy of lateral Ventricle segmentation methods and we project it can be extended to other anatomical structures. J. Magn. Reson. Imaging 2012;36:476–482. © 2012 Wiley Periodicals, Inc.
-
A novel MRI‐compatible Brain Ventricle phantom for validation of segmentation and volumetry methods
Journal of magnetic resonance imaging : JMRI, 2012Co-Authors: Amanda F. Khan, John Drozd, Robert Moreland, Michael Borrie, Robert BarthaAbstract:Purpose: To create a standardized, MRI-compatible, life-sized phantom of the Brain Ventricles to evaluate Ventricle segmentation methods using T1-weighted MRI. An objective phantom is needed to test the many different segmentation programs currently used to measure Ventricle volumes in patients with Alzheimer's disease. Materials and Methods: A Ventricle model was constructed from polycarbonate using a digital mesh of the Ventricles created from the 3 Tesla (T) MRI of a subject with Alzheimer's disease. The Ventricle was placed in a Brain mold and surrounded with material composed of 2% agar in water, 0.01% NaCl and 0.0375 mM gadopentetate dimeglumine to match the signal intensity properties of Brain tissue in 3T T1-weighted MRI. The 3T T1-weighted images of the phantom were acquired and Ventricle segmentation software was used to measure Ventricle volume. Results: The images acquired of the phantom successfully replicated in vivo signal intensity differences between the Ventricle and surrounding tissue in T1-weighted images and were robust to segmentation. The Ventricle volume was quantified to 99% accuracy at 1-mm voxel size. Conclusion: The phantom represents a simple, realistic and objective method to test the accuracy of lateral Ventricle segmentation methods and we project it can be extended to other anatomical structures. J. Magn. Reson. Imaging 2012;36:476–482. © 2012 Wiley Periodicals, Inc.
Yao Wang - One of the best experts on this subject based on the ideXlab platform.
-
deep mouse an end to end auto context refinement framework for Brain Ventricle body segmentation in embryonic mice ultrasound volumes
International Symposium on Biomedical Imaging, 2020Co-Authors: Ziming Qiu, William Das, Chuiyu Wang, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:The segmentation of the Brain Ventricle (BV) and body in embryonic mice high-frequency ultrasound (HFU) volumes can provide useful information for biological researchers. However, manual segmentation of the BV and body requires substantial time and expertise. This work proposes a novel deep learning based end-to-end auto-context refinement framework, consisting of two stages. The first stage produces a low resolution segmentation of the BV and body simultaneously. The resulting probability map for each object (BV or body) is then used to crop a region of interest (ROI) around the target object in both the original image and the probability map to provide context to the refinement segmentation network. Joint training of the two stages provides significant improvement in Dice Similarity Coefficient (DSC) over using only the first stage (0.818 to 0.906 for the BV, and 0.919 to 0.934 for the body). The proposed method significantly reduces the inference time (102.36 to 0.09 s/volume ≈1000x faster) while slightly improves the segmentation accuracy over the previous methods using slide-window approaches.
-
deep mouse an end to end auto context refinement framework for Brain Ventricle and body segmentation in embryonic mice ultrasound volumes
arXiv: Image and Video Processing, 2019Co-Authors: Ziming Qiu, William Das, Chuiyu Wang, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:High-frequency ultrasound (HFU) is well suited for imaging embryonic mice due to its noninvasive and real-time characteristics. However, manual segmentation of the Brain Ventricles (BVs) and body requires substantial time and expertise. This work proposes a novel deep learning based end-to-end auto-context refinement framework, consisting of two stages. The first stage produces a low resolution segmentation of the BV and body simultaneously. The resulting probability map for each object (BV or body) is then used to crop a region of interest (ROI) around the target object in both the original image and the probability map to provide context to the refinement segmentation network. Joint training of the two stages provides significant improvement in Dice Similarity Coefficient (DSC) over using only the first stage (0.818 to 0.906 for the BV, and 0.919 to 0.934 for the body). The proposed method significantly reduces the inference time (102.36 to 0.09 s/volume around 1000x faster) while slightly improves the segmentation accuracy over the previous methods using slide-window approaches.
-
automatic mouse embryo Brain Ventricle body segmentation and mutant classification from ultrasound data using deep learning
Internaltional Ultrasonics Symposium, 2019Co-Authors: Ziming Qiu, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:High-frequency ultrasound (HFU) is well suited for imaging embryonic mice in vivo because it is non-invasive and real-time. Manual segmentation of the Brain Ventricles (BVs) and whole body from 3D HFU images is time-consuming and requires specialized training. This paper presents a deep-learning-based segmentation pipeline which automates several time-consuming, repetitive tasks currently performed to study genetic mutations in developing mouse embryos. Namely, the pipeline accurately segments the BV and body regions in 3D HFU images of mouse embryos, despite significant challenges due to position and shape variation of the embryos, as well as imaging artifacts. Based on the BV segmentation, a 3D convolutional neural network (CNN) is further trained to detect embryos with the Engrailed-1 (En1) mutation. The algorithms achieve 0.896 and 0.925 Dice Similarity Coefficient (DSC) for BV and body segmentation, respectively, and 95.8% accuracy on mutant classification. Through gradient based interrogation and visualization of the trained classifier, it is demonstrated that the model focuses on the morphological structures known to be affected by the En1 mutation.
-
Automatic Mouse Embryo Brain Ventricle & Body Segmentation and Mutant Classification From Ultrasound Data Using Deep Learning
2019 IEEE International Ultrasonics Symposium (IUS), 2019Co-Authors: Ziming Qiu, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:High-frequency ultrasound (HFU) is well suited for imaging embryonic mice in vivo because it is non-invasive and real-time. Manual segmentation of the Brain Ventricles (BVs) and whole body from 3D HFU images is time-consuming and requires specialized training. This paper presents a deep-learning-based segmentation pipeline which automates several time-consuming, repetitive tasks currently performed to study genetic mutations in developing mouse embryos. Namely, the pipeline accurately segments the BV and body regions in 3D HFU images of mouse embryos, despite significant challenges due to position and shape variation of the embryos, as well as imaging artifacts. Based on the BV segmentation, a 3D convolutional neural network (CNN) is further trained to detect embryos with the Engrailed-1 (En1) mutation. The algorithms achieve 0.896 and 0.925 Dice Similarity Coefficient (DSC) for BV and body segmentation, respectively, and 95.8% accuracy on mutant classification. Through gradient based interrogation and visualization of the trained classifier, it is demonstrated that the model focuses on the morphological structures known to be affected by the En1 mutation.
-
deep bv a fully automated system for Brain Ventricle localization and segmentation in 3d ultrasound images of embryonic mice
IEEE Signal Processing in Medicine and Biology Symposium, 2018Co-Authors: Ziming Qiu, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:Volumetric analysis of Brain Ventricle (BV) structure is a key tool in the study of central nervous system development in embryonic mice. High-frequency ultrasound (HFU) is the only non-invasive, real-time modality available for rapid volumetric imaging of embryos in utero. However, manual segmentation of the BV from HFU volumes is tedious, time-consuming, and requires specialized expertise. In this paper, we propose a novel deep learning based BV segmentation system for whole-body HFU images of mouse embryos. Our fully automated system consists of two modules: localization and segmentation. It first applies a volumetric convolutional neural network on a 3D sliding window over the entire volume to identify a 3D bounding box containing the entire BV. It then employs a fully convolutional network to segment the detected bounding box into BV and background. The system achieves a Dice Similarity Coefficient (DSC) of 0.8956 for BV segmentation on an unseen 111 HFU volume test set surpassing the previous state-of-the-art method (DSC of 0.7119) by a margin of 25%.
John Drozd - One of the best experts on this subject based on the ideXlab platform.
-
Association between gait variability and Brain Ventricle attributes: a Brain mapping study.
Experimental gerontology, 2014Co-Authors: Cédric Annweiler, Manuel Montero-odasso, Robert Bartha, John Drozd, Vladimir Hachinski, Olivier BeauchetAbstract:Abstract Background It remains unknown which Brain regions are involved in the maintenance of gait dynamic stability in older adults, as characterized by a low stride time variability. Expansion of lateral cerebral Ventricles is an indirect marker of adjacent Brain tissue volume. The purpose of this study was to examine the association between stride time variability and the volume of sub-regions of the lateral cerebral Ventricles among older community-dwellers. Methods One-hundred-fifteen participants free of hydrocephalus from the GAIT study (mean, 70.4 ± 4.4 years; 43.5% female) were included in this analysis. Stride time variability was measured at self-selected pace with a 10 m electronic portable walkway (GAITRite). Participants were separated into 3 groups based on tertiles of stride time variability (i.e., 2.8%). Brain Ventricle sub-volumes were quantified from three-dimensional T 1 -weighted MRI using semi-automated software. Age, gender, Cumulative Illness Rating Scale for Geriatrics, Mini-Mental State Examination, Go-NoGo, Brain vascular burden, 4-item Geriatric Depression Scale, psychoactive drugs, vision, proprioception, body mass index, muscular strength and gait velocity were used as covariates. Results Participants with the highest (i.e., worst) tertile of stride time variability exhibited larger temporal horns than those with the lowest (P = 0.030) and intermediate tertiles (P = 0.028). They also had larger middle portions of ventricular bodies than those with the intermediate tertile (P = 0.018). Larger temporal horns were associated with increase in stride time variability (adjusted β = 0.86, P = 0.005), specifically with the highest tertile of stride time variability (adjusted OR = 2.45, P = 0.044). Conclusions Higher stride time variability was associated with larger temporal horns in older community-dwellers. Addressing focal neuronal losses in temporal lobes may represent an important strategy to prevent gait instability.
-
a novel mri compatible Brain Ventricle phantom for validation of segmentation and volumetry methods
Journal of Magnetic Resonance Imaging, 2012Co-Authors: Amanda Khan, Robert Moreland, John Drozd, Michael Borrie, Robert BarthaAbstract:Purpose: To create a standardized, MRI-compatible, life-sized phantom of the Brain Ventricles to evaluate Ventricle segmentation methods using T1-weighted MRI. An objective phantom is needed to test the many different segmentation programs currently used to measure Ventricle volumes in patients with Alzheimer's disease. Materials and Methods: A Ventricle model was constructed from polycarbonate using a digital mesh of the Ventricles created from the 3 Tesla (T) MRI of a subject with Alzheimer's disease. The Ventricle was placed in a Brain mold and surrounded with material composed of 2% agar in water, 0.01% NaCl and 0.0375 mM gadopentetate dimeglumine to match the signal intensity properties of Brain tissue in 3T T1-weighted MRI. The 3T T1-weighted images of the phantom were acquired and Ventricle segmentation software was used to measure Ventricle volume. Results: The images acquired of the phantom successfully replicated in vivo signal intensity differences between the Ventricle and surrounding tissue in T1-weighted images and were robust to segmentation. The Ventricle volume was quantified to 99% accuracy at 1-mm voxel size. Conclusion: The phantom represents a simple, realistic and objective method to test the accuracy of lateral Ventricle segmentation methods and we project it can be extended to other anatomical structures. J. Magn. Reson. Imaging 2012;36:476–482. © 2012 Wiley Periodicals, Inc.
-
A novel MRI‐compatible Brain Ventricle phantom for validation of segmentation and volumetry methods
Journal of magnetic resonance imaging : JMRI, 2012Co-Authors: Amanda F. Khan, John Drozd, Robert Moreland, Michael Borrie, Robert BarthaAbstract:Purpose: To create a standardized, MRI-compatible, life-sized phantom of the Brain Ventricles to evaluate Ventricle segmentation methods using T1-weighted MRI. An objective phantom is needed to test the many different segmentation programs currently used to measure Ventricle volumes in patients with Alzheimer's disease. Materials and Methods: A Ventricle model was constructed from polycarbonate using a digital mesh of the Ventricles created from the 3 Tesla (T) MRI of a subject with Alzheimer's disease. The Ventricle was placed in a Brain mold and surrounded with material composed of 2% agar in water, 0.01% NaCl and 0.0375 mM gadopentetate dimeglumine to match the signal intensity properties of Brain tissue in 3T T1-weighted MRI. The 3T T1-weighted images of the phantom were acquired and Ventricle segmentation software was used to measure Ventricle volume. Results: The images acquired of the phantom successfully replicated in vivo signal intensity differences between the Ventricle and surrounding tissue in T1-weighted images and were robust to segmentation. The Ventricle volume was quantified to 99% accuracy at 1-mm voxel size. Conclusion: The phantom represents a simple, realistic and objective method to test the accuracy of lateral Ventricle segmentation methods and we project it can be extended to other anatomical structures. J. Magn. Reson. Imaging 2012;36:476–482. © 2012 Wiley Periodicals, Inc.
Jonathan Mamou - One of the best experts on this subject based on the ideXlab platform.
-
deep mouse an end to end auto context refinement framework for Brain Ventricle body segmentation in embryonic mice ultrasound volumes
International Symposium on Biomedical Imaging, 2020Co-Authors: Ziming Qiu, William Das, Chuiyu Wang, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:The segmentation of the Brain Ventricle (BV) and body in embryonic mice high-frequency ultrasound (HFU) volumes can provide useful information for biological researchers. However, manual segmentation of the BV and body requires substantial time and expertise. This work proposes a novel deep learning based end-to-end auto-context refinement framework, consisting of two stages. The first stage produces a low resolution segmentation of the BV and body simultaneously. The resulting probability map for each object (BV or body) is then used to crop a region of interest (ROI) around the target object in both the original image and the probability map to provide context to the refinement segmentation network. Joint training of the two stages provides significant improvement in Dice Similarity Coefficient (DSC) over using only the first stage (0.818 to 0.906 for the BV, and 0.919 to 0.934 for the body). The proposed method significantly reduces the inference time (102.36 to 0.09 s/volume ≈1000x faster) while slightly improves the segmentation accuracy over the previous methods using slide-window approaches.
-
ISBI - Deep Mouse: An End-to-End Auto-Context Refinement Framework for Brain Ventricle & Body Segmentation in Embryonic Mice Ultrasound Volumes
2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI), 2020Co-Authors: Ziming Qiu, William Das, Chuiyu Wang, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. KetterlingAbstract:The segmentation of the Brain Ventricle (BV) and body in embryonic mice high-frequency ultrasound (HFU) volumes can provide useful information for biological researchers. However, manual segmentation of the BV and body requires substantial time and expertise. This work proposes a novel deep learning based end-to-end auto-context refinement framework, consisting of two stages. The first stage produces a low resolution segmentation of the BV and body simultaneously. The resulting probability map for each object (BV or body) is then used to crop a region of interest (ROI) around the target object in both the original image and the probability map to provide context to the refinement segmentation network. Joint training of the two stages provides significant improvement in Dice Similarity Coefficient (DSC) over using only the first stage (0.818 to 0.906 for the BV, and 0.919 to 0.934 for the body). The proposed method significantly reduces the inference time (102.36 to 0.09 s/volume ≈1000x faster) while slightly improves the segmentation accuracy over the previous methods using slide-window approaches.
-
deep mouse an end to end auto context refinement framework for Brain Ventricle and body segmentation in embryonic mice ultrasound volumes
arXiv: Image and Video Processing, 2019Co-Authors: Ziming Qiu, William Das, Chuiyu Wang, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:High-frequency ultrasound (HFU) is well suited for imaging embryonic mice due to its noninvasive and real-time characteristics. However, manual segmentation of the Brain Ventricles (BVs) and body requires substantial time and expertise. This work proposes a novel deep learning based end-to-end auto-context refinement framework, consisting of two stages. The first stage produces a low resolution segmentation of the BV and body simultaneously. The resulting probability map for each object (BV or body) is then used to crop a region of interest (ROI) around the target object in both the original image and the probability map to provide context to the refinement segmentation network. Joint training of the two stages provides significant improvement in Dice Similarity Coefficient (DSC) over using only the first stage (0.818 to 0.906 for the BV, and 0.919 to 0.934 for the body). The proposed method significantly reduces the inference time (102.36 to 0.09 s/volume around 1000x faster) while slightly improves the segmentation accuracy over the previous methods using slide-window approaches.
-
automatic mouse embryo Brain Ventricle body segmentation and mutant classification from ultrasound data using deep learning
Internaltional Ultrasonics Symposium, 2019Co-Authors: Ziming Qiu, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:High-frequency ultrasound (HFU) is well suited for imaging embryonic mice in vivo because it is non-invasive and real-time. Manual segmentation of the Brain Ventricles (BVs) and whole body from 3D HFU images is time-consuming and requires specialized training. This paper presents a deep-learning-based segmentation pipeline which automates several time-consuming, repetitive tasks currently performed to study genetic mutations in developing mouse embryos. Namely, the pipeline accurately segments the BV and body regions in 3D HFU images of mouse embryos, despite significant challenges due to position and shape variation of the embryos, as well as imaging artifacts. Based on the BV segmentation, a 3D convolutional neural network (CNN) is further trained to detect embryos with the Engrailed-1 (En1) mutation. The algorithms achieve 0.896 and 0.925 Dice Similarity Coefficient (DSC) for BV and body segmentation, respectively, and 95.8% accuracy on mutant classification. Through gradient based interrogation and visualization of the trained classifier, it is demonstrated that the model focuses on the morphological structures known to be affected by the En1 mutation.
-
Automatic Mouse Embryo Brain Ventricle & Body Segmentation and Mutant Classification From Ultrasound Data Using Deep Learning
2019 IEEE International Ultrasonics Symposium (IUS), 2019Co-Authors: Ziming Qiu, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey A. Ketterling, Yao WangAbstract:High-frequency ultrasound (HFU) is well suited for imaging embryonic mice in vivo because it is non-invasive and real-time. Manual segmentation of the Brain Ventricles (BVs) and whole body from 3D HFU images is time-consuming and requires specialized training. This paper presents a deep-learning-based segmentation pipeline which automates several time-consuming, repetitive tasks currently performed to study genetic mutations in developing mouse embryos. Namely, the pipeline accurately segments the BV and body regions in 3D HFU images of mouse embryos, despite significant challenges due to position and shape variation of the embryos, as well as imaging artifacts. Based on the BV segmentation, a 3D convolutional neural network (CNN) is further trained to detect embryos with the Engrailed-1 (En1) mutation. The algorithms achieve 0.896 and 0.925 Dice Similarity Coefficient (DSC) for BV and body segmentation, respectively, and 95.8% accuracy on mutant classification. Through gradient based interrogation and visualization of the trained classifier, it is demonstrated that the model focuses on the morphological structures known to be affected by the En1 mutation.