检索结果-内蒙古大学图书馆

Personalized and uncertainty-aware coronary hemodynamics simulations: From Bayesian estimation to improved multi-fidelity uncertainty quantification

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Menon, Karthik Zanoni, Andrea Khan, Owais Geraci, Gianluca Nieman, Koen Schiavazzi, Daniele E. Marsden, Alison L. Stanford School of Medicine StanfordCA United States Institute for Computational and Mathematical Engineering Stanford University StanfordCA United States Department of Electrical Computer and Biomedical Engineering Toronto Metropolitan University TorontoON Canada Center for Computing Research Sandia National Laboratories AlbuquerqueNM United States Division of Cardiovascular Medicine Stanford School of Medicine StanfordCA United States Department of Radiology Stanford School of Medicine StanfordCA United States Department of Applied and Computational Mathematics and Statistics University of Notre Dame Notre DameIN United States Department of Bioengineering Stanford University StanfordCA United States

Background: Non-invasive simulations of coronary hemodynamics have improved clinical risk stratification and treatment outcomes for coronary artery disease, compared to relying on anatomical imaging alone. However, simulations typically use empirical approaches to distribute total coronary flow amongst the arteries in the coronary tree, which ignores patient variability, the presence of disease, and other clinical factors. Further, uncertainty in the clinical data often remains unaccounted for in the modeling pipeline. Objective: We present an end-to-end uncertainty-aware pipeline to (1) personalize coronary flow simulations by incorporating vessel-specific coronary flows as well as cardiac function;and (2) predict clinical and biomechanical quantities of interest with improved precision, while accounting for uncertainty in the clinical data. Methods: We assimilate patient-specific measurements of myocardial blood flow from clinical CT myocardial perfusion imaging to estimate branch-specific coronary artery flows. Simulated noise in the clinical data is used to estimate the joint posterior distributions of the model parameters using adaptive Markov Chain Monte Carlo sampling. Additionally, the posterior predictive distribution for the relevant quantities of interest is determined using a new approach combining multi-fidelity Monte Carlo estimation with non-linear, data-driven dimensionality reduction. This leads to improved correlations between high- and low-fidelity model outputs. Results: Our framework accurately recapitulates clinically measured cardiac function as well as branch-specific coronary flows under measurement noise uncertainty. We observe substantial reductions in confidence intervals for estimated quantities of interest compared to single-fidelity Monte Carlo estimation and state-of-the-art multi-fidelity Monte Carlo methods. This holds especially true for quantities of interest that showed limited correlation between the low- and high-fidelity model

关键词： Monte Carlo methods

A Generative Modeling Approach for Interpreting Population-Level Variability in Brain Structure 23rd

学校读者我要写书评

暂无评论

A Generative Modeling Approach for Interpreting Population-L...

23rd International Conference on Medical Image Computing and computer-Assisted Intervention, MICCAI 2020

作者： Liu, Ran Subakan, Cem Balwani, Aishwarya H. Whitesell, Jennifer Harris, Julie Koyejo, Sanmi Dyer, Eva L. School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta United States Mila - Quebec Artificial Intelligence Institute Montreal Canada Neuroanatomy Division Allen Institute for Brain Science Seattle United States Department of Computer Science University of Illinois at Urbana Champaign Urbana United States Department of Biomedical Engineering Georgia Institute of Technology Atlanta United States

ISBN: (纸本)9783030597214

Understanding how neural structure varies across individuals is critical for characterizing the effects of disease, learning, and aging on the brain. However, disentangling the different factors that give rise to individual variability is still an outstanding challenge. In this paper, we introduce a deep generative modeling approach to find different modes of variation across many individuals. Our approach starts with training a variational autoencoder on a collection of auto-fluorescence images from a little over 1,700 mouse brains at 25μ m resolution. We then tap into the learned factors and validate the models expressiveness, via a novel bi-directional technique that makes structured perturbations to both, the high-dimensional inputs of the network, as well as the low-dimensional latent variables in its bottleneck. Our results demonstrate that through coupling generative modeling frameworks with structured perturbations, it is possible to probe the latent space of the generative model to provide insights into the representations of brain structure formed in deep networks. © 2020, Springer Nature Switzerland AG.

关键词： Taps

Dynamic nitrogen vacancy magnetometry by single-shot optical streaking microscopy

学校读者我要写书评

暂无评论

Photonics Research 2022年第9期10卷 2147-2156页

作者： MARK A.KEPPLER ZACHARY A.STEELMAN ZACHARY N.COKER MILOŠNESLÁDEK PHILIP R.HEMMER VLADISLAV V.YAKOVLEV JOEL N.BIXLER Department of Biomedical Engineering Texas A&M UniversityCollege StationTexas 77843USA SAIC JBSA Fort Sam HoustonTexas 78234USA National Research Council Research Associateship Program WashingtonDC 20001USA IMOMEC Division IMECB-3590 DiepenbeekBelgium Institute for Materials Research(IMO) Hasselt UniversityB-3590 DiepenbeekBelgium Czech Technical University in Prague 27201 KladnoCzech Republic Department of Electrical and Computer Engineering Texas A&M UniversityCollege StationTexas 77843USA Bioeffects Division Airman System DirectorateAir Force Research LaboratoryJBSA Fort Sam HoustonTexas 78234USA

Nitrogen vacancy diamonds have emerged as sensitive solid-state magnetic field sensors capable of producing diffraction limited and sub-diffraction field ***,for the first time,to our knowledge,we extend those measurements to high-speed imaging,which can be readily applied to analyze currents and magnetic field dynamics in circuits on a microscopic *** overcome detector acquisition rate limitations,we designed an optical streaking nitrogen vacancy microscope to acquire two-dimensional spatiotemporal *** demonstrate magnetic field wave imaging with micro-scale spatial extent and~400μs temporal *** validating this system,we detected magnetic fields down to 10μT for 40 Hz magnetic fields using single-shot imaging and captured the spatial transit of an electromagnetic needle at streak rates as high as 110μm/*** design has the capability to be readily extended to full 3D video acquisition by utilizing compressed sensing techniques and a potential for further improvement of spatial resolution,acquisition speed,and *** device opens opportunities to many potential applications where transient magnetic events can be isolated to a single spatial axis,such as acquiring spatially propagating action potentials for brain imaging and remotely interrogating integrated circuits.

关键词： vacancy utilizing overcome

Deep networks as logical circuits: Generalization and interpretation

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Snyder, Christopher Vishwanath, Sriram Department of Biomedical Engineering University of Texas Austin United States Department of Electrical and Computer Engineering University of Texas Austin United States

Not only are Deep Neural Networks (DNNs) black box models, but also we frequently conceptualize them as such. We lack good interpretations of the mechanisms linking inputs to outputs. Therefore, we find it difficult to analyze in human-meaningful terms (1) what the network learned and (2) whether the network learned. We present a hierarchical decomposition of the DNN discrete classification map into logical (AND/OR) combinations of intermediate (True/False) classifiers of the input. Those classifiers that can not be further decomposed, called atoms, are (interpretable) linear classifiers. Taken together, we obtain a logical circuit with linear classifier inputs that computes the same label as the DNN. This circuit does not structurally resemble the network architecture, and it may require many fewer parameters, depending on the configuration of weights. In these cases, we obtain simultaneously an interpretation and generalization bound (for the original DNN), connecting two fronts which have historically been investigated separately. Unlike compression techniques, our representation is exact. We motivate the utility of this perspective by studying DNNs in simple, controlled settings, where we obtain superior generalization bounds despite using only combinatorial information (e.g. no margin information). We demonstrate how to "open the black box" on the MNIST dataset. We show that the learned, internal, logical computations correspond to semantically meaningful (unlabeled) categories that allow DNN descriptions in plain English. We improve the generalization of an already trained network by interpreting, diagnosing, and replacing components within the logical circuit that is the DNN. Copyright © 2020, The Authors. All rights reserved.

关键词： Deep neural networks

Combining imitation and deep reinforcement learning to accomplish human-level performance on a virtual foraging task

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Giammarino, Vittorio Dunne, Matthew F. Moore, Kylie N. Hasselmo, Michael E. Stern, Chantal E. Paschalidis, Ioannis Ch. Division of Systems Engineering Boston University BostonMA02446 United States Dept. of Electrical and Computer Engineering Boston University BostonMA02446 United States Dept. of Biomedical Engineering Boston University BostonMA02215 United States Cognitive Neuroimaging Center Boston University BostonMA02215 United States Graduate Program for Neuroscience Boston University BostonMA02215 United States Center for Systems Neuroscience Boston University BostonMA02215 United States

We develop a simple framework to learn bio-inspired foraging policies using human data. We conduct an experiment where humans are virtually immersed in an open field foraging environment and are trained to collect the highest amount of rewards. A Markov Decision Process (MDP) framework is introduced to model the human decision dynamics. Then, Imitation Learning (IL) based on maximum likelihood estimation is used to train Neural Networks (NN) that map human decisions to observed states. The results show that passive imitation substantially underperforms humans. We further refine the human-inspired policies via Reinforcement Learning (RL) using the on-policy Proximal Policy Optimization (PPO) algorithm which shows better stability than other algorithms and can steadily improve the policies pretrained with IL. We show that the combination of IL and RL can match human results and that good performance strongly depends on combining the allocentric information with an egocentric representation of the environment. © 2022, CC BY.

关键词： Reinforcement learning

Super-resolution Radial Fluctuations Enables Polarization-resolved Nonlinear Optical Nanoscopy

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Harvey, MacAulay Cisek, Richard Al-Turk, Sarry Ruda, Harry E. Kreplak, Laurent Tokarz, Danielle Department of Chemistry Saint Mary’s University 923 Robie Street HalifaxNSB3H 3C3 Canada Centre for Advanced Nanotechnology University of Toronto 170 College Street TorontoONM5S 3E4 Canada Department of Materials Science and Engineering University of Toronto 184 College Street TorontoONM5S 3E4 Canada Department of Electrical and Computer Engineering University of Toronto 10 Kings College Road TorontoONM5S 3G4 Canada Department of Physics and Atmospheric Science School of Biomedical Engineering Dalhousie University HalifaxNSB3H 4J5 Canada

Second harmonic generation microscopy (SHG) is a powerful imaging modality which has found applications in investigating both biological and synthetic nanostructures. Like all optical microscopy techniques, the resolution of SHG is limited to approximately half the wavelength of the excitation light. Because of this several groups have proposed techniques to enable super-resolution SHG imaging. However, these techniques often involve quite complicated optical setups compared to standard SHG microscopes, a major impediment towards more widespread utilization. Here we apply super-resolution radial fluctuations (SRRF), a commonly used technique for super-resolution fluorescence imaging, to enable super-resolution SHG microscopy. By imaging individual nanostructures, we demonstrate that SRRF can provide resolution enhancement of up to 3× compared to a laser scanning SHG microscope, which is comparable to the best resolution enhancement reported in the literature. Additionally, we show that SRRF maintains the polarization dependence of SHG, therefore enabling super-resolution polarization SHG imaging. Finally, we perform SRRF processing on third harmonic generation images to demonstrate the significant potential of SRRF for other super-resolution nonlinear optical microscopy. Importantly, since SRRF can achieve super-resolution purely through image processing, the technique demonstrated here could be used to enhance the resolution of images obtained using a wide variety of nonlinear optical microscopy setups including both laser scanning and widefield configurations. Copyright © 2025, The Authors. All rights reserved.

关键词： Nonlinear optics

A Bioinformatic Study of Genetics Involved in Determining Mild Traumatic Brain Injury Severity and Recovery

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Tajik, Mahnaz Noseworthy, Michael D. School of Biomedical Engineering McMaster University 1280 Main St W HamiltonON Canada Imaging Research Centre St. Joseph’s Healthcare Hamilton 50 Charlton Ave E HamiltonON Canada Department of Medical Sciences McMaster University 1280 Main St W HamiltonON Canada Department of Electrical and Computer Engineering McMaster University 1280 Main St W HamiltonON Canada Department of Radiology McMaster University 1280 Main St W HamiltonON Canada St Joseph's Healthcare Hamilton Fontbonne Building Room 130A 50 Charlton Ave E HamiltonONL8N 4A6 Canada

Aim: This in silico study sought to identify specific biomarkers for mild traumatic brain injury (mTBI) through the analysis of publicly available gene and miRNA databases, hypothesizing their influence on neuronal structure, axonal integrity, and regeneration. Methods: This study implemented a three-step process: (1) Data searching for mTBI-related genes in Gene and MalaCard databases and literature review;(2) Data analysis involved performing functional annotation through GO and KEGG, identifying hub genes using Cytoscape, mapping protein-protein interactions via DAVID and STRING, and predicting miRNA targets using miRSystem, miRWalk2.0, and mirDIP (3) RNA-sequencing analysis applied to the mTBI dataset ***: Eleven candidate hub genes associated with mTBI outcome were identified: APOE, S100B, GFAP, BDNF, AQP4, COMT, MBP, UCHL1, DRD2, ASIC1, and CACNA1A. Enrichment analysis linked these genes to neuron projection regeneration and synaptic plasticity. miRNAs linked to the mTBI candidate genes were hsa-miR-9-5p, hsa-miR-204-5p, hsa-miR-1908-5p, hsa-miR-16-5p, hsa-miR-10a-5p, has-miR-218-5p, has-miR-34a-5p, and has-miR-199b-5p. The RNA sequencing revealed 2664 differentially expressed miRNAs post-mTBI, with 17 showing significant changes at the time of injury and 48 hours post-injury. Two miRNAs were positively correlated with direct head hits. Conclusion: Our study indicates that specific genes and miRNAs, particularly hsa-miR-10a-5p, may influence mTBI outcomes. Our research may guide future mTBI diagnostics, emphasizing the need to measure and track these specific genes and miRNAs in diverse cohorts. © 2025, CC BY.

关键词： Gene encoding

Stimulated Raman scattering signal generation in a scattering medium using self-reconstructing Bessel beams

学校读者我要写书评

暂无评论

Photonics Research 2020年第6期8卷 929-939页

作者： XUELI CHEN XINYU WANG LIN WANG PENG LIN YONGHUA ZHAN JI-XIN CHENG Engineering Research Center of Molecular&Neuro Imaging Minitry of EducationXi'an 710126China School of Life Science and Technology Xidian UniversityXi'an 710126China School of Information Sciences and Technology Northwest UniversityXi'an 710127China Department of Electrical and Computer Engineering&Biomedical Engineering Boston UniversityBostonMassachusetts 02215USA

Scattering is a huge challenge for microscopic ***,it is difficult to observe target chemicals in scattering media by means of the current Gaussian beam-based stimulated Raman scattering(SRS)microscopy,since the tight focus of the Gaussian beam is destroyed after propagating through a certain *** beams,featuring self-reconstructing property,may bring a solution to this *** combining Bessel beams with SRS microscopy,we can probe the SRS signal from a scattering *** this paper,using the beam propagation method,we first simulate the propagation of the Bessel beam as well as the generation and self-reconstruction of SRS *** adding glass beads along the beam propagation path in order to simulate scattering,the propagation of the Bessel beams and the generation of the SRS signals will ***,we design a series of simulations to investigate the influence of the size,position,number,and distribution of the added glass beads on the generation of the SRS signals.A preliminary experiment is also carried out to confirm the simulation *** demonstrate that the SRS signals can be generated or be recovered at a certain depth in scattering media,and that such signals are greatly affected by the parameters of the scatters.

关键词： Bessel beads destroyed

Diagnostic Tools for Detecting Autism Spectrum Disorder: A Review

学校读者我要写书评

暂无评论

Diagnostic Tools for Detecting Autism Spectrum Disorder: A R...

electrical, Energy, and Artificial Intelligence (EICEEAI), International engineering Conference on

作者： Khosro Rezaee Mohammad R. Khosravi Hossein Ghayoumi Zadeh Mohammad Kazem Moghimi Ghassan Samara Hani Attar Sattam Almatarneh Deprtement of Biomedical Enginering Meybod University Meybod Iran Weifang University of Science and Technology Weifang Shandong China Department of Electrical Engineering Vali-E-Asr University of Rafsanjan Rafsanjan Iran Department of Communication Engineering University of Sistan and Baluchestan Zahedan Iran Department of Computer Science Zerqa University Zerqa Jordan Department of Energy Engineering Faculty of Enineering Zarqa University Zerqa Jordan Data Science and Artificial Intelligence Department Faculty of Information Technology Zarqa University Zerqa Jordan

The autism spectrum disorder (ASD) is a developmental disability caused by abnormalities in the brain. Different methods have been used to diagnose ASD in the past, each with its own advantages and disadvantages. Prior research has focused mainly on algorithms and processing methods, rather than identification tools, to improve the automated diagnosis of ASD. This article identifies and describes diagnostic tools in the ASD literature, including facial features, EEG recordings, speech signals, and neuroimaging, in order to assist researchers interested in developing statistical, computational, and sound clinical approaches to ASD data mining. Accordingly, based on the responses obtained, this review study intends to assess several aspects of the key automatic diagnostic tools for autism spectrum disorders, including diagnostic accuracy, privacy protection, uncertainty, cost, efficiency, absence of methodological interference, and prospective clinical and therapeutic conditions.

关键词： Neuroimaging Autism Privacy Costs Uncertainty Brain modeling Electroencephalography