A5025843483- Glaucoma and retinal disorders
- Retinal Diseases and Treatments
- Retinal Imaging and Analysis
- Liver Disease Diagnosis and Treatment
- Hepatocellular Carcinoma Treatment and Prognosis
- Radiomics and Machine Learning in Medical Imaging
- Chronic Obstructive Pulmonary Disease (COPD) Research
- EEG and Brain-Computer Interfaces
- Lung Cancer Diagnosis and Treatment
- MRI in cancer diagnosis
- Ultrasound in Clinical Applications
- AI in cancer detection
- COVID-19 diagnosis using AI
- Inhalation and Respiratory Drug Delivery
- Phonocardiography and Auscultation Techniques
- Neuroscience and Neural Engineering
- Neural dynamics and brain function
- Advanced MRI Techniques and Applications
- Cerebral Venous Sinus Thrombosis
- Epilepsy research and treatment
- Retinal and Optic Conditions
- Respiratory Support and Mechanisms
- Brain Tumor Detection and Classification
- Cardiovascular Health and Disease Prevention
- Advanced Clustering Algorithms Research
University of California, San Diego
2017-2024
San Diego State University
2020-2024
University of Liverpool
2023
Medical University of South Carolina
2023
University of Bonn
2023
Emory University
2023
Rush University
2023
Cleveland Clinic
2023
National Jewish Health
2021
University of Colorado Denver
2021
To assess feasibility of training a convolutional neural network (CNN) to automate liver segmentation across different imaging modalities and techniques used in clinical practice apply this enable automation biometry.We trained 2D U-Net CNN for two stages using 330 abdominal MRI CT exams acquired at our institution. First, we the with non-contrast multi-echo spoiled-gradient-echo (SGPR)images 300 provide multiple signal-weightings. Then, transfer learning generalize additional images from 30...
Abstract We propose a random forest classifier for identifying adequacy of liver MR images using handcrafted (HC) features and deep convolutional neural networks (CNNs), analyze the relative role these two components in relation to training sample size. The HC features, specifically developed this application, include Gaussian mixture models, Euler characteristic curves texture analysis. Using outperforms CNN smaller sizes with increased interpretability. On other hand, enough data, combined...
Statistical learning is characterized by detection of regularities in one's environment without an awareness or intention to learn, and it may play a critical role language social behavior. Accordingly, this study we investigated the electrophysiological correlates visual statistical young children with autism spectrum disorder (ASD) using event-related potential shape paradigm, examined relation between cognitive function. Compared typically developing (TD) controls, ASD group as whole...
Purpose To develop a deep learning–based algorithm to stage the severity of chronic obstructive pulmonary disease (COPD) through quantification emphysema and air trapping on CT images assess ability proposed stages prognosticate 5-year progression mortality. Materials Methods In this retrospective study, an using co-registration lung segmentation was developed in-house automate from inspiratory expiratory images. The then tested in separate group 8951 patients COPD Genetic Epidemiology study...
The electroencephalography (EEG) data created in event-related potential (ERP) experiments have a complex high-dimensional structure. Each stimulus presentation, or trial, generates an ERP waveform which is instance of functional data. are made up sequences multiple trials, resulting longitudinal and moreover, responses recorded at electrodes on the scalp, adding electrode dimension. Traditional EEG analyses involve simplifications this structure to increase signal-to-noise ratio,...
Convolutional neural networks (CNNs) are increasingly being explored and used for a variety of classification tasks in medical imaging, but current methods post hoc explainability limited. Most commonly highlight portions the input image that contribute to classification. While this provides form spatial localization relevant focal disease processes, it may not be sufficient co-localized or diffuse processes such as pulmonary edema fibrosis. For latter, new required isolate texture features...
BACKGROUND. Histologic fibrosis stage is the most important prognostic factor in chronic liver disease. MR elastography (MRE) accurate noninvasive method for detecting and staging fibrosis. Although accurate, manual ROI-based MRE analysis complex, time-consuming, requires specialized readers, prone to methodologic variability suboptimal interreader agreement. OBJECTIVE. The purpose of this study was develop an automated convolutional neural network (CNN)-based analysis, evaluate its...
The confounder-corrected chemical shift-encoded MRI (CSE-MRI) sequence used to determine proton density fat fraction (PDFF) for hepatic quantification is not widely available. As an alternative, can be assessed by a two-point Dixon method calculate signal (FF) from conventional T1-weighted in- and opposed-phase (IOP) images, although FF prone biases, leading inaccurate quantification.
<h3>Importance</h3> Certain features of the lamina cribrosa may be associated with increased risk glaucoma progression. <h3>Objectives</h3> To compare rates retinal nerve fiber layer (RNFL) thinning in patients open-angle or without (LC) defects and to evaluate factors rate progression eyes LC defects. <h3>Design, Setting, Participants</h3> This longitudinal cohort study designed September 2017 conducted at a tertiary center California included 51 43 83 68 followed up for mean (SD) 3.5 (0.8)...
Summary Differential brain response to sensory stimuli is very small (a few microvolts) compared the overall magnitude of spontaneous electroencephalogram (EEG), yielding a low signal-to-noise ratio (SNR) in studies event-related potentials (ERP). To cope with this phenomenon, are applied repeatedly and ERP signals arising from individual trials averaged at subject level. This results loss information about potentially important changes form over course experiment. In article, we develop...
A new frontier in diagnostic radiology is the inclusion of machine-assisted support tools that facilitate identification subtle lesions often not visible to human eye. Structural neuroimaging plays an essential role patients with epilepsy, which coincide seizure focus. In this study, we explored potential for a convolutional neural network (CNN) determine lateralization onset epilepsy using T1-weighted structural MRI scans as input.
Liver alignment between series/exams is challenged by dynamic morphology or variability in patient positioning motion. Image registration can improve image interpretation and lesion co-localization. We assessed the performance of a convolutional neural network algorithm to register cross-sectional liver imaging series compared its manual registration.Three hundred fourteen patients, including internal external datasets, who underwent gadoxetate disodium-enhanced magnetic resonance for...
Explainability of convolutional neural networks (CNNs) is integral for their adoption into radiological practice. Commonly used attribution methods localize image areas important CNN prediction but do not characterize relevant imaging features underlying these areas, acting as a barrier to the CNNs clinical use. We therefore propose Semantic Exploration and using Style-based Generative Adversarial Autoencoder Network (SEE-GAAN), an explainability framework that uses latent space manipulation...
To compare optical coherence tomography angiography (OCTA) measured macular vessel density and spectral domain (SDOCT) ganglion cell complex (GCC) thickness in primary open-angle glaucoma eyes with without focal lamina cribrosa (LC) defects.
Abstract Motivated by a study on visual implicit learning in young children with Autism Spectrum Disorder (ASD), we propose robust functional clustering (RFC) algorithm to identify subgroups within electroencephalography (EEG) data. The proposed RFC is an iterative based principal component analysis, where cluster membership updated via predictions of the trajectories obtained through non-parametric random effects model. We consider data resulting from event-related potential (ERP) waveforms...
To develop a convolutional neural network (CNN)-based deformable lung registration algorithm to reduce computation time and assess its potential for lobar air trapping quantification.
Introduction: The legalization of cannabis products has increased their usage in the United States. Among ∼500 active compounds, this is especially true for cannabidiol (CBD)-based products, which are being used to treat a range ailments. Research ongoing regarding safety, therapeutic potential, and molecular mechanism cannabinoids. Drosophila (fruit flies) widely model factors that impact neural aging, stress responses, longevity. Materials Methods: Adult wild-type melanogaster cohorts...
Purpose: To assess if a novel automated method to spatially delineate and quantify the extent of hypoperfusion on multienergy CT angiograms can aid evaluation chronic thromboembolic pulmonary hypertension (CTEPH) disease severity. Materials Methods:Multienergy obtained between January 2018 December 2020 in 51 patients with CTEPH (mean age, 47 years ± 17 [SD]; 27 women) were retrospectively compared those 110 controls no imaging findings suggestive vascular abnormalities 16; 81...