Search for: All records

IntroductionRecent efforts including the U.S. Department of Education’sRaise the Bar: STEM Excellence for All Students, designed to strengthen Science, Technology, Engineering and Mathematics (STEM) education, typify the development of effective outreach programs implemented in high school settings to increase STEM achievement and literacy and to promote future careers in STEM. Specifically, artificial intelligence (AI) and machine learning (ML) are topics of great importance and interest but are often reserved for higher-level education. Introductions of complex subjects in high school promotes student efficacy, enthusiasm, and skill-development for STEM careers. Establishing strong partnerships between universities and high schools is mutually beneficial for the professional development of students, teachers, and professors. In this paper, we detail immersive outreach efforts and their effectiveness in a high school setting. MethodsFrom Spring 2021 to Spring 2024, we conducted eight data-science and analysis-coding style workshops along with two data science units, with 302 students participating in the data science workshops and 82 students in the data science units. All students who participated in the data science lessons completed a comprehensive final project. Surveys measuring knowledge and appeal to data science and coding were conducted both retrospectively and prospectively, before and after each workshop and the data science units. A 1 year follow up survey was conducted for students in the 2023 data science lessons (n= 23). ResultsOverall, average student interest significantly increased from 2.72 ± 1.08/5.0 (n= 205) to 3.15 ± 1.18/5.0 (n= 181,p= 0.001) during the data science workshops, while 70% of students expressed desire to continue with coding. Interest modestly increased in the data science lessons from 3.15 ± 0.65/4.0 to 3.17 ± 0.77/4.0 (n= 82,p= 0.8571), while knowledge significantly increased from 64.16% to 88.5% (% correct out of six questions) in the 2023 data science lessons and from 52.62% to 60.79% (% correct out of 29 questions) in the 2024 data science lessons. DiscussionIncreasing STEM exposure through outreach programs and a modified curriculum can positively alter students’ career trajectory and prepare them for the evolving technologically advanced world and the careers within it. 

Given the pronounced impact COVID-19 continues to have on society—infecting 700 million reported individuals and causing 6.96 million deaths—many deep learning works have recently focused on the virus’s diagnosis. However, assessing severity has remained an open and challenging problem due to a lack of large datasets, the large dimensionality of images for which to find weights, and the compute limitations of modern graphics processing units (GPUs). In this paper, a new, iterative application of transfer learning is demonstrated on the understudied field of 3D CT scans for COVID-19 severity analysis. This methodology allows for enhanced performance on the MosMed Dataset, which is a small and challenging dataset containing 1130 images of patients for five levels of COVID-19 severity (Zero, Mild, Moderate, Severe, and Critical). Specifically, given the large dimensionality of the input images, we create several custom shallow convolutional neural network (CNN) architectures and iteratively refine and optimize them, paying attention to learning rates, layer types, normalization types, filter sizes, dropout values, and more. After a preliminary architecture design, the models are systematically trained on a simplified version of the dataset-building models for two-class, then three-class, then four-class, and finally five-class classification. The simplified problem structure allows the model to start learning preliminary features, which can then be further modified for more difficult classification tasks. Our final model CoSev boosts classification accuracies from below 60% at first to 81.57% with the optimizations, reaching similar performance to the state-of-the-art on the dataset, with much simpler setup procedures. In addition to COVID-19 severity diagnosis, the explored methodology can be applied to general image-based disease detection. Overall, this work highlights innovative methodologies that advance current computer vision practices for high-dimension, low-sample data as well as the practicality of data-driven machine learning and the importance of feature design for training, which can then be implemented for improvements in clinical practices. 

BackgroundCOVID-19 is constantly evolving, and highly populated communities consist of many different characteristics that may contribute to COVID-19 health outcomes. Therefore, we aimed to (1) quantify the relationships between county characteristics and severe and non-severe county-level health outcomes related to COVID-19. We also aimed to (2) compare these relationships across time periods where the Delta (B.1.617.2) and Omicron (B.1.1.529 and BA.1.1) variants were dominant in the U.S. MethodsWe used multiple regression to measure the strength of relationships between healthcare outcomes and county characteristics in the 50 most populous U.S. counties. ResultsWe found many different significant predictors including the proportion of a population vaccinated, median household income, population density, and the proportion of residents aged 65+, but mainly found that socioeconomic factors and the proportion of a population vaccinated play a large role in the dynamics of the spread and severity of COVID-19 in communities with high populations. DiscussionThe present study shines light on the associations between public health outcomes and county characteristics and how these relationships change throughout Delta and Omicron’s dominance. It is important to understand factors underlying COVID-19 health outcomes to prepare for future health crises. 

Abstract: In 2019, a series of novel pneumonia cases later known as Coronavirus Disease 2019 (COVID-19) were reported in Wuhan, China. Chest computed tomography (CT) has played a key role in the management and prognostication in COVID-19 patients. CT has demonstrated 98%sensitivity in detecting COVID-19, including identifying lung abnormalities that are suggestive of COVID-19, even among asymptomatic individuals. Methods: We conducted a comprehensive literature review of 17 published studies, including focuses on three subgroups, pediatric patients, pregnant women, and patients over 60 years old, to identify key characteristics of chest CT in COVID-19 patients. Results: Our comprehensive review of the 17 studies concluded that the main CT imaging finding is ground glass opacities (GGOs) regardless of patient age. We also identified that crazy paving pattern, reverse halo sign, smooth or irregular septal thickening, and pleural thickening may serve as indicators of disease progression. Lesions on CT scans were dominantly distributed in the peripheral zone with multilobar involvement, specifically concentrated in the lower lobes. In the patients over 60 years old, the proportion of substantial lobe involvement was higher than the controlgroup and crazy paving signs, bronchodilation, and pleural thickening were more commonly present. Conclusion: Based on all 17 studies, CT findings in COVID-19 have shown a predictable pattern of evolution over the disease. These studies have proven that CT may be an effective approach for early screening and detection of COVID-19. 

Laboratory tests seeking to improve detection of COVID-19 have been widely developed by laboratories and commercial companies. This review provides an overview of molecular and antigen tests, presents the sensitivity and specificity for 329 assays that have received US FDA Emergency Use Authorization and evaluates six sample collection methods – nasal, nasopharyngeal, oropharyngeal swabs, saliva, blood and stool. Molecular testing is preferred for diagnosis of COVID-19, but negative results do not always rule out the presence of infection, especially when clinical suspicion is high. Sensitivity and specificity ranged from 88.1 to 100% and 88 to 100%, respectively. Antigen tests may be more easy to use and rapid. However, they have reported a wide range of detection sensitivities from 16.7 to 85%, which may potentially yield many false-negative results. 

Abstract As data sharing has become more prevalent, three pillars - archives, standards, and analysis tools - have emerged as critical components in facilitating effective data sharing and collaboration. This paper compares four freely available intracranial neuroelectrophysiology data repositories: Data Archive for the BRAIN Initiative (DABI), Distributed Archives for Neurophysiology Data Integration (DANDI), OpenNeuro, and Brain-CODE. The aim of this review is to describe archives that provide researchers with tools to store, share, and reanalyze both human and non-human neurophysiology data based on criteria that are of interest to the neuroscientific community. The Brain Imaging Data Structure (BIDS) and Neurodata Without Borders (NWB) are utilized by these archives to make data more accessible to researchers by implementing a common standard. As the necessity for integrating large-scale analysis into data repository platforms continues to grow within the neuroscientific community, this article will highlight the various analytical and customizable tools developed within the chosen archives that may advance the field of neuroinformatics.