22689 research outputs found
Sort by
An Evaluation of the Impact of Social and Structural Determinants of Health on Forgone Care during the COVID-19 Pandemic in Baltimore, Maryland
Get PDFThe Coronavirus 2019 (COVID-19) pandemic led to widespread disruptions in healthcare utilization. This included forgone care, defined as someone who perceives a need for healthcare but does not receive it. These disruptions exacerbated the morbidity and mortality associated with the pandemic and disproportionately impacted those who experience inequities across the social and structural determinants of health (SSDoH). Existing literature on the impacts of the pandemic on healthcare utilization predominately describe outpatient and hospital trends. However, very few studies have captured patient-reported forgone care.
The purpose of this dissertation was to investigate the impacts of the COVID-19 pandemic on healthcare access and utilization by specifically looking at forgone chronic and preventive care, emergent care, and elective and dental procedures among adults living in Baltimore, MD. Cross-sectional survey data were abstracted from two parent studies that used different, yet complimentary sampling strategies that increased representation of minoritized and under-resourced populations. The resulting combined analytic sample provided a platform to explore forgone care within one urban city that has historically suffered from systemic and structural racism, leading to widespread disparities across the SSDoH.
Several cross-cutting themes emerged as important considerations in this exploration of forgone care during the COVID-19 pandemic. First, individuals experiencing housing instability had higher rates and odds of forgone care when comparing those who reported not experiencing housing instability. Second, under-resourced and marginalized individuals who require frequent engagement with the health system suffered higher rates and odds of forgone care. Third, the correlates of forgone care are likely indicative of existing health disparities. Finally, community-level determinants of health were found not to account for forgone care, above and beyond individual-level factors.
Understanding the overall rate of forgone care during the COVID-19 pandemic and its intersections with the SSDoH provides a more comprehensive view of the health impacts of the pandemic. It can also inform the development of models of care that can help dismantle systems that perpetuate inequities across the SSDoH and that can be leveraged during future public health emergencies to maintain individual and community health
COMPUTATIONAL STUDY OF TRANSCRIPTIONAL LANDSCAPES FROM RNA-SEQ DATA
Get PDFSince the inception of genetic science in the days of Gregor Johann Mendel, the major focus of genetics has been to identify functional units, genes, passed through generations and to determine how variation affects the development of an organism. While genome projects allowed us to collect complete and accurate DNA sequences of individual organisms, novel functional elements are being routinely discovered. When RNA sequencing technologies appeared in the late 2000s, they opened a new view over all transcriptional activity of cells at varying levels of resolution. However, the imperfections of both technical and biological processes, the growing amounts of data as well as the overall complexity of eukaryotic genomes underscore the need for novel analytical approaches to discover new and improve understanding of known genes and isoforms.
This work begins with an overview of the status of human gene annotation and presents a comprehensive discussion of existing and new methods for creating complete and accurate gene catalogs, via comparative genomics and RNA sequencing. Through careful analysis of large RNA-seq datasets, we annotate effects of transcriptional artifacts and inaccuracies in gene expression on downstream analysis. Moreover, we identify defining properties of validated transcription that distinguish it from the effervescent noise. To address the challenges presented by noisy transcription and improve gene expression analysis, we introduce TieBrush, a comprehensive suite of tools for efficient processing of multi-sample sequencing datasets. This suite allows for the creation of condensed representations of data, facilitating the identification of shared transcriptional motifs and enhancing downstream analysis. Furthermore, to enhance our understanding of alternative splicing and distinguish functional isoforms from noise at protein-coding loci, we develop ORFanage. This highly efficient and accurate system assigns open reading frames (ORFs) to gene transcripts, thereby improving gene annotations. Finally, we employ all presented techniques to design a complete sample-to-annotation protocol for annotating genes and transcripts. We apply this protocol to create CHESS 3 - an improved human genome annotation, identifying multiple novel tissue specific isoforms while increasing consistency and reliability of known transcript models.
In addition to advancements in gene annotation, this thesis briefly explores the critical role of large representative datasets of viral genomes in acquiring novel insights into diseases. More specifically, we discuss how pangenomic analysis of HIV-1 facilitated novel insights into viral persistence and how the challenges of mass sequencing of SARS-CoV-2 genomes required a novel approach for identification of first emerging recombinant lineages
Microstructure Optimization and Melt Pool Control in Powder Bed Fusion towards the Development of Architected Lattice Materials
Get PDFPowder bed fusion (PBF) metal additive manufacturing (AM) techniques can shape materials from the micron to centimeter scale and presents the ability to manufacture parts with greater topological complexity than previously allowed by subtractive techniques. The solidification environment varies between electron beam and laser PBF (PBF-EB and PBF-LB, respectively) but is generally characterized by anisotropy, high thermal gradients, and inconsistency in material properties across the scale of the part. This work describes the unique thermal processing conditions of these methods and the microstructures they produce, showing that this anisotropy can manifest through different mechanisms (e.g., residual strains, grain sizes, microstructural morphology) and is driven by part geometry in addition to PBF processing parameters or scan strategy. The relationships between residual strains, microstructural morphology, and laser processing conditions are explored and this enables methods for controlling and influencing the microstructure of a part in real time as it is being manufactured.
This has implications towards the application of fine-featured, highly architected, and topologically optimized metallic micro lattices (MML) materials; with the desirable feature resolution (100-500um) overlapping with the length scale of the PBF melt pool. PBF enables the manufacture of fine features, but the melt pool environment is shown to vary across the build volume of a MML cell which leads to microstructural anisotropy across it. Therefore, the solution proposed and described utilizes control of common laser processing parameters (e.g., power or speed) to control the microstructure across the volume of a given part. The topology of a given MML will vary significantly across its volume, and this research indicates that the microstructure of the lattice will as well; unless a variable scan strategy is adopted which can allow for a more real-time control of microstructure. Using in-situ melt pool sensing, a digital twin of the part is created, and provides a method of mapping how changes to laser process variables can influence the solidification environment and microstructure as a function of this changing topology. This has implications for the functional grading, control, validation, and prediction of microstructures within both MML materials and traditionally engineered parts
SELECTIVE INTERCEPTION OF MAMMALIAN RAB VESICLES BY THE INTRACELLULAR PARASITE TOXOPLASMA GONDII
No full textMembrane trafficking is pivotal for cell function and interactions with microorganisms. Upon infection with the intracellular protozoan Toxoplasma gondii, mammalian cells acquire a novel dynamic compartment – the parasitophorous vacuole (PV). T. gondii incorporates its own proteins, lipids, and selected components from the host cell's plasma membrane making the PV a distinct, specialized ‘organelle’. The unique composition of the PV membrane (PVM) precludes its fusion with mammalian organelles. However, despite the nonfusogenic nature of the PV, we showed that T. gondii salvages lipids from various organelles, prompting the intriguing question: how does the parasite obtain nutrients from host cellular organelles? Chapter 2 shows that Toxoplasma intercepts vesicular trafficking in infected cells and engulfs host organelles and Rab vesicles intact into the PV. These host organelles penetrate the PV through deep invaginations of PVM that are created by the parasite using an intravacuolar network of membranous tubules that fuse with the PVM. Chapter 3 describes a novel method to identify and characterize properties of host Rab vesicles that are sequestered into the Toxoplasma PV. Finally, Chapter 4 examines uptake of specific subsets of Rab11a vesicles by Toxoplasma. Of the class I Rab11-Family Interacting Proteins (FIP), Toxoplasma predominantly targets in a Rab11-dependent fashion FIP2, which regulates the transport of vesicles from the plasma membrane to the recycling endosome. Of Class II effectors, FIP3 and FIP4 are recruited to endosomes in interphase and govern the transport of Rab11 vesicles to the cleavage furrow during cytokinesis. Unlike FIP2, both WT and Rab11-binding domain mutants of FIP3 and FIP4 are intercepted by the parasite. Similarly, FIP3 and FIP4 association with the midbody is Rab11-independent while association with endosomes is Rab11-dependent. This suggests that the PVM contains proteins that selectively recognize subsets of host Rab11a vesicles, prior to PV internalization
ONLINE DECISION MAKING FOR DYNAMICAL SYSTEMS: MODEL-BASED AND DATA-DRIVEN APPROACHES
Get PDFThe widespread availability of data sources and their increased speed compared to the past decade have created both new opportunities and challenges for developing decision-making algorithms for data streams. The ability to process data streams and make real-time decisions that align with system dynamics is a crucial aspect in the development of online decision-making algorithms. This thesis leverages tools from control theory, optimization, and learning to address the problem of online decision-making for dynamical systems, considering streaming data and dynamically changing information.
Two online decision-making frameworks are presented in this thesis, depending on the availability of system dynamic information. In the first scenario, where the system can be represented by ordinary differential equations using a state-space model, a time-varying convex optimization framework is introduced. This framework combines motion planning and control to design control signals that lead the dynamical system to asymptotically track optimal trajectories implicitly defined through constrained time-varying optimization problems. Consequently, the nonlinear dynamical system is effectively transformed into an optimization algorithm that seeks the optimal solution to the optimization problem. Global asymptotic convergence of the optimization dynamics to the minimizer of the time-varying optimization problem is proven under sufficient regularity assumptions.
In the second scenario, when system dynamics are not available, a data-driven approach called constrained reinforcement learning is adopted. Constrained reinforcement learning deals with sequential decision-making problems where an agent aims to maximize its expected total reward while interacting with an unknown environment and receiving sequentially available information over time. The constrained reinforcement learning framework further includes safety constraints or conflicting requirements during the learning process through secondary expected cumulative rewards. To address the limitations of the learning process in constrained reinforcement learning problems, a novel first-order stochastic gradient descent-ascent (GDA) algorithm is proposed: the stochastic dissipative GDA algorithm. This algorithm almost surely converges to the optimal occupancy measure and optimal policy, overcoming the issue of policy oscillation and convergence to suboptimal policies often encountered in C-RL problems
Using Metaphor Analysis to Understand Individuals’ Personal and Familial Experiences with Behavioral Variant Frontotemporal Dementia
No full textFrontotemporal dementia (FTD) is a disease that causes changes in behavior, personality, and language. As researchers better understand the genetic component of FTD, rates of asymptomatic and symptomatic genetic testing are increasing, making it more important to understand the impact of genetic diagnosis on lived experience. Individuals with genetic conditions often use metaphors when they describe their illness experiences. This qualitative study explores how individuals use metaphors to describe personal and familial experiences with behavioral variant FTD (bvFTD) in order to understand what themes may arise in genetic counseling sessions and to accordingly tailor clinical care. The data for this analysis comes from two study populations: asymptomatic individuals with a confirmed genetic risk for bvFTD (n=16) and diagnosed individuals (n=9). Secondary analysis of 25 semi-structured interviews was completed. Abductive analysis and metaphor analysis were used to identify the primary metaphors participants used to describe their experiences with bvFTD. Two of the main themes expressed by metaphors, lack of control over the disease and uncertainty of inheritance and symptoms, consistently conveyed that participants were grappling with how to understand their pasts, presents, and futures. This study suggests that metaphors can provide valuable insight into how patients are revising their life stories when bvFTD causes biographical disruption. These findings have important implications for clinical and research genetic counselors working with individuals with neurodegenerative conditions
Teacher Voice on Social Media: #teachersofTikTok
Get PDFThis dissertation addresses what teachers say they need. It includes a literature review on student trauma, trauma-related classroom behaviors, and teachers’ knowledge and skills in addressing those behaviors. The dissertation also includes two studies: A mixed method needs assessment aimed at identifying teacher concerns and needs and a content analysis of what teachers are saying on TikTok. The needs assessment provides an analysis of survey results, interviews, and secondary data from a charter elementary school in an urban area. Key findings highlighted teachers’ desire for more information related to proactive and reactive strategies to respond to students’ challenging behaviors and a disconnect between teachers’ and school-based staff’s perceptions of what teachers need. A secondary data analysis illustrated a disconnect between school policies and practices, particularly regarding student behavior. The second study used content analysis to examine what teachers are saying on TikTok, a social media platform. Key findings were very specific concerns related to student behavior and professional learning, but not pay, which is often referenced in the literature. Teachers on TikTok relied on humor or hyperbole to discuss student behavior and professional learning. The finding that teachers are deeply concerned with excessive behavior of emotionally and physically abusive students and the lack of professional development and policy to help teachers address it in their classrooms was consistent across the two studies as were teachers’ perceptions that their highly trained professional voices are not heard by stakeholders who both blame them for educational dysfunction and expect them to solve it
Design for Additive Manufacturing: Rapid Response to COVID-19 and Topology Optimization for Materials with Stress-dependent Anisotropies
No full textDuring the initial stages of the COVID-19 pandemic, critical design and engineering challenges quickly arose in response to the needs of the medical community combatting this complex public health crisis. As part of a multidisciplinary team of designers, medical clinicians, and researchers, we looked to address endemic shortages of life-preserving equipment by designing and rapidly prototyping an on-demand multiplex solution to enhance existing ventilator capacity for patients with respiratory complications from SARS‑CoV‑2. Using 3D-printing technology, we helped develop an ad-hoc ventilator splitter concept as an open-source “final-option” treatment approach to this crucial translational design problem.
Time-sensitivity of our emergency collaboration leant itself to informal optimization techniques, but the work referenced above bridges engineering design and 3D-printing/additive manufacturing investigation to the main research focus of this dissertation which is computational design using Topology Optimization. Specifically, conventional topology optimization methods for structures often use isotropic linear elastic material assumptions, which do not consider anisotropies often observed in additively-manufactured materials, including stress state-dependent material properties. Inspired by previous work in our research group, we developed orthotropic stress-dependent material model topology optimization formulations with non-linear iterative finite-element solvers and regularized model-sensitivity function methods. Design examples focusing on tension-compression asymmetries defined in principal and global coordinate system spaces are used to compare this framework with traditional topology optimization methods. Current limitations of these methods, as well as future opportunities for investigation into expanding these formulations to a wider range of applications and/or with a deeper look into material complexity, are also discussed
Place Cells to Path Planning: a Neural Study of Complex Locomotor Behaviors as an Inspiration for Robotics
No full textEven with the advancements of mobile robots in recent years, robots still lag behind animals such as squirrels in performing agile behaviors. Drawing inspiration from animal cognitive planning and navigational strategies has proven beneficial in enhancing robot functionality. Particularly, examining animals' spatial decision-making strategies can aid in advancing robots that can perform complex locomotor tasks.
This dissertation delves into how neuronal activity in the hippocampus, a brain region crucial for spatial cognition, is involved during complex locomotor behaviors.
Previous studies have explored decision-making processes in the hippocampus, predominantly during animal navigation on the surfaces of the experiment rigs. This research builds on that by looking at voluntary animal navigation in 3D spaces. It investigates how hippocampal place cell activity encodes and predicts different 3D trajectories based on the routes taken (retrospective coding) or will be taken (prospective coding) by the animals.
The study explores Long-Evans rats navigating a linear track with an adjustable gap, where they must choose between 'jumping' (crossing over the gap with a single leap) or 'ditching' (jumping into and out of the gap) to cross the gap. Neuropixels 2.0 silicon probes recorded neural activity from the hippocampus in sessions with both jumping and ditching behaviors. Recordings revealed place cell activity during the airborne phase of jumping. Moreover, in sessions involving both jumping and ditching behaviors, place cells exhibited 'splitter-like' behavior by encoding these trajectories differently. For example, some place cells showed strong selectivity for jumping, while others exhibited a strong preference for ditching. Place cells even discriminated between trajectories at locations beyond the gap, indicating retrospective coding. These findings provide evidence that place cells adjust their firing properties to reflect the complex behavioral choices made by animals.
This research also investigates the predictive nature of place cell activity. A Bayesian decoder was trained to predict the animal's behavior based on the average firing rates of place cells in the time interval preceding takeoff. The decoder achieved accuracies ranging from 60% to 85%, significantly surpassing the chance level. This finding demonstrated that place cells encode anticipatory information during the complex locomotor task, enabling the prediction of complex locomotor behaviors solely based on firing rates before takeoff.
In conclusion, this dissertation enhances the understanding of the role of place cells in spatial navigation. It highlights their ability to adapt their firing properties to reflect the structure of complex locomotor tasks. By gaining insights from animals, we can deepen our understanding of spatial cognition and, ultimately, use these findings to create bioinspired algorithms that enhance the functionality of robots in solving complex navigational challenges
Middle School Teachers' Self-Efficacy for Teaching Mathematics to Students with Special Needs
Get PDFStudents with special needs often receive most of their instruction in the general education setting, necessitating coordination of instruction between general education mathematics teachers and special education teachers. Mathematics education and special education teachers approach instruction with different beliefs, content knowledge, pedagogical knowledge, and background experiences. This research sought to understand the efficacy of middle school mathematics and special education teachers to instruct students with special needs. Results of a needs assessment in the spring of 2022 suggested an intervention based on pedagogical content knowledge for teaching mathematics as a viable foundation for an intervention aimed at improving teacher self-efficacy to teach mathematics to students with special needs. This dissertation used a mixed methods research design to study the implementation and evaluation of professional learning on teacher self-efficacy to teach mathematics to middle school students with special needs