50107 research outputs found
Sort by
FEEDING BEHAVIOR AND LIPID REGULATION IN Caenorhabditis elegans: A FUNCTIONAL AND STRUCTURAL APPROACH TO BIOACTIVE COMPOUND SCREENING
No full textFeeding behavior and lipid metabolism are tightly regulated biological processes that are essential for organismal health and are conserved across species. This project employs Caenorhabditis elegans as a model organism to investigate behavioral and metabolic responses to various bioactive compounds, with an emphasis on food preference and lipid regulation. In Chapter 1, a comprehensive review highlights how C. elegans serves as a powerful model for studying diverse feeding behaviors, such as food preference, avoidance, satiety, and social feeding, while emphasizing its translational relevance to human health. Chapter 2 introduces a high-throughput food preference assay in C. elegans to evaluate the effects of opioid and cannabinoid receptor modulators, demonstrating that the nematode’s homologous receptors (npr-17 and npr-19, respectively) respond similarly to human μ-opioid receptor and cannabinoid 1 receptor, making it a promising protocol for screening substances related to substance use disorder. Chapter 3 explores the fat-reducing effects of esculetin, a natural coumarin derivative, showing that esculetin significantly reduces triglyceride accumulation in C. elegans through insulin/insulin-like growth factor-1 and AMP-activated protein kinase (AMPK) signaling pathways without impairing locomotion activity. Finally, chapter 4 would expand feeding preference-based approach by examining the structure-activity relationship of coumarin derivatives on food preference behaviors, revealing that specific substitutions, particularly hydroxyl and methyl groups at defined positions, can modulate behavior in a structure-dependent manner. Collectively, this project highlights the utility of C. elegans for screening bioactive compounds and uncovering molecular mechanisms relevant to feeding behavior and lipid metabolism, providing translational insights for therapeutic development in metabolic and behavioral disorders.Doctor of Philosophy (Ph.D.
Secrecy and Privacy in Communications and Distributed Learning Systems
No full textWith the rapid development of network and distributed systems, security and privacy have become primary concerns. Various data encryptions and network infrastructures have been proposed to improve security and privacy performance. For example, such mechanisms were built upon cryptography, security protocols, or decentralized architectures to decrease the risk of the centralized option. Information theory is the fundamental quantity that measures the information sharing between different parties. In this dissertation, we discuss how information theory can be combined with the proposed encryption algorithms to guarantee the secrecy and privacy of the network and distributed systems.
Secrecy in communication has been considered to ensure safe transmission of information considering a passive eavesdropper. Traditional information-theoretic security methods require an advantage of the main channel over the eavesdropper channel to achieve a positive secrecy rate. However, such an advantage cannot always be guaranteed in a realistic scenario of wireless communications. This dissertation develops a receiver for a novel approach that exploits information sharing between a communication system and a co-existing radar to instantiate a technique for everlasting security and considers improving the secrecy performance with receiver impairment scenarios by employing multi-look channel and machine learning (ML)-based decoding framework.
In addition to security, privacy has recently attracted attention in communication and network systems. Complying with the privacy rules becomes important as we make a transition to a data-driven world. The risk of a user identity being disclosed increases as more data is collected or used improperly. Simultaneously, maintaining utility for application purposes imposes challenges for privacy compliance. In this dissertation, we focus on designing data privacy protection mechanisms without necessarily hurting utility in distributed systems. User privacy can be compromised by matching user data traces to records of their previous behavior, called adversarial pattern matching attacks. Our goal is to thwart such an adversary by applying small artificial distortions to data traces such that each potentially identifying pattern is shared by a large number of users. In addition, we examine optimal randomization methods for obfuscating reports of distributed learning systems to preserve the privacy of client attributes while maintaining utility at the central server.Doctor of Philosophy (Ph.D.
Mapping Cultural Signatures: A Cross-national Study of K-12 Student Math Game Design and Computational Thinking Education
No full textThe integration of culture into computing education remains an underexplored frontier, particularly in contexts where students are asked to design their own educational games as expressions of computational thinking. This doctoral dissertation investigates how culturally informed elements—termed “cultural signatures”—influence students’ design of educational games in computing and programming education contexts.
In the context of this dissertation, cultural signatures are defined as the visible traces of cultural manifestations—artifacts of meaning, memory, and identity that students consciously or unconsciously encode into their game designs through visual, narrative, and structural choices. These signatures may appear in the form of familiar gameplay metaphors, localized symbols, collaborative norms, linguistic expressions, or aesthetic conventions. As such, cultural signatures act as bridges between students’ lived experiences and the educational technologies they are invited to shape, offering insight into how computational thinking is enacted in culturally situated ways.
Grounded in socio-cultural theory, constructionism, and culturally responsive pedagogy, the study explores how students draw on cultural background, identity, and community practices to create meaningful game-based learning experiences. These lenses guide an inquiry into how cultural narratives and localized knowledge systems shape engagement with computing concepts.
A comprehensive literature review situates this work within broader conversations on multicultural education, game-based learning technologies, and the challenges of cross-cultural computing education. The study employs a mixed-methods framework, combining quantitative analysis of student-created game design artifacts (“booklets”) with qualitative focus group interviews. A novel Cultural Signatures Coding Scheme identifies and quantifies cultural values, aesthetic elements, and communication styles in student work across three global contexts: the United States, India, and Argentina.
Statistical techniques such as frequency distributions and chi-square tests reveal meaningful cross-cultural patterns in how students embed culture into their computational artifacts. Thematic insights from qualitative data further demonstrate how students remix familiar media, personal histories, and collaborative processes to craft unique educational game narratives. These adaptations illustrate that deeply held beliefs and lived experiences can serve as assets in computational design tasks.
The findings suggest that culturally grounded game design enhances student engagement, fosters creativity, and lowers barriers to participation in STEM. By placing culture at the heart of computing education, this dissertation introduces a replicable framework for analyzing cultural influences in student-designed learning technologies. It offers both theoretical and practical guidance for educators and curriculum designers seeking to build more inclusive, human-centered computing classrooms—challenging the assumption that computational thinking is culturally neutral and paving the way for globally responsive pedagogies.Doctor of Philosophy (Ph.D.
The First-Year Experience: A Critical Narrative Inquiry of the First College Year for Neurodivergent Students
No full textThis dissertation explores the meaning-making of neurodivergent students as they consider their lived experiences navigating their first year of college in an ableist post-secondary environment. This study sought to understand the challenges they faced academically, personally, and interpersonally and how their experiences with campus programs, policies, procedures, faculty, staff, and peers helped or hindered their success. Additionally, I explored the salience and intersection of multiple identities with disability and how students perceive the impact on their first-year experience. Using a critical qualitative approach, this study was influenced by the field of disability studies and three models of disability that provide a critical framework: the social model, the critical theory of disability or DisCrit, and the social justice model. Primarily grounded in the social justice model, which focuses on pervasiveness of ableism that permeates throughout society, existing on our campuses in structures, policies, and procedures, this study seeks to foreground the voices of students who have been marginalized, bringing to light their experiences to serve as counternarratives, highlighting the danger of doing things in typical ways. Drawing from data collected through narrative inquiry, conducting 22 semi-structured interviews with 11 participants who self-identified as neurodivergent, I examined what students expected their first year of college to be like and what informed those expectations, what students perceived as positively and negatively contributing to their success in the first year, the salience of their disability identity and the impact of holding more than one minoritized identity on their first year experience, and the degree to which students felt a sense of belonging on campus and what contributed to that positively or negatively.
My findings reveal that cultural norms surrounding what college should be like reproduced in media and film, and experiences shared by family members influenced participants' expectations of college. Previous experiences in the K-12 system, whether they involved stigma, stereotyping and bullying, or failure to secure adequate administrative support led participants to desire more positive experiences in college. Participants expected that college (1) will be much better socially than high school; (2) will afford them more freedom and independence; and (3) will be the same or more challenging academically than high school.
Participants discussed factors that influenced their first-year experience across five domains: (1) the interpersonal; (2) the intrapersonal; (3) the curricular; (4) the co-curricular; and (5) the institutional or structural. Factors that contributed to success included having positive and supportive relationships with family, developing meaningful relationships with peers, possessing self-advocacy and efficacy skills, engaging in the co-curricular, supportive faculty and classroom design, enrolling in interesting and relevant coursework, and registering with disability services. Factors that hindered success included less supportive family relationships, feelings of fear, anxiety, and inability to ask for help, large classes sizes, impersonal academic environments with ableist classroom design, a neurotypical approach to student engagement, and structural ableism evident in university processes and practices.
Participants often minimized, made light of, or spoke negatively about having a disability, reflecting an internalization of the deficit-oriented paradigm surrounding disability. The impact of previous stigma and traumatic experiences led to participants being anxious and hyper-aware of their actions and behaviors, often masking their disability with others. Participants rejected the notion of being defined by their disability, seen as “only” disabled and often highlighted other aspects of their identity that were more salient to them, making meaningful connections with peers based on those identities.
Recommendations included infusing anti ableist practices by (1) engaging in regular assessment of the campus climate and disabled student experience; (2) intentionally designing the onboarding and training of faculty and staff; (3) providing proactive support for the college transition; (4) offering intentional and proactive coaching and advising; (5) infusing universal design in co-curricular opportunities for engagement; (6) considering the physical structure of the campus; (6) developing intuitive systems and accessible campus resources; (7) utilizing universal design for learning principles in classroom design; (8) and offering an intersectional approach to support, de-siloing support services.
Future scholarship should take an intersectional approach; the study illustrates a need for more research related to the intersection of gender and disability, particularly the experiences of neurodiverse trans students. Future research should explore the salience of disability identity for those with invisible disabilities; the impact of social exclusion for disabled students; and the significance of a diagnosis for neurodiverse students and the subsequent timing of the diagnosis and the impact on student’s self-advocacy and efficacy skills, and persistence. Likewise, a longitudinal approach, following neurodivergent students throughout their college experience could illustrate further identifying factors that help or hinder their pathway to graduation. Students described factors that contributed to their success like learning communities, smaller class size, and course design. Further research could explore specifically the impact of such programs on neurodiverse student success and belonging. Likewise, more research could be conducted into the relationship between advising and neurodivergent student success and persistence as well as the role of mental health and proactive approaches to supporting disabled students' health and wellbeing in college. From a systemic view, research could be conducted with institutions who intentionally seek to be anti-ableist. Focusing on specific aspects of the university that contribute to the student experience may illustrate practices that are universally designed in a manner that elevates the student experience for all students.Doctor of Philosophy (Ph.D.
SOCIAL DETERMINANTS OF QUALITY OF LIFE IN THE LAST YEAR OF LIFE AMONG COMMUNITY-DWELLING OLDER ADULTS WITH MULTIMORBIDITY: A COMPARATIVE MIXED METHODS STUDY
No full textBackground: Quality of life (QOL) in the last year of life declines among community-dwelling older adults with multimorbidity, with social determinants playing a critical role. However, social capital remains underexamined in nursing research, particularly in end-of-life care. Objective: This study explores how social determinants, especially social capital, influence QOL in this population. Design: A convergent mixed-methods design was used. The quantitative component analyzed data from 230 participants in the National Health and Aging Trends Study (NHATS) using linear regression and path analysis to examine social capital’s mediating role between social determinants and QOL. The qualitative component involved thematic analysis of semi-structured interviews with seven family caregivers. Results: Findings were integrated through joint displays and meta-inferences. Results showed that QOL in the last year of life was generally low, as was social capital. Mental conditions and social capital were key predictors of QOL, with social capital partially mediating the relationship between mental conditions and QOL. Thematic analysis identified six key themes: financial well-being, access to resources through social networks, supportive relationships, acceptance of death, positive emotions, and manageable physical symptoms. Meta-inferences revealed that mental health and social capital significantly influence QOL. Social capital not only mediates the relationship between mental conditions and QOL but also extends to socioeconomic factors, physical symptoms and disability. Family caregivers, acting as bridging social capital, connect patients to broader support networks, positively impacting QOL and highlighting the need for care strategies that enhance social capital to improve end-of-life QOL for both patients and caregivers. Conclusions: Social capital and mental health are critical determinants of QOL in the last year of life. In this crucial period, patients draw upon accumulated bonding social capital—close, supportive relationships with family and caregivers—as well as bridging social capital, which connects them to broader social networks. Family caregivers, acting as both bonding and bridging social capital, play a pivotal role in supporting patients. Care strategies should prioritize the dynamic interplay between patients and caregivers, focusing on enhancing social capital to improve QOL outcomes during end-of-life care.Doctor of Philosophy (Ph.D.)2026-09-0
Neural Operators for Rarefied Gas Dynamics: BGK relaxation problem, Polyatomic Shock waves, and Hypersonic Cylinder Flow
No full textThis work introduces a suite of targeted methodologies utilizing neural operators—including physicsinformed, constrained, and data-driven approaches—for creating computationally efficient and robust surrogate models of rarefied gas dynamics, a domain where high-fidelity kinetic solvers are often prohibitively expensive. We present three key contributions demonstrating stability, physics-discovery capabilities, and powerful generalization across distinct challenges. First, for the BGK kinetic relaxation
problem, we introduce a novel perturbation ansatz that ensures numerical stability. We leverage this stabilized physics-informed neural network (PINN) framework to solve a challenging combined forward and inverse problem, demonstrating its capability as a physics-discovery tool by successfully inferring the unknown, velocity-dependent relaxation time using only the initial conditions and governing equations. Second, for one-dimensional shock waves in polyatomic gases, we develop a physics-constrained Deep Operator Network (DeepONet) that accurately captures complex non-equilibrium structures for unseen viscosity ratios by embedding monotonicity constraints directly
into the learning process, eliminating non-physical oscillations. Finally, we demonstrate significant data efficiency and powerful generalization for two-dimensional hypersonic flow over a cylinder. A data-driven DeepONet ensemble, trained on a dataset spanning a wide hypersonic regime from Mach 5 to 15, accurately predicts the flow field for multiple unseen conditions. The resulting surrogate excels at interpolation (e.g., predicting Mach 7, 9, 12, and 14), and the extrapolation case (predicting Mach 15 while trained on Machs 5-14) with quantitative validation confirming high accuracy across the parametric range. This work establishes a powerful, flexible methodology for building neural operator surrogates that ensure physical consistency, perform physics-discovery, and generalize robustly, significantly lowering the computational barrier for design and analysis in high-speed aerodynamics
Transformative Encounters: Metamorphoses in Eastern and Central European Media
No full textMetamorphosis–the change of form–has enjoyed a long and ubiquitous existence, originating in ancient mythological and folkloric traditions, including premodern traditions of present-day Eastern and Central Europe. In such traditions, metamorphosis warns its audience of irresistible forces, nominally supernatural, whose arbitrary imposition could have irreversible repercussions. In recent centuries, metamorphosis has since been relegated to the genres of science fiction and fantasy; accordingly, critical thinking engaging with metamorphosis has been speculative and limited to the domains of idealist-materialist debate. “Transformative Encounters” contends that metamorphosis is owed a more complex understanding. To this end, “Transformative Encounters” takes up the motif of metamorphosis in Eastern and Central European media of the 20th and 21st century. It argues that the metamorphoses represented therein have philosophical, ecocritical, and political implications. As such, metamorphosis reprises its didactic function by indicating a shift from supernatural force to human imposition, particularly at the level of the state–a timely consideration, in light of an increasingly dehumanizing geopolitical landscape, and, under the global climate crisis, an increasingly fragile Earth.Doctor of Philosophy (Ph.D.)2026-09-0
Exploring Loneliness in Cambodian Immigrants Through an Intergenerational Lens
No full textLoneliness is a well-documented risk factor for poor health outcomes. Yet, little is known about how it affects Cambodian refugees and immigrants in the United States (U.S.) – a historically underserved and under-researched population with high rates of cultural trauma stemming from the Khmer Rouge Genocide (1975–1979). This study explored experiences of loneliness among first-generation (i.e., genocide survivors who migrated to the U.S.) and second-generation (i.e., descendants of the first-generation) Cambodian immigrants. Four aims guided the investigation: (1) to explore participants’ definitions of loneliness; (2) to examine whether participants identified intergenerational conflict as contributing to loneliness; (3) to assess the role of trauma in loneliness, with the hypothesis that first-generation participants would cite pre- and post-migration stressors, while second-generation participants would identify intergenerational trauma; and (4) to identify other perceived sources of loneliness, such as systemic racism. Seventeen participants (8 first-generation; 9 second-generation) completed questionnaires and semi-structured interviews. Data were analyzed using content analysis and frequency coding. Findings revealed that first-generation themes reflected loneliness rooted in their experiences during pre-migration, post-migration, and resettlement. Second-generation themes centered around contextual factors, intergenerational transmission of trauma, and feelings of cultural and social isolation. All hypotheses were partially supported. These findings offer insight into the complex and layered nature of loneliness in trauma-affected immigrant populations. Results may inform culturally responsive clinical practices and support services for both first- and second-generation Cambodian immigrants, as well as for other refugee communities with similar histories.This work was supported in part by a Graduate School Fieldwork Grant from the University of Massachusetts Amherst (UMass), a Community Engagement Dissertation Award from the Center for Research on Families at UMass, and a Dissertation Research Grant from the Asian American Psychological Association.Doctor of Philosophy (Ph.D.)2030-09-0
Adaptive Techniques for Low Power, Thermal-Aware, Fault-Tolerant and Safe Cyber-Physical Systems
No full textThis dissertation addresses challenges in Cyber-Physical Systems (CPS), emphasizing improvements in lifetime, efficiency, and anomaly detection through advanced adaptive methodologies. Given the critical necessity for reliability in CPS, particularly in safety-critical applications, this research explores innovative techniques that manage power consumption and thermal stress, workload adjustment, and anomaly detection effectively.
To reduce thermal stress and enhance energy efficiency, a novel Dynamic Voltage and Frequency Scaling (DVFS) algorithm tailored to workload characteristics is presented, which preferentially adjusts task execution speeds based on their individual power profiles. Additionally, a new task allocation and scheduling strategy is proposed specifically for heterogeneous computing architectures, significantly reducing thermal impacts through optimized utilization of power-efficient cores.
Moreover a physical state adaptive fault-tolerance and workload adjustment technique is presented. This approach segments the operational state-space into distinct sub-spaces, each assigned tailored fault-tolerance requirements and dynamically adjustable control task frequencies, effectively balancing redundancy with system safety. Lightweight machine learning algorithms are employed to efficiently classify these sub-spaces in real-time, facilitating rapid system adaptation.
Furthermore, this research introduces a deep reinforcement learning strategy designed to optimize control task dispatch decisions based on physical state conditions, explicitly accounting for long-term operational impacts. This technique ensures significant workload reductions, minimizes degradation of control quality, and maintains stringent safety criteria.
Finally, the dissertation presents a robust physical-state-based anomaly detection mechanism utilizing Recurrent Neural Networks, capable of detecting subtle anomalies induced by external factors, such as malicious attacks, even when conventional cyber metrics indicate normal operation.
Collectively, these comprehensive advancements significantly enhance CPS reliability, operational efficiency, and application safety.Doctor of Philosophy (Ph.D.
Modeling TBCK Syndrome in Zebrafish
No full textTBCK Syndrome is a rare, neurologic pediatric disorder caused by mutations in TBC1 Domain Containing Kinase (TBCK). The disorder is characterized by epilepsy, severe intellectual disability, and hypotonia. Unfortunately, there are no cures or treatments for TBCK Syndrome and the disorder is often fatal. Animal models for a disorder are crucial to probe disease pathogenesis and screen potential drug candidates. In this work, we explore several different strategies for creating a zebrafish model of TBCK Syndrome. In our initial efforts, we obtained and generated tbck loss-of-function alleles. Surprisingly, tbckumz520 mutants are nearly indistinguishable from their wild-type siblings in all parameters we assayed, which includes behavioral, electrophysiological, and cellular and molecular outputs. The notable exception is that tbckumz520 mutants are subtly hyperactive following changes in illumination during the visuomotor response assay. We sought to determine if the mild tbckumz520 phenotype was due to transcriptional adaptation, a newly described genetic compensation method. We performed bulk RNAseq on tbckumz520 mutants and on tbck crispants, a model that we predicted should not demonstrate compensation. We describe differentially expressed pathways in tbckumz520 mutants, and identify three genes that could potentially compensate for loss of tbck. Finally, we explored additional genetic disruption strategies
in an attempt to make a more robust model of TBCK Syndrome. We phenotyped fish directly injected with CRISPR-Cas9, known as crispants, and found that they are largely normal. We also started work toward generating a tbck promotorless allele, which is yet to be characterized. Lastly, we disrupted tbck using Morpholino oligonucleotides, an antisense knockdown strategy. tbck Morpholino injected embryos have survival deficits, severe morphological defects, and reduced locomotion in the touch-evoked escape response. As a control, we demonstrated that tbckumz520 mutants are more resilient to tbck Morpholino. Taken together, this work has demonstrated that simple, premature termination codon-generating mutations are insufficient to model TBCK Syndrome in zebrafish. However, we have also highlighted that the use of Morpholinos or other techniques that may evade transcriptional adaptation represent a promising path forward to study tbck in zebrafish.Doctor of Philosophy (Ph.D.