50107 research outputs found
Sort by
MULTIMODAL MASS SPECTROMETRY-BASED METHODS FOR ASSESSMENT OF STRUCTURE AND HETEROGENEITY OF BIOPOLYMERS AND THEIR COMPLEXES
No full textOver the last decades mass spectrometry has evolved from the elemental composition analysis method into a key bioanalytical technique for studies of biopolymers. However, the significant structural heterogeneity of some of these biopolymers presents a formidable analytical challenge for conventional mass spectrometry-based techniques. To address this problem, a set of novel multimodal approaches has been developed to assess the structural heterogeneity at different levels of complexity. They include cross-path reactive chromatography-based methods for analysis of disulfide-linked multisubunit proteins, MALDI-TOF/TOF based methods for analysis of biomolecules containing large polycationic segments, and charge detection FT-ICR mass spectrometry for analysis of biomolecules of elevated heterogeneity. Taken together, these methods can dramatically expand the capabilities of mass-spectrometry for the analysis of biopolymers extracted from complex biological matrixes or produced in the various (bio)manufacturing processes.NIH R01GM132673
NIH R01GM112666
NIH R01HL174310
NSF CHE-1709552
IALS Translational Graduate AssistantshipDoctor of Philosophy (Ph.D.)2030-09-0
Assessing Utility of Wearable Devices for Monitoring Health of Women and Women Cagivers for Dementia
No full textBackground:
Dementia caregiving is a demanding and often prolonged role that disproportionately affects older women. While critical for care recipients, caregiving can impose significant psychological, behavioral, and physiological strain on caregivers. This dissertation investigates how caregiving relates to physical activity patterns, autonomic function, and psychological and cardiovascular health, using a multi-method approach that integrates literature synthesis, wearable sensor data, and predictive modeling.
Methods:
The dissertation comprises three interconnected studies. Study 1 is a systematic review and meta-analysis of studies comparing physical activity between informal caregivers and non-caregivers. Random-effects meta-analyses estimated pooled odds of physical inactivity and insufficient moderate-to-vigorous physical activity (MVPA), with heterogeneity assessed via the I² statistic. Study 2 is an observational study comparing device-measured physical activity, sedentary behavior, and heart rate variability (HRV) between older women dementia caregivers (n = 15) and non-caregivers (n = 16), alongside survey-based assessments of psychological health and caregiving burden. Study 3 used logistic and stepwise regression models to evaluate associations between wearable-derived metrics and psychological and cardiovascular health, including bootstrapping for internal validation.
Results:
Study 1 found caregivers had slightly higher odds of physical inactivity (OR = 1.11, 95% CI: 0.80–1.54) and insufficient MVPA (OR = 1.13, 95% CI: 0.92–1.41), though associations were not statistically significant. Substantial heterogeneity (I² > 90%) highlighted variability in caregiving definitions and physical activity measures. In Study 2, caregivers reported significantly poor psychological health including depression (rₚb > 0.5), and anxiety. Caregivers also reported significantly higher stress and poor sleep quality and showed more fragmented activity patterns. Despite higher daily step counts, they accumulated fewer sustained walking bouts and more prolonged sedentary time. HRV day-to-night Root Mean Square of Successive Differences (RMSSD) metrics indicated impaired autonomic recovery, particularly among those with higher caregiving burden. In Study 3, several wearable-derived metrics including prolonged sedentary bouts and lower HRV were associated with poor psychological health. Time in activity bouts > 5 minutes, prolonged sedentary bouts and HRV predicted poor cardiovascular health in stepwise regression. However, predictive modeling performance declined after bootstrapping, underscoring the need for caution when interpreting these models due to the small sample size and limited generalizability.
Conclusion:
Older women and women caregivers for individuals with dementia experience measurable psychological and physiological strain, which can be feasibly and objectively captured using wearable technologies. While caregivers often accumulate more daily steps, they also demonstrate less favorable movement patterns, such as fewer sustained walking bouts, more prolonged sedentary periods, and greater fragmentation of activity. These findings support the utility of real-world monitoring psychological and physiological dysregulation. Wearable-derived metrics show promise as informative indicators of caregiver health; however, caution is warranted when applying predictive models in small samples, given their limited generalizability and reduced performance after bootstrapping. Collectively, these results highlight the need for scalable, flexible interventions that address caregiver burden, reduce risky sedentary accumulation, and enhance autonomic recovery to promote long-term health and resilience.American Association of University Women International Fellowship & Priscilla Clarkson AwardDoctor of Philosophy (Ph.D.
Evaluating global soil weathering data through a solute production framework
No full textThe silicate weathering feedback, thought to be Earth’s primary long-term climate regulator, is hypothesized to be most active in rapidly uplifting mountainous settings, dissolving freshly-exposed silicate rocks to draw down elevated atmospheric pCO2 concentrations. Silicate weathering rates are influenced by the interaction of many environmental factors, including temperature, lithology, hydrology, vegetation, and erosion, but disentangling the influence of each on a site-by-site and global basis yields conflicting conclusions and therefore remains an active area of research. Silicate weathering rates are estimated from both soil and dissolved solutes; however, competing theoretical frameworks, centered around solid-phase and aqueous-phase geochemical data, have yet to be reconciled. In this study, we assessed the influence of erosion, lithology, and hydrology on global weathering and analyzed weathering rates from soils through the lens of a solute production weathering framework. To examine the correlation between environmental factors and silicate weathering, we expanded existing global soil compilations to include additional published soil weathering rates and extracted site-specific environmental parameters from global reanalysis, satellite remote-sensed, and regional meteorological raster datasets. To reduce variability in weathering due to mineralogy, we calculated sodium fluxes as a proxy for Na-plagioclase weathering, resulting in a strengthened relationship between mineral-specific weathering fluxes and erosion. Normalizing for mean annual precipitation, we observe potential hydrologic controls on fluid residence time limiting solute production, indicating that all soils in the weathering compilation classify as kinetically limited according to the solute production framework. To assess how soil weathering rates align with conceptions of solute generation, we superimpose soil data onto the Maher and Chamberlain (2014) solute production model, testing the fit of model scenarios against soil weathering trends. We find that scenarios including a flow path length dependence on regolith thickness, controlled by erosion and precipitation rates, best capture denudation and weathering trends, yielding the highest predictive power in capturing modeled Na weathering concentrations. Although a shortage of weathering rate data in areas with high denudation rates limits our ability to assess the best fit model scenario with certainty, our analysis highlights the coupled role of hydrology and erosion as primary drivers of silicate weathering in soils globally.Master of Science (M.S.
Culture and Inequality on Well-Being: Meritocratic Culture Moderates the Relation between Income Inequality and Subjective Well-Being
No full textThis study investigates whether meritocratic culture moderates the relationship between income inequality and subjective well-being using cross-national data from the World Values Survey (1990–2023). Building on research that connects culture to the amplification of inequality’s social impact, I argue that meritocratic narratives may either dampen or intensify the psychological burden of inequality, depending on how they shape individuals’ preferences and lived experiences. Multilevel regression analyses reveal that income inequality is negatively associated with subjective well-being. More importantly, both personal and public forms of meritocratic culture significantly moderate this relationship—intensifying its negative effects. Meritocratic culture, far from neutralizing the harms of inequality, may act as a cultural amplifier—deepening the emotional and social damage inflicted by unjust structural conditions. The results highlight a previously underexplored mechanism by which inequality can harm well-being, extending beyond economic conditions to the cultural meanings attached to success and failure. This study contributes to the literature by introducing meritocratic culture as a key moderator in the inequality–well-being link and emphasizing the value of cultural analysis in social epidemiology. It further calls for greater attention to how cultural narratives interact with structural inequality, especially in societies with strong meritocratic ideals. The findings have implications for both theory and policy, suggesting that addressing inequality requires engaging with the cultural frameworks that legitimize it.Master of Arts (M.A.)2026-09-0
MODELING THE SPATIAL ECOLOGY OF YELLOW-THROATED MARTENS: OCCUPANCY AND SPECIES DISTRIBUTION ANALYSES IN NORTHERN MYANMAR
No full textThe yellow-throated marten (Martes flavigula) is a widely distributed meso-carnivore in Southeast Asia, but they are poorly studied, particularly within the ecologically understudied landscapes of Northern Myanmar. Despite its role as a facultative top predator and indicator species of forest ecosystem health, limited empirical data exists on its habitat preferences and distribution in the region. This study addresses these knowledge gaps by integrating camera trap surveys with statistical modeling to investigate ecological, anthropogenic, and topographic drivers of yellow-throated marten distribution in Hkakabo Razi National Park (HKNP). Camera trap data collected between 2014 and 2017 (4,647 trap nights) were analyzed using single-season occupancy models and species distribution models (SDMs). Predictor variables included ecological (NDVI, canopy height, distance to water), topographic (elevation, slope, aspect), and anthropogenic (distance to roads, villages, and guard posts) factors. Occupancy models identified NDVI and canopy height as the most influential predictors, with higher vegetation density positively associated with occupancy, though coefficients were not statistically significant. For the SDMs, I used an ensemble approach incorporating Generalized Linear Models (GLM), Maximum Entropy (MaxEnt), and Random Forest (RF) algorithms. MaxEnt yielded the highest predictive performance (AUC = 0.897, TSS = 0.622), with distance from roads and canopy height as the most important variables. Ensemble model outputs revealed core habitat zones in remote, vegetated areas with minimal human disturbance. These findings offer insights into the habitat ecology of yellow-throated Martens and provide a scientific foundation for conservation planning in HKNP. By identifying key habitat features and spatial distribution patterns, this study contributes to broader conservation efforts aimed at conserving mid-sized carnivores and maintaining ecological integrity in Southeast Asia’s montane forest ecosystems.USAID, UMassMaster of Science (M.S.)2026-03-0
The Role of Cis-acting Elements, RNA Modification, and Cellular Localization of mRNA Transcripts During KSHV Lytic Infection
No full textExploitation and control of the cellular environment is the cornerstone of successful viral infection, with struggle for control over gene expression machinery being a driving force for the evolutionary arms race between viruses and their hosts. During infection, different viruses utilize a wide range of strategies to prioritize viral transcripts over host transcripts. In both Coronavirus and Kaposi’s sarcoma associated herpesvirus (KSHV) this is driven by a wide-spread RNA decay event which leads to the destruction of host RNA as driven by viral proteins. In KSHV, this process is driven by multiple factors during the duration of KSHV lytic infection with a viral endoribonuclease, SOX, being the primary trigger and a process known as hyperadenylation, triggered by SOX activity and SOX nuclear relocalization. In Coronavirus, this process is driven by the nsp1 protein, which has variable activity between different Coronaviruses, but specifically targets and degrades host transcripts, primarily through interactions with the ribosome. However, in both conditions, there exists subsets of transcripts that are resistant to these viral decay events, of both host and viral origin. These resistant transcripts are protected from decay through a myriad of different mechanisms related to RNA secondary structure, RNA modifications, and cellular localization. In this dissertation, we will explore the role of these different factors in two separate viral systems to investigate the impact of viral activity on RNA stability, to better our understanding of viral-host interplay in the RNA landscape. In the first chapter, we will explore the impact of four different Coronavirus nsp1 proteins and how the different classifications of viruses utilize their homologs in a variety of ways to impact the host transcriptional landscape. In the second chapter, we will explore work dedicated to understanding our primary escapee, human Interleukin-6 (IL-6), and the impact of RNA secondary structure, nuclear export, and virally induced RNA modification on this transcript. Finally, in the third chapter, we extend exploration
into the polyA landscape of transcripts during KSHV lytic reactivation to better understand the impact of hyperadenylation on the host transcriptome. To achieve this we performed PolyA-seq to explore the changes in polyA tail length as well as subcellular localization of transcripts as well as experiments exploring the impact of CRM1 inhibition on the stability and localization of these transcripts.Doctor of Philosophy (Ph.D.
Whither Pragmatism? A Critical Realist Reconstruction of Dewey's 'Naturalistic Humanism'
No full textI make three interrelated arguments in this dissertation. The first argument challenges a tendency to kick away the ladder of metatheory. Because metatheoretical critique is an integral moment of inquiry, we should pull up the ladder after us and use it to address the problems we still face. Reading John Dewey against the grain of contemporary sociology suggests that articulating the methodological significance of reflexive knowledge is one such metatheoretical problem. Although Dewey argued that solving social problems requires communicative knowledge, many pragmatist sociologists focus exclusively on the role of instrumental knowledge.
The second argument identifies another problem where the ladder of metatheory is useful. Problem-solving sociologists direct us to solve social problems by generating new causal knowledge. However, the nature of social causality suggests that we should not expect all new causal knowledge to track resolutions to social problems. This is because causal factors are practically infinite, and social problems are moving targets. Consequently, the project of problem-solving sociology needs a methodological strategy for orienting research towards root causes. More generally, sociologists should move towards a less dichotomous understanding of methods that balances the problem of how to systematically answer research questions against the equally important problem of how to systematically ask them.
The third argument advocates for reconstructing theories of history along postpositivist lines as one possible way of orienting problem-solving sociology towards root causes. Karl Popper’s influential philosophy of science marginalized theories of history within sociology. Although sociologists have rejected the basic principles of Popper’s philosophy of science, the presupposition of scientism (i.e., the sidelining of social epistemology) continues to obstruct reconstructive efforts. This is especially evident among contemporary constructivists, who reduce their theory of science to the question of representation but ignore the equally important question of experimental activity. A balanced approach clarifies the important role theories of history can play in sociological research. Theorizing the history of society orients research towards questions of causal depth by making the relational interdependence of society the focus, thereby counterbalancing the fracturing drive of Mertonian middle-range theory to narrow the problematics of sociology to substantive, disconnected foci.Doctor of Philosophy (Ph.D.)2026-03-0
Jazz Malungaje: Understanding the Cultural Identity Implications in African American Jazz from the South African Jazz Tradition
No full textJazz is a historically contested terrain in the United States public narrative and in jazz historiography. Using Dave Brubeck’s Downbeat Magazine published origination theory of the music’s “mixed parentage of primitive African and highly developed European cultures” (1951), scholarly and journalistic attention has focused on a hegemonic discourse that privileges Europe in claiming the music’s supposed hybridity as a Euro-American and African American created form. Conversely, Langston Hughes claims the music as African American culture and heritage, saying that it is “one of the inherent expressions of Negro life in America” (1921). The transnational connection between Black South Africa and African-America expressed in the shared jazz tradition between the two groups offers a response to these contested origins. Jazz has been indigenized by Black South African musicians since the music’s early development in the 1920s, as a form that, from their perspective, was created by Africans in America via the Trans-Atlantic Feedback Cycle (Ballentine, 1991/Collins, 1989). Building on the work of Coplan (1985), Ballentine (1991), Ansell (2004), Dlamini (2010), Washington (2012) and Kubik (2017), this dissertation is an ethnomusicological inquiry into the way Black South African jazz musicians in Johannesburg hear the Africanity of jazz with what I call their ‘bi-musical ear.’ The term ‘bi-musical ear’ indicates the bi-musicality that many Black South Africans have in both traditional, indigenous South African musics and in jazz.REAL Fellowship, University of Massachusetts Amherst
Graduate School Field Research Grant, University of Massachusetts Amherst
Diversity Scholars Predoctoral Fellowship, Ithaca CollegeDoctor of Philosophy (Ph.D.)2030-09-0
Detecting Frailty Using Wearable Device-Measured 24-Hour Movement Behaviors in Older Adults
No full textFrailty is a common geriatric syndrome associated with an increased risk of adverse health outcomes, yet its assessment in clinical and research settings remains limited due to reliance on subjective measures and time constraints. The goal of this dissertation was to improve frailty assessment in older adults by leveraging wearable device–measured 24-hour movement behaviors through two approaches: (1) detecting frailty using machine learning models based on accelerometer-derived movement behaviors, and (2) refining the Fried Frailty Phenotype (FFP) by replacing the self-reported low physical activity criterion with accelerometer-measured physical activity.
Study 1 was a cross-sectional study involving 44 older adults who wore a thigh-worn accelerometer (ActivPAL) for 10 consecutive days. Frailty was assessed using both the FFP and the Comprehensive Geriatric Assessment–Frailty Index (CGA-FI). Machine learning models were developed to classify frailty status, detect individual FFP components, and estimate CGA-FI scores based on 24-hour movement behavior features. Results showed that the Random Forest model had higher AUC for FFP-defined frailty, while the K-Nearest Neighbors model performed best for CGA-FI-defined frailty. The Support Vector Machine showed higher accuracy for predicting CGA-FI continuous scores. Key predictors for detecting frailty included mean steps/day and variability of stepping and standing time, and time in stepping cadence ≥100 steps/min.
Study 2 was a prospective cohort study using data from 38,429 UK Biobank participants aged ≥60 years with valid wrist-worn accelerometer data, modified FFP data (FFP-Mod), and mortality follow-up. Two revised FFP definitions were created by substituting the self-reported low physical activity criterion with accelerometer-measured physical activity including the lowest quintile of: 1) overall mean acceleration (FFP-Acc) and 2) time in MVPA (FFP-MVPA). Results revealed that while individuals classified as frail or prefrail by all three FFP definitions had significantly higher all-cause mortality risk compared to robust individuals, these associations were stronger for FFP-Acc and FFP-MVPA than for FFP-Mod.
In conclusion, this dissertation supports the feasibility and utility of integrating wearable-derived movement behavior metrics into frailty assessments. Machine learning models using 24-hour movement behaviors can detect frailty, and replacing self-reported physical activity with accelerometer-measured physical activity improves the predictive validity of FFP. This work contributes to the development of objective and clinically relevant frailty assessments for use in both research and public health applications.Study 1 was supported by the National Institute on Aging (P30AG073107) through the MassAITC pilot project. Study 2 was supported by the Mutual Mentoring Grant from the Office of Faculty Development at the University of Massachusetts Amherst.Doctor of Philosophy (Ph.D.
DRIVERS OF GALAXY EVOLUTION AROUND COSMIC NOON: GAS, DUST, AND DISK INSTABILITIES
No full textGalaxies are the building blocks of our Universe. Understanding their formation and evolution provides a coherent picture of the past, present, and future of the Universe. The evolution of a galaxy is shaped by a complex interplay of physical processes that influence its star formation activities and observable characteristics such as stellar mass, metallicity, dust attenuation, and morphology. This dissertation investigates the roles of gas, dust, and disk instability in understanding galaxy evolution around the Cosmic Noon, when star formation rates and AGN activities across the Universe peaked. Through high-resolution multi-band imaging and spectral energy distribution modeling, we show the reciprocal influence between star formation and gas dynamics, dust obscuration, and structural transformation in three interconnected chapters. In the early Universe, rapid changes in gas dynamics can quickly change the distribution of metals in galaxies, while the observed metallicity gradients only capture a snapshot of their evolutionary history. In terms of dust obscuration, current UV/optical surveys might significantly underestimate the stellar mass and star formation in the early Universe, owing to the common presence of optically thick dust. The production and dissipation of the optically thick dust is closely related to local star-forming activities, and could be a common evolutionary phase of massive galaxies around the probed redshift. We also report evidence of the bar-driven suppression of star formation and the inside-out galaxy formation scenario as early as z~3. The presence of a stellar bar just ~2 Gyr after the Big Bang suggests a low dark matter fraction and the rapid growth of disk structures at high redshift. Together, these studies provide a relatively comprehensive framework for understanding the cycle of baryons between gas, dust, and stars, and highlight the key mechanisms that govern galaxy evolution in the early Universe.Doctor of Philosophy (Ph.D.)2026-09-0