Boston University Theses & Dissertations

Permanent URI for this collection

https://hdl.handle.net/2144/8520

This is the master collection of contemporary BU theses and dissertations. We plan to consolidate school- and college-specific collections into this one, and add school- and college-specific metadata to enable users to browse appropriately.

Browse

Now showing 1 - 20 of 9827

State structures and interventions for acute infectious disease outbreaks: a comparative politics perspective
(2024) Kim, Elvis Heyun; Fewsmith, Joseph; Boas, Taylor C.
Through three interconnected articles, my dissertation examines government responses to COVID-19 from the lens of state-society relations and political institutions. Extraordinary events and national emergencies such as war and pandemics help to amplify the existing state-society relations and accentuate how institutions structure the behaviors and expectations of political actors. Using large-N analyses based on revised and updated global datasets, Paper I investigates how political institutions affect the effectiveness of pandemic responses measured by excess mortality rates. The spatial autoregressive models suggest that among all countries, greater bureaucratic quality and higher degrees of liberal democracy predict lower excess mortality rates, and among electoral democracies, higher levels of institutionalized centripetalism are associated with lower excess mortality rates. Paper II examines how distinct institutional structures in mainland China and Hong Kong led to the divergence of nonpharmaceutical interventions after the surge of the Omicron variant. I argue that China’s authoritarian and Leninist institutions feature a high level of social control and a low level of social inclusion, which facilitated its unwavering pursuit of “zero COVID” in the form of total mobilization but contributed to the rising social instabilities and snap reopening in December 2022. In contrast, Hong Kong’s hybrid and Weberian institutions are characterized by a lower level of social control but a higher level of social inclusion, leading to an early desertion of “zero COVID” when the cost-effectiveness ratio of interventions increased. Paper III continues to focus on mainland China and Hong Kong. Based on surveys conducted between November 2022 and January 2023, the analyses suggest that both instrumental and normative assessment of NPIs motivates voluntary compliance through individuals’ psychological processes, and at the aggregate level, the government in mainland China has been more effective in bringing cognitive and behavioral changes in support of nonpharmaceutical interventions than its counterpart in Hong Kong.
Rhythm and character in Homer's Iliad
(2024) Kotiuga, Peter; Scully, Stephen
This dissertation provides a methodology by which to study the relationship between a character’s rhetoric and the narrative context in rhythmical terms. In the first chapter, I remark on the differences between the narrator’s voice and that of his characters. In order to assess such distinctions in rhythm, I study the traditional hexameter verse in Chapter 2 to determine the “basic” metric of the hexameter. The basic metric involves many aspects of versification, including scansion, prosody, and syntax. My methodology for analyzing Homer’s “rhetorical” metric, given in Chapter 3, involves a combined study of those aspects of the verse, both within individual verses and across runs of verses. In Chapter 4, I analyze the eight single-verse speeches of the Iliad to demonstrate how dramatically effective the rhythm of even a single verse can be. In Chapter 5, I explore how the nearly identical rhythmic variation in the greetings of Charis (18.385-7), Hephaestus (18.424-6/7), and Aphrodite (14.194-6) expresses the same sense of surprise despite the disparate circumstances of these speakers. In contrast to such variation, Achilles’ public and quasi-legalistic announcement of the iron-toss competition (23.831-5) conveys no surprise, only a steady rhythm. In Chapters 6–8, I apply such structural rhythmic analysis to specific narrative contexts. In Chapter 6, I find that Odysseus’ words for restraining other kings (2.190-197), which the narrator describes as “gentle,” are in fact rhythmically balanced and steady; the rhythms of his “harsh” words directed at boisterous commoners (2.198-205/6), however, are much more erratic and jolting. Similarly, in Chapter 7, I find that Thersites’ “disorderly” and “indecipherable” words (2.225-42) fail because his thoughts are not rhythmically delineated. Thersites’ failure is especially evident in comparison with Achilles’ oath of withdrawal (1.225-44, from which Thersites takes his inspiration) and Odysseus’ response to Thersites (2.246-64). In Chapter 8, I explore how Theano’s prayer (6.305-10) reflects the traditional rhythmical style of archaic Greek prayer, the shift toward which is also visible in Agamemnon’s progression from rebuke to prayer (8.228-244). I argue that the failure of Theano’s prayer is made ironic due to its near-perfect and “strophic” rhythm and in light of Glaukos’ successful—though more rhythmically inconsistent—prayer (16.514-26). In the concluding chapter, these observations are applied to the transition from narrative to speech discussed in the first chapter. I argue that archaic epic poets could align or contrast verse rhythms for rhetorical and dramatic effect without breaching the established norms of the hexameter verse. In this way, the Homeric poet suited the verse rhythm to his characters and the context.
Assembly of interferometric scattering microscopy for discerning plasmonic nanoparticles and cell receptors
(2025) Velasco, Leslie; Reinhard, Björn M.
Noble metal nanoparticles (NPs) have unique photophysical properties, including non-blinking and non-bleaching characteristics, making them highly compatible with smaller NP sizes. These features make them especially appealing for studying cellular dynamics, as their smaller size will not interfere with biological activity. When compared to bulk metal particles, excitation of the conduction band of nanoparticles by an external field transforms the behavior of the electrons into an oscillating behavior. This phenomenon is known as localized surface plasmon resonance (LSPR). Under optical excitation, the combined near-field enhancements lead to coupled interactions observed as spectral shifts in the red region. Interferometric scattering microscopy (iSCAT) is an interference-based imaging platform that has high sensitivity of weak scatterers and indefinite limit on observation time. While iSCAT alone is diffraction limited, combining it with metal nanoparticles allows us to break the diffraction limit, offering enhanced resolution and sensitivity. We developed a two-color iSCAT microscope with ratiometric detection to monitor size differences in Ag NPs. Here, we explain the selection of optical components for this home-built system and share insights gained from addapting iSCAT with a commercial upright microscope. We chose 405 and 445 nm laser diodes for the microscope since they are located at the higher and lower plasmon resonance of Ag. Furthermore, laser diodes are advantageous because they have shorter coherence lengths, which help minimize speckle patterns in the imaging field of view. We demonstrated how two-color iSCAT can distinguish differences in Ag nanoparticle size and monitor trends with respect to change in ambient refractive index. Additionally, we validated the sensitivity of two-color iSCAT as small as 5 nm in size, which may be useful as labels for future studies on EGFR structural conformations. Using two-color iSCAT, we were able to observe the assembly of PEG-tethered Ag NP dimers of varying lengths by streptavidin-biotin interactions. The spectral shifts observed on both wavelength channels allowed us to monitor the dimers formed by from dimers formed by 0.4 kDa and 3.4 kDa PEG tethers. To validate iSCAT observations, we imaged the monomer and dimer samples using scanning electron microscopy (SEM). We were able to correlate phenomena observed in iSCAT with the SEM by showing probability plots of formed dimers. In iSCAT, monomers (10 and 20 nm) appeared as dark contrast, with 10 nm dimers showing a larger intensity, and 20 nm dimers with bright contrast. The spectral shifts observed in both wavelength channels enabled us to study the assembly of small nanoparticle dimers. Cells have a complex background which makes them too difficult to study with a standard iSCAT microscope, this is what led to the addition of a third color to our microscope. In the final section of this dissertation, we describe the calibration of three-color iSCAT microscope by imaging Au and Ag NPs. While Ag has greater optical properties compared to Au, they are toxic and less stable for nano conjugation purposes. Therefore, we incorporated a 520 nm laser line, which is close to the peak plasmon resonance of gold. We demonstrated how three-color iSCAT can discern between Au and Ag NPs and how this can pave the way for a new biosensing tool to distinguish different cell receptors.
Essays on the guidance of the direction of innovation
(2024) Donald, Eric Robert Paul; Restrepo, Pascual
My dissertation studies how policy should guide, and respond to, technological change. The first chapter addresses the "guide" side of the question, while chapters two and three address the "respond to" side of the question.In the first chapter, I argue that cross-technology knowledge spillovers are critical for understanding policy's role in the transition to clean technology. I develop an endogenous growth model with clean and dirty technologies and a network of cross-technology spillovers. I derive formulas for the size and speed of technological transition, following a policy reform, which show that greater spillovers across technologies induce a faster transition but at the expense of a smaller long-run impact of policy. Such spillovers also prevent the lock-in of dirty technology. The economy's spillover structure can be summarized by a sufficient statistic matrix, which I estimate using patent citation data. Applying my model to US transportation and electricity generation, I find that cross-technology spillovers are mid-sized: they prevent lock-in but imply a slow transition with a high long-run impact of policy. I conclude by examining how cross-technology spillovers affect optimal clean innovation subsidies, deriving an innovation subsidy formula that holds for arbitrary carbon prices. Quantitatively, I find that optimal clean innovation subsidies are small, reflecting the low centrality of clean technologies in the spillover network. Hence, a "big push" of temporary clean innovation subsidies is not warranted. In the second chapter, I start with the observation that the standard reaction to the problem of automation, by both lay people and the economics literature, follows a Pigouvian intuition: robots harm workers, so they should be taxed. I argue that this Pigouvian intuition is misguided, or at least oversimplified. As shown by the recent literature modeling automation within the task framework, capital only exerts a negative pecuniary externality on labor at the extensive margin of automation. At the intensive margin, more capital producing a task that has already been automated raises wages for everyone via capital deepening. To formalize this point, I present a model with heterogeneous agents where the Planner can tax income from capital and labor as well as target the extensive margin of automation by stipulating how much more expensive labor must be than capital before automation can occur. I show, via an envelope argument, that capital taxation should ignore automation when the extensive margin tool is set optimally. In a quantitative application to the US economy, I find that labor should be 3.4% more expensive than capital before automation can occur. In the third chapter, Masao Fukui, Yuhei Miyauchi, and I study optimal transfer policy in dynamic spatial equilibrium models with frictional migration and incomplete financial markets. A key policy trade-off is to provide consumption insurance while minimizing the distortion of migration flows. We derive a recursive formula for optimal spatial transfers that strikes this balance. We calibrate our model to U.S. states and find that the U.S. economy would benefit from increased transfers to low-income-growth states. Welfare gains from optimal transfers are substantial but smaller than in a framework abstracting from slow migration adjustment.
Ontological exiles and returns: memory and mourning in 21st century Francophone literatures and film
(2024) Ducey, Joanna; Cazenave, Odile
This dissertation examines the role of 21st-century Francophone transnational fictional literature and films in constructive processes of decolonial remembrance that mobilize occulted stories and histories of exiles and returns. The corpus represents de facto exiles and returns, as well as ontological exiles resulting from the transgenerational repercussions of diasporic traumas like the Atlantic slave trade, colonialism, the Shoah, and the world wars. I analyze how such decolonial texts foster a necessary rewriting of history, memory, trauma, and literature that promotes remedial reflection, global alliance, and collective action. This work is divided into three sections: Exiles and Returns, Memorial Excavation, Mourning and Remembrance, and Looking Forward in Retrospect. Contemporary decolonial theoretical texts from Achille Mbembe, Françoise Vergès, and Malcolm Ferdinand, and the memory work of Marianne Hirsch, Michael Rothberg, and Debarati Sanyal, provide contextual support. I first explore how authors Marie-Célie Agnant, Dany Laferrière, and Fatou Diome, and director Mati Diop create a reciprocal yet distinctive poetics of nostalgia and loss that circumnavigates temporal and spatial boundaries to promote healing and reconciliation. In Section II, the work of Louis-Philippe Dalembert, Alice Zeniter, Boualem Sansal, and Abdourahman Waberi informs the personal and collective excavation of trauma and transmission as catalysts to solidarity and justice. The final section draws upon the representation of interspecies, interracial, and decolonial feminist alliances in Bessora and David Diop’s fiction as impactful movements against modern coloniality and ecocidal crises issued from the epoch of slavery and beyond. Finally, recent works by filmmaker Raoul Peck and author Mohamed Mbougar Sarr emphasize the productive, collaborative force of creative work to flesh out “true” history and commemorate the past. The unique memorial cartographies intrinsic to these works of literature and film constitute crucial therapeutic efforts to confront collective traumas and potentiate a more just and equitable future.
Influence of the Kuroshio Extension on the Pacific Decadal Precession and Marine environment in the Northeast Pacific
(2024) Etige, Nishchitha Sandeepana Silva; Anderson, Bruce T.
The North Pacific’s Western Boundary Current’s extension – the Kuroshio Extension (KE) – has decadal variability of both sea surface temperature and sea surface height. The variability of the KE has been linked to a quasi-decadal mode of climate variability in the North Pacific’s atmosphere called the Pacific Decadal Precession (PDP). Research suggests that on decadal time scales, the PDP both responds to and forces the KE. This research investigates the link between the KE and the PDP and how this link extends to the occurrence of Marine Heatwaves (MHWs) in the Northeast Pacific. This dissertation addresses three research questions: 1. What scale of the KE variability influences the downstream atmosphere that links with the Pacific Decadal Precession? 2. What is the influence of Kuroshio Extension variations on marine environmental extremes in the Northeast Pacific Ocean? and 3. How does the Kuroshio Extension’s influence on the North Pacific atmosphere and marine environments change as the global climate changes? To answer these questions, high-resolution ocean and atmospheric data and climate model data were used. In addition, multiple analyses are used, including empirical orthogonal function analysis, spatial correlation and regression analysis, and causality analysis. Through these analyses, it was determined that the second mode of the KE’s large-scale variability links well with the PDP. In particular meridional variability of the KE’s sea surface temperature supports a north-south atmospheric pressure dipole over the North Pacific. This atmospheric modification then modifies the downstream atmospheric pressure patterns through a reduction of zonal propagation of stationary wave energy and an enhancement of the climatological zonal wave heights over North America, which in turn gives rise to the east-west phase of the PDP. Next, it was found that this large-scale KE variability links with the occurrence of MHWs in the northeast Pacific region. An atmospheric teleconnection is proposed between the KE variability and MHWs whereby the KE variability modifies the sea-level pressure in the northeast Pacific region. The physical and biogeochemical changes that the region undergoes during an MHW are linked to the KE variability. Finally, the presence of these links in the high-resolution Community Earth System Model simulation was investigated. Through this analysis, it was determined the link between the KE and the PDP is present in the model’s pre-industrial, historical, and future simulations. This enabled the project to study the future of the KE and the PDP relationship and its influence on the Northeast Pacific MHWs in a changing climate.
Machine learning on induced geometries
(2024) Douthit, Scott; Kon, Mark
Many high performance machine learning techniques exploit the geometry of the dataset for efficient feature extraction. We produced induced geometry to allow for the use of these high performance methods on general datasets. The induced geometry is based on intrinsic measures from the feature network of the dataset. Convolutional neural networks use the geometry of the dataset to apply re-used filters. Using the correlation structure of the feature network we attempted to create a minimally sufficient geometry for convolution. The technique first created small receptive fields based on highest correlation. It then used the Isomap algorithm to project correlations into the plane, this provided insight into the effectiveness of the method. The method was general and could apply to any dataset including a bag-of-words model. When used on an unstructured sentiment analysis dataset the results were mixed. While the technique was superior when compared against an identity filter, there was a significant computation cost and the effect size of the performance gain was marginal, due to the partial recovery of the underlying feature geometry. However, the principle of recovery of an underlying feature geometry using feature networks was partially successful and speaks to future improvements. Graph networks exploit the graph structure of a dataset. The application of a graph structure using correlation has a significant drawback; the graph will be fully connected. Every node will connect to all other nodes. This creates an incredible computational cost on a network and we found that it also produced very poor inference. We developed two methods of addressing this issue. First the use of a correlation threshold value prevented many edges from forming. This reduced overhead and improved accuracy. The second method was the use of a graph coarsening system, merging connected nodes in the training set. These techniques performed well resulting in accuracy beyond a similarly complex multi-layer perceptron. Visual transformers use the internal geometry in the classification of images. These transformers have achieved state of the art accuracy but require extreme amounts of data to train. Convolutional feature maps for each image patch were included in the input to the visual transformer in an attempt to reduce the the required amount of data and training time. Results were mixed with little advantage to the technique while computational cost increased.
Essays on bank heterogeneity and monetary policy
(2024) Fernandez Rojas, Angel David; King, Robert G.
In this dissertation, I study how bank heterogeneity and the marginal propensity to lend affect the transmission of monetary policy. In the first chapter, I develop a banking model with heterogeneous banks to study how heterogeneity in marginal propensities to lend and responses of deposits to monetary shocks affect the monetary transmission to bank lending. The marginal propensity to lend (MPL) measures how much lending increases after an idiosyncratic one unit increase in deposits. Banks face financial frictions to substitute deposits with wholesale funding, which exposes bank lending to idiosyncratic deposit shocks. When banks are heterogeneous in the degree of financial frictions they face, the aggregate response of bank lending to monetary shocks depends on a deposit heterogeneity channel, which comes from the covariance of MPLs and responses of deposits to monetary shocks. I use U.S. bank-level data to calibrate the model and I find that heterogeneity in the degree of financial frictions dampens monetary policy by at least 17%. In the second chapter, I study how heterogeneity in the volatility of deposit withdrawal shocks affects the monetary transmission to bank lending. I develop a general equilibrium model where banks differ in their size and small banks are endowed with a riskier distribution of deposit withdrawal shocks, consistent with the data. In the model, small banks experience a larger decline in deposits and lending after an increase in the policy rate. Moreover, bank size heterogeneity dampens monetary policy. I use U.S. bank-level data and I find that banks at the 90th percentile of the withdrawal risk distribution reduce lending by an extra 1% and deposits by an extra 0.7-0.9% relative to banks at the 10th percentile after a monetary shock that raises the Fed funds rate by 100 basis points. Moreover, aggregate lending falls by 0.9% due to withdrawal risk. In the third chapter, I study the role of MPLs in the transmission of monetary policy in a general equilibrium model. I incorporate banks into a standard New Keynesian DSGE model. Banks face frictions to substitute deposits with wholesale funding. I use U.S. bank-level data to calibrate the model and I find that higher financial frictions that increase the aggregate MPL by 66% amplify the response of bank lending and investment to monetary shocks by 11% and 16%, respectively. Moreover, if the sensitivity of the marginal cost of funds also increases, the loan pass-through increases by 20%, which amplifies the response of bank lending and investment by 31% and 54%, respectively. Higher MPLs do not amplify the response of production in the short run but they do at longer horizons, due to the decline in investment.
Essays on political economy: theory, surveys, and experiments
(2024) Ferroni, Matteo F.; Fisman, Raymond
In this dissertation, I present three distinct essays where I use online survey data and experiments to answer questions in political economy. Factors such as perceptions, expectations, beliefs, and policy preferences are critical determinants of social, economic, and political outcomes. As these factors are unobservable, surveys are an essential approach for eliciting them. By relying on well-designed surveys and experiments, I shed light on the role that perceptions and expectations play in shaping institutional and policy preferences.In Chapter 1, I study how shocks to economic expectations induced by elections contribute to democratic discontent in polarized societies. Using new large-scale survey data collected throughout the 2022 Brazilian presidential election, I show that highly polarized voters who assign a large probability to their candidate’s victory experience a larger negative shock to their economic expectations in case their candidate loses. As predicted by a stylized model, this expectation shock then leads to an increase in violent and anti-democratic sentiments, an effect that is particularly strong among the most extreme supporters. In an additional survey experiment, I provide complimentary evidence in which I positively update respondents’ expectations about the economy and find that this information treatment reduces their violent and anti-democratic sentiments. In Chapter 2, joint work with Alberto Alesina and Stefanie Stantcheva, we investigate, using new large-scale survey and experimental data, how respondents perceive racial inequities between Black and white Americans, what they believe causes them, and what interventions they think should be implemented to reduce them. We document a stark partisan gap among white respondents, particularly in the perceived causes of racial inequities and what should be done about them. White Democrats and Black respondents are much more likely to attribute racial inequities to adverse past and present circumstances and want to act on them with race-targeted and general redistribution policies. White Republicans are more likely to attribute racial gaps to individual actions. A policy decomposition shows that the perceived causes of racial inequities correlate most strongly with support for race-targeted or general redistribution policies, a finding confirmed by the experimental results. In Chapter 3, joint work with Raymond Fisman and Miriam Golden, we conduct parallel surveys of legislators and citizens in three countries to understand tolerance of corruption. In Colombia, Italy, and Pakistan, legislator and citizen respondents share similar views: both groups perceive corruption as prevalent and as undesirable. These novel descriptive data further reveal that political elites generally have accurate beliefs about public opinion on corruption and appreciate its policy-relevance. Overall, results suggest that barriers to effective anti-corruption policies do not lie with lack of information by legislators.
Identity in the ecosystem: a study of international students, belonging, and culturally relevant foods
(2025) Kurtz, Gretchen; Metheny, Karen
This thesis challenges longstanding anthropocentric identity theories using an ecological approach to the study of identity and belonging. If ecology studies relationships between organisms and their environment, an ecological lens considers interactions among all actors—human and nonhuman—in a broader conception of society as ecosystem. Ecological Theory of Identity (ETI) posits that identity is symbiotically co-created between the self and other agents by examining interactions between human actors and nonhuman actors, in this case international students and culturally relevant foods (CRFs). ETI’s respect for foods’ power represents an important shift in scholarly emphasis. Identity is not merely mental: materiality matters. Students and foods have power to shape identity and well-being. Drawing from 38 surveys and 11 semi-structured interviews with international graduate and undergraduate students at a large public U.S. university, this qualitative study has important ramifications. Findings show that culturally relevant foods boost students' sense of belonging, with CRFs themselves responsible for this important uptick in well-being. Of the students who rated belonging differently when eating foods from their home country versus overall, 80% reported increased belonging when eating CRFs, at times creating a positive sense of belonging where there had been a negative one.Additionally, as students cook, shop, and increase access to foods from home, they develop resiliency and problem-solving skills. ETI's recognition of temporal dynamics reframes these critical life management skills, gained over time in the very places that Communication Theory of Identity might find personal-enacted or personal-relational identity gaps, as important parts of students’ self-concept. Through ETI, foods are scrutinized for what they are literally communicating to students, just as scholars have analyzed what individuals have communicated to others with foods, with practical ramifications for how universities can better support their student populations. What students eat, with whom they eat, and CRFs’ sensorial agency to impact memories and emotions cultivate identity and belonging in ways overlooked by traditional approaches.
Novel statistical methods for the assessment and development of polygenic scores in multi-ancestry cohorts
(2024) Gunn, Sophia; Lunetta, Kathryn L.
Polygenic scores (PGS) have incredible potential to advance biological research and precision medicine. PGS estimate an individual’s genetic liability to traits or diseases and have been widely used to identify individuals at high risk of disease. However, there are limitations to current polygenic score development and applications. The most important limitation is that PGS performance often declines when they are applied to populations different from which they were derived. Most currently available PGS were developed using primarily European-ancestry populations, and when applied to populations underrepresented in genetic research, their performance is substantially worse. The evaluation of PGS is therefore challenging because performance of PGS can vary considerably across different populations. While methods have been proposed to build PGS using multi-ancestry data that can perform better in underrepresented populations, how to best develop PGS for multi-ancestry populations is still unknown. Further, while PGS can identify individuals at high risk of disease, they do not provide insight into the sources of genetic risk for these individuals. In this dissertation, we aim to offer guidance for navigating and addressing these limitations. Our second chapter introduces methodology to evaluate the performance of multiple polygenic scores in multiple populations using correlation-based tests. We show in simulations that our method has appropriate type 1 error and reasonable power at appropriate sample sizes. We then apply our methods to height and low-density lipoprotein cholesterol, providing two examples of how our methods can be used to analyze the performance of multiple polygenic scores in multiple populations. Our third chapter compares methods for building polygenic scores for multi-ancestry populations with GWAS from multiple populations. Using population-specific GWAS results from the Million Veterans Program, we build polygenic scores for five binary and five continuous traits using both ancestry-specific and multi-ancestry approaches and evaluate the scores in three populations in the All of Us (AoU) cohort. With the statistical framework introduced in the second chapter, we compare the various approaches and find multi-ancestry scores built with PRS-CSx outperform the other approaches in the three AoU populations. Our fourth chapter introduces a method for building pathway-specific polygenic scores (PPGS) with the Bayesian method PRS-CS. In simulations we demonstrate this method outperforms the previously proposed PPGS approach PRSet. We use our proposed methods to derive PPGS for AF in the UK Biobank and demonstrate the heterogeneity of genetic risk.
Switching functional network models of oscillatory brain dynamics
(2024) Hsin, Wan-Chi; Eden, Uri T.; Stephen, Emily P.
Functional brain networks can change rapidly as a function of stimuli or cognitive shifts. Tracking dynamic functional connectivity is particularly challenging as it requires estimating the structure of the network and its changes moment to moment. In this dissertation, we describe a general modeling framework and a set of specific models that provide substantially increased statistical power for estimating rhythmic dynamic networks, based on the assumption that for a particular experiment or task there are a discrete set of networks that are expressed at any moment. Each model is comprised of three components: (1) a set of latent switching states that represent transitions between the expression of each network mode; (2) a set of latent oscillators, each characterized by an estimated mean oscillation frequency and an instantaneous phase and amplitude at each time point; and (3) an observation model that relates the activity at each recorded network node to a linear combination of the latent oscillators. We develop an expectation-maximization procedure to estimate the network structure for each switching state and the probability of each state being expressed moment to moment. We conduct a set of simulation studies to illustrate the application of these models and quantify their statistical power, even in the face of model misspecification. Additionally, we demonstrate the application of this model framework through an analysis of EEG data from a subject undergoing general anesthesia.
Usage of hypothesis testing methods for the equivalence of covariance matrices in dementia outcome analysis
(2024) Guan, Calvin; Gangopadhyay, Ashis
Dementia and its most common form Alzheimer's disease (AD) are urgent yet complicated problems to decipher and as such predictive modeling for AD outcomes attracted researchers' attention for years. Numerous works in literature have identified categories of predictors that are linked to AD such as inflammatory cerebrospinal fluid (CSF) biomarkers and brain MRI measurements. A major contribution of the thesis is introducing covariance structures as a tool to signify AD. In clinical settings, it is often imperative to account for demographic covariates as potential confounding factors. We explore two ways to account for these covariates: one by removing their effects via partial covariance and the other by calculating the covariance matrices for given values of covariates via function covariance matrix estimators. For both covariance estimating methods we use hypothesis testing methods to determine if the covariance estimates are significantly different between AD outcome groups. These methods are the parametric Tracy-Widom, the semi-parametric Forkman's test, and the nonparametric Permutation method. We evaluate the utility of the covariance estimation as well as the hypothesis testing methods via extensive simulation studies. Additionally, we apply these methods to real world data studies such as the FHS and the ADNI. Moreover, we explore the scenario where the cases are rare compared to the controls in a binary outcome (\textit{aka} rare event), as often the case in bio-medical application such as AD data. The imbalance between the outcomes has been shown to introduce bias in the estimation of the model parameters, which in turn affects the predictive probabilities. The problem becomes more severe as the imbalance becomes starker, therefore methods that adjust for the imbalance could be beneficial in such situations. As part of the thesis, we explore the adequacy of logistic regression model which is known to suffer from the problem of bias in rare event cases. Additionally, we evaluate derivative methods that aim to compensate for rare event cases such as prior correction, weighting, Firth’s logistic regression, FLIC, and FLAC using the ADNI data set. Our investigation into the performances of the various methods show that the weighting method provides a significant improvement in the predictive utility of the regression model.
“As if it had Growed There”: resort architecture and the New England landscape, 1875-1915
(2024) Granston III, David W.; Abramson, Daniel M.
This dissertation reconsiders the late nineteenth-century architecture of Northeastern resort and country destinations and situates the development of a new style - later named the Shingle Style - within period interpretations of the region’s landscape. Contextualizing designs, buildings, and material choices with environmental, landscape, and art histories, and building on existing architectural scholarship, I argue that understandings of, and values surrounding, the landscape impacted the selection of sites, reshaped the design and construction processes, and influenced the interpretation of finished buildings as well.My first chapter considers how the New England landscape was interpreted at the end of the nineteenth century, and what factors figured into the marketing of property for seasonal development. The second chapter analyzes changing architectural tastes and recognizes new practices that evolved in resort architecture during this period. This chapter argues that architects approached the design process pictorially, envisioning completed buildings within their surroundings from the start. The third chapter considers how construction materials were selected and used to foster tangible connections with sites, and also how finishes were employed to conceal the role of humans in the building process. The raw materials – stone and wood – were not unlike those found in urban contexts, but in this chapter I contend that architects and builders approached the construction process differently in resort areas, consciously obscuring the means by which materials were produced and buildings were realized. Chapter four considers how completed buildings were presented and discussed, and addresses the ways these presentations were curated and manipulated to suggest new buildings were part of their sites, or even organic, instead of products produced by humans. Writing to his mother in 1887, the Boston architect Alexander Wadsworth Longfellow, Jr. proudly relayed comments he had heard about one of his recently realized designs. Admirers, he claimed, felt the new summer residence “kind of looked as if it had growed there.” With its rough stone chimney and olive-brown stain, the building “goes beautifully with the surroundings,” he wrote. Anxious to atone for the effects new structures had on the landscape, my dissertation argues that architects like Longfellow were involved in the creation of a new style of architecture and new approaches to the design process. To minimize visual impact and foster congruity between new resort buildings and their surroundings, materials were selected and employed in novel ways and, to obscure the newness of these buildings, they were represented and discussed as though they were parts of the landscape itself.
Development of sustainable and biocompatible polymers for medical applications
(2024) Fitzgerald, Danielle M.; Grinstaff, Mark W.
The development of sustainable polymers that participate in a circular economy, from manufacture through disposal, is imperative for the next generation of novel performance materials. Specifically for medicine, additional requirements of biocompatibility and degradation necessitate thoughtful material design at the molecular level. Polycarbonates and polyesters are promising biomaterials for fulfilling these requirements. Herein, the development of poly(glycerol carbonate)s for use as pressure sensitive adhesives and drug delivery vehicles in medicine are discussed. This green polymerization, featuring low catalyst loadings, neat reaction conditions, and carbon dioxide incorporation, generates a sustainable alternative to commodity polyacrylic-based materials. The development of poly(glycerol carbonate) for applications in thoracic surgery and drug delivery are discussed, with an emphasis on defining the structure-property relationships between polymer composition and adhesive performance. The influence of both intermolecular interactions and intramolecular cohesion are probed, leading to a greater understanding of adhesive design on the molecular level. Applications of poly(glycerol carbonate) are then extended beyond adhesion to anti-cancer therapy, where a polycarbonate-based block copolymer and a polyester mesh formulation are each explored for their efficacy in localized chemotherapeutic delivery.
RNA synthesis initiation and termination by the non-segmented negative-strand RNA virus polymerases
(2024) Kleiner, Victoria Anne; Fearns, Rachel
Non-segmented negative-strand RNA viruses (nsNSVs) make up an order of viruses that range in prevalence and severity. For example, respiratory syncytial virus (RSV) causes respiratory tract infections and countless hospitalizations annually, while Ebola virus (EBOV) causes hemorrhagic fever which is less widespread but more deadly. A key commonality of the nsNSVs is their RNA-dependent RNA polymerase that can replicate and transcribe their genome. These viral polymerases (L) have five domains that are each responsible for different functions and need to coordinate with each other and other cellular and viral factors for RNA synthesis. Due to the essential nature of this enzyme, it is an attractive candidate for antiviral research. We sought to further study the polymerase and identify aspects that are important for different stages of RNA synthesis. The priming loop of the polymerase has been recognized to be involved in many different facets of polymerase function. Here, we identified priming residues that are involved in RSV initiation and characterized a non-nucleoside inhibitor that impedes the conformation changes necessary for the RSV polymerase priming loop to transition from initiation into elongation mode. In addition, mutation of a separate residue of the priming loop caused changes in mRNA polyadenylation. Besides the priming loop, we identified another region of the polymerase, the template exit channel, that contains essential residues for transcription termination/ mRNA release. Finally, we studied a cellular protein for involvement during mRNA transcription/ polyadenylation but did not find any conclusive evidence. The work in this dissertation widens our understanding of the polymerase and necessary interactions for RNA synthesis, which could be key targets for therapeutic design.
Host-pathogen cartography: predictive network mapping of Ebola virus-host cell interactions reveals mechanisms of viral control of the cell
(2024) Donahue, Callie Jeanne; Davey, Robert A.
Ebola virus causes severe disease with high case-fatality rates and is a concerning public health issue. Current therapeutics are only capable of treating acute aspects of disease and require administration soon after exposure to be effective. A better understanding of virus-host dependencies will help in development of more effective treatments. For productive infection, the virus must both counteract and co-opt different host factors to successfully replicate. To identify host factors important for virus infection, we conducted full genome siRNA knockdown and CRISPR knockout screens with infectious Ebola virus as well as a protein-protein interaction screen using virus structural proteins. Like other host-pathogen interaction screens, we found little overlap in the individual gene hits identified in each screen but predicted that application of network analysis would reveal overlaps between each dataset and help in prioritizing hits for follow up and for suggesting mechanism. Prior to this work, limited tools were available to model host-pathogen interactions in the context of both the cell and virus. Hits were mapped onto existing host genetic and protein interaction networks using several algorithms including the Prize-Collecting Steiner Forest network analysis. We investigated whether network topology from the resulting network maps could predict the identification of key genes regulating viral replication. While we found no significant relationship between network topology and hit identification, topology guided network clustering and gene-set enrichment was effective at predicting the mechanistic role of genes. The orphan gene, SPNS1, was one example of a cell entry factor for Ebola virus where its interactions with other host genes within a network cluster suggested its function in entry-relevant pathways. We additionally found that protein-protein interaction data from BioID labeling of host proteins interacting with tagged virus proteins was highly productive in identification of proteins involved in replication. Annotation of this network with those virus proteins interacting with each host protein appeared to reveal instances of virus proteins interacting with host protein complexes whose functions could be deduced by corresponding gene enrichment. We then tested this outcome by studying the relationship between viral protein VP35 and members of the host mRNA decapping complex. VP35 was found to interact with the host protein EDC4, which forms a scaffold for other members of the decapping complex and controls processing body formation. We found that Ebola virus replication depended on multiple components of the complex and demonstrated that depletion of these components resulted in a reduction in viral infection. Ultimately, our work shows that during infection, Ebola virus depends on and interacts with multiple host factors in concert rather than individually, and that the functional relationships in host factors should be considered when identifying targets for therapeutic development.
Long-acting HIV-mimicking nanoparticles for enhanced antiretroviral delivery to lymph nodes
(2024) Fofana, Josiane; Gummuluru, Suryaram
HIV remains a global threat despite the tremendous progress of combination antiretroviral therapy (cART). Major therapeutic challenges persist, marked by chronic inflammation increasing the risk of non-AIDS co-morbidities in people with HIV (PWH), and the emergence of HIV drug resistance, leading to treatment failure. These conditions are attributed to various factors including viral persistence in secondary lymphoid tissues such as lymph nodes (LNs). Notably, viral persistence in LNs has been linked to low and heterogeneous antiretroviral (ARV) distribution in these reservoirs. Furthermore, uptake of daily ARVs remains a burden, resulting in poor adherence and subsequent viral rebound. To address these challenges, we developed a novel approach using dual ARV-loaded long-acting membrane wrapped nanoparticles that recapitulate HIV trafficking to subcapsular sinus (SCS) CD169+ macrophages in LNs for improved ARV delivery and retention in these tissues. Previous studies have shown that monosialo di-hexosylganglioside (GM3) incorporated in the lipid membrane bilayer of HIV-1 particles binds to the lectin receptor CD169 on macrophages, thereby triggering formation of non-degradative virus-containing-compartments (VCCs) that preserve virus infectivity. Importantly, VCCs remain surface accessible and promote efficient HIV-1 trans-infection of bystander CD4+ T Cells. Consequently, in collaboration with the Reinhard lab, we designed biomimetic GM3-expressing membrane wrapped polymeric nanoparticles (GM3-PLA-NPs), that mimic HIV-1 capture and trafficking pathways by binding to CD169 and forming durable VCC-like compartments that we termed nanoparticle containing compartments (NPCCs). I hypothesized that by loading ARVs in GM3-PLA-NPs and the subsequent establishment of NPCCs in CD169+ macrophages, sustained viral inhibition can be achieved in macrophages and bystander CD4+ T cells through cell-to-cell ARV transfer. Therefore, I used GM3-PLA-NPs harboring a non-nucleoside reverse transcriptase inhibitor, Rilpivirine (RPV) and integrase inhibitor, Cabotegravir (CAB) and demonstrated that sustained preservation of GM3-PLA-NPs in CD169+ NPCCs prolonged antiviral potency against cis- and trans-infection in macrophages and CD4+ T cells, respectively, for one month. I next sought to characterize the capacity of GM3-PLA-NPs (also referred to as GM3-NPs) to enhance ARV distribution and retention in LNs via targeting of SCS CD169+ macrophages. Subcutaneous injection of GM3-NPs in BALB/c mice displayed specific targeting to SCS CD169+ macrophages and enhanced dissemination in LNs compared to ligand-deficient (BLK) or phosphatidylserine (PS) displaying NPs. In addition, GM3-NPs persisted in CD169+ macrophages for over 35 days. Interestingly, GM3-NPs infiltrated follicular regions of LNs and were found in close proximity to CD4+ T follicular helper cells (Tfh) which are major HIV targets in the LNs. Finally, we designed GM3-expressing lipid wrapped mesoporous silica nanoparticles (LMSNs) as an alternative to polymeric NPs, due to their enhanced stability at ambient temperatures and higher payload capacity for ARV delivery to CD169+ macrophages. I demonstrated that LMSNs extended the shelf-life of RPV and CAB and promoted antiviral synergism against HIV-1 infection. Additionally, encapsulation of mesoporous silica nanoparticles by GM3 lipid coating conferred “stealth” properties to suppress innate immune activation by encapsulated ARVs and NP cores. These studies elucidated a new approach for safe and effective delivery of long-acting therapeutics to LNs which could be beneficial for HIV prevention and treatment. Collectively, my findings demonstrated that leveraging viral intracellular capture and trafficking routes offers a promising avenue for targeted delivery of therapeutics.
Dynamic functional network connectivity and neuroplasticity in post-stroke aphasia
(2026) Falconer, Isaac; Kiran, Swathi
The recovery of language abilities post-stroke follows diverse trajectories influenced by various factors including lesion characteristics, overall brain health, demographics, and social factors. While many studies have examined static functional connectivity (sFC, i.e., time-invariant inter-regional correlations of intrinsic neural activity), little attention has been given to the influence of short timescale brain dynamics on recovery. While static functional connectivity provides a snapshot of brain function, dynamic functional connectivity (dFC) analyses allow for the detection of transient states and examination of moment-to-moment fluctuations in functional connections. Very few studies have applied this type of approach to post-stroke aphasia populations, but one such study found a specific connectivity state associated with greater treatment response, suggesting that these may be informative techniques for post-stroke aphasia research. In this thesis, I aim to address this gap by investigating (1) spatial and temporal patterns of dFC in people with aphasia (PWA) compared to healthy controls and its relationships with aphasia severity, (2) the potential of dFC to predict treatment-induced recovery, and (3) the relationship between temporal patterns of dFC and functional reorganization during recovery, using both empirical data and computational simulations to test a hypothesized mechanism for these relationships. A major focus of this work is on temporal metrics of dFC, rather than specific spatial features or network properties, particularly temporal variability (TV), which specifically measures the magnitude of fluctuations over time. Neural variability, measured using numerous approaches including TV of dFC, has recently emerged as an important factor related to cognition, behavior, and mental health. Spatial patterns and network properties of dFC were investigated as well and discussed in the context of previous findings. The first study (Chapter 1) investigated alterations in dFC due to stroke by comparing TV and fractional occupancy of (i.e., time spent in) connectivity states between PWA and healthy controls. Additionally, relationships between each of these measures and aphasia severity were investigated. PWA were found to have reduced TV in language network regions compared to healthy controls and spent more time in a highly integrated state (state 3) with low modularity (i.e., little segregation between locally specialized communities). Higher TV was also associated with less severe aphasia, particularly in PWA with larger lesions. Although state 3 represents altered connectivity in PWA, those who spent more time in this state were no more severe than those who spent less time in it, suggesting that it is not simply a state defined by stroke-related dysfunction but may also (or instead) represent compensatory and adaptive changes.Given that dFC is altered in PWA and relates to aphasia severity, the second study (Chapter 2) aimed to determine whether it is also predictive of response to aphasia therapy. Consistent with the main finding of Chapter 1, PWA with higher TV at baseline were found to have greater treatment-induced gains in picture naming accuracy. A second temporal metric, community stability, which measures the tendency of the brain to maintain a given dFC configuration for longer lengths of time, was also positively associated with treatment response. For the dFC states analysis, participants with higher fractional occupancy of a higher modularity state showed greater improvement with treatment, consistent with the previous study mentioned above. Results of both the TV and dFC states analyses are consistent with the findings of Chapter 1, with PWA who are more normal-like (i.e., having higher TV and higher modularity) showing greater treatment gains. Given the apparent benefit of higher TV suggested by the findings of Chapters 1 and 2, we propose a mechanism linking higher TV to improved recovery: (1) Transient inter-regional synchronization facilitates plasticity in the synaptic connections between the respective regions, and (2) greater diversity of these transient synchronizations (i.e., higher TV) provides a greater variety of opportunities for plasticity mechanisms to reshape functional networks. The third study (Chapter 3) sought to investigate this proposed mechanism by testing the hypothesis that PWA with higher TV have a greater capacity for functional network reorganization, measured here as treatment-induced changes in sFC. These changes were quantified using global and node-level graph metrics computed from pre- and post-treatment scans of the same sample of PWA used in the second study. Node strength, a measure of a region’s overall connectivity with the rest of the brain, was found to decrease from pre- to post-treatment, and greater decreases were associated with greater behavioral treatment response. This decreased node strength may indicate a subtle shift toward segregation and local specialization that was not adequately captured by global-level segregation measures, which were not found to change significantly. Additionally, PWA with higher baseline TV showed greater decreases in node strength, supporting the hypothesis that TV facilitates functional network changes underlying recovery. Brain dynamics simulations tested two Hebbian-like plasticity rules and showed that only one of these was able to produce increases in network properties thought to be associated with better function (i.e., modularity and small-worldness). According to this rule, changes in average synaptic weights between a given pair of regions was inversely related to simultaneous coactivation of the respective regions with all other regions (referred to in this thesis as “mutual coactivation”). Additionally, simulations using this rule in which brain dynamics had greater TV showed greater increases in these network properties, consistent with the hypothesized mechanism. Ultimately, this work demonstrates a robust relationship between greater TV and better outcomes in post-stroke aphasia and provides support for an underlying mechanism of TV facilitating plasticity in PWA.
Principles of microbiome structure and their implications for climate change mitigation
(2025) Silverstein, Michael Robert; Segrè, Daniel; Bhatnagar, Jennifer M.
Microorganisms assemble into diverse communities of various ecological structures across virtually all of Earth’s environments, where they drive biogeochemical processes that range from microscopic to climatic scales. While most studies of these communities, or microbiomes, have been traditionally focused on demographic surveys of what microbes reside where, there is increasing interest in gaining a more systematic perspective and identifying general principles that govern microbiome structure, i.e. the way in which the resident taxa or the functions enacted by those taxa are organized within a community. Uncovering these principles could enable unprecedented control over the ecological structure of microbiomes as well as their surrounding environments, paving the way to microbiome engineering. In this dissertation, I first reviewed the legacy of prior attempts at employing environmental microbiome engineering towards sustainability and climate change mitigation, including proposed approaches for overcoming outstanding challenges to its implementation. A promising approach consists of using directed evolution to design microbial communities as inocula to boost the carbon stabilization capacity of soils. Importantly, this avenue would require further research into the principles that govern the establishment of new microbial communities into existing ecosystems. Second, I used experimental and computational approaches to specifically address this last aspect of microbiome engineering, focusing on the question of whether different microbiomes states are, in principle, possible under a given environment. This study uncovered a novel principle of microbiome structure: that environmental metabolic complexity drives the taxonomic divergence of microbial communities, or how taxa differ between communities. This suggests that complex environments may be more susceptible to microbiome engineering since these environments can host a larger diversity of types of microbial communities, which may include communities with higher capacity than the resident one to perform climate change mitigating activities, for example. Finally, I expanded the previous analysis to explore how taxonomic divergence relates to functional divergence — an essential step to ensure that communities displaying distinct taxonomic composition in a given environment do not always converge to identical functions, which would reduce the impact of the engineering effort. To address this question, I developed a novel metric called functional response to understand what environments can host communities that vary in function. I then measured the functional response of microbial communities in natural microbial communities from an existing dataset and in synthetic communities from a newly generated dataset. Ultimately, these projects contribute to growing efforts to understand microbiome structure and inform engineering efforts to overcome challenges in microbially-regulated systems, such as with human disease and climate change.

Browse

Recent Submissions