22689 research outputs found
Sort by
SIMULATING FUNDAMENTAL FIELDS IN STRONG GRAVITY: ASTROPHYSICAL IMPLICATIONS OF ULTRALIGHT DARK MATTER
Get PDFThe nature of dark matter remains one of the biggest mysteries in modern-day physics. Ultralight dark matter models hypothesize that dark matter is composed of one or possibly several bosonic particles with their mass between approximately and .
In this thesis, we use numerical relativity to understand the behavior of ultralight dark matter under the influence of strong gravity in the nonlinear regime. We consider three astrophysical scenarios: black hole superradiance, gravitational Magnus effect, and polarized Proca stars.
We first study the phenomenon of black hole superradiance, where bosonic particles can extract energy from spinning black holes. While ultralight dark matter can trigger black hole superradiance, we investigate a competing mechanism in which photons in a diffuse plasma gain an effective mass and thus can also trigger superradiance. We carry out relativistic simulations of a Proca field evolving on a Kerr background, with modifications to account for the spatially varying effective mass. We find that a constant asymptotic mass or a shell-like plasma structure is required for superradiant growth.
Next, we study a rotating black hole moving at relativistic velocities through scalar dark matter. We simulate the system numerically and extract the Magnus force on the black hole perpendicular to its motion.
We confirm that the force scales linearly with the dimensionless spin parameter of the black hole up to . We also reveal a significant nonlinear correction at relativistic speeds of the black hole up to 0.55 of the speed of light.
Lastly, we study the ground state of a spatially localized massive vector field (Proca stars) as a model for ultralight dark matter. We include general relativistic effects and numerically investigate the stability of compact polarized Proca stars. We find that these initial conditions lead to stable configurations. However, they can collapse to black holes at sufficiently large initial compactness. Polarized Proca stars collapse at higher compactness compared to non-polarized states
BRING NEW CHEMISTRY TO NONHEME ENZYMES
No full textRepurposing natural enzymes to catalyze synthetic transformations absent in biology has emerged as a significant research field bridging chemistry and biology. A key challenge in this pursuit is the introduction of synthetic reaction mechanisms into natural protein scaffolds. Over the past decades, substantial breakthroughs have been achieved, with many enzymatic systems developed to catalyze critical chemical transformations not previously observed in biology. However, much of this progress has focused on proteins or enzymes containing heme or organic cofactors. In this context, this thesis explores the potential of diverse, yet vastly underexplored, nonheme enzymes to unlock synthetic reaction mechanisms currently unprecedented in biology. The key design concept is to identify mechanistic connections between synthetic reactions and natural nonheme enzyme systems, which will be leveraged to discover initial activities for non-natural transformations and to further expand the scope of related abiotic transformations. In this context, the first focus of this thesis is to enable synthetic radical-relay reaction mechanisms in nonheme enzymes by generating both nitrogen-centered and oxygen-centered radical species via the reductive activation of various radical precursors by the iron(II) center of nonheme enzymes. The second focus centers on interfacing synthetic photoredox strategies for radical generation with the stereoselective interception of carbon-centered radical species by the iron center of nonheme enzymes. Altogether, this thesis demonstrates that an array of abiological radical transformations can be achieved by natural nonheme enzymes, which paves the way for the future development of this important class of enzymes for diverse synthetic applications. Additionally, mechanistic studies of these non-natural systems could deepen our understanding of fundamental enzymatic catalysis and bioinorganic chemistry
EXPLORING THE RELATIONSHIPS BETWEEN WEIGHT MANAGEMENT HOPE AND WEIGHT-RELATED COGNITIONS, BEHAVIORS, AND OUTCOMES IN PEOPLE WITH OBESITY
Get PDFBackground: Obesity has reached epidemic levels in the US. Questions persist about which psychological factors can be targeted to support weight management (WM). Grounded in hope theory, this dissertation introduces weight management hope (WMH) as a novel construct reflecting individuals’ beliefs in their abilities to generate strategies and sustain motivated effort in pursuit of WM goals.
Methods: A cross-sectional online survey was administered to US adults with current or former obesity pursuing WM goals. Study 1 developed the Weight Management Hope Scale (WMHS) and used exploratory factor analyses, correlations, and hierarchical linear regressions to analyze its psychometric properties (N = 395). Study 2 used bivariate correlations and hierarchical linear regressions to examine associations between hope constructs and behavioral automaticity across five weight-related behaviors (N = 385). Study 3 employed mixed-methods to analyze the content of participant-generated WM goals and test their associations with WMH (N = 365).
Results: Study 1 supported a two-factor WMHS structure (agentic and pathways thought). The scale demonstrated internal consistency (α = .79) and was significantly associated with dispositional hope (DH), goal commitment, and weight loss self-efficacy (r = .50–.62, Ps < .001). WMHS scores were inversely associated with BMI (rs = –.24, P < .001) and lifetime weight loss attempts (rs = –.14, P = .006) and accounted for additional variance in BMI beyond established constructs (ΔR² = .01–.03, Ps < .05). In Study 2, both WMH (r = .25–.37, Ps < .001) and DH (r = .21–.32, Ps < .001) positively associated with behavioral automaticity across all five behaviors. WM agentic thought associated more strongly with automaticity than WM pathways thought for three behaviors. The WMHS accounted for additional variance in automaticity beyond DH (Step 3: ΔR2 = 1%–7%, Ps < .05). In Study 3, most goals focused on body composition, were outcome-oriented, and specific. WM agentic thought associated with specific (OR = 1.14, P = .018) and behavioral (OR = 1.17, P = .008) goals, and WM pathways thought associated with the number of goal strategies (IRR = 1.03, P = .031).
Conclusions: This dissertation extends hope theory into the domain of WM and establishes WMH as a novel latent construct. Across all three studies, findings provided consistent evidence for the validity, reliability, and potential utility of the WMHS in obesity research and practice
Multidimensional Imaging Reveals Morphogenetic Heterogeneity in Embryogenesis and Early Pancreatic Tumorigenesis
No full textNovel multidimensional imaging approaches can visualize previously unappreciated heterogeneities in complex biological events such as embryonic development and cancer initiation. Time-lapse imaging of transgenic reporters can capture transient biochemical activity patterns during dynamic tissue change. In addition to developing several new transgenic mouse strains with real-time signaling reporters, we characterized of spatiotemporal heterogeneity in JNK pathway signaling in early preimplantation mammalian embryogenesis and uncovered evolutionary links to animal multicellularity. By integrating a 3D tissue reconstruction workflow with in situ assays for oncogenic mutants, we spatially mapped intralesional genetic heterogeneity within pancreatic precancers. Based on novel structural phenotypes seen in three-dimensional tissue reconstructions, we identified histological correlates for morphological phenotypes of pancreatic intraepithelial neoplasia, ultimately developing a deep-learning model capable of extrapolating histological features from immunohistochemistry-based annotations. Our analyses suggest that these 3D morphological phenotypes resemble cell populations seen in normal gastric mucosa and gastric metaplasia, raising the possibility that gastric phenotypic differentiation programs may contribute to pancreatic precancer growth and progression. Overall, these tools and the resultant discoveries will expand our understanding of multidimensional heterogeneity in both physiological and pathological tissue development and morphogenesis
Oral History of R.T.
No full text“RT” discusses her parents and grandmother, immigrants from El Salvador, her older brother, her time growing up in Palmdale, California, and her medical program in high school that allowed her to gain early clinical experience, including being at a clinic when the COVID-19 pandemic hit. She talks about coming to Baltimore to attend Johns Hopkins University and realizing that she loved the combination of humanities with healthcare, majoring in Medicine, Science, and Humanities with a minor in film. She also discusses the impact of the 2024 US presidential election on Johns Hopkins students and researchers, her love of Baltimore and a film that she made for a class featuring the Hispanic community in Baltimore’s Highlandtown neighborhood, and her plans to take the MCAT and attend medical school
A Theoretical Investigation Into Adversarially Robust Learning
Get PDFMachine learning models based on neural networks power the state-of-the-art systems for various real-world applications. However, systems based on neural networks admit vulnerabilities in the form of adversarial attacks. One such limitation is due to arbitrary adversarial corruption of data at the time of training, commonly referred to as data poisoning. Another vulnerability is due to arbitrary adversarial corruption of data at the time of prediction, commonly referred to as inference time attack. These attacks involve carefully designed input perturbations that may deceive machine learning models. There has been much empirical interest in finding robust predictors against various attacks, and in this dissertation, we focus on the theoretical perspective from the following three aspects.
The first part investigates the robustness of stochastic approximation approaches against data poisoning attacks. We focus on two-layer neural networks with ReLU activations and show that under a specific notion of separability in the RKHS induced by the infinite-width network, training (finite-width) networks with stochastic gradient descent is robust against data poisoning attacks.
The second part characterizes the robustness properties of neural networks. We start by explaining why adversarial examples exist for neural networks trained using stochastic gradient descent. Our analysis reveals that over-parametrized neural networks in the lazy training regime, which enjoy strong generalization and computational guarantees, remain vulnerable to a single-step gradient ascent attack. This negative result highlights that robustness is unattainable for all networks within the lazy regime, irrespective of the training algorithm, thus failing to explain the practical robustness achieved through adversarial training. To address this, we extend beyond the lazy training regime and demonstrate that robust networks do exist and can indeed be found using adversarial training under mild distributional assumptions.
The third part focuses on robust transfer learning, a practical scenario where training (source domain) and test (target domain) data come from different distributions. This challenge is particularly relevant in several settings. For instance, due to the expensive cost of labeling, unsupervised domain adaptation considers the scenario where the learner has access to a labeled source domain and an unlabeled target domain. In other cases, privacy concerns may restrict access to the original training data, leaving the learner dependent on pre-trained models, as discussed in the hypothesis transfer learning setting. Across these settings, our ultimate goal is to obtain a robust predictor that performs effectively on the target domain while ensuring robustness against adversarial attacks. To this end, we design various defense algorithms capable of handling both distribution shifts and adversarial threats. We establish provable robustness guarantees for the resulting model and provide a detailed characterization of the statistical and computational complexity inherent in adversarially robust learning
Single Particle ICP-MS for Characterizing Metals Associated with E-Cigarette Use
No full textElectronic cigarettes (e-cigs) are the most commonly used tobacco products among youth. While initially marketed as a harm-reduction tool, e-cigs emit toxic substances, including heavy metals, raising serious public health concerns. In this dissertation, we assessed the metal and metal particle content in aerosols from popular disposable e-cig brands, including FDA-approved devices and those with ceramic coils, and evaluated the cellular metal and metal particle uptake.
Aerosols from popular devices were collected as condensates from Vuse Alto, EB Create, and Orion Bar, and were analyzed using bulk and single-particle inductively coupled plasma mass spectrometry (SP-ICP-MS) for the presence of As, Cr, Cd, Fe, Ni, Pb, Zn. All aerosols contained all metals tested, with Cr, Fe, Ni, Pb, and Zn present as particles ranging from 14.2–281.3 nm. Vuse Alto consistently exhibited the highest particle counts, except for Cr, which was most abundant in EB Create devices. Metals and metal particles were also detected in ceramic coil condensate investigated, with Fe having the highest number concentrations.
To examine biological uptake, A549 lung epithelial cells and RAW 264.7(RAW) macrophages were exposed to e-cig aerosol using an In Vitro chamber system. Single-cell ICP-MS revealed accumulation of Ni ions and Cr particles in individual RAW cells but
not A549. Gene expression analysis indicated significant changes in oxidative stress
genes, suggesting potential for cellular dysfunction.
These findings demonstrate the presence of metals and metal particles in e-cig aerosols—including in FDA-approved products—and highlight their potential to enter and affect lung cells. The use of SP/SC-ICP-MS provides novel insights into metal uptake and cellular heterogeneity. Together, these results emphasize the need for further toxicological evaluation and regulation of e-cigarette devices, especially concerning metal emissions and their health implications
Finding Their Footing at the Front of the Classroom: The Emerging Professional Identities of Pre-Service Teachers
No full textPre-service teachers’ resilience is dependent on the development of teacher self-efficacy as part of their emerging teacher identity. This process is complicated by teacher preparation programs that perpetuate systemic inequities and complicate the sociocultural identities of preservice teachers. The study in this dissertation used the Phenomenological Variant of Ecological Systems Theory (PVEST) as a theoretical framework to examine the influences of various social interactions and support systems that nurture a preservice teacher’s identity. The resulting conceptual framework explored how a sense of autonomy, competence, and relatedness within the profession leads to increased teacher self-efficacy, motivation, and persistence as an extension of teacher identity. The convergent parallel mixed-methods study sought to better understand the formation of teacher identity and self-efficacy as an interaction of the identities and experiences of preservice teachers. Additionally, the study examined how various systems embedded in a teacher education program at a small private university on the West Coast could foster adaptive strategies that strengthen preservice teachers’ resilience in the profession. Quantitative data were collected using the Teacher Identity Measurement Scale. Qualitative data were collected through interviews and storyline illustrations. Findings revealed that preservice teachers entered the program with strong motivations towards teaching and a clear teacher identity but initially struggled with low teacher self-efficacy. As they progressed through the program, their teacher self-efficacy increased, and their teacher identity and self-efficacy became more closely aligned. Mentor teachers emerged as the most significant source of confidence and support, offering encouragement and constructive feedback that communicated their belief in the preservice teachers’ abilities to be successful. Despite the difficulties of the program, the preservice teachers in the study developed adaptive coping mechanisms that included leaning into meaningful relationships with each other, focusing on their love for students, and developing a reflective stance towards their practice. The participants’ teacher self-efficacy grew through student teaching experiences, influenced by both their personal and professional identities. These findings suggest that teacher preparation programs must acknowledge the emotional journey of preservice teachers, encourage reflective practices, and provide robust support systems that promote teacher identity formation, self-efficacy development, and ultimately, resilience in the teaching profession
Oral History Interview with Daniel I. Wikler
No full textThis interview with Daniel I. Wikler, PhD, is part of “Moral Histories: Voices and Stories from the Founding Figures of Bioethics,” an oral history project of the Johns Hopkins Berman Institute of Bioethics. Professor Wikler is the Mary B. Saltonstall Professor of Ethics and Population Health in the Department of Global Health and Population at the T.H. Chan School of Public Health at Harvard University. His early research projects and publications addressed many issues in bioethics, including reproduction, transplantation, and end-of-life decision making. Later work focused on measuring global health disparities, ethical issues in population and international health, the allocation of global health resources, and dilemmas in public health practice, including the ethical dimensions of global tobacco control policy.
Wikler recounts his upbringing in Lexington, Kentucky on the grounds of the U.S. Narcotic Farm (later known as the U.S. Public Health Service Hospital), where his psychiatrist father worked. He described his education at Oberlin College and his graduate work in philosophy at the University of California Los Angeles. While at UCLA, he described a medical school symposium regarding the treatment of children with spina bifida as shaping his initial interest in bioethics. Wikler describes joining the University of Wisconsin-Madison's medical ethics program and his tenure there. He discussed his time during the Carter administration when he served as Staff Philosopher for the President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research.
Prof. Wikler describes his visits to China, his growing interest in global bioethics, and how he came to be the co-founder of the International Association of Bioethics. He recounts his position as the first staff ethicist for the World Health Organization in Geneva, Switzerland and his involvement with the Global Burden of Disease Studies. Wikler discussed his move to Harvard and the establishment of the Harvard Program in Ethics and Health. The interview concludes with a discussion of the ethical challenges in public health and the importance of addressing global health disparities
Oral History Interview with Gilbert Meilaender
No full textThis interview was conducted with Gilbert Meilaender as part of “Moral Histories: Voices and Stories from the Founding Figures of Bioethics,” an oral history project of the Johns Hopkins University Berman Institute of Bioethics. Professor Meilaender is a Senior Research Professor at Valparaiso University. His areas of expertise include theological ethics, Christian ethics, human dignity, the philosophy of friendship, and adoption. He is the author of several books, including Bioethics: A Primer for Christians, Not by Nature but by Grace: Forming Families through Adoption, and Friendship: A Study in Theological Ethics. Professor Meilaender discusses his upbringing as the son of a Lutheran pastor, his education at Concordia Senior College and his path to academia after being ordained as a Lutheran minister. He discusses his graduate studies at Princeton University with mentor Paul Ramsey. He talks about his identity as a theological ethicist in a time when higher education was trying to distinguish the academic study of religion from theological study. He also discusses his experience with foster care and adoption, which shapes his view on reproductive technologies and the implications of the unquestioned use of such technologies. Professor Meilaender talks about his involvement with The Hastings Center and his work on the President’s Council on Bioethics during the George W. Bush administration. He notes that the Council’s work was philosophical, in contrast to law- and policy-oriented bioethics. He discusses the influence of the Council’s work on his thinking about human dignity, as well as the limits of bioethics, his approach to politics, and his belief in including religious views in public debate. The conversation concludes with reflections on his influences, friendships, correspondence with readers, and views on end-of-life care