545298 research outputs found
Sort by
Rapid Discrimination of High-Grade Prostate Cancer Using Label-Free Fluorescence Lifetime Measurements
Get PDFPurpose: Histologic evaluation of prostatic needle biopsies is essential for prostate cancer (PCa) diagnosis and treatment planning, yet tissue targeting remains suboptimal despite MRI-guided Bx procedures. This pilot study investigates the use of label-free Fluorescence Lifetime Imaging (FLIm) for real-time biopsy guidance. Using ex vivo specimens, we assess FLIm's preliminary efficacy in discriminating malignant from benign prostate tissue.
Materials and Methods: Twenty patients undergoing prostate biopsy were enrolled. FLIm measurements were performed immediately after sample collection using a custom fiber-optic probe. Optical parameters from 4 spectral bands associated with distinct endogenous fluorophores including structural proteins and metabolic cofactors (e.g. NADH, FAD) were extracted and labeled based on histological annotation. Data were analyzed to characterize tissue-type differences and train and evaluate a classifier to distinguish malignancy.
Results: Separation between benign tissue and Gleason grade ≥4 PCa was achieved using just 2 of 56 FLIm-derived parameters. A Support Vector Machine classifier using all parameters achieved a ROC of 0.88 in identifying grade 4 PCa. A reduced lifetime value in the NADH-associate band, likely due to increased free NADH from upregulated glycolysis, supports the biochemical basis for optical differentiation.
Conclusions: FLIm shows significant potential for high-grade PCa identification. The single-fiber approach requires minimal modification for integration into current biopsy tools, supporting its feasibility for clinical translation
FOVDA: A Federated Architecture for Overcoming Data Silos in Water Domain [Vision]
Get PDFEffective water management relies on integrating data from diverse sources, including both static and dynamic datasets. However, the challenge of data silos, especially in cities and agencies with disparate systems, has hindered progress in this domain. To address this issue, we introduce FOVDA (Federated Ontology View Data Access), an ontology-driven federated system designed to overcome data silos in the water domain. FOVDA enables seamless data integration and querying across heterogeneous data stores by leveraging a federated architecture and a domain-specific ontology. This system supports both local and global data interoperability, allowing agencies to exchange critical water-related data while maintaining data sovereignty. FOVDA’s federated query engine facilitates complex queries across distributed datasets, enabling decision-makers to access comprehensive insights for tasks such as water resource management, infrastructure resilience analysis, and disaster response. By bridging data silos, FOVDA empowers stakeholders with actionable insights, improving efficiency and collaboration in water management
Essays on Experimental Finance
Get PDFBehavioral biases affect investment decisions, leading to opportunity costs for investors and inefficiencies for the economy as a whole. These biases can be problematic to study “in the wild" due to numerous confounding factors that make clear-cut causal identification difficult. Therefore, a laboratory setting is often needed to draw robust conclusions about behavioral finance.In this dissertation, I present three chapters exploring the results of experiments I conducted at the UC Santa Cruz Learning & Experimental Economics Projects (LEEPS) laboratory, which shed new light on how investors' biases affect asset valuation.The first chapter examines the tendency of investors to overvalue growing assets when such assets might be superstars. Superstars are assets that will sustain strong growth far into the future. Superstar assets readily come to mind when investors see growing assets. Therefore, when investors see growing assets that might be superstars, they are primed to believe that such assets are likely to be superstars and, therefore, overstate their payoffs and pay too high a price. Outside of the laboratory, such assets often have much hype around them, and it is difficult to disentangle the overvaluation due to this hype and the overvaluation due to overextrapolation of past performance. By presenting subjects with simple assets in a lab that do not have a romantic story or an exciting personality behind them, I can determine that growing assets are overvalued even in the absence of hype, provided that such assets have the chance of being a superstar. When the possibility of being a superstar is removed, I do not find evidence of overvaluation.The second chapter focuses on finding an estimate for the key parameter in a popular model of distorted expectations known as the diagnostic expectations model. In this model, people overreact to news. For instance, if an asset has been growing, it is more likely to be a superstar. However, they update the probability that the asset is a superstar too much. The extent to which people overreact is given by the representativeness parameter θ. While other papers have estimated this parameter using observational data, to my knowledge, my work is the first to use laboratory evidence, which has fewer confounding factors, to estimate θ. The nature of the data requires multiple judgment calls, and the various specifications yield different estimates of θ. However, my best estimates are broadly in line with those contained in the prior literature.The third and final chapter introduces a novel phenomenon called the Winner's Markdown. To my knowledge, this has not yet been documented in the literature. The markdown of a risky asset is the difference between an investor's guess for the asset's payoff and the investor's maximum willingness to pay for the asset. This markdown is greater for assets that have increased in fundamental value than for assets that have decreased, hence the name the Winner's Markdown. The Winner's Markdown can explain the Disposition Effect, which is the tendency of investors to sell winners too early and hold on to losers too long. The Disposition Effect has also been explained by the tendency for investors to be risk averse in gains and risk seeking in losses as well as by the false belief that assets will display a reversion to the mean. In my laboratory experiment, I show that a Disposition Effect tendency manifests itself even when the structure of the experiment eliminates these as motivations
Learning-based Framework for Heart Arrhythmia Classification and Forecasting
Get PDFHeart arrhythmia is a chronic medical disorder disease characterized by irregular heartbeats. It is typically not identified until it reaches a severe stage, potentially resulting in high mortality rates. Electrocardiogram (ECG), one of the most popular cardiac tests, is a quick and painless tool for early diagnosis. Machine learning algorithms have been advocated as a promising tool to facilitate the analysis of ECG signals. Prompt and precise identification of arrhythmia is essential for timely intervention and the avoidance of fatal consequences.This thesis addresses significant challenges in the identification of heart disease using learning-based frameworks on electrocardiograms, which are data imbalance, domain discrepancies across patients, model accuracy and model complexity, and generalization capability, leading to a reduced reliability of the learning-based framework.To close the gaps, including high model complexity, time-consuming training, and low accuracy, the developed convolutional neural network (CNN) model takes responsibility for classifying ECG signals under the intra-patient paradigm by utilizing the data collected from the Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology (MIT-BIH Arrhythmia Database). To significantly improve arrhythmia classification, the proposed framework uses a preprocessing denoising approach prior to classification, which is capable of enhancing the Signal-to-Noise Ratio (SNR) by around 13.6 dB. The proposed model is designed to classify 3 super-classes, 5 sub-classes, and 6 sub-classes, to achieve an average classification accuracy of 99.11%, 99.1%, and 98.88%, respectively.To address significant intra- and inter-patient variability in cardiac signals comprehensively, which cause notable domain shift and class imbalance issues that hinder model generalization across diverse patient cohorts and limit the accuracy and robustness of existing classification approaches, we propose a novel distributional adaptive learning framework capable of simultaneously performing ECG classification and synthetic ECG generation, addressing these challenges comprehensively. Our approach integrates a transformer-based generator with a convolutional discriminator to generate synthetic ECG signals, effectively alleviating data imbalance issues. To further mitigate domain discrepancy and improve generalization across unseen patient recordings, we introduce single-heartbeat test-time adaptation (TTA). We rigorously evaluate our model using the MIT-BIH Arrhythmia Database, MIT-BIH Supraventricular Database, and INCART Arrhythmia Database. Results outperform state-of-the-art models across intra- and inter-patient settings, highlighting the effectiveness of synthetic ECG augmentation and the robustness of our model in real-world clinical scenarios. It paves the way for personalized and adaptive ECG-based diagnostics, underscoring the potential of generative models for advancing digital health solutions in cardiovascular medicine.To close the knowledge gap regarding efficient forecasting models that handle multirate-sampled data while capturing long-term dependencies along with good generalization, we propose a CNN-Informer Framework for long-term time series forecasting of ECG signals. Through rigorous evaluations, the model achieves a long-term prediction accuracy of 94.89%, a short-term prediction accuracy of 97.07% on the database from Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology (MIT-BIH Arrhythmia Database), and 97.93% on the 12-lead ECG arrhythmia database under the auspices of Chapman University and Shaoxing People’s Hospital. By accurately predicting arrhythmia onset, the CNN-Informer model has the potential to significantly enhance patient care and management.In concert, these contributions of this thesis represent the significant achievements in the design of heart disease identification frameworks, which are capable of low model complexity, high accuracy performance, well generalizability, and good stability
Player Engagement with Idle Games: A Mixed-Methods Exploration with Design Implications
Get PDFIdle games are a relatively recent type or genre of game, characterized by minimal player interaction. Due to their less interactive nature, there has been a research gap in the understanding of player engagement within idle games. This thesis presents preliminary research examining the engaging aspects of idle game design through two studies: 1) a semi-structured group interview with idle game players, and 2) a 2 x 2 (idle game vs. casual game x PC vs. mobile) experimental diary study. Findings revealed that idle games are as engaging as casual games since there was no statistically significant difference in engagement and gameplay time. Key characteristics of favored idle games are also outlined, highlighting notable gameplay patterns within idle games and exploring players’ motivation. Overall, this thesis provides game designers and researchers with a more nuanced understanding of idle games, their relationship to casual games, and how they can be designed for more effective engagement across PC and mobile devices
Ecology of Speciation and Adaptive Divergence in Tropical Plants
Get PDFSpeciation is driven by the evolution of reproductive barriers that reduce gene flow between diverging populations, often promoted by ecological adaptation. In this dissertation, I investigate how ecological and floral divergence contribute to reproductive isolation and speciation in Neotropical plants, with a focus on the genus Costus. Chapter One documents the origin of C. flammulus, a high-elevation peripheral isolate, as a case of budding speciation supported by phylogenetic, morphological, and ecological evidence. Chapters Two through Four examine a recent pollination shift between sister species, bee-pollinated C. kuntzei and hummingbird-pollinated C. wilsonii. I assess floral trait divergence and pollinator specificity, and test hypotheses on the drivers of pollination shifts along elevational gradients. Reciprocal translocations demonstrate that adaptation to hummingbird pollination in montane habitats is driven by increased visitation and per-visit efficiency, not by declining bee activity. Field experiments further reveal that floral isolation is context-dependent and partially confounded by ecogeographic and habitat isolation. Together, these studies underscore the role of ecological and floral divergence in shaping reproductive barriers and illuminate the processes by which ecological adaptation and pollination shifts contribute to tropical plant speciation
Reassessing Sub-Neptune Structure and Evolution: From the Deep Interior to the Escaping Upper Atmosphere
Get PDFSub-Neptunes are a class of low-mass, large-radius planets that are ubiquitous in our galaxy. They exhibit unique characteristics that distinguish them from both terrestrial and giant planets. In particular, their radii are strongly influenced by the mass of their convective envelopes, which are, however, highly susceptible to mass loss over time through atmospheric escape. Modeling efforts have shown that it may be responsible for sculpting the observed ``radius valley'', a dearth of planets separating the sub-Neptune and super-Earth populations. Over the past decades, several atmospheric escape mechanisms have been proposed, many of which successfully reproduced the observed bimodal distribution. However, the relative contributions of these processes still remain unclear. Moreover, most existing models of sub-Neptune evolution and mass loss have been adapted from frameworks originally developed for giant planets, potentially overlooking key differences in structure and evolution.
In this thesis, I developed a state-of-the-art Python-based sub-Neptune evolution model that advances previous frameworks by incorporating more comprehensive interior physics, self-consistent physical assumptions, and refined numerical methods. Using this improved numerical tool, I first analytically re-examined the role of core-powered mass loss and found it to have a negligible impact on the observed population distribution. Second, my analysis revealed that previous models systematically underestimated planetary radii and overestimated the influence of photoevaporation. I argue that boil-off is the dominant mass loss mechanism, responsible for the majority of total envelope loss. Several key physical processes, previously overlooked, are shown to contribute to the discrepancies in earlier models. In this thesis, I present comprehensive mass-radius relationships, improved initial conditions, and new analytical mass loss scaling laws that can be readily incorporated into other evolution frameworks. Third, by comparing our model predictions with the observed properties of super-puff planets, I find strong agreement, supporting a scenario in which these planets were born H/He-rich and experienced both boil-off and photoevaporation, likely in-situ
Flooding the Zone: An Agent-based Exploration
Get PDFOnline public discourse faces many threats such as human and bot networks spreading disinformation or harassment campaigns aimed at excluding certain voices. One such threat is the strategy of 'flooding the zone': intentionally pumping into the discourse information that is irrelevant to, or distracting from, an important issue. This technique is employed by both individual and state actors with seeming success. How and why that technique is successful, by contrast, is less well understood. In this paper we use agent-based modelling to help elucidate the disruptive impact of flooding the zone on communication itself. Specifically, we probe the ways in which flooding hampers the spread of relevant information and show consequences of this even for idealized, rational, actors
Persistent eye contact by a patient with temporary blindsight followed by visually conscious prosopagnosia: A case report
Get PDFA single case study of a female patient with temporary blindsight is reported. The patient experienced traumatic brain injury from a fall and experienced blindsight for approximately 61 hours. While hospitalized, the patient exhibited unusually persistent eye contact with hospital staff and visitors that was maintained at varying distances. Follow-up interviews during partial recovery of conscious vision revealed a complete failure to recognize familiar faces. Without individual recognition, facial Gestalts were perceived only by progressive scanning of facial features. Later CT-scans show bilateral scaring in the optic-radiation fibers projecting from the lateral geniculate nucleus to the primary visual cortex that includes scaring in grey matter. A perimetry test showed degraded vision in the upper visual fields and vision tests showed intact foveal perception sufficient for projected chart-letter recognition. The ability to see the eyes of other people using blindsight, as characterized by intense visual fixation of the eyes of others, suggests continued functionality of the superior colliculus, an evolutionarily ancient neural structure capable of recognizing the schema of two-facing eyes
In-beam spectroscopy reveals competing nuclear shapes in the rare isotope 62Cr
Get PDFIn recent decades, rare-isotope facilities have enabled the study of short-lived, neutron-rich nuclei. Their measured properties indicate that shell structure changes in the regime of unbalanced neutron-to-proton ratios compared with that of stable nuclei. In the so-called islands of inversion in the nuclear chart—around the neutron-rich nuclei 32Mg, 42Si and 64Cr, for example—the textbook shell model predicts spherical shapes due to the respective magic neutron numbers of 20, 28 and 40 of these nuclei. However, nuclei in these regions turn out to be deformed in their ground states. Another hallmark of these islands is shape coexistence, where a nucleus assumes different shapes with excitation energy. Here we present evidence for this phenomenon from the observation of an excited 0+ state in 62Cr, two neutrons away from the heart of the island of inversion around neutron number N = 40. We use large-scale shell-model calculations to interpret the results, and we report extrapolations for the doubly magic nucleus 60Ca