19304 research outputs found
Sort by
Annotation-free Large-scale Overlapping Nuclear Segmentation on Multiplexed Fluorescence Images Using Foundation Model and Weak to Strong Learning
No full textWe present a weak to strong generalization methodology for fully automated training of a multi-head extension of the Mask-RCNN method with efficient channel attention for reliable segmentation of overlapping cell nuclei in multiplex cyclic immunofluorescent (IHC) whole-slide images (WSI), and present evidence for pseudo-label correction and coverage expansion, the key phenomena underlying weak to strong generalization. This method can learn to segment de novo a new class of images resulting from a new instrument and/or a new protocol without the need for human annotations. We present metrics for automated self-diagnosis of segmentation quality in production environments, where human visual proofreading of massive WSI images is unaffordable. The proposed method was benchmarked against five current widely used methods and showed a significant improvement. We extend our study by adapting and applying the framework to two additional datasets to evaluate its generalizability across different imaging domains. The code, sample WSI images, and high-resolution segmentation results are provided in open form for community adoption and adaptation to other image analysis tasks
Examining Leadership Efforts to Disrupt the Disproportional Discipline of Black Students in Special Education in a Suburban-Urban District
No full textBackground: Nationally, Black students with disabilities are nearly three times as likely to receive out-of-school suspensions as their White peers with and without disabilities. In 2022, the Results-Driven Accountability (RDA) report revealed a troubling disparity within Seaway Independent School District (SISD, a pseudonym): Black students in special education were three times more likely than students of other ethnicities to face out-of-school suspension and expulsion. This significant disparity highlights the urgent need for targeted interventions to address and mitigate inequities in the district’s disciplinary practices. This study was conducted in SISD to explore how educators understand and respond to these patterns of inequity. Purpose: This study aimed to examine the perceptions, strategies, and systemic responses related to the discipline of Black students in special education. Framed by Disability Critical Race Theory (DisCrit), the study investigated how racism and ableism shape educator beliefs, disciplinary practices, and institutional decision-making. Methods: Through a qualitative case study approach and practitioner inquiry, I collected data through interviews with district leaders, campus administrators, and special and general education teachers. Additional data sources included district strategic plans and related documents. The thematic analysis involved inductive and deductive coding, using a priori codes aligned with DisCrit to surface patterns in perceptions and practices. Findings: Participants acknowledged race as a factor in disciplinary outcomes for Black students in special education. However, few identified formal, district-led initiatives aimed at addressing these disparities. Most participants described their efforts as informal and self-initiated rather than part of a coordinated, equity-focused system. Educators reported limited access to professional development, inconsistent use of data, and minimal integration of culturally responsive frameworks like PBIS or MTSS with fidelity. Conclusion: Findings reveal a disconnect between equity goals and actual implementation at the district and campus levels. To address discipline disproportionality, systems must embed culturally responsive practices, provide ongoing professional learning, and establish accountability structures prioritizing marginalized students' lived experiences. This study discusses implications for practice, policy, leadership, and future research
Simulation Study of Magnetic Field Effects in the Pair Spectrometer Luminosity Detector for ePIC Experiment at the Electron Ion Collider
No full textWe aim to address the challenge of optimizing the magnetic field configuration that guides pair-produced e+e− from the conversion foil to the pair spectrometer trackers and calorimeters. This involves systematically exploring a range of tunable parameters that influence the acceptance rate, the efficiency with which both members of the pair are detected. Key parameters include the distance from the dipole magnet, the dimensions of the pair spectrometer, the magnetic field strength within the dipole, and the width of the photon beam. The study is conducted within the DD4hep framework, which enables detailed simulation of the detector geometry and associated components. This facilitates analysis of particle trajectories and tracker acceptance in a realistic experimental setup. A significant challenge arises from the wide energy spectrum of Bethe-Heitler photons (0–18 GeV), which affects the pair acceptance for any given tracker configuration at the Electron-Ion Collider (EIC). The ultimate goal is to develop a flexible suite of scripts and algorithms capable of fine-tuning the tracker acceptance based on adjustable design parameters. This approach allows adaptation to evolving experimental requirements and maximizes detection efficiency across the full photon energy spectrum. As a result, it contributes to reducing systematic uncertainties in luminosity measurements across different beam energies at the EIC
Solubility of HIV1 Intergrase and Interaction With LEDGEF
No full textLentiviral vectors derived from HIV-1 show promise for gene therapy, but are limited by integrase solubility issues and potential oncogenic risks from non-specific genomic integration. While the F185K mutation was proposed to improve integrase solubility, it compromises viral assembly and infection capability. This study investigates how this mutation affects the interaction between HIV-1 integrase and its primary cellular binding partner, Lens Epithelium-Derived Growth Factor (LEDGF), which is essential for chromatin association. Using GST and MBP fusion protein pull-down assays in E. coli, we examine the integrase-LEDGF interaction. Sequence analysis revealed unexpected additional mutations in the integrase gene. Understanding these molecular interactions is crucial for developing more precisely targeted lentiviral vectors for safer gene therapy applications.Biology and Biochemistry, Department ofHonors Colleg
Transforming Health Outcomes for Special Olympic Athletes: A Family-Centered Approach
No full textUnhealthy lifestyle behaviors take root early in life, resulting in high obesity rates among youth. In response to this epidemic, numerous intervention programs have been developed to enhance nutrition and physical activity (PA) behaviors. However, these have been tailored for children undergoing typical development and often exclude children with intellectual disabilities. This leads to a gap in the literature when researching the effects of these interventions in this population. To address this critical gap, we adapted an effective school-based intervention program with Special Olympics (SO) that addresses nutrition and PA, aiming to promote healthy lifestyle behaviors in children with intellectual disabilities. First, we took our intervention program to SO and experts in the field (n=10) to identify which components are or are not consistent with the needs of children with intellectual disabilities. After implementing changes based on expert review, we sought out feedback from families who have children in SO (n=12).With this feedback, we finalized an 8-week lifestyle intervention program for children with intellectual disabilities. The main modification implemented included the following: (1) shifting the focus from weight to lifestyle behaviors, (2) targeting the entire family rather than the child alone, (3) supporting positive parent-child interactions regarding nutrition and PA, and (4) involving an expert outside of the family's system to support lifestyle changes to help discuss sensitive issues. Parents also asked for specific skills in navigating mealtime behaviors. We have developed a family-based lifestyle intervention to address nutrition and PA behaviors in children with intellectual disabilities.Health and Human Performance, Department ofHonors Colleg
Effects of Matched and Mismatched Visual Flow and Gait Speeds on Human Electrocortical Spectral Power
No full textBackground/Objectives: Visuomotor integration relies on synchronized proprioceptive and visual feedback during visually guided locomotion. How the human brain processes unimodal or asynchronous multimodal inputs during locomotion is unclear. Methods: Using high-density mobile electroencephalography (EEG) and motion capture in a virtual reality environment, we investigated electrocortical responses during altered treadmill gait speeds (0.5 and 1.5 m/s) and visual flow speeds (0.5×, 1×, and 1.5× gait speed) among 13 healthy human subjects. Experimental conditions included passive viewing of a moving virtual environment, walking in a stationary virtual environment, and walking in a moving environment with synchronous and asynchronous visual flow. Results: At faster gait speed, we identified reduced premotor, sensorimotor, and visual electrocortical beta-band spectral power (13–30 Hz) and greater premotor cortex theta power (4–8 Hz). At faster visual flow speeds, we identified reduced sensorimotor electrocortical beta-band spectral power, reduced alpha (8–13 Hz) and beta power, and greater gamma-band power (30–50 Hz) from the visual cortex. During visual flow and gait speed mismatches, sensorimotor and parietal alpha- and beta-band electrocortical spectral power decreased at faster gait speed. During treadmill walking at 1.5 m/s, parietal electrocortical spectral power increased when visual flow exceeded gait speed. Conclusions: Electrical brain dynamics during human gait identified distinct neural circuits for integrating kinesthetic and visual information during visuomotor conflicts, gated by the parietal cortex
Impact of Affirmative Action Policy on Public University Campus Diversity
No full textThis research investigates the effectiveness of affirmative action in promoting diverse student bodies. The 2023 SSFA v. Harvard ruling, which banned race-based affirmative action in college admissions, sparked significant controversy among prospective students. Considering this, the study aims to explore whether affirmative action has played a crucial role in fostering campus diversity over the past three decades.Political Science, Department ofHonors Colleg
Optimal Approximation of High-dimensional Functions on Smooth Manifolds Using Deep ReLU Neural Networks
No full textThe expressive power of deep neural networks is manifested by their remarkable ability to approximate multivariate functions in a way that appears to overcome the curse of dimensionality. This ability is exemplified by their success in solving high-dimensional problems where traditional numerical solvers fail due to their limitations in accurately representing high-dimensional structures. In this dissertation, we focus on the approximation theory of neural networks to explain this phenomenon. First, we construct a version of the Johnson-Lindenstrass Lemma on the smooth manifold to analyze the approximation of -smooth H\"older functions defined on a -dimensional smooth manifold embedded in , with , using deep neural networks. We prove that the uniform convergence estimates of the approximation and generalization errors by deep neural networks with ReLU activation functions do not depend on the ambient dimension of the function but only on its lower manifold dimension , in a precise sense. Our result improves existing results from the literature where approximation and generalization errors were shown to depend weakly on . Since the aforementioned method is not constructive and does not provide explicit knowledge on the structure of the deep neural networks involved, we next introduce a constructive approach to achieve a similar approximation error. For this analysis, we consider the approximation of times continuously differentiable functions defined on a -dimensional smooth manifold embedded in , with . We prove that a deep ReLU neural network with fixed width depending polynomially on and depth can approximate an times continuously differentiable function with approximation rate which is completely independent of the ambient dimension . This result improves previous approximation estimates where the constant of the approximation rate depends on . This dissertation also derives the theoretical optimal approximation rate of deep ReLU neural networks for -smooth Hölder functions with and for times continuously differentiable functions with defined on
Polyethylene-based Polar Block Copolymers Preparation using Organometallic-Mediated Radical Polymerization
No full textThe synthesis of polyethylene-polar block copolymers presents significant challenges due to the different reactivity of the monomers involved. The difference in reactivity required several polymerization techniques. The Co(salen) offers reactivities in mediating the polymerization of various monomers. This study addressed these challenges by employing polymer-Co(salen) complexes and conducting the synthesis of several polyethylene-b-polar polymers using only the radical approach. The research was divided into six chapters. Chapter 1 underscored the importance of polyethylene-polar block copolymers and introduced controlled radical polymerization techniques, including organocobalt-mediated radical polymerization. Chapter 2 explored the preparation of a polymer-Co(salen) complex with chain-end fidelity. Kinetic studies of Co(salen) mediated radical polymerization and propagation resumption reactions concluded that the polymer-Co(salen) complexes with chain-end fidelity were produced for methyl acrylate, vinyl acetate, and N, N- dimethyl acrylamides. Chapter 3 examined the reactivity of polymer-Co(salen) complexes, polar monomers, and ethylene monomers. TEMPO trapping experiments implied that the polymer-Co(salen) dormant species could initiate a radical species. Electron donors were found to coordinate with Co(salen), accelerating bond homolysis. Random copolymerization experiments suggested that radical transfer between polar monomer and ethylene was possible in the presence of Co(salen). Chapters 4, 5, and 6 reported the synthesis of PMA-b-PE, PVAc-b-PE, and PDMA-b-PE copolymers using PMA-Co(salen), PVAc-Co(salen), and PDMA-Co(salen) complexes, respectively. It was found that the degree of polymerization of ethylene increased with higher ethylene pressure, longer reaction duration, and the introduction of an electron donor, which aligned with the kinetics study. In summary, this research demonstrated the potential of preparing polyethylene-polar block copolymers using polymer-Co(salen) dormant species prepared through Co(salen) mediated radical polymerization. This method facilitated the preparation of PMA-b-PE, PVAc-b-PE, and PDMA-b-PE using one type of organometallic complexes, opening the door for an expanded scope of monomers by using Co(salen) mediated radical polymerization
La Mujer En Acción: Latina Contributions and Connections in Early 20th Century Press
No full textDuring the early 20th century, it was uncommon for Latina women to have access to avenues for publishing their writings. In response, Latinas collaborated with each other, shared opportunities and fellow connections, and made spaces of their own to write. Still, much of their work has been ignored or erased. This project aims to highlight the recovery of these voices through the Recovering Hispanic Literary Heritage Program. This poster presents "La Mujer en Accion," a newspaper column. The column, directed by Concha Romero James, ran in "La Nueva Democracia" (ND). ND was a Spanish Language periodical printed in New York City from 1920-1948. The column ran 1921-1924. Dr. Romero James' work on "La Mujer en Acción" and her overall advocacy is demonstrative of a culture of Latina communication, collaboration, and connection among Latinas during this time. Gathering and analyzing her writings and scattered mentions of her life presents the opportunity to illustrate the connections and networks among Latina writers that were active participants of the intellectual and feminist circles during this time. Additionally, the wealth of Latina writers found in ND shows the possibilities for future research recovering Latina writings and contributions.Honors CollegeSociology, Department o