545298 research outputs found
Sort by
Micromobility Services
Get PDFThis project reviews and summarizes empirical evidence for a selection of transportation and land use policies, infrastructure investments, demand management programs, and pricing policies for reducing vehicle miles traveled (VMT) and greenhouse gas (GHG) emissions. The project explicitly considers social equity (fairness that accounts for differences in opportunity) and justice (equity of social systems) for the strategies and their outcomes. Each brief identifies the best available evidence in the peer-reviewed academic literature and has detailed discussions of study selection and methodological issues. VMT and GHG emissions reduction is shown by effect size, defined as the amount of change in VMT (or other measures of travel behavior) per unit of the strategy, e.g., a unit increase in density. Effect sizes can be used to predict the outcome of a proposed policy or strategy. They can be in absolute terms (e.g., VMT reduced), but are more commonly in relative terms (e.g., percent VMT reduced). Relative effect sizes are often reported as the percent change in the outcome divided by the percent change in the strategy, also called an elasticity
Functional Anatomy of Ocular Counter-Rolling in Superior Oblique Palsy
Get PDFPurposeSimulations suggest that displacement of rectus extraocular muscle pulleys in superior oblique (SO) palsy accounts for incomitant strabismus patterns even without postulating SO contractile weakness. We asked how rectus extraocular muscle pulleys reorient during head tilt in SO palsy.MethodsIn 13 subjects with unilateral SO palsy, supine magnetic resonance imaging (MRI) in 2-mm-thick quasi-coronal planes in target-controlled central gaze was repeated in both lateral decubitus positions equivalent to 90° head tilts. From extraocular muscle centroids, we computed oculocentric pulley coordinates and compartmental posterior partial volumes (PPVs) of the rectus and SO muscles.ResultsValidating atrophy, PPV of the palsied SO was smaller than its fellow (P < 10-4). In fellow orbits, the array of all four rectus pulleys exhibited counter-rotation during head tilt (P < 0.03), averaging 5.9°. The palsied pulley array was in both tilts excyclorotated relative to the fellow orbit, particularly by 4° to 5° for horizontal rectus pulleys (P < 0.03), and also counter-rotated with head tilt similarly to the fellow orbit. Differential compartmental changes in PPV were significant in the lateral and superior rectus and SO muscles of the fellow orbit that were consistent with observed torsion, but were absent in the palsied orbit.ConclusionsSimilar counter-rotation of the rectus pulley array during head tilt occurs in both eyes in unilateral SO palsy, but superimposed on excyclorotation of the array in the palsied orbit. Differential compartmental change in PPV occurs during head tilt in the lateral and superior rectus muscles of the fellow but not palsied orbit and could augment ocular counter-rolling
Remote‐Controlled Wireless Bioelectronics for Fluoxetine Therapy to Promote Wound Healing in a Porcine Model
Get PDFWound healing is a significant opportunity in biomedical science, requiring precise therapeutic delivery and real-time monitoring. Bioelectronic devices address this need and require further validation in large animal models that reflect human healing dynamics. This work introduces a wireless, remotely controlled bioelectronic platform with an iontophoretic pump to deliver fluoxetine, a selective serotonin reuptake inhibitor that modulates inflammation and promotes wound repair. Treatment over 3 days in a porcine model increased re-epithelialization by 37% and reduced inflammation, as shown by a 33% decrease in the ratio of pro-inflammatory M1 macrophages to pro-reparative M2 macrophages compared to controls. The platform accelerated early wound healing and stimulated neuronal growth at the wound site
Imaging Spectroscopy of Surface Soil Mineralogy and Vegetation Cover: Sensitivity and Validation
No full textImaging spectroscopy has enabled humanity to perceive regions of the electromagnetic spectrum that lie beyond the bounds of natural human vision, revealing information about the Earth’s surface once hidden from sight. With the launch of the Earth surface Mineral dust source InvesTigation (EMIT) mission in July of 2022 and NASA’s commitment to the Surface Biology and Geology (SBG) Designated Observable set to launch in 2032, spaceborne imaging spectroscopy is poised to make major breakthroughs in our understanding and modeling of the Earth System as we enter the dawn of a new era of Earth imaging spectroscopy. Drylands provide many societal, cultural, and environmental benefits; therefore, these factors make drylands especially ripe for the application of imaging spectroscopy. Additionally, the composition of the vegetation and soil in drylands is critical to the role these lands play in the Earth system. This research integrates spectral unmixing algorithms, simulations, and mineral detections algorithms across different scales to evaluate the capability of imaging spectroscopy to simultaneously and accurately quantify vegetation fractional cover and soil mineral composition, with a focus in Earth’s drylands
"Breaking the Cycle of Corporal Punishment: Global Evidence and Chilean Community-Level Insights"
Get PDFCorporal punishment (CP) is one of the most common yet under-addressed forms of violence against children worldwide. This three-part paper integrates global, theoretical, and local perspectives to understand what drives the persistence of CP and provides evidence to reduce it effectively.The first study presents a scoping review of 56 parenting programs from 19 countries aiming to reduce CP. Through content analysis and a Random Forest model, we identified seven core components that were consistently present in successful interventions. Emotional regulation for caregivers and cultural adaptation emerged as the strongest predictors of program success, emphasizing the need for interventions that go beyond behavior management to address the emotional and contextual realities of parenting.The second paper focuses on the Chilean context, using nationally representative data from the Encuesta Longitudinal de Primera Infancia (ELPI, 2017; n ≈ 17,000) to examine whether community-level crime is associated with the use of CP by caregivers. A LASSO-based propensity score model was used to adjust for confounding, followed by logistic and multilevel models. While no overall association was found between crime and CP in basic models, multilevel analyses revealed significant geographic variability. A random slope model showed that the effect of crime on CP use differed substantially across comunas, underscoring the importance of place-based dynamics, showing that in some areas, higher crime was associated with increased CP use. In contrast, in others, the relationship was null or even inverse, highlighting the importance of local context.The third paper evaluates the impact of Chile Crece Contigo (ChCC), a universal early childhood support program, on reducing CP. Using observational data from ELPI and an emulated trial design, we applied LASSO for variable selection and built propensity scores to estimate treatment effects. Results indicate that exposure to ChCC was associated with a statistically significant reduction in caregiver-reported use of CP. These findings suggest that universal parenting programs embedded in social protection systems can play a crucial role in violence prevention.Together, these studies underscore that ending corporal punishment requires a multi-level approach: culturally responsive parenting programs, context-sensitive public health analyses, and evidence-based social policy. By integrating global lessons with national insights, this research advances the movement to eliminate violence against children through informed, scalable, and systemic interventions.
El castigo corporal (CC) es una de las formas más comunes pero menos abordadas de violencia contra los niños a nivel mundial. Este documento, compuesto por tres partes, integra perspectivas globales, teóricas y locales para comprender qué impulsa la persistencia del CC y proporciona evidencia para reducirlo de manera efectiva.
El primer estudio presenta una revisión exploratoria de 56 programas de crianza provenientes de 19 países que buscan reducir el uso del CC. A través de un análisis de contenido y un modelo de Random Forest, se identificaron siete componentes clave presentes de forma consistente en las intervenciones exitosas. La regulación emocional de los cuidadores y la adaptación cultural surgieron como los predictores más sólidos del éxito de los programas, lo que enfatiza la necesidad de intervenciones que vayan más allá del manejo conductual y aborden las realidades emocionales y contextuales de la crianza.
El segundo artículo se centra en el contexto chileno, utilizando datos representativos a nivel nacional de la Encuesta Longitudinal de Primera Infancia (ELPI, 2017; n ≈ 17.000) para examinar si el nivel de criminalidad comunitaria se asocia con el uso de CC por parte de los cuidadores. Se utilizó un modelo de puntaje de propensión basado en LASSO para ajustar por factores de confusión, seguido de modelos logísticos y multinivel. Si bien en los modelos básicos no se encontró una asociación global entre crimen y CC, los análisis multinivel revelaron una variabilidad geográfica significativa. Un modelo con pendiente aleatoria mostró que el efecto del crimen sobre el uso de CC varía sustancialmente entre comunas, lo que resalta la importancia de las dinámicas locales. En algunas comunas, una mayor criminalidad se asoció con un aumento en el uso del CC, mientras que en otras, la relación fue nula o incluso inversa, evidenciando la relevancia del contexto local.
El tercer estudio evalúa el impacto de Chile Crece Contigo (ChCC), un programa universal de apoyo a la primera infancia, en la reducción del uso del CC. Utilizando datos observacionales de la ELPI y un diseño de ensayo emulado, se aplicó LASSO para la selección de variables y se construyeron puntajes de propensión para estimar los efectos del tratamiento. Los resultados indican que la exposición al ChCC se asoció con una reducción estadísticamente significativa en el uso del CC reportado por los cuidadores. Estos hallazgos sugieren que los programas de crianza universales, integrados en sistemas de protección social, pueden desempeñar un rol clave en la prevención de la violencia.
En conjunto, estos estudios subrayan que eliminar el castigo corporal requiere un enfoque multinivel: programas de crianza culturalmente pertinentes, análisis de salud pública sensibles al contexto, y políticas sociales basadas en evidencia. Al integrar lecciones globales con conocimientos nacionales, esta investigación aporta al movimiento para erradicar la violencia contra la niñez mediante intervenciones informadas, escalables y sistémicas
A Grammar of Nomlaki
No full textThis thesis is a grammar of Nomlaki, a Wintuan language of northern California preserved exclusively through archival documents. Many early documenters did not consider Nomlaki sufficiently differentiated from its sister language Wintu to justify separate investigation. As a consequence, Nomlaki is the only Wintuan language without a grammatical description. This grammar is a preliminary answer to this gap in the Californianist literature.This work is divided into eight chapters. The introduction includes an orientation of Nomlaki in its genetic context, a finding guide for its documentation, and a discussion of the limitations and methods involved in obtaining a grammatical description from archival records. The second chapter concerns Nomlaki phonetics, including studies on the vowel space, stress, voice onset time, locus equations for stops, and moment spectra for fricatives and affricates. The third chapter discusses Nomlaki phonology, including phonotactics, syllable structure, basic phonological processes, the behavior of loan words, lexical stress, and intonation. Chapters 4-7 discuss Nomlaki verbs, word classes, nouns, and syntax. Here I highlight several Nomlaki innovations, including possibly an ablative case marker which is not present in either sister language or reconstructed for Penutian, its unique typology of hortatives, the innovative use of the negative suffix -mena as a possibility marker, and semantic variations in the use of 'particular' and 'generic' noun marking. The final chapter presents a summary of conclusions and directions for future work.This thesis is the first work to provide a detailed description of Nomlaki. While by no means complete, this work greatly expands not only our descriptive knowledge of Nomlaki itself, but of Wintuan and California language typology. The results not only confirm the close relation between Wintu and Nomlaki, but indicate key areas in which they differ. This in turn highlights Nomlaki as a unique language area, meriting independent study and consideration. Most importantly, this work enables higher quality learning for ongoing Nomlaki revitalization work
A High-Efficiency Transmitter Architecture Based on Envelope Tracking Supply Modulation
No full textEnvelope tracking has been widely used to improve the average efficiency of power amplifiers (PAs). It tracks the envelope voltage and modulates the supply voltage of PAs. A linear-switching hybrid architecture has been utilized for analog envelope tracking; however, it faces intrinsic bandwidth limitations for emerging communications, such as 5G and 6G. This is because analog envelope tracking uses negative feedback to track the envelope voltage sophistically, but its loop compensation should be broadband while supporting large currents with high efficiency. The primary limiting factor is the second pole location, generated by the decoupling capacitor of the PA.To overcome this issue, a discrete-level supply modulator (SM) is suggested, which is required to generate a multi-level supply voltage with fast transitions and minimal voltage ringing. This thesis proposes an integrated SM-PA system to support these requirements, and a suggested adaptive biasing network can mitigate linearity degradation under multi-supply PA operations. The design is implemented using a CMOS 28nm process.Additionally, a hybrid capacitor-inductor supply modulator design is proposed and analyzed for the heterogeneous integration of the supply modulator and power amplifier in a board-level design. It supports soft-switching operations for a soft transition with small voltage ringing when a fast transition is required. The proposed architecture can also be operated as analog envelope tracking, so the architecture can support multi-band and multi-standard with average power tracking, analog envelope tracking, and discrete-level power tracking.To further enhance power efficiency under low-power conditions, a zero-current sensing mechanism is implemented in the inductive buck converter. To achieve immunity from PVT variations, an auto-calibration scheme is applied, ensuring the robustness of the output voltage variations under multi-supply operations.The final section of each chapter presents the measured and simulated results of the designs
Rollout of Hydrogen Refueling Stations in Support of Heavy-Duty Fuel Cell Electric Truck Adoption in California
Get PDFClimate change and degraded air quality have led to governmental and societal pressure for zero emission technology. Heavy-duty vehicles (HDV), fueled by diesel, are a major contributor to the emission of both greenhouse gases (GHGs) and criteria air pollutants (CAPs). As such, California has set aggressive decarbonization and clean transportation goals for HDV that require zero-emission vehicle (ZEV) adoption. Heavy-duty fuel cell electric vehicles (HDFCEV) offer a zero-emission solution for many HDV vocations due to their ability to support long range trips and quick refueling times. The main challenge inhibiting HDFCEV adoption is the lack of hydrogen refueling infrastructure. Effective planning of HDV hydrogen refueling stations (HRS) is important to enable HDFCEV adoption and minimize infrastructure costs.
This dissertation presents an HDV HRS siting model to locate HDV stations in California in support of the forecasted HDFCEV demand. The model locates HDV stations such that drive distance is minimized and refueling demand is maximized given capacity and service range restrictions on HDV HRS. The siting model is supported by (1) a HDV HRS capital cost model for economic evaluation, and (2) a light-duty vehicle (LDV) refueling demand forecast and allocation model for improved market understanding through HDV/LDV collocated hydrogen refueling. A scoring methodology is developed to evaluate the coverage and resilience of an HDV HRS network evaluation, and to compare results and the impacts of siting decisions across scenarios. Design of experiments (DOEx) is applied to guide the siting model input sensitivities by statistically evaluating the HDV HRS siting model and identifying the inputs (factors) that are most influential in affecting the score (response). Recommendations resulting from the study are structured to assure effective and resilient refueling coverage for both the nascent (2030) and developed (2045) HDFCEV markets concomitant with minimizing capital infrastructure investment costs
Applications of Bayesian Inverse Modeling and Deep Learning for Mapping Urban CO2 Emissions
No full textCarbon dioxide (CO2) is an important greenhouse gas emitted into the atmosphere during combustion of fossil fuels. As municipalities seek to reduce their CO2 emissions to meet climate targets, there is large interest in improved methods for quantifying anthropogenic CO2 emissions. Given the role of cities, as both emitters of large amounts of CO2 and policy leaders, this work focuses on improved methods for developing measurement-informed CO2 emission estimates at sub-city scales.Improving measurement, reporting, and verification (MRV) of urban CO2 emissions is a major goal for deployments of networks of CO2 measuring sensors. However, transforming measured CO2 concentrations (mixing ratios) into estimates of anthropogenic CO2 emissions (fluxes), is non-trivial. Without sufficient temporal and spatial resolution of observations, the problem of mapping urban emissions is under-constrained. Multiple emission scenarios can result in the same observed concentrations. As such, much previous work treats urban emission estimation as a Bayesian optimization problem: coupling observations with atmospheric transport models to update existing emission inventories.In this work, we apply two methodologies, an established Bayesian inverse modeling framework and a novel machine learning approach, for mapping urban CO2 emissions based on sensor observations. First, we use CO2 measurements from the Berkeley Environmental Air Quality and CO2 Network (BEACO2N) to constrain interannual trends in CO2 emissions from 2018-2022 for the San Francisco Bay Area. An established Bayesian method is used and a decreasing emission trend at a rate of 1.8 ± 0.3%/year is found for the region. Using multiple linear regression, we attribute emissions to three emission sectors: traffic, seasonal, and constant emissions. The decrease in emissions is interpreted as primarily due to passenger vehicle electrification, reducing on-road emissions per vehicle mile traveled at a rate of 2.6 ± 0.7%/year.Having established the sensitivity of the Bayesian inversion system to interannual emission trends and vehicle fleet emission factors, we next test the sensitivity to home heating emissions. It is well-established that home heating behavior (and in particular the outdoor temperature at which natural gas home heating systems are turned on) has large regional variation. By withholding information about emission seasonality from the prior emission inventory, we show that the inversion correctly identifies the critical temperature of home heating and the seasonal patterns of emissions in three US cities.Finally, we develop a general machine learning benchmark for mapping urban emissions using sparse surface measurements. We create a first-of-its-kind 1 km simulation of CO2 concentrations over the continental US, develop a train-validation-test split, and establish baseline model performance. The resulting datasets and baseline models are released as a machine learning benchmark to encourage future innovation on this problem
ADHD, Personality, and Key Academic Outcomes: A Prospective Longitudinal Investigation in Girls
No full textThis dissertation investigated whether adolescent personality traits mediate the long-term association between childhood attention-deficit/hyperactivity disorder (ADHD) and academic outcomes in young adulthood. Drawing on the Berkeley Girls with ADHD Longitudinal Study (BGALS)—a racially, ethnically, and socioeconomically diverse, all-female sample both with and without ADHD, followed prospectively from childhood through early adulthood—I tested a developmental mediation model—divided into paths a, b, and c. I pre-registered the hypotheses that childhood ADHD and its symptom dimensions would predict adolescent Conscientiousness (path a), which in turn would predict (path b) academic achievement (standardized test scores), attainment (years of education), and performance (GPA), thereby mediating ADHD-related academic risk.
Path a of the mediation model was established in prior work (see Bell et al., 2024), showing that childhood ADHD predicted lower Conscientiousness and Agreeableness and higher Neuroticism in adolescence. Such effects in that study and the present dissertation emerged even when identical or nearly identical items pertaining to both ADHD and personality were removed. In the present study, path b analyses revealed that, as expected, adolescent Conscientiousness robustly predicted all three domains of academic outcomes, although predictions only extended to math but not reading test scores. Conscientiousness remained a robust predictor of academic outcomes even after adjusting for IQ. The other two potential mediators, Agreeableness and Neuroticism, showed limited predictive power after adjusting for other Big Five traits, so they were not included in subsequent mediation models. Path c analyses demonstrated that childhood ADHD—particularly the overall diagnosis and the inattention dimension—significantly predicted poorer academic performance, lower standardized achievement, and reduced educational attainment in young adulthood. Multiple mediation models indicated that adolescent Conscientiousness partially mediated these links, providing evidence that personality development helps explain links between ADHD and negative long-term academic outcomes in girls. Robustness analyses suggested that family income moderated predictions from ADHD symptoms to attainment via Conscientiousness, with stronger negative predictions observed in higher-income families. These findings illuminate developmental pathways through which early ADHD contributes to later academic difficulties via adolescent personality traits, particularly Conscientiousness