203435 research outputs found
Sort by
Artificial Intelligence in Healthcare and Science
Full text linkThis section explores AI\u27s impact on science and healthcare, including medical diagnostics and treatments, and scientific research
Autonomous Vehicle Supported Mobility Services for Rural Areas
Full text linkAutonomous Vehicles (AVs) are emerging as a promising mobility solution to address growing transportation demands, particularly unmet traffic demand in rural areas where conventional transit systems often fall short. Rural areas face unique mobility challenges due to dispersed populations, limited infrastructure, and low/no transit coverage. Existing transit services tend to concentrate around dense urban cores, leaving low-demand areas outside underserved. This spatial undercoverage leads to significant mobility gaps for rural residents. This study examines the potential of AV-based ride-sharing services to address these mobility needs by providing flexible, demand-responsive mobility services where fixed-route transit services are either economically or operationally unfeasible. To evaluate this potential, we modeled a multimodal transportation system that integrates AV-based mobility service with the existing fixed-route transit network in a rural setting. Using a traffic simulation tool and a multi-objective optimization framework, we evaluated the performance of AV-based ride-sharing services across various travel demand scenarios, AV fleet sizes, and service-priority schemes, with current transit services. The AV-based ride-sharing service optimization aimed to minimize both travel time and passenger trip request rejection. A passenger prioritization mechanism was introduced, which ranked ride requests based on the current zonal transit accessibility to ensure that AV-based mobility services were directed toward passengers with the fewest or no other mobility alternatives. Results demonstrated that 100% of the trips in transit-unserved areas (known as the Transit Desert) in Monongalia County, West Virginia, can be served during peak hours with a fleet of 583 AVs. Even with significantly smaller AV fleets (10 or 20 vehicles), 0.27%–0.54% of total peak hour demand could be served. These findings highlight the value of integrating AV-based mobility services with existing transit systems to extend service coverage and improve accessibility, particularly in areas currently unserved by transit
Characterization of a descending neuron that modulates flight performance and provides input onto a flight-related predictive motor circuit
Full text linkDescending neurons (DNs) play a critical role in sensorimotor integration by relaying motor information to the ventral nerve cord (VNC). Within the VNC, local interneurons, other DNs, and ascending neurons (ANs) make circuits with sensory and motor neurons to modify flies’ behavior. One important AN pair in Drosophila Melanogaster is mesothoracic ascending histaminergic neurons (MsAHNs), which are known to convey predictive wing motor signals to the brain. DNg32 is the primary upstream synaptic partner of MsAHNs, integrating DNg32 input with other inputs to generate predictive motor signals. The goals of this study were to 1) map the synaptic connectivity of DNg32, 2) determine its neurotransmitter identity, and 3) assess its functional role in flight behavior. Using electron microscopy volume reconstruction of the brain and VNC, we reconstructed DNg32’s main pre- and postsynaptic partners. We found that DNg32 receives bilateral input from auditory-related commissural neurons in the brain, suggesting they may contribute to bilateral fusion of mechanosensory information. DNg32 also projects contralaterally and unilaterally to the wing and haltere tectulum within the VNC, positioning it to provide flight steering signals to flight control circuits. Using molecular genetic markers with immunohistochemistry, we confirmed that DNg32 is cholinergic. To examine DNg32’s role in behavioral performance. The FlyTrap assay quantifies odor-guided flight performance by measuring flies’ capture in an attractive odor trap in 10-minute intervals over 60 minutes. We genetically manipulated DNg32 function using targeted ablation (UAS-rpr) or synaptic silencing (UAS-BoNT-C). Here, we found that ablation or silencing of DNg32 results in a significant increase in the trapping rate, suggesting an increased rate of flights of equal success and/or increased odor-guided performance. To resolve these possibilities, we used the FlyFall assay, which captures high-speed (600 fps) videos of voluntary takeoffs. We again disrupted DNg32 activity and quantified the number of flights per recording and flight quality. Both DNg32 ablation and silencing led to increased takeoff frequency; however, they also resulted in flight instability as indicated by increased rate of crashing. As a final experiment, we used an optogenetic tool (UAS-CsChrimson) to photoactivate DNg32 in the FlyFall assay. We find that DNg32 activation significantly suppressed the rate of takeoffs, but flight quality was not significantly affected. Together, these results suggest that DNg32 plays a role in flight behavior and may provide steering information for flies to modify their flight path
Dynamically Scaling Biomimetic Robots Through Parallel Viscoelastic Actuators
Full text linkRobots provide a unique proving ground for testing scientists’ understanding of neuromuscular systems. Synergistically, mimicking biological control laws and mechanics could offer robotics the robust and adaptive locomotion of animals. Previous work has utilized the wealth of research on insect biomechanics to develop scaled robotic models but doing so presented a problem: while these robots’ larger scale has simplified manufacturing, the scale has also shifted its mechanics from those of the modeled organisms. While animals walk, forces are generated that describe their mechanics and influence motor control. The musculoskeletal structure and stretch of muscles result in viscous and elastic forces, while the raising and motion of limbs result in gravitational and inertial forces. These forces are dependent on the scale and speed of an animal. Larger, faster animals experience greater inertial forces due to their mass. At smaller scales viscoelastic forces dominate. How these forces balance determines the phase between active muscle force and displacement of limbs which has descending repercussions on control. This phase represents an animal’s dynamic scale. To make more accurate biomechanical models and grant legged robots’ animal- like mechanics, I developed a Parallel Viscoelastic Actuator (PVA) using 3D printed torsional springs. The PVA successfully reduced the phase between motor actuation and limb displacement. Through the introduction of a PVA, a robotic limb resembling an inertially dominated animal was able to operate at the same speed with the dynamic scale of an insect. Furthermore, this introduction im- proved responses to perturbations and enabled faster motions to be produced using less active force. PVAs will enable robotic models of insects to be built on a scale of convenience while maintaining the dynamic scale of their biomechanical model. A robot fully equipped with PVAs may offer unprecedented biomechanical accuracy and the ability to integrate biological control strategies, furthering understanding in biol- ogy and performance in robotics
Implementation of an Evidenced Based Enteral Nutrition Protocol for Mechanically Ventilated Patients: A Quality Improvement Initiative
Full text linkAbstract
Implementation of an Evidenced Based Enteral Nutrition Protocol for Mechanically Ventilated Patients: A Quality Improvement Initiative.
Hannah Nuzum
Background: Mechanically ventilated patients require supplemental nutrition as an important factor for their overall recovery. Malnutrition in hospitalized patients is reported as high as 35% (Ikram et al., 2022).
Purpose: Lack of an enteral nutrition (EN) protocol has contributed to variability in initiation of EN for post-intubated mechanically ventilated patients.
Intervention: The project interventions are threefold: the development of an EN protocol, EN initiated within 24 to 28 hours for post-intubation patients, and staff understands that EN should be initiated within 24 to 48 hours after post-intubation for better patient outcomes.
Methods: This change project was developed utilizing the Lewin’s Theory of Change theoretical framework to implement an EN protocol for adult intensive care unit patients. Collaboration with the information technology team will be done to integrate clinical notifications within the electronic health record, and education will be provided to staff to appropriately implement the enteral nutrition protocol into patient care. Staff will also take post intervention surveys to look at potential barriers and feasibility of the intervention.
Results: EN protocol was developed and integrated within the medical intensive care unit at Ruby Memorial Hospital. The median time for initiation of EN for pre-intervention patients was 79 hours, and the median time for initiation of EN for post-intervention patients was 17 hours.
Conclusions: Development of an electronic EN protocol aids in making initiation of EN for mechanically ventilated patients more consistent and improves the time to initiation of EN for critically ill patients
Advancements in technology integration, screening methodologies, and modeling of fundamental behaviors of gunshot residues
No full textUnderstanding the collection, preservation, and analysis of gunshot residue (GSR) is crucial for making informed decisions when interpreting evidence. As gun violence continues to persist across the United States, gunshot residue evidence plays a key role in providing investigative leads and reconstruction of events in criminal investigations. Current GSR standards have a strong scientific foundation in elemental detection and classification of gunshot residues from the primer (pGSR) of ammunition. However, key information is also held in the organic (OGSR) components that are released during the discharge of a firearm. Furthermore, the GSR discipline can benefit from knowledge of collection, preservation, transfer, and persistence to provide best practices for comprehensive GSR workflows and interpretation. This research aims to fill these gaps through 1) increasing knowledge of pGSR and OGSR preservation, transfer, and persistence, 2) developing reliable screening methods for GSR detection, and 3) developing interpretation tools for forensic laboratories.
First, three studies were conducted to understand and model GSR behavior for collection, preservation, transfer, and persistence. The first study assesses the persistence of pGSR and OGSR over time using synthetic skin and GSR standards with known concentrations and particle counts to model GSR behavior. The synthetic skin model and GSR standards have been previously validated by our research group and tested in a similar study. One hundred and eighty samples were collected at various post-deposition times (i.e., 0, 2, 4, 8, 10, and 12 hours), and both pGSR and OGSR were detected up to 12 hours post-deposition.
The collection and preservation of pGSR and OGSR are crucial when addressing laboratory backlogs and understanding OGSR evidence, as immediate analysis of GSR samples is not realistic. Storage time, storage condition, and analytical workflow were assessed to determine the stability and preservation of pGSR and OGSR evidence. The findings demonstrate that the detection of pGSR and OGSR was unaffected by time for up to six months and workflow, and also recommend freezer storage for the preservation of OGSR.
The transfer and persistence of GSR in arrest situations complicate the interpretation of results for forensic analysts, since police often participate in arrests and are exposed to GSR. Therefore, this study aimed to understand the behaviors of pGSR and OGSR during arrests and evaluate possible solutions to reduce GSR exposure risks for arrestees. Screening and confirmatory methods were used to gain a thorough understanding of GSR in low-, medium-, and high-contact arrests, where transfer may occur either after firing or during handling between the officer and the arrestee GSR loss was more likely than transfer as arrest activities increased; additionally, wearing nitrile gloves during arrests helped decrease secondary GSR transfer.
Second, an investigation into using a Raman spectroelectrochemical method (Raman-SEC) as an emerging screening technique for OGSR analysis. A method was developed and validated for the analysis of diphenylamine (DPA), a common stabilizer, in smokeless powder. The application of the method was proven to be selective for DPA, as characterized by the unburnt and partially burnt smokeless powder, corroborated by LC-MS/MS and electrochemical techniques.
Finally, Bayesian networks (BN) were developed as an interpretation tool for activity level propositions using in-house data as a proof of concept. Additionally, a survey was conducted to assess how everyday activities (e.g., washing your hands) affect the transfer and persistence of GSR. The BNs were evaluated by analyzing the probabilities for each hypothesis and the likelihood ratios of mock case scenarios.
Overall, the development of rapid screening tools, the application of statistical methods using Bayesian Networks, and growing knowledge of GSR collection, preservation, transfer, persistence, and analysis are expected to revolutionize current methods for analyzing and interpreting gunshot residues. Additionally, it offers forensic laboratories more tools to make informed decisions about implementing screening methods and interpretation tools
Using Probabilistic Risk Assessment in Drilling R&D Projects
No full textWell drilling, since its inception, has always carried an intrinsic quantity of risk. Economically, the evolution of data analytics and the tools available to operators provide the opportunity to transform the lessons learned from previous well drilling endeavors into bottom line improvements. Additionally, methods for forecasting and risk management of potential trouble areas have significantly advanced due to the abundance of empirical data from prior drilling operations. Probabilistic Risk Assessment is a viable method through which risk can be tracked and mitigated in project management. Coupled with this data, the usage of Probabilistic Risk Assessment (PRA) provides a quantified, trackable approach to risk management and mitigation, with respect to safety and economic risk. As projects are budgeted, it is difficult to price in risk mitigation to outside vendors. By integrating PRA into the process, it will provide an internal tracking mechanism that can be tracked and refined as more data is provided in the project process. Faults and the probability of their occurrence can be determined by subject matter experts within the project team. Fault trees analyze the causes of failure. These faults are then coupled together to create event trees, which analyze the respective consequences of failure. Event trees occur in sequence, leading to a variety of final end states, each having a unique set of success and failures. Combining fault tree analysis with event tree sequencing provides a comprehensive look at possible end states, and thus, the number of different economic or safety outcomes of a project. With each successive project, previous data can be integrated to refine fault predictions and identify new sequences of failure or lack thereof. Using PRA can aid investors without deep technical knowledge to leverage risk assessment to forecast economics of a drilling project accurately and identify safety hazards
Increasing Preparedness for Malignant Hyperthermia Crises: An Evidence Based Quality Improvement Project
No full textIntroduction/Background: MH is a low frequency, but high risk potentially fatal genetic mutation if left untreated. A rural Pennsylvania Hospital was found to have an incidence rate of 12/4,033 MH related occurrences within the last 6 months. Delayed treatment of MH can result in increased mortality rates and poor patient outcomes.
Purpose: Given the rarity and high mortality rate without prompt treatment, implementation of a standardized cognitive aid to be kept on the Malignant Hyperthermia (MH) cart, coupled with operating room staff education on early recognition and treatment of malignant hyperthermia can assist in proper perioperative MH management steps for staff at the rural Pennsylvania hospital.
Intervention: Provide educational training in conjunction with implementation of an MH cognitive aid to increase staff knowledge and cognitive aid utilization in management of MH. Gathered provider baseline knowledge via a pre-educational survey, follow up survey with formulated educational PowerPoint and MH cognitive aid presentation/introduction, and distribute a post knowledge/cognitive aid utilization survey.
Methods: Lewin’s Theory of Change guided implementation.
Results: Analysis of pre and post survey results were compared respectively using a dependent t test to gauge impact of the intervention on increased knowledge/ cognitive aid utilization.
Conclusions: The project highlighted the effectiveness of targeted training and cognitive aids in improving provider knowledge, confidence, and preparedness for managing malignant hyperthermia crises, ultimately enhancing patient safety
Code Blue Narrator Documentation: An Evidence Based Quality Improvement Initiative
Full text linkBackground: The code documenter plays a crucial role in ensuring accurate, real-time documentation during Code Blue events, which is critical for patient care and compliance with established guidelines and protocols. However, nurses have reported discomfort with the Code Narrator tool in EPIC, which has hindered effective documentation during these high-pressure situations.
Purpose: The lack of comfort in using the Code Narrator tool has led to inconsistent and inaccurate documentation during Code Blue events, potentially impacting the effectiveness of the response and patient outcomes, such as the achievement of return of spontaneous circulation (ROSC).
Intervention: An educational competency program was implemented, incorporating high-fidelity simulation training to improve nurse comfort with the Code Narrator tool. Two pilot units received training and participated in mock Code Blue events designed to improve comfort and documentation accuracy. The intervention aimed to enhance nurse comfort, improve documentation quality, and increase ROSC rates post-Code Blue.
Methods: A retrospective chart review of Code Blue events assessed baseline documentation practices, while a staff questionnaire evaluated nurse comfort with the Code Narrator tool and identified areas for further education and support.
Results: Post-intervention analysis demonstrated significant improvements in nurse comfort with both the code documenter role and with real-time documentation. Comfort levels increased by 119.5% for nurse comfort within the documenter role, surpassing the projects 50% target. Comfort with real-time documentation also improved by 45.3%, exceeding the 30% goal. However, the targeted 15% increase in the proportion of patients achieving ROSC for at least 20 minutes post-code was not met, with the post-implementation ROSC rate remaining at 75%. Despite this, improvements in medication administration adherence were observed, through pulse check delays persisted.
Conclusion: This quality improvement initiative improved nursing comfort in the code blue documenter role, enhancing documentation accuracy, team communication, and patient outcomes. The initiative emphasized the importance of ongoing training and support for sustained impact across healthcare settings
A Data-Driven Framework for Energy and Emissions Quantification in Electronics Manufacturing: A Case Study of a U.S.-Based PCB Facility
No full textGlobal mandates such as the Paris Agreement to limit global warming to 1.5°C have placed industries under significant pressure to reduce greenhouse gas emissions. Given the high energy intensity of electronics manufacturing, where semiconductor facilities alone consume over 1 billion kWh annually, and projections showing electronics-related emissions could double by 2030, with the industry using 14% of global electricity by 2050, carbon management is now more critical than ever.
This study focuses on a U.S.-based printed circuit board (PCB) facility as a representative case of the electronics manufacturing sector. Equipment-level energy consumption was measured for 21 major energy-consuming machines, selected based on predefined research criteria, to assess load patterns and operational behavior. The first phase of results presents key findings from a week-long real-time monitoring campaign, followed by data extraction, feature engineering, and visualization using Python and Energy Star Portfolio Manager. Power usage correlation analyses, intra- and inter-department comparisons, and unsupervised K-means clustering are used to classify machines into three operational load bands.
In the second part of the study, monitored current values are translated into power consumption and emissions calculations. Scope 1, Scope 2, and selected Scope 3 emission sources are identified, and the U.S. EPA emissions calculator is employed to determine total carbon output. The study further quantifies emissions intensity in terms of kgCO₂e per PCB, offering a replicable and data-driven framework to support sustainability efforts in electronics manufacturing. This study locates critical emission hotspots, offers an actionable and scalable methodology to bridge the gap between theoretical carbon estimation and real-world industrial uses