22484 research outputs found
Sort by
Facilitating collaboration to advance technological agency in underrepresented people
No full textThe LRDC team is working on projects which support underrepresented communities in the Pittsburgh area while gathering research. One of our projects has welcomed youth from the Homewood community to Pitt’s campus to explore and learn more about data as a means of empowering themselves and their community. Another project is continuing at a senior center in the Hill District in which we practice using and discussing artificial intelligence with voice assisted technology. We also have multiple projects working with black girls in the north side’s Manchester community as well as supporting their educators. We aim to guide these participants as they navigate bias in artificial intelligence (AI) and learn to engage with AI in ways that inspire their self-agency and joy through self-expression
Strategies for Enhancing the Biocompatibility and Chronic Recording Performance of Neural Interface Devices
No full textInvasive neural microelectrode arrays (MEAs) enable high-resolution bidirectional
communication with nervous tissue and can record neural activity with high spatial selectivity and
resolution. However, their long-term efficacy is hindered by inflammatory responses to
implantation. This dissertation investigates strategies to enhance the biocompatibility and chronic
recording performance of implantable MEAs.
In the first part of this dissertation, we explored the novel application of a chondroitin
sulfate bioactive coating for MEAs. Both in vitro and in vivo studies demonstrated the coating’s
stability, anti-fouling properties, and ability to promote neurite growth and reduce microglial
activation. Single-unit neural recordings and endpoint histology revealed improved recording
performance and reduced inflammation in animals implanted with coated MEAs compared to the
control, showcasing the effectiveness of the coating at acute time points.
In the second part, we examined the impact of MEA substrate rigidity on chronic neural
recording quality and inflammatory tissue response. We compared the performance of stiff silicon-
based and flexible polyimide-based MEAs with similar geometries in the mouse striatum. Flexible
MEAs maintained more stable impedance, noise, peak-to-peak amplitude, and signal-to-noise ratio
(SNR), whereas stiff MEAs showed a progressive decline in recording performance. Compared to
the stiff probes, a milder immune response, and an elevated neural population were observed
around the flexible probes. Increased expression of the mechanosensitive ion channel
correlated with microglial activation and astrocytic reactivity near the stiff probes, suggesting the
involvement of mechanosensitive channels in the tissue response.
Finally, as the first step to investigate bipolar disorder using neural interface devices, we
characterized the neural recording performance of implanted MEAs in the striatum of the bipolar
mania ClockΔ19 mutant mice. We compared the chronic neural recording metrics and histological
markers to wild-type control mice. SNR and channel yield were comparable between the groups,
however, mutant mice exhibited lower impedance, noise levels, and amplitude. A significantly
reduced neuronal population was observed around the implant in mutant mice accompanied by
lower microglia and astrocytic responses, indicating a compromised inflammatory system likely
due to the ClockΔ19 mutation.
These studies collectively contribute to understanding the neural electrode-tissue interface
and present promising strategies for improvement
Large Language Model For Mental Health: Attributional Style Transfer And Data Generation
No full textAccording to the reformulated version of the Learned Helplessness theory, an individual who experiences uncontrollable negative events may subsequently develop a negative attributional style, whereby the person exhibit greater susceptibility to helplessness deficits in response to negative events compared to their optimistic counterparts. This attributional style not only contributes to depressive symptoms but also represents a malleable target for cognitive therapy. Through attention to patients' attributional style, therapists may be able to alter patients' perceptions and experiences of negative events and thereby decrease the likelihood of depressive symptoms. In an effort to better interpret mental health issues and assist individuals with depressive tendencies in overcoming negative attributions, we introduce the Attributional Style Transfer Dataset (ASTD), which features paragraphs of events described in six distinct attributional styles. Additionally, we offer an open-source benchmark library comprising datasets and baseline methods, designed to support and advance future research and applications in this domain. At the same time, we have experimentally demonstrated that by introducing information from other datasets, rather than simply generating data through LLM, we are able to obtain additional information and improve the diversity of the generated data, which is a breakthrough for the use of LLM for data generation
Environmental and Genetic Influences on Low-density Lipoprotein Cholesterol Levels and Cholesterol-lowering Medication Use in Asian Americans:Findings from the All of Us Research Program
No full textCardiovascular disease (CVD) remains the leading cause of death globally and in the US. Around 7% of Asian Americans have CVD, posing a significant health burden. One of the major risk factors for CVD is elevated Low-density lipoprotein (LDL) cholesterol. Although extensive research has examined LDL cholesterol and its environmental and genetic determinants in the general population, studies focused specifically on Asian Americans are limited. Statins, essential for reducing LDL cholesterol and preventing CVD, are often underused in the US, with about 40% of eligible individuals not receiving treatment. Research on racial disparities in statin use has primarily highlighted Black/African American and Hispanic American populations, leaving gaps in understanding for Asian Americans. This study aims to identify environmental and genetic factors affecting LDL cholesterol levels and assess statin use and barriers among Asian Americans.
We analyzed data from self-reported Asian American adults aged 18 and older in the All of Us Research Program. Environmental factors were examined using surveys, physical measurements, and lab data. We also conducted a genome-wide association study (GWAS) to test the association between genetic markers and LDL cholesterol levels. Statin use was assessed based on the 2013 ACC/AHA (American College of Cardiology/American Heart Association) Blood Cholesterol Guideline. Our findings indicate that Asian Americans have higher LDL cholesterol levels compared to other racial/ethnic groups, with no significant lifestyle factors linked to high LDL cholesterol. GWAS analysis identified six suggestive SNPs associated with decreasing the risk of high LDL cholesterol. Among them, rs5748554 and rs6485549 were located near ZDHHC8 and DKK3, which are indirectly related to lipid metabolism. Despite eligibility, only two-thirds of Asian Americans received cholesterol-lowering medications, and statin use was 17% lower in women than in men with ASCVD (atherosclerotic cardiovascular disease), indicating a notable gender disparity.
The public health significance of this study is identifying novel genetic and non-genetic risk factors for high LDL cholesterol and underutilizing lipid-lowering medication. These findings will help to understand population-specific factors affecting LDL cholesterol and inform tailored interventions for its management by improving medication adherence, ultimately leading to improved cardiovascular health in this population
Developing a Multi-Component Mobile Health App to Provide Ongoing Support for Family Caregivers of People with Chronic Conditions and Disabilities
No full textOver 20% of Americans, approximately 53 million adults, are family caregivers providing long-term care for an adult or a child, particularly those with chronic conditions and disabilities. While providing a valued service to relatives, family caregivers are at considerable risk of adverse physical, mental, and social health outcomes. Often starting their caregiving roles with minimal training and being tired of getting familiar with the role, they tend to neglect their own health to satisfy caregiving demands. Comprehensive and continuous support for family caregivers remains scarce. Mobile health apps have the potential to bridge the gap by providing multi-component portable support on caregiving skills, self-care, and socialization. This dissertation employs user-centered approaches throughout the process to develop and evaluate a mobile health app to support family caregivers of individuals with chronic conditions and disabilities on self-care and caregiving skill training. The app is tailored to assist with general and condition-specific caregiving and is adaptable for various family caregiver populations. Aligning with user-centered approach principles, one focus group, two co-design workshops, and one pilot test focusing on usability, accessibility, and feasibility were employed to examine the results. New FCG challenges and unmet needs were revealed from the dissertation studies. Additionally, feedback from participating family caregivers suggested design criteria and development guidelines, which took into account family caregiver needs, usability, accessibility, and security to inform the future development and iterations of this and similar family-caregiver-focused mobile health supports. After the pilot test, family caregivers indicated that the app is user-friendly, holds the potential to expand in future studies, and supports family caregivers in real-world settings
Building Bridges to Medicine: A Community-based Approach to Empowering Under-resourced Premedical Students
No full textDiversity in medicine enhances healthcare quality, yet systemic barriers often impede under-resourced students from pursuing medical careers. To address this, Giving a Boost (GAB)—a student-led initiative—provides free, individualized medical school application support to aspiring physicians from underserved backgrounds. Over four application cycles, GAB supported 341 mentees through essay writing, mock interviews, and more, services that can cost up to $10,000. In the current cycle of 154 mentees, 24.7% identify as underrepresented in medicine, and 88.3% cannot afford paid application services, underscoring GAB’s role in assisting aspiring medical students. Historically, surveyed mentees rated GAB’s support as significantly more effective (8.4/10) than traditional resources, with qualitative feedback highlighting its value. An overall 87% medical school acceptance rate, more than double the national average, further emphasizes GAB’s impact. This presentation highlights how community-engaged scholarship and outreach can reduce barriers while fostering equitable opportunities for future healthcare leaders
Pittsburgh Policy Initiative (PPI): the University of Pittsburgh’s first student-run think tank
No full textThe Pittsburgh Policy Initiative (PPI), sponsored by the Office of Community Engaged Learning in the Frederick Honors College, represents an innovative approach to civic education as the University of Pittsburgh’s first student-run think tank. Its flagship Policy Analyst Program (PAP) exemplifies engaged scholarship by integrating policy analysis with community-based research to address pressing local issues. Through interdisciplinary teams, faculty collaborations, and direct community engagement, students address critical challenges facing Pittsburgh neighborhoods—including limited healthcare access, affordable housing shortages, and food insecurities—while amplifying underrepresented voices in the policymaking process. Grounded in residents’ lived experiences, the PAP empowers students to become civic leaders while rebuilding trust between institutions and communities. The program allows students to drive meaningful change by presenting actionable solutions to policymakers at PPI's annual forum. This workshop will showcase PPI’s model, offering strategies to design community-engaged research initiatives that strengthen civic preparedness, build institutional trust, and uplift community voices
Predicting Perceived Lasting Benefits and Future Digital Detox Willingness from a Mood Management Perspective
No full textThis study investigates the psychological outcomes of digital detox among graduate students, focusing on factors influencing future detox engagement and perceived lasting benefits. It examines how post-detox psychological changes in mood, anxiety, sleep, relationships, and productivity predict willingness to engage in future detox activities and perceptions of lasting benefits. It also considers how demographic factors and prior social media experiences shape these perceptions. Mood Management Theory was employed as a theoretical framework for understanding the willingness to engage in future digital detox and its lasting benefits. Data were collected through an online survey from 114 graduate students at an American public university, recruiting social media users who had voluntarily undergone a social media detox. The findings offer implications for developing programs and interventions aimed at enhancing mood, productivity, and promoting mindful social media use. These insights can help educators and academic librarians support students during digital detox and promote well-being
Efficient Computational Method for Three-Dimensional Thermally- and Electrically-Coupled Performance Prediction of GPHS-RTG
No full textThis paper introduces a novel analytic model for thermally and electrically coupled General Purpose Heat Source–Radioisotope Thermoelectric Generators (GPHS-RTGs). The model simplifies the RTG’s complex structure by leveraging symmetry arguments, capturing critical radiative, conductive, and electrical interactions while significantly reducing computational costs. The model includes many segmented unicouples and multi-foil insulation (MFI) sections, which enable it to assess spatial variations in thermal and electrical performance. Results demonstrate that unicouples closer to the RTG’s midspan heat source support exhibit approximately 1\% higher power output than those near the end caps due to radiation view factor disparities. Scaling up the model to a full RTG configuration, the total predicted beginning-of-mission (BOM) electrical power output is 295.5 W, aligning well with the 296 W average reported for the Cassini mission’s RTGs. The model reveals that MFI layers limit heat loss to only 0.1\% of the total heat from the GPHS, ensuring thermal efficiency. The surface temperatures of the GPHS bricks range between 1,310.6 K and 1,311.6 K, with a nearly uniform MFI surface temperature from 1,307.3 K to 1,307.9 K. This framework also provides granular insights into unicouple performance: unicouples on the diagonal relative to the GPHS produce less power due to unfavorable radiation view factors, resulting in slight asymmetries across the RTG’s electrical output. The model’s ability to resolve unicouple-level behavior while maintaining RTG-wide thermal and electrical balances makes it a powerful tool for analyzing RTG configurations