Gemstone Team Research

Permanent URI for this collection

The Gemstone Program at the University of Maryland is a unique multidisciplinary four-year research program for selected undergraduate honors students of all majors. Under guidance of faculty mentors and Gemstone staff, teams of students design, direct and conduct significant research, often but not exclusively exploring the interdependence of science and technology with society. Gemstone students are members of a living-learning community comprised of fellow students, faculty and staff who work together to enrich the undergraduate experience. This community challenges and supports the students in the development of their research, teamwork, communication and leadership skills. In the fourth year, each team of students presents its research in the form of a thesis to experts, and the students complete the program with a citation and a tangible sense of accomplishment.

Browse

Recent Submissions

Now showing 1 - 20 of 177
  • Item
    Challenging the Fast Fabric Industry: A Zero Fabric Waste Campus
    (2023) Bloch, Margot; Chang, Mia; Esuke, Alexis; Fallen, Kira; Gallagher, Kayleigh; Senthil, Mina; Zheng, Stephanie; Furst, Mary Beth
    The fast fashion industry creates environmental, humanitarian, and economic problems by selling inexpensive, poor-quality products, causing pollution during and after manufacturing, and encouraging throw-away culture. This research aimed to understand industry forces and consumer behavior while developing a process for diverting textile waste from landfills. Data gathered through surveys revealed cost as the leading factor among college students in clothing purchases. Additionally, students were deterred from sustainable habits such as upcycling and mending their own clothes due to a lack of time and skill. However, they were willing to make changes to behaviors as long as it was convenient. We also tested the feasibility of establishing a zero-fabric waste campus by collecting textiles and sorting them for redistribution for upcycling, donation, and recycling. The goal was to create a comprehensive blueprint for residential communities like universities to recreate a system as convenient as curbside recycling. More than 700 pounds of textiles were collected and diverted from landfills by donating them back to community organizations and giving them a second chance. As a result, we provided a channel for college students to act on their knowledge of fast-fashion clothing. This zero-fabric waste system has the potential to be highly successful given the attitudes of students determined in our research, who will drive change as a more environmentally conscientious generation.
  • Item
    CONTROLLING A NON-INVASIVE UPPER- LIMB PROSTHETIC DEVICE VIA A MACHINE-LEARNING ASSISTED BRAIN- COMPUTER INTERFACE
    (2023) Fadul, Nora; Ghaemi, Seyed Mohammad; Gresh, Nicholas; Gurwitz, Zachary; Liu, Wensen; Lopez, Daniel; Mittu, Nikhil; Patel, Abhay; Shah, Khadija; Wolcott, Micaela; Deane, Anil
    The goal of our research was to improve the accessibility of current upper-limb prostheses. We aimed to maintain non-invasive aspects of an Electroencephalography (EEG), use affordable material and resources, and match the accuracy and control of conventional prostheses alongside improved training methods. Our process began with us designing a method of data collection that uses a 3D-printed headset with dry electrodes to record brain signal data through EEG software. We then analyzed the signals, applied preprocessing to reduce noise, and used machine learning (ML) models to classify EEG signals with respect to specific actions such as the opening and closing of a hand. Finally, we constructed a 3D-printed hand that is actuated by servos with Arduino to demonstrate the physical actions interpreted through analysis, and we leveraged novel techniques to build a virtual reality (VR) environment to serve as a tool for prosthetic rehabilitation. We successfully met the goals set for data collection and prosthetic arm actuation. Additionally, we have created a functional algorithm for action prediction but were not able to achieve the desired accuracy. Overall, we achieved our primary goal of collecting brain signal data, analyzing that data through an algorithm, and actuating a prosthetic arm with actions interpreted from the brain signals all in real-time. Moving forward, there is room to increase accessibility and quality of prostheses through further development of non-invasive brain-computer interface (BCI) based technology for 3D-printed prostheses and VR environment prosthetic models.
  • Item
    An Investigation of the Effect of Curved Cantilever Geometry on Piezoelectric Power Output
    (2023) Brady, Ian; Esuke, Mikaila; McCarthy, Liam; Samson, Catherina; Sebastian, Rithik; Wedzielewski, Mark; Yadav, Sanjali; Blanton, Rick
    Team PISO recognizes an urgent need for new niches of environmentally sustainable energy, and believes that piezoelectric materials offer a small part of a solution. With this consideration, Team PISO has focused on the process of optimizing the application of strain on a piezoelectric component by altering the dimensions of a cantilever beam. Future research in this avenue could focus on a computational model based off of the experimental data collected herein, or more focused optimization of a selected cantilever profile. In the long term, PISO’s research could be applied to energy harvesters to reclaim power from vibrations and deformations, such as sounds and footsteps, as a novel source of renewable energy for implementation in public, heavily-traveled areas. This paper investigates the relationship between the geometric shape of a piezoelectric cantilever structure and its power output from discrete impulses. To this end, Team PISO created several curved cantilevers to examine the impact of the geometry of a piezoelectric cantilever on its output. These cantilevers were tested on an apparatus that simulated the movement of a footstep and their power outputs compared. PISO concluded that convex cantilevers were the most effective, with the convex spherical geometry outputting 24% more power than the control rectangular geometry.
  • Item
    ORBIT: Orbital Repairs By Innovative Technology
    (2023) Allee-Press, Lacey; Bindra, Navkaran; Cha, Leo; Dacey, William; Akin, David L.
    The University of Maryland’s Space Systems Laboratory has a long history with multiple design projects for small single-person spacecraft (SPS) intended for extravehicular operations. Since the SPS concept has been criticized as differing from teleoperators, suited-missions, and other space utility vehicle (SUV) models only by the ability of the operator to have direct vision (“eyes-on”) of the worksite, “hands-on” interaction with the worksite, and simultaneous use of robotic arms. Testing will focus on identifying performance differences between the methods. To quantify effectiveness for each option, we performed a Fitts’ Law tapping task in three hands-on environments: shirt-sleeve, space-suit arms, and SPS suit-arms, as well as two robotic control environments: one with direct eyes-on vision, and one using video screens for teleoperation. After each series of tasks, participants completed a short survey including the NASA Task Load Index (TLX) as well as a Cooper Harper rating. We hypothesized that a combination of hands-on control and eyes-on robotic control will be the most suitable design for an SUV. Experimentally, we found that within the hands-on control environments (shirt-sleeve, suit arms, SPS arms), there is not much variation in task difficulty, but that these environments were much easier to control than the robotic control environments. Among the robotic control environments, subjects performed better overall when using direct eyes-on vision, as opposed to teleoperation. Our findings suggest that combining eyes-on and hands-on interaction is both important and significant in SPS design and handleability.
  • Item
    Evaluating Physicochemical Properties of Poly-vinyl Pyrrolidone (PVP) Hydrogels for Local Delivery of Lipoproteins in Glioblastoma Multiforme (GBM)
    (2023) Liu, Cindy; Evensen, Ella; Patel, Aashka; Thales-Mogo, Nandi; Asfaw, Deborah; Bui, Julia; Deslouches, Jakobi; Lee, Isaiah; Kapoor, Ria; Al-Sheikhly, Mohamad
    Glioblastoma multiforme (GBM) is a fast growing, malignant tumor that develops from the glial cells that support the health of the nerve cells that affect the brain and spine. GBM has a low survival rate with minimal drugs/treatment options available. The purpose of our research was to seek potential innovative solutions to the challenges surrounding the treatment of GBM; particularly, we were interested in exploring a novel drug-delivery treatment that could be used in future GBM therapy. Through the use of Poly-vinyl Pyrrolidone (PVP), we aimed to develop a hydrogel with specific biomechanical properties that would contain lipoprotein-encased chemotherapeutic drugs; the idea being that these hydrogels would provide an ideal sustained released over a set amount of time. Using varying polymer concentration, molecular weights, and radiation doses, we tested the physical and chemical properties while simultaneously testing the kinetics of release of lipoproteins in each hydrogel.
  • Item
    DISRUPTION OF E-CADHERIN PROMOTES INTESTINAL STEM CELL PROLIFERATION IN COLONOIDS
    (2023) Arpon, Aaliyah; Bell, Michaela; Brown, Katie; Forsberg, Alisa; Krishnan, Ananya; Margolis, Ryan; Marhefka, Laura; Yang, Jerry; Yardi, Isha; Jin, Younggeon
    Reconstitution of the wounded epithelium is integral to achieve the full healing of the gut mucosa in treating Inflammatory Bowel Disease (IBD). The ability of intestinal stem cells (ISCs) to indefinitely self-renew while generating new functional epithelia makes them a potential therapeutic tool for IBD. Transmembrane protein E-cadherin, a calcium dependent cell-to-cell adhesion protein at adherens junctions, also regulates the Wnt signaling pathway. The canonical Wnt (β-catenin dependent) pathway is vital for the ISC homeostasis and regeneration. However, the role of E-cadherin in ISCs is an important yet notably understudied phenomenon. Disruption of E-cadherin increases unbound cytosolic β-catenin levels, which go to the nucleus and increase transcription of Wnt target genes. We hypothesize that disrupted E-cadherin will increase proliferation of ISCs. In our experiments, we disrupt E-cadherin with different concentrations of EGTA, a calcium chelator, and see the effect it has on colonoid growth and development. Our experiments showed that with EGTA there was greater proliferation; 1 mM EGTA experimental groups had larger colonoids than vehicle control colonoids on day 6 after seeding. This indicates that EGTA treatment may induce proliferation of the organoid with E-cadherin disruption. For future study, we will check and confirm the disruption of E-cadherin/β-catenin complex and Wnt target genes by real-time PCR and immunofluorescence studies. Ultimately our study will open novel therapeutic applications for patients living with IBD and other clinic inflammatory gut disorders.
  • Item
    Expansion of Polyurethane Foam in Low Pressure Environment for Space Debris Removal Applications
    (2023) Cai, Edward; Green, Ethan; Kavuturu, Aravind; O'Neill, Finnegan; Shedleski, Joseph; Yang, John; Young, Robert J.; Zakaria, Muhamad; Sedwick, Raymond
    As the privatization of space flight has led to an increased number of rocket launches and a new era of space exploration, the issue of space debris is becoming a well-researched phenomenon. With more emphasis on future extraterrestrial missions, such as NASA’s Artemis and SpaceX’s Mars missions, there is a growing concern surrounding the unsustainable practices that create space debris. Although there is a plethora of papers discussing the sources of space debris, as well as its negative impacts on the future of space flight, there are comparatively fewer papers discussing active debris removal methods. This study focuses on one such method - namely, using spray foam to remediate space debris. Spray foam has the advantage of being low-cost and multi-use, which differs from most other active removal methods. To determine the viability of spray foam in space debris removal, this study tests the expansion of polyurethane foam in vacuum, a poorly documented characteristic in current literature. From a sample of 20 tests, a maximum volumetric expansion ratio of 53 was found. The resulting discussion focuses on spray foam’s efficacy as an active debris removal method from these observations.
  • Item
    Exploration of a Viral Protein for Cancer Therapy
    (2023) Carter, Victoria; Funk, McKenzie; Johnson, Jordan; Lanasa, Dominic; Loewenstein, Eva; Luo, Katherine; Patel, Grishma; Shih, Eileen; Sofola, Rotimi; Srinavasan, Sruthi; Zhang, Yanjin
    Cancer is a group of malignant diseases and is one of the leading causes of death worldwide. Current treatments can be invasive and nonspecific, therefore killing healthy cells along with cancerous cells. In many types of cancers including lymphoma, signal transducer and activator of transcription 3 (STAT3) is upregulated and regarded as a risk factor for its enhancing tumorigenicity. Thus, STAT3 is a target for cancer therapy. In this project, we explored a viral protein called nsp5 that induces the degradation of STAT3 to develop cancer therapeutics against lymphoma. We cloned the nsp5 gene into a retroviral expression system and determined its expression. Replication-defective retrovirus particles were packaged and used to deliver nsp5 gene into the lymphoma-derived cells. The nsp5 effect on downregulation of STAT3 and tumor cell growth were determined. These results demonstrate that the viral protein can be explored for further preclinical development for potential tumor therapeutics.
  • Item
    Examining the Binding Mechanism and Public Health Implications of Alcohol Consumption and Poisoning
    (2023) Akosa, Chiamaka; Kim, Abigail; Kleptich, Emma; Majadly, Ahlam; Shi, Tesia; Wasson, Diana; Matysiak, Silvina
    Alcohol is a commonly abused drug, especially in the United States, which contributes to a death rate of over 2,000 people a year from alcohol poisoning. In particular on college campuses, social binge drinking is prevalent, which can lead to alcohol poisoning, defined as excessive alcohol diffusion into the brain. In the case of alcohol poisoning, ethanol molecules diffuse through the blood brain barrier (BBB) and bind to extrasynaptic gamma aminobutyric acid type A (GABA-AR) receptors which causes neuronal inhibition, nausea, breathing difficulties, and potential death. However, specific binding mechanisms remain unknown. To analyze the biological impacts of alcohol poisoning, we created computational models of the lipid bilayer with an inserted GABA protein and the corresponding ethanol concentration of 0.3 g/dL using molecular dynamics. To examine the social impacts of alcohol consumption and college students’ drinking habits, we conducted a survey of undergraduate students at the University of Maryland. After examining the relationship between family history, parental knowledge of alcohol consumption, and drinking frequency and quantity, we elucidated relationships between these variables. By understanding both the biological and social implications of alcohol consumption, future researchers may be able to develop campaigns to prevent deaths from overconsumption of alcohol.
  • Item
    Analyzing Unconscious Bias in Indeed’s Employee Resume Search
    (2023) Antony, Rachel; Bral, Benjamin; Gleason, Seth; Ihm, Joanna; Malhotra, Aarushi; Mathew, Philip; Nagaokar, Soham; Rajala, Johnny; Truong, Kyle; Zhu, Daniel; Sin, Steve
    This project analyzes if artificial intelligence (AI) hiring systems demonstrate racial bias as measured by prestige bias against graduates of historically black colleges and universities (HBCUs), measured in a number of different metrics, and how that bias can be mitigated. The metrics we used to measure prestige were: size, university rankings, public vs. private universities, and attendance of HBCUs vs. attendance of non-HBCUs. We examined how the hiring site Indeed utilizes AI to list candidate resumes by relevance and measured the relationship between candidates’ resume rankings and the universities they attended. While we found no significant difference between the overall average rankings of applicants from HBCUs and applicants from non-HBCUs, we did find significant differences between these applicants when we made comparisons based on variables such as major, experience level, and most recent company size. Future research on this topic includes training an AI model on the collected resumes to see if the same results are generated and adjusting the model to mitigate said biases. This research will shed light on the bias embedded in human hiring departments. With businesses considering AI as a tool for hiring, companies must understand that AI hiring systems can perpetuate the same biases found in human hiring on a larger scale.
  • Item
    Intra-haptic Bayesian Integration between frequency and intensity in humans: Evidence of super-optimal integration in haptic signals
    (2023) Weiner, Matthew J.; Tumkur, Medha V.; Blurton Jones, Martin X.; Murshed, Zaheen; Koh, Kyung; Shim, Jae Kun
    Understanding how the Central Nervous System (CNS) integrates multiple forms of information is crucial for the field of neuromechanics, as it can inform the development of optimal feedback systems. Such systems could benefit individuals with various limitations by easing their difficulties and enhancing their capabilities. This study presents an experimental procedure that investigates how the CNS integrates changes in intensity and frequency of haptic feedback, and tests whether the results align with the Bayesian integration theory. Participants were asked to match reference signals using a custom device and software for three conditions: changing intensity, changing frequency, and simultaneous changes in both. Overall, this study investigated the integration of changes in intensity and frequency of haptic feedback within the somatosensory modality, and found that combining different types of haptic information may improve perceptual precision. While the results were supported by Bayesian integration theory, the data also suggests the need for further investigation into potential interactions between inter-sensory modalities, such as audio-haptic feedback. Furthermore, the results showed that the subjects had the least errors in perception when both intensity and frequency conditions are changing, suggesting that the implementation of both conditions is optimal. The study opens up avenues for future research, such as exploring how exactly the variation of both conditions should be from two instantaneous points in time, and whether linearly increasing the levels of both conditions is the best way to communicate information to a user. Additionally, the study suggests that there may be intra-haptic integration ability in humans, which could be explored in future research.
  • Item
    Developing A Broadly Protective MRNA Influenza Vaccine: A Review
    (2022) Acquah, Wellington; Amini, Cameron; Buddula, Saharsh; Chen, Michelle; Chintala, Navya; Dang, Quinn; Ferziger, Noa; Hollis, Grace; Jameison, Devin; Jayaram, Jyostna; Manus, Joseph Anthony; Rosenberg, Jacob; Zhiteneva, Julia; Yarwood, Stephanie
    Current influenza vaccines are limited in their efficacy due to antigenic drift of the hemagglutinin target; advances in mRNA vaccines in response to the COVID-19 pandemic may provide a new direction for influenza vaccine development. Existing literature shows that mRNA vaccines have higher efficacy in preventing illness, hospitalizations, and death. We evaluated eleven influenza A viral proteins as potential targets for an mRNA vaccine under the following criteria: degree of conservation, ability to elicit a robust immune response, and ability to prevent illness and death. We recommend future researchers direct their efforts towards developing an annually administered tri-sequence mRNA vaccine targeting hemagglutinin head (HA1), the matrix 2 ectodomain (M2e), and nucleoprotein (NP). Development of a highly effective influenza mRNA vaccine would be significant for prevention of disease burden worldwide.
  • Item
    Analysis of Whole-Brain Resting-State MRI Using Multi-Label Deformable Offset Networks and Segmentations Based Attention with Explorations into the Ethical Implications of Artificial Intelligence in Clinical Psychiatry Settings and Care
    (2022) Agarwal, Vatsal; Ayoroa, Evan; Burdick, Ryerson; Ganeshan, Aravind; Paliyam, Madhava; Wood, Sam; Lee, Caitlin; Akhtarkhavari, Sepehr; Inala, Shika; Matharu, Sagar; Mupparapu, Neelesh; Deane, Anil
    Due to the poor understanding of the underlying biological mechanisms of psychiatric disorders, diagnoses rely upon symptomatic criteria and clinicians’ discretion. Reviews of these criteria have revealed issues of heterogeneity, over and under specificity, and symptom overlap between disorders. Deep learning provides a method to produce quantifiable diagnostic labels based upon biological markers such as specific features of brain anatomy or functionality. In practice, these methods fail to indicate how a particular result was determined, raising major obstacles for clinical implementation.To improve the efficiency and interpretability of existing deep networks, we have developed a novel atlas-based attention module to more easily capture global information across different areas of brain function. Our model can be extended to symptom level classification using NIMH data to give clinicians usable information outside of broad disorder classification. We have compared our model against leading 3D deep learning frameworks and have shown that our novel atlas-based attention module achieves 88% F1 and 91% accuracy on the UCLA Consortium for Neuropsychiatric Phenomics dataset. We have embedded our model with elements like deformable convolutions, gradient activation visualizations, and occlusion testing to show model attention and function. In addition to the lack of explainability, addressing the ethical issues surrounding clinical implementation of artificial intelligence is necessary before usage can become a reality. We identified a series of regulatory recommendations to address pertinent ethical concerns of equity and bias during both model development and clinical usage. We propose a standardized protocol for developing a clinical reference standard, the development of diversity reports regarding data used by models, and regulation of usage scenarios to reduce contextual bias.
  • Item
    Exploration of the Security and Usability of the FIDO2 Authentication Protocol
    (2022) Breit, Zachary; Dean, Hunter; Generrette, Tai-Juan; Howard, Samuel; Kodali, Balaji; Kong, Jim; Tash, Jonah; Wang, Phillip; Wu, John; Baras, John
    Fast IDentity Online (FIDO) is a passwordless authentication protocol for the web that leverages public key cryptography and trusted devices to avoid shared secrets on servers. The current version of FIDO, FIDO2, has become widespread and is directly integrated into popular systems such as Windows Hello and Android OS. This thesis details two contributions to the advancement of FIDO2. The first is a modification to the protocol which uses Trusted Execution Environments to resolve security vulnerabilities in the Client To Authenticator Protocol Version 2 (CTAP2), which is a component of FIDO2. It is formally demonstrated that this modification provides a stronger security assumption than CTAP2. The second contribution is an outline of procedures and resources for future researchers to carry out a study of the usability of FIDO2 authenticators via a within-subjects experiment. In the study, subjects register an account on a custom web app using both passwords and FIDO2 credentials. The web app collects metrics about user behavior such as timing information for authentication sessions. Over the course of a week, subjects log in to the same web app every day using both authentication methods. Subjects complete entrance and exit surveys based on the System Usability Scale (SUS) according to their experiences. The surveys and user metrics would then be analyzed to determine whether users perceive FIDO2 as more usable than passwords.
  • Item
    Knowledge, Attitudes, and Practices Towards Teaching of Menstruation and Sexual Health Among Parents of Middle School Students
    (2022) Adkins, Aaliyah G.; Barrett, Christina S.; Fano, Gabby M.; Hospes, Katrina C.; Kilby, Christina F.; Mareno, Michael C.; Ollila, Elizabeth L.; Pettit, Jessica C.; Savoy, Jayme G.; Rowe, Tatiana; Wilkerson, Lucy A.; Mittal, Mona
    Menarche (the onset of menstruation), along with puberty in general, presents as a trying time for adolescents as they adjust to changes occurring in their bodies. Family life and sexual education are imperative during this transitional stage as they set the foundation for future reproductive health decisions adolescents may make. Prior research on menstruation and menstrual health has primarily focused on rural populations in developing countries; few studies on this topic have been conducted in the United States (U.S.). The findings of these studies show disparities in knowledge related to menstruation and menstrual health among different racial and socio-economic groups in the U.S. We added to current literature by conducting a mixed-method study to investigate knowledge, attitudes, beliefs, and parenting practices related to menstruation and sexual health education among the parents of middle school students in the U.S. We conducted an online study and collected survey data from parents of middle school students, followed by qualitative interviews with select parents (those who opted-in for this portion) to gain further insight into the attitudes and sentiments regarding menstruation and menstrual health.
  • Item
    Fabrication of Soft, Ferromagnetic Films and Devices and their Properties, Printability, and Applications
    (2022) Carlson, Elizabeth; Chen, Anson; Chung, Stephen; Dhamsania, Anjali; Mah, William; Mueller, Lillian; Sivarajan, Arjun; Ting, John; Das, Siddhartha
    Materials enabling fabrication of multifunctional devices are a cornerstone of present-day materials science and engineering. Multifunctionality enables use for novel applications in fields like energy, health, sensing, etc. We conducted an extensive literature review into the development of one material capable of multifunctional device fabrication, elastic magnetic films and devices, which have two notable properties: magnetizability, and physical softness and compliance. We highlighted materials, fabrication, characterization, and resulting interactions harnessed to develop inks used to fabricate these films, as well as broadscale applications. We also experimented with an Fe3O4-PDMS compliant magnetic film to characterize magnetic properties under modes of deformation. Through bending and twisting, the magnetic saturation, coercivity, and retentivity were measured. Results revealed that bent configurations preserved magnetic characteristics better than twisting configurations; out of tested twisting angles, a 180° rotation displayed properties closest to the undeformed state. We concluded by describing the potential of future research endeavors.
  • Item
    Analysis of Gene Targeting Techniques for Huntington’s Disease and Gene Expression in Human Cells
    (2022) Fields, Eric; Tripu, Deepika; Vaughan, Erik; Lim, Isabelle; Conway, Jessica; Salib, Nicole; Jacobsen, Michael; Lee, Yubin; Dhamsania, Akash; Woo, Ashley; Shrout, Katie; Cao, Kan
    Huntington’s disease (HD) is an inherited neurodegenerative disorder that is caused by a CAG trinucleotide repeat expansion in the huntingtin (HTT) gene. Our team performed a literature analysis to investigate the current state of research for treating HD and identified a new technology called prime editing that could be applied to HD in combination with single nucleotide polymorphisms (SNPs). We found that at least 729 SNPs within the HTT gene are compatible with our proposed approach. Experimentally, we performed preliminary studies using Western Blots and RT-qPCR to examine the differences in expression of HTT in a variety of cell lines. Our literature-based work suggests that prime editing is a promising tool for addressing the basis of a variety of genetic disorders. Our experimental-based work confirms that human fibroblast cells express HTT and therefore may be used in proof of concept studies of gene targeting techniques to address HD.
  • Item
    Cognitive Testing, Neuroimaging, and Blood Biomarkers in the Development and Progression of Alzheimer's Disease
    (2022) Cieslak, Zofia; Hemani, Danny; Kubli, Anjali; Lee, So Min; Mgboji, Rejoyce; Nallani, Madhulika; Park, Michael; Samson, Mahalet; Wu, Benjamin; Smith, J. Carson
    Alzheimer’s disease (AD) is a progressive neurodegenerative disorder, characterized by significant loss of memory and cognitive dysfunction. It has a significant impact on an individual’s health and may financially and socially burden these individuals and their loved ones. Although the disease has been researched extensively, there is still no clear understanding of the proposed mechanisms behind the development of AD and factors aside from genetics which potentially influence the risk of developing AD. The purpose of this research is to compile and analyze data on cognitively healthy participants, participants with MCI, and participants with AD to better understand the importance of genetic risk and changes in cognitive function, bioimaging and biomarker levels, as recorded on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. There are complex and significant relationships among these three variable groups with genetics and time. Executive function scores in healthy participants and participants with MCI were decreased with age and increased with education. In participants with AD, scores decreased over time. Language scores in healthy participants decreased with age, increased with education and for women. In participants with MCI, scores decreased with risk and time, and there was an interaction between these two variables. They also decreased with age and increased with education. In participants with AD language scores decreased over time. Memory scores in healthy participants increase with time and education and for women. In participants with MCI, scores increased with education and decreased with risk and time, and there was an interaction between these two variables. For participants with AD, there was a decrease over time. Visuospatial ability scores in healthy participants decreased with education. In participants with MCI, scores decreased with genetic risk and increased with education. In participants with AD, scores decreased over time and increased with age. Left hippocampal volume in healthy participants decreased with time, age, and education, and is increased in women. In participants with MCI, volume decreased with risk, time, age, and education. In participants with AD, volume decreased with time and age. Right hippocampal volume in healthy participants decreased with time, age, and education. In participants with MCI, volume decreased with risk and time, and there was an interaction between these two variables. Volumes also decreased with age. For participants with AD, volume decreased with risk, time, and age. Total hippocampal volume in healthy participants decreased with time, age, and education, and was increased for women. There was also an interaction between risk and time. In participants with MCI, volumes decreased with risk and time, and there was an interaction between these two variables. Volumes also decreased with age and education. For participants with AD, volumes decreased with risk, time, and age. Aβ42 levels in healthy participants decreases with risk and increased with time. In participants with MCI, levels increased with time and age, and were lower in women. In participants with AD, levels increased with time. Aβ40 levels in healthy participants increased with time and were lower for women. For participants with MCI, levels increased with time and age, and were lower for women. In participants with AD, levels increased over time. The Aβ42/40 ratio in healthy participants decreased with risk and time, and decreased with time in participants with MCI. The findings give insight into AD development and contribute to a greater understanding of longitudinal changes in AD progression. In relation to the study of AD includes the perpetuation of racial inequalities. People of color have an increased risk of developing AD and are disproportionately affected by the disease, yet are severely underrepresented in most research studies, including the research collected in the ADNI database. Racial minorities also often do not have the same access to healthcare as white people, thus contributing to the decreased possibility of early detection and treatment of AD. Black Americans, specifically, often face socio-economic barriers, which further renders the burden of AD development and progression more serious for minority families. In order to promote awareness of AD among underrepresented communities, Team Brain virtually presented to the African American Health Program, a local community of minority elders, via virtual presentations. Overall, this research concluded that hippocampal atrophy and cognitive tests appear to be the most consistent factors in the progression of MCI and AD. The analysis of blood biomarkers produced inconclusive results. This research indicates a clear set of imaging and cognitive factors that can be used to create less invasive and novel diagnostic methods for AD as well as supports the need for further research on blood biomarkers to understand their relationship with cognitive decline and progression of AD.
  • Item
    LEMMA: A Data-Driven Approach to Modeling the Spread of Extremism Over Online Platforms
    (2022) Fream, Mitchell; Hayes, Nathan; Kochar, Sahil; Kolbeck, Paul; Schneider, Charlie; Schwartz, Russell; Sharon, Olivia; Shen, Yuang; Weiss, Winslow; Wolle, Robert; Jabin, Pierre-Emmanuel
    The online spread of extremist ideas has been a growing problem. Team LEMMA has worked to quantitatively model the spread of extremist ideas over Reddit in order to gain insight into how they may spread. A modest dataset of Reddit comments were manually rated on the level of extremist rhetoric present and used to train a machine learning algorithm to automatically classify large swaths of Reddit data. These ratings were then fit to a predictive agent-based model with the hopes of better understanding past trends and potentially forecasting future spread of extremism.
  • Item
    Measuring Mental Workload and Brain Dynamics in Prosthesis Motor Learning over Multi-Session Practice
    (2022) Asenso, Maxine; Brown, McCauley; Dayanim, Gabriel; Doyle, Erin; Greenbaum, Maya; Lavarias, Gabrielle; Mercado, Natalia Nava; Nguyen, Christina; Russell, Ashley; Still, Alexys; Subramaniam, Carolyn; Varma, Anagha Rama; Gentili, Rodolphe
    The capability of humans to adapt their motor behavior and learn new motor skills is critical to interact with their changing environment as well as for integration with new machine interfaces, such as assistive technology (Casadio, Ranganathan, & Mussa-Ivaldia, 2012; Kitago & Krakauer, 2013; Mussa-Ivaldi et al., 2011). Such learning capability depends on the recruitment of cognitivemotor resources (Wickens, 2002). Mental workload (MWL), which is an important component in understanding learning, can be defined as the relationship between the deployment of neural resources and imposed task demands (Sharples & Megaw, 2005; Young et al., 2015). Although a large body of work has examined the behavior and cortical dynamics underlying the motor learning processes, most of this prior effort generally did not examine changes in mental workload through multiple practice sessions and did not consider individuals with upper limb (UL) loss (Marchand, de Graaf, & Jarrassé, 2021; Park & Zahabi, 2022). In this work, UL amputees were approximated by considering healthy individuals using bypass prostheses (Bloomer, Wang, & Kontson, 2018; Wang et al., 2021). Based on the work by Bloomer and Wang, able-bodied individuals can serve as a reasonable proxy for amputees while using these bypass prostheses. From a methodological standpoint, the use of human-body interfaces such as a bypass prosthesis is interesting since it requires participants to acquire a novel and unusual sensorimotor mapping, mitigating the influence of prior motor experiences and ultimately offering a fairly unbiased learning paradigm (Casadio, Ranganathan, & Mussa-Ivaldia, 2012; Mussa-Ivaldi et al., 2011). Thus, we employed this approach here, along with electroencephalography (EEG), which was used to assess the cortical dynamics as participants completed the learning task in order to objectively assess mental workload. In addition, surveys were employed to subjectively assess the level of workload perceived by the participants along with performance (e.g., time, smoothness, number of blocks transported within a fixed time period) collected via an inertial measuring unit. Overall, the aim of this research was to examine the concomitant changes in performance (e.g., number of blocks transported within a fixed time period) and in mental workload (by means of surveys and the cortical dynamics indexed by EEG) that occur when healthy individuals learn to operate a bypass prosthetic device via multi-session practice to perform a variety of motor tasks of daily living. This work can inform not only the human cognitive-motor processes underlying mental workload and performance during learning but also, to some degree, the rehabilitation/training of UL amputees, as well as the design and evaluation of prosthetic devices.