Simantika Bhattacharjee Dristi, Computer Science | University of Virginia

"Analyzing and Mitigating Surface Bias in Code Evaluation Metrics"

Advisor: Dr. Matthew B. Dwyer

With the increasing popularity of large language models (LLMs), reliable and effective code evaluation metrics (CEMs) have become crucial for several software engineering tasks. While popular benchmarks often provide test cases to assess the correctness of generated code, crafting and executing test cases is expensive. Reference-based CEMs provide a cheaper alternative by scoring a candidate program based on its functional similarity to a reference. Though prior research has reported the weak correlation between these CEMs and functional correctness, plausible solutions remain unexplored. In this work, we critically evaluate four popular reference-based CEMs, revealing their strong bias towards surface-level features rather than code functionality. Despite this surface bias, current evaluation datasets rarely include code pairs that are surface-similar yet functionally dissimilar, or functionally similar yet surface-dissimilar. To mitigate this gap, we propose LoCaL (Looks Can Lie), a CEM evaluation benchmark, with 3117 code pairs at both the method and program levels. Each pair is labeled with a functional similarity score and aims to target the critical regions for CEMs. The functional similarity scores are computed without predefined test cases, with reliability improved by executing 10x more tests than prior work. We find that all four CEMs show significant performance degradation on LoCaL, compared to the baselines. Finally, our findings imply that exposing CEMs to LoCaL-like data might facilitate the development of metrics that are robust to surface bias. 

Simantika Bhattacharjee Dristi is a PhD student in Computer Science at the University of Virginia (UVA), advised by Prof. Matthew B. Dwyer. She is a member of the LESS Lab research group. Her research lies at the intersection of software engineering and generative AI, with a focus on leveraging Large Language Models (LLMs) to enhance the testing, evaluation, and reasoning capabilities of software systems.

Dipin Khati, Computer Science

"Mapping the Trust Terrain: LLMs in Software Engineering-Insights and Perspectives"

Advisor: Dr. Denys Poshyvanyk

Co-Authors: Y. Liu, D.N. Palacio, Y. Zhang

The application of Large Language Models (LLMs) in Software Engineering (SE) continues to grow rapidly across both industry and academia. As these models become integral to critical SE processes, ensuring their reliability and trustworthiness becomes essential. Achieving this requires a balanced approach to trust: excessive trust can introduce security vulnerabilities, while insufficient trust may hinder innovation. However, the conceptual landscape of trust in LLMs for SE(LLM4SE) remains unclear. Key concepts such as trust, distrust, and trustworthiness lack precise definitions, factors that shape trust formation remain underexplored, and metrics for trust in LLMs remain undeveloped. To clarify the current research landscape and identify future directions, we conducted a comprehensive review of 88 articles: a systematic review of 18 studies on LLMs in SE, supplemented by an analysis of 70 articles from the broader trust literature. Furthermore, we surveyed 25 domain experts to gather practitioners’ perspectives on trust and identify gaps between their experiences and the existing literature. Our findings provide a structured overview of trust-related concepts in LLM4SE, outlining key areas for future research. This study contributes to building more trustworthy LLM-assisted software engineering processes, ultimately supporting safer and more effective adoption of LLMs in SE.

Dipin Khati is a fifth-year Ph.D. candidate in Computer Science at William & Mary. His research lies at the intersection of Artificial Intelligence, software engineering, and human–computer interaction, with a focus on trustworthy AI for software engineering. His work investigates how trust shapes developers' interactions with AI-assisted tools and how it influences developer productivity and adoption.

Daniel Rodriguez-Cardenas, Computer Science

"Evaluating and Explaining Large Language Models for Code Using Syntactic Structures"

Advisor: Dr. Denys Poshyvanyk

Trustworthiness and interpretability are closely connected concepts for LLMs. The more interpretable an LLM is, the more trustworthy it becomes. However, current techniques for interpreting LLMs in code-related tasks focus on accuracy, responses to changes, or individual task performance rather than the detailed explanations needed at prediction time for higher interpretability and trust. To improve this situation, this paper introduces ASTxplainer, an interpretability method for LLMs of code that creates explanations based on the relationship between model confidence and the syntactic structures of programming languages. ASTxplainer explains generated code using syntax categories derived from Abstract Syntax Trees (ASTs) and helps practitioners understand model predictions. By assigning and grouping model confidence scores with well-known syntactic structures in ASTs, our approach goes beyond previous methods, providing a view of model confidence that relates to familiar programming language concepts. We developed an automated visualization that shows the combined model confidence scores overlaid on sequence, heat-map, and graph-based visuals of syntactic structures from ASTs. We examine the benefits of ASTxplainer through a study on 12 popular LLMs using a curated set of GitHub repos and the usefulness of the approach through a human study. Our findings illustrate a causal connection between learning error and an LLM's ability to predict different syntax categories; our approach can be used to interpret model effectiveness in the context of its syntactic categories.

Daniel Rodriguez-Cardenas is a fifth-year Ph.D. candidate in Computer Science at William & Mary. His research focuses on the evaluation of large language models for software engineering, including benchmark construction, standardized data pipelines, and reproducible assessment of code-related tasks. His work has been published at ICSE and ICSME.

Kathryn Gour, Psychological Sciences

"Neural Resonance Profiling: A Psychometric Reliability Analysis"

Advisor: Dr. Paul Kieffaber

In a previous study, we systematically mapped neural resonance profiles (NRP) across the full oscillatory spectrum (i.e., 1–50 Hz). The present study aims to assess the internal consistency of NRPs, in the hopes of establishing NRPs as a reliable, transdiagnostic biomarker. This study unifies findings in entrainment research by mapping the relationships between oscillatory responses to a binaural auditory stimulus across the frequency spectrum (1-75 Hz), Autism Spectrum Disorder (ASD) traits, and cognitive abilities. By bridging traditionally siloed research domains, we hope to identify emergent properties of neural function that cannot be fully captured by focusing on individual oscillatory frequencies or isolated social and cognitive processes. Approximately 60 younger adults from the College of William & Mary were recruited to participate in the study. Electrophysiological responses were recorded after participants completed the Broad Autism Phenotype Questionnaire, the Big-Five Personality Inventory, and the Montreal Cognitive Assessment. The results aim to indicate whether the internal consistency of neural resonance can be quantified for future replication studies.

Kathryn Gour is a second-year M.S. student in the Psychological Sciences Department at William & Mary. Her research areas include cognitive and behavioral neuroscience. Her project investigates the reliability of a novel neuronal oscillatory biomarker for Autism Spectrum Disorder. She holds a B.S. in Neuroscience from William & Mary.

Javiera A. Hernandez Puelma, Psychological Sciences

"Cross-National Longitudinal Insights into College Alcohol Beliefs and Consequences: The Role of Food and Alcohol Disturbance"

Advisor: Dr. Adrian J. Bravo

The present study examined the mediational role of FAD-intoxication (i.e., restricting calories to increase or quicken alcohol intoxication) in the relationship between college alcohol beliefs (CABs; i.e., internalized college drinking norms about the importance of drinking to the college experience) and negative alcohol-related consequences. A sample of 2,543 past-month student drinkers (Mage = 20.87, SD = 3.98; 70.2% female) from the U.S., Argentina, Spain, South Africa, England, and Canada, completed the baseline survey, and 505 drinkers completed at least one follow-up. A cross-sectional model was first employed to examine whether frequency of FAD-intoxication mediates the relationship between CABs and alcohol-related problems, and structural invariance testing was then conducted to see if findings were consistent across sex and countries. A longitudinal model was then conducted to examine if FAD-intoxication prospectively mediates the relationship between CABs and negative alcohol-related consequences. FAD-intoxication was concurrently and prospectively associated with greater endorsement of alcohol-related consequences and significantly mediated the effects of CABs on alcohol-related consequences. The relationship of FAD-intoxication on negative alcohol related consequences was stronger for females compared to males, and stronger in the U.S., South Africa, England, and Canada, compared to Argentina and Spain. Intervention programs for individuals engaging in FAD-intoxication may benefit from targeting CABs to reduce negative alcohol-related consequences.

Javiera Hernandez is a second year M.S. student in Psychological Sciences at William & Mary. Her research focuses on predictors and consequences of substance use among college students. Her thesis examines an online personalized normative feedback intervention for individuals who engage in food and alcohol disturbance. She holds a B.S. in Psychology from the University of Iowa.

Samantha J. Kline, Psychological Sciences

"Determining if Alcohol Use and Impulsivity Alter Cognitive Control Processes in Event-Related Potentials"

Advisor: Dr. Paul Kieffaber

Longstanding theoretical models define cognitive control (CC) as a construct that regulates behavior by suppressing impulsive responses. Neural measures also assume this, where the P3 event-related potential (ERP) amplitude is used to quantify CC. However, whether CC moderates the link between automatic responses and behavior remains untested. It is also unclear how this presumed association is altered in traits linked to low CC, such high impulsivity and heavy drinking. The present study is recruiting college participants (N=120) to complete a decision-making task while EEG activity is recorded. A combined Simon-Flanker task will observe two aspects of CC: response inhibition and interference control to clarify how CC drives multiple mechanisms at once. The Alcohol Use Disorder Identification Test will measure alcohol use and the Impulsive Behavior Short Scale to index trait impulsivity. ERPs reflecting CC and automatic response preparation will be compared to task performance (reaction time) to test the hypotheses that (1) CC (P3 amplitude) moderates the link between automatic response activation and behavior, and (2) CC is reduced in high binge drinking and impulsivity, weakening the moderation, leading to a strengthened link between automatic responses and unsuccessful inhibition (poor performance). This will be tested using a multilevel path analysis. Data collection is ongoing, and preliminary results will be available soon. Results may inform of altered CC mechanisms among non-clinical populations and confirm a longstanding assumption in cognitive neuroscience.

Samantha Kline is a second year master's student in the department of Psychological Sciences at William & Mary. Her research interests include cognitive control, substance use and impulsive behavior. She holds a B.S. in Psychology from West Chester University of Pennsylvania. 

Jackie Wo, Psychological Sciences

"The Role of Regulation in Intraindividual Dynamics of Daily Depressive Symptoms"

Advisor: Dr. Meghan Quinn

Co-Authors: Q. Liu

Regulation—the process of managing thoughts, behaviors, and emotions to achieve goals—is linked to depression, yet the complex and dynamic nature of depressive symptoms challenges simple associations between specific strategies and depressive severity. This study elucidates the role of five regulation strategies (rumination, reappraisal, problem-solving, acceptance, distraction) in the dynamics of daily depressive symptoms uncovered from multilevel hidden Markov models (mHMM). A secondary analysis will be conducted on 28-day daily diaries from 419 undergraduates. Regulation strategies were assessed at baseline using the Brief-COPE and Ruminative Response Scale. For 28 mornings, participants completed the Clinically Useful Depression Outcome Scale regarding the prior day. mHMMs will model the daily dynamics of nine depressive symptoms as probabilistic transitions between latent states. Multivariate multiple regression will link baseline regulation strategies to transition probabilities in subject-specific mHMMs. A three-state model is expected: low-depression, elevated-depression, and cognitive/physical states. We also hypothesize: rumination predicts higher persistence of the elevated-depression state; reappraisal predicts higher persistence of the low-depression state and transitions from higher to lower depression states; problem-solving and acceptance predict higher transitions from higher to lower depression states; no specific hypothesis is made for distraction. Findings will inform depression interventions on when and what regulation strategy should be delivered.

Jackie Wo is a first-year Master's student in Psychology at William & Mary. Her research focuses on the temporal dynamics of internalizing disorders and the transdiagnostic factors that shape these dynamics. Passionate about leveraging computational approaches and intensive assessments, she is exploring the role of regulation strategies in the dynamics of daily depressive symptoms uncovered from hidden Markov models.

Kelly Govain Leffel, Education

"Automotive Technicians’ Journey from Program to Practice: A Look at Self-efficacy and Longevity"

Advisor: Dr. Chris Gareis

The United States is experiencing a labor shortage in several critical industries across the country, including the automotive industry. As the baby boomer generation retires, there are not enough skilled workers to replace them. Additionally, it appears that Generation Z is looking towards opportunities in trades as artificial intelligence (AI) is leading to the elimination of office jobs and that the return on investment in college education is no longer what it used to be. However, according to surveys conducted in Virginia, only 2% of students that complete an automotive technology program in high school stay in the industry after five years. How do we ensure that we have enough automotive technicians to fill the gaps as the older generation retires, and how do we ensure those students that graduate from an automotive program stay in the industry beyond 5 years? Through mixed methods embedded case study research, I will examine the automotive technology program at an education center in Virginia using content analysis of program documents and also interviewing a selected sample of early-career automotive technicians who graduated from the program in the last five years to understand their self-efficacy in the workplace and their intent to stay in the automotive industry. I hope to determine which program-level experiences contribute to participants’ self-efficacy and their predicted longevity in the industry. 

Kelly Govain Leffel is in her final year of the Curriculum & Learning Design program in the Education, Policy, Planning & Leadership department at the School of Education. Kelly has been an educator for over 20 years and found a passion for CTE now that she owns an automotive repair shop. Her dissertation examines self-efficacy and longevity for automotive technicians. She holds a B.A. and M.Ed. from Rutgers University. 

Xiaoya Yang, Anthropology and Gender Studies | Brandeis University

"Becoming Global Elites: Labor and Aspiration of Chinese International Students in the Contemporary Time"

Advisor: Dr. Janet McIntosh

While Elon Musk defends Chinese international students in U.S. high-tech industries against the Trump administration’s mass visa cancellations, they have positioned these students respectively as either exploitable elite labor or threats to national security. Such narratives obscure the lived realities of Chinese international students who navigate their own aspirations of global elitism with everyday struggles of labor in the temporary period after graduation. My research then asks: how do Chinese international students negotiate global elite identities under restrictive immigration regimes? Drawing on participant observation, in-depth interviews, and shadowing, I analyze the embodied labor students perform to sustain legal status and daily life. My fieldwork indicates that the pursuit of global elitism motivates continuous self-making while simultaneously being unsettled by immigration policies and political dynamics across borders. I argue that global economic systems generate seductive images of elite success that compel specific forms of labor while marginalizing others, facilitating inequality and capitalistic extraction. My research contributes to anthropological debates on immigration temporality and elitism, while engaging feminist and queer theory on embodied desire and performance. By narrating students’ embodied labor during temporal immigration, the study reveals how young immigrants of China’s single-child policy generation negotiate precarious futures within current global capitalism.

Xiaoya is a second-year master student in the Anthropology and Gender Studies program at Brandeis University. The research examines the post-graduation lives of Chinese women international students in the U.S. In addition to gender, labor, immigration, and globalization, their interest include critical elite studies, neoliberalism, and Sinophone media.

Brittany Young, Economics | University of Virginia

"Closing the Gap: Compensation, Demand Shocks, and the Future of Men in Nursing"

Advisor: Dr. Emma Harrington

The United States’ nursing workforce is projected to increase by 5% from 2024-2034 with an average of 189,100 openings each year. Will this increase in demand for nursing also lead to more men entering the profession, or will men continue to make up about a tenth of nurses? I will use the National Sample Survey of Registered Nurses (NSSRN) and Bureau of Labor Statistics (BLS) data from 2018-2022 to study how the workforce responds to local shocks to nurse demand, such as from the aging of the local population. I predict that these shocks will increase nurse compensation, thereby recruiting more nurses, with particularly large effects for men who require more compensation to enter a “female” profession. After the initial demand shock wanes, I expect persistent changes in gender composition due to shifts in local norms.

Brittany Young is a second year Ph.D. student in Economics from UVA. Her research interests include health, labor, and public policy. She holds a B.A. from William & Mary and M.A. from UVA.

*Maho Inomata, History

"Islanders in Uniform: Micronesian Enlistment in the U.S. Armed Forces, Citizenship, and the Empire"

Advisor: Dr. Hiroshi Kitamura

*GRS Award for Excellence in Scholarship -- College of Arts & Sciences

This research examines U.S. military empire in the Pacific not through territorial expansion, but through the legal and social aspects of citizenship and nationality. It focuses on the origins of the Compact of Free Association (COFA) from the 1960s to the 1980s, an agreement that that enabled the Micronesian nations independence while preserving deep U.S. strategic influence. Through COFA, citizens of the Federated States of Micronesia (FSM), the Marshall Islands, and Palau may live and work in the United States without a visa and may enlist in the U.S. armed forces, while the United States retains full defense authority over these nations. Legally, COFA nationals are classified as a distinct category of lawfully present nonimmigrants. Despite this “second-class” legal status, the COFA countries have more Army recruits per capita than any U.S. state and their casualties in Iraq and Afghanistan far exceed the U.S. national average. Drawing on U.S. government records and local newspapers in the Pacific, it examines why and how the military service clause was written into COFA, who supported it, and how Micronesians responded. It highlights Micronesians’ military service—often overlooked in scholarship on the U.S. military presence in the Pacific. It also analyzes the legal singularity of COFA as both an international treaty and a statute incorporated into U.S. law, giving Congress the authority to alter COFA citizens’ rights. By centering Micronesia in the history of U.S. empire, it shows how law—rather than land— became a tool and venue of the U.S. imperial expansion.

Maho Inomata is a first-year Ph.D. student in History at William & Mary. Her research explores twentieth-century transnational history in the Pacific, especially U.S. and Japanese imperialism in Micronesia through the League of Nations and the UN. She previously received her M.A in International Relations from Waseda University after serving as a researcher at the Embassy of Japan in the Federated States of Micronesia.

Anja Keil, American Studies

"Memoria (De)Colonial – U.S. Imperialism and the Material Culture of Independence in Puerto Rico"

Advisor: Dr. Simon Stow

“People from the continental United States tend to imagine their country as a nation of states and citizens,” writes Alvita Akiboh in the introduction to Imperial Material (2023). In her work, Akiboh traces the myths and narratives that surround the shaping of the U.S. empire, amplified by the widespread circulation of national symbols such as the Stars and Stripes and other material cultures. Regardless of its territories, and the presence of U.S. bases overseas, the United States has an interest in containing the public and political image of its nation to the continental U.S. They thus enforce a demarcation between the “legitimate” nation (continental U.S.) and “second-class” nations (territories). These territories are simultaneously expected to adopt U.S. material cultures and a sense of U.S.-American nationalism without interfering or seeking to shape the United States’ political landscape. As a United States territory, Puerto Rico is continuously confronted with this tension. In my talk, I explore to what extent Puerto Rico adopts and resists the implementation of U.S. material cultures. Based on my field research in San Juan, Puerto Rico, in the fall of 2025, I will interweave photojournalism and interviews conducted with community members of the independence movement with theoretical approaches to U.S. nationalism. I highlight to what extent imperial materials are manifested in San Juan’s urban landscape and how they are continuously re-negotiated in the light of Puerto Rico’s vibrant independence movement. 

Anja Keil is a second year Ph.D. student in American Studies at William & Mary. Her research areas include migration studies, borderland studies, and states of exception. She holds a B.A. and M.A. from Münster University, Germany. 

Rafael Tamae, History

"The United States and the Guerra Grande in the Río de la Plata Basin (1839-1851)"

Advisor: Dr. Fabricio Prado

During the mid-nineteenth century, the Río de la Plata Basin in South America was a contested borderland, where the presence of foreign powers—Britain, France and the United States—intertwined with processes of state formation in Argentina, Brazil, Paraguay, and Uruguay to produce a volatile regional setting. This resulted in a number of regional wars, like the Guerra Grande (1839-1851), a Uruguayan civil war that escalated into a broader confrontation. On one side, Manoel Oribe, the Uruguayan general vying for the presidency, found the support of Juan Manuel de Rosas, leader of the Argentine Confederation. On the other, Fructuoso Rivera, a former ally of Oribe who had ousted him from the presidency, formed a coalition with the anti-Rosas Argentinian opposition. The Montevideo government also found support from Britain and France, leading to an intervention between 1845 and 1850. But what about the United States? This is the focus of this investigation. Though studies have long emphasized the British and French presence, the US was also an important actor in the region. Recent researches have highlighted American involvement in the Río de la Plata Basin, but the crucial decade of 1840 remains underexamined. Relying on the diplomatic and consular communications from US agents accredited to Argentina, Brazil, Paraguay, and Uruguay, this work’s objective is to discuss the American presence in the region and to present how US representatives dedicated themselves to foster trade and advance their government’s interest during the Guerra Grande (1839-1851).

Rafael Tamae is a first-year Ph.D. student in the Department of History at William & Mary. His research focuses on the diplomatic and political history of the Río de la Plata basin in South America during the nineteenth century, examining how nation-state building processes were characterized by the intersection of domestic and foreign policies in the. He earned his BA and MA from the University of São Paulo, Brazil.

Daniel Otten, Computer Science

"Evaluating the Impact of Generative AI on Software Engineering Education"

Advisor: Dr. Denys Poshyvanyk

Artificial Intelligence in the form of Large Language Models (LLMs) has exploded in popularity during the last 3 years, helping many Computer Science students to design programs, write code and papers, and analyze data as part of their coursework. This has raised significant concerns among parents and teachers about how these new tools might impact education, as well as how to properly integrate them into existing courses to ensure that students are prepared to use these tools in real-world jobs. In order to understand the necessary changes to the current system, we plan to answer the question “What is the relationship between AI usage and learning outcomes in software engineering class?" To accomplish this, we have obtained consent from all 36 students in W&M’s Fall 2025 semester section of CS435 to collect and analyze the data they produce as part of their normal classwork. We created 6 groups of 6 students, each with roughly the same overall programming experience and AI habits, and assigned each group the same starting codebase and assignments. By periodically surveying every student about their AI usage habits and opinions in addition to collecting their code, tests, homework, and AI log submissions, we gain consistent insight into student decisions about whether/how to use AI as part of their coursework. When the class has concluded, we will analyze the trends of AI use throughout the semester to gain insight into its effects on code quality, student efficiency, and team collaboration, allowing us to make actionable recommendations about how to improve CS education.

Danny Otten is a third year Ph.D. candidate in Computer Science at William & Mary. His research interests center around the social impacts of AI and other emerging technologies, and he is currently studying the effects of AI integration in software engineering industry and education.

Enze Xu, Computer Science

"DaisenBot: Human-AI Collaboration in GPU Performance Analysis with Multi-Modal AI Assistant"

Advisor: Dr. Yifan Sun

Co-Authors: J. Coonley, D. Xu

Graphics Processing Units (GPUs) play a critical role in accelerating applications across artificial intelligence, physical simulations, and medical imaging. Analyzing simulator-generated execution traces is essential for understanding GPU behavior and identifying performance bottlenecks, but users often face challenges interpreting complex visualizations and hierarchical data structures. To address this, we present DaisenBot, an interactive AI assistant that leverages large language models to provide accurate, context-specific answers from multimodal inputs, including text, images, simulation traces, and source code. DaisenBot helps users clarify questions, navigate relevant simulation data and subpages, and better understand and analyze simulation results, offering practical support for both novice and experienced users. 

Enze Xu received the B.S. and M.S. degrees in Computer Science from Peking University and Wake Forest University, respectively. He is currently a third-year Ph.D. student in the Department of Computer Science at William & Mary. His research focuses on GPU architecture and performance analysis.

Susan Zehra, Computer Science | Old Dominion University

"Building Trust on the Move: A Cryptographic Framework for Secure Data Exchange in Intelligent Vehicular Networks"

Advisor: Dr. Stephan Olariu

As vehicles evolve into mobile computers, future transportation systems will depend on secure data sharing and collaborative computation between cars and roadside infrastructure. Ensuring security, privacy, and accountability in such dynamic networks is a major challenge, especially as vehicles constantly join and leave. This paper introduces VERA-VANET (Verifiable Encrypted Routing and Attestation), a core algorithm that makes vehicular communication both secure and mathematically provable. It embeds cryptographic proofs of delivery and proofs of correct computation into each data exchange. When an Access Point (AP) distributes encrypted job chunks, each vehicle processes only the data it is authorized to handle. Packets are digitally signed and acknowledged, producing compact proofs that can be verified by the AP or cloud. To ensure correctness, VERA-VANET uses lightweight zero knowledge proofs that confirm computations without revealing data. Even with lost connectivity, vehicles can safely store, forward, and trace encrypted chunks through cryptographic receipts. The protocol relies on asymmetric encryption for confidentiality, aggregate signatures for authenticity, and zero knowledge proofs for verifiable computation, making it nearly impossible to forge or alter results. VERA VANET integrates with existing vehicular standards such as IEEE 802.11p and C V2X, using trusted hardware for secure key storage. Scalable and efficient, it supports real time applications, enabling private, provable, and tamper resistant collaboration across intelligent vehicular networks.

Susan Zehra is a Ph.D. candidate in Computer Science at Old Dominion University. She is a Senior Lecturer in the Computer Science Department and has over 12 years of experience teaching computer science and information technology courses. Her interests include computer architecture, intelligent systems, vehicular networks, cybersecurity, and cryptography.

Catherine Bare, Applied Science

"Correlating Calcium Dynamics with Gene Expression in the Developing Nervous System"

Advisor: Dr. Margaret Saha

Early in the development of vertebrate embryos, nerve cells exhibit spontaneous bursts of electrical activity that cause brief rises in calcium inside cells. These calcium signals can be tiny and limited to single cells or large, coordinated events that spread across many cells, and they have been implicated in fundamental developmental decisions including neural induction, progenitor cell differentiation, neurotransmitter specification, and coordinated morphogenetic movements. Yet the irregular, heterogeneous nature of embryonic calcium transients and wide variation in calcium event‑detection practices have obscured which features of calcium dynamics are reliably decoded into transcriptional gene expression responses. Utilizing an amphibian (X. laevis) neural plate model, this project seeks to bridge that gap by directly linking quantitative calcium activity phenotypes to gene expression programs during the neural plate stage. By comparing activity features across developing embryos during controlled perturbations, we identify genes and coexpressed gene modules whose expression varies with distinct activity patterns. We characterize how detection choices influence the discovery of activity‑associated transcripts, clarifying which candidate genes are robust to analytical variation and which require targeted validation. The outcome will be a defined set of activity-regulated genes and modules that shed light on how calcium dynamics influence early neural fate and tissue patterning and will provide a clear foundation for functional follow‑up in vivo.

Catherine Bare is a first year M.S. candidate in the Applied Science Department at William & Mary. Her research focuses on the bioinformatics and molecular analysis of neuronal cell fate, especially related to patterns of calcium activity in the developing embryo. She holds a B.S. also from W&M in Neuroscience and has a broad interest in nervous system abnormalities throughout the lifespan.

Eva Kalajian, Applied Science

"µ-Opioid Receptor Activation Increases Kir3 Activity in Glutamatergic preBötC Neurons via Membrane-Delimited Signaling"

Advisor: Dr. Christopher Del Negro

Co-Authors: M. Stettler, G. Smith

Opioid drugs like oxycodone and fentanyl reduce pain and produce euphoria but also stop breathing, which is the leading cause of opioid overdose death. Opioids act on μ-opioid receptors (μORs) in a key brainstem site that generates breathing rhythm, the preBötzinger Complex (preBötC). It’s unclear which types of neurons in this area are affected by opioids and which ion channels are involved. Using fluorescent labeling (in situ hybridization), we found that most excitatory and inhibitory preBötC neurons express μOR mRNA, suggesting they all may respond to opioids. Using a preparation that keeps brain tissue alive, we found that only excitatory neurons were opioid-responsive; the μOR drug DAMGO increased potassium channel activity, and the effect was precluded by barium, pointing to a specific class of channels called Kir3. On-cell single Kir3-channel activity increased when DAMGO was inside the recording pipette, but not in the bath solution, suggesting that DAMGO’s effect occurs via membrane-delimited signaling. Despite having μOR mRNA, inhibitory neurons did not respond to DAMGO, possibly due to post-transcriptional modifications such as mRNA degradation or relocalization from the cell body. By clarifying opioid targets in the preBötC, there is potential for new therapies that act on molecules other than the μOR itself, preventing opioid-induced respiratory depression without compromising clinical pain relief or causing renarcotization, when residual opioids in the body reactivate the µOR after conventional therapies like Narcan wear off.

Eva Kalajian is a 4th-year Ph.D. candidate in Applied Science at William & Mary, specializing in respiratory neurobiology and ion channel biophysics. Her dissertation addresses how opioid signaling causes fatal apnea. She holds a B.S. in Neuroscience and an M.S. in Applied Science from William & Mary. She graduates in May and will pursue science communication to translate complex research and enhance public understanding.

Diego Morandi Zerpa, Biotechnology | Brandeis University

"Hippocampal-Cortical Coordination is Altered During Sleep in a Rat Model of Fragile X Syndrome"

Advisor: Shantanu Jadhav

Co-Authors: M. Ding, J. D. Shin

Sleep disturbances affect up to 80% of children with autism spectrum disorder (ASD). While these deficits have been extensively documented, the underlying sleep neurophysiological changes remain largely unknown, limiting our ability to develop targeted interventions. Human investigations are limited to EEG recordings, but rodents provide a tractable mammalian model to probe cellular and network level neuronal activity, since key sleep rhythms, such as cortical slow oscillations (SO), thalamocortical spindles, and hippocampal sharp-wave ripples (SWR), are conserved. Synchronization of these rhythms/oscillations during non-rapid eye movement sleep supports memory consolidation, and this multi-region network coordination supports sleep stability and continuity. Thus, to test if sleep deficits in ASD are associated with dysfunctional circuit coordination, we conducted high-density electrophysiological recordings in the hippocampus (CA1 area) and prefrontal cortex (PFC) of wild-type (WT), Fmr1 knockout (Fmr1), and Shank3 knockout (Shank3) rats during sleep. Fmr1 rats, a model of Fragile X syndrome, displayed shorter sleep bouts, more micro-arousals, and aberrant coupling between SOs and spindles compared to WT rats. Single neuron activity of Fmr1 rats in both regions showed reduced modulation by SOs, and co-activation of hippocampal and PFC neurons was also diminished. With the future incorporation of Shank3 recordings into our dataset, these findings can elucidate mechanisms for disordered sleep in ASD and identify biomarkers to guide the development of effective interventions.

Diego Morandi Zerpa is a second year M.S. student in Biotechnology at Brandeis University. In the Jadhav Lab at Brandeis, he conducts research on the systems neuroscience of sleep, cognitive function, and memory consolidation. He is currently studying sleep architecture and brain circuit coordination in rat models of autism spectrum disorder. An alum of William & Mary, Diego is excited to return to present his research!

Sarah Sakly, Applied Science

"Phase separation dynamics of SynGAP & PSD-95 in post-synaptic densities"

Advisor: Dr. Greg Conradi Smith

The post-synaptic density (PSD), is a protein rich structure that organizes receptions and signaling molecules. SynGAP is a protein found within the density, and interactions with a scaffolding protein PSD-95. Mutations in SynGAP reduce interactions with PSD-95 and are strongly associated with intellectual disability, epilepsy, and autism. To investigate possible mechanisms of the SynGAP mutation, we develop and analyze a model including three species, SynGAP, PSD-95, and a solvent, to simulate how SynGAP and PSD-95 may self-organize through liquid-liquid phase separation. Based on Flory-Huggins energetics proposed by Lin et al. [Biophys. J. 121:1, 2022], we develop a Cahn-Hillard system representing the total concentrations of SynGAP and PSD-95 under a rapid equilibrium. This model utilizes partial differential equations to track how the concentrations of both proteins evolve over space and time. Numerical simulations reveal a favorable interaction between SynGAP and PSD-95, driving the formation of a PSD-95 condensate resembling condensates observed experimentally. Furthermore, our model reproduces the experimentally observed effect of long-term potetniation, a form of synaptic strengthening, in which SynGAP escapes from the post-synaptic density, providing a physical framework for understanding how mutations in this key synaptic protein can lead to altered brain signaling and neurological diseases.

Sarah Sakly is a first-year Ph.D. candidate in Applied Science at William & Mary. Her research applies principles from polymer physics to study the post-synaptic density, with a focus on the intrinsically disordered protein SynGAP. Using computational and theoretical approaches, she investigates how multivalency, phase separation, and molecular interactions govern SynGAP organization and dynamics.

Claudia Garcia Mendoza, American Studies

"Severed Selves and Emotional Data"

Advisor: Dr. Elizabeth Losh

The Apple TV+ series Severance depicts Lumon Industries as an absurd and opaque workplace, where the purpose of production is as unclear to the viewer as it is to the employees themselves. At Lumon, employees undergo a surgical procedure that severs their memory between their personal and their professional lives, producing split identities. 

The show centers on the Macrodata Refinement department (MDR), where employees must cluster and identify threatening numbers based on their affective intuition. The nonsensical nature of this labor not only obscures its purpose but also invites us to consider how, under contemporary regimes of surveillance capitalism, emotion is not only managed, but it is extracted and commodified. The show’s refusal to clarify the purpose of MDR mirrors the real-world algorithmic systems and data-driven management practices increasingly dependent on affective relationships. Drawing on affect theory and surveillance studies literature, this paper employs critical media analysis to argue that Severance exposes how the relationship between bodies and data is increasingly emotional and suggests that the managing of labor is not only about productivity but also about shaping emotional subjects. 

Claudia Garcia Mendoza is a Ph.D. candidate in the American Studies program at William & Mary. Her research focuses on different approaches to technology, including media representation, counternarratives in digital spaces, and affective technology. Her work primarily draws from media studies, science and technology studies, disability studies, and cultural studies. 

Teddy Hogerhuis, American Studies

"The Village People’s ‘Native American’ Member: Nationalism, Resistance, and Black, Brown, & Queer Performance"

Advisor: Dr. Elizabeth Losh

The Village People’s place in queer cultural memory is rife with tension between American nationalism and gay pride, both deeply entangled with race. Emerging from 1970s disco subculture, the group adapted sexualized, macho American archetypal cosplays from gay nightlife into costumes. In the original lineup, Black uniformed officers stood alongside working class garb, a cowboy, and the “Native American” archetype embodied by Felipe Rose, a gay man who claims Lakota, Taíno/Puerto Rican, and Two Spirit identities. This project asks how the figure of Felipe Rose is represented and contested in popular media discourse. I close read depictions of Rose in the decades following the group's commercial peak, juxtaposing a portrayal of Rose on an episode of That 70’s Show against his contemporary solo performances and digital presence. I situate Rose among two contexts: the US history of suppressing queer Black, brown, and Indigenous performance—from Ghost Dance (1890) to discophobia (late 70s-80s)—and contemporary resistance, particularly two spirit perspectives on nation, from the “Two Spirit Nation” which occupied Standing Rock during the Dakota Access Pipeline protests, to an Instagram page dedicated to two-spirit Taíno activism. Foremost amongst the theories of nationalism I draw on is Jasbir Puar’s theorization of queer and racialized identities to examine how tensions around multiplicity service the American nationalist project. Ultimately, this study finds that resistance lies in rejecting these imposed tensions to instead encourage a plural national identity.

Teddy Hogerhuis is a first year MA/PhD student in American Studies at William & Mary. They research contemporary popular culture and transformative works. Their current project examines how the disco group Village People reflects and narrates gay American cultural history through its production, reception, and reinterpretation over time. They hold a BA in American Studies from CSU Fullerton.

Maryam Iftikhar, Genocide and Mass Atrocity Prevention | Binghamton University

"The Chains We Forge: Complicity as a Tool of Oppression"

Advisor: Dr. Kerry Whigham 

What makes ordinary people complicit in systems of extraordinary violence? This presentation examines the architecture of oppression through an interdisciplinary lens, expanding on critical theory and atrocity prevention scholarship. Using the video game character Sylas from League of Legends as a case study, I argue that effective systems of violence rely not only on coercion but on moral frameworks that recruit the oppressed to enforce their own subjugation. Sylas’s transformation from mage-hunter to revolutionary reveals the paradox of moral responsibility and agency under oppression. His story exemplifies weaponized complicity, a process through which oppression reframes self-betrayal as virtue. This unfolds through four stages: (1) pathologizing difference, (2) selective recruitment through conditional inclusion, (3) cultivating grateful servitude via moral duty, and (4) strategic expendability once compliance loses value. This framework synthesizes Arendt’s banality of evil, Fanon’s internalized oppression, and Gramsci’s hegemonic consent, showing how virtue itself becomes weaponized to sustain tyranny. Beyond fiction, similar dynamics appear when colonial powers, state agencies, or digital platforms incentivize marginalized actors to police their own communities. In an age of democratic backsliding and technological amplification of coercion, understanding weaponized complicity is not only theoretical, but essential to resisting co-optation and strengthening collective solidarity.

Maryam Iftikhar is a first-year MS student in the Institute for Genocide and Mass Atrocity Prevention at Binghamton University. Her interdisciplinary work explores how identities and institutions shape the contours of complicity, resistance, and justice in contexts of mass violence. As a first-gen scholar, she is committed to elevating historically underrepresented voices. She holds a B.A. from Hood College in Maryland.

David Choate, American Studies

"How to Read Zé Carioca: Disney’s Cartoon Parrot as Inter-American Ambassador and National Icon "

Advisor: Dr. Charles McGovern

From 1942 to 1945, Disney cooperated with the Office for Inter-American Affairs to produce shorts and films promoting inter-American cooperation during World War II, part of a propaganda effort known as the Good Neighbor policy. The most successful creation of Disney’s Good Neighbor period was the character José Carioca, a Brazilian cartoon parrot modeled after the malandro stereotype of Rio hustlers; the anthropomorphic bird proved a brief hit in the United States before establishing durable popularity in Brazil in his comic book Zé Carioca. This talk explores the evolving symbology of José Carioca, tracing the development of the character from an American portrayal of Brazil, embodying a mix of ethnographic research and stereotype, to an iconographic stand-in for Brazilians’ perception of themselves. The talk will first examine the character’s initial appearance in Good Neighbor omnibuses like The Three Caballeros, and his run in a U.S. newspaper strip from 1942-1944. Subsequently it will look at the character’s use and eventual reclamation in Brazilian Disney comics, typified by the character’s reinvention by Renato Cavini and Ivan Saidenberg in the 1970s as a more authentically Brazilian character. Zé Carioca moved from symbol of U.S. propaganda in Latin America, pace Ariel Dorfman and Armand Mattelart in their classic critique How to Read Donald Duck, to protest subject of Brazilian installation artist Rivane Neuenschwander, who criticized the character as epitomizing Brazilian nationalism; the central question of this talk is how such a shift occurred. 

David Choate is a third year MA/Ph.D. student in American Studies at William & Mary. His primary research area is American comic books and strips of the the early twentieth century through the early post-WWII era.

Luma Mousa, Cultural Studies | George Mason University

"Writing on the West Bank Wall: Visual Culture of Resistance, Commodification, and Imaginaries of the Future"

Advisor: Dr. Alison Landsberg 

This paper examines the Israeli West Bank Wall as both a physical divider and a symbolic canvas, exploring the graffiti on the Wall as a site of resistance and global solidarity. By analyzing the interplay of local and global representations, this paper investigates the dual role of graffiti as a tool for both solidarity and imagination. Drawing on Rancière’s The Politics of Aesthetics and scholarship on photography and visual culture, this paper explores how the Wall’s graffiti constructs an alternative narrative that challenges its intended role as a divider. I ask in this paper: How do the universal themes expressed in the graffiti, such as solidarity with global struggles, critiques of oppression, and calls for justice, transform the Wall into a site of collective imagination? Secondly, this paper considers the commodification of the Wall through graffiti tourism and workshops, which complicates its role as a site of resistance. Does the graffiti on the Wall resist the oppressive function, or does it risk normalizing the Wall’s presence through commodification? I propose in the paper that while the Wall itself embodies a dystopian present for Palestinians, the graffiti challenges its existence by projecting visions of hope and alternative futures. Ultimately, this paper highlights how cultural productions on the Wall speak to both the current dystopian realities of the Israeli occupation and the utopian aspirations for a more just future, locating the West Bank Wall’s graffiti within broader debates on decolonization. 

Luma Mousa is a doctoral candidate in Cultural Studies at George Mason University. Currently, her research focuses on visual culture and social media, memory studies, and Middle Eastern studies, with an emphasis on Palestinian resistance studies.

Jarred Loughran, Physics

"Effects of Magnetic Perturbations on Plasma Particle Transport"

Advisor: Dr. Saskia Mordijck

Fusion reactor performance is highly dependent on radial density and temperature profiles. Operating with a “pedestal” region where the density and temperature gradients are steep in a narrow edge region of the plasma is a promising way to maximize performance for a fixed machine size. The shape of the density pedestal is set by the balance between transport flows and the particle ionization source. In high-performance operation, the pedestal is susceptible to abrupt edge instabilities which lead to spikes in the heat flowing to the wall, and for high energy density devices these heat loads will be intolerable. The leading approach for mitigating these instabilities is by applying small perturbations in the magnetic field which break the symmetry of the system and enhance outwards transport. Quantifying how transport changes when these perturbations are applied is an open question, which we investigate by analyzing particle transport in three experiments by utilizing a time-dependent forward modeling framework with Bayesian inference to optimize for transport flows which return a density evolution that best matches experimental data. We find that not only is there a substantial difference in transport between cases with and without magnetic perturbations, but that the higher degree of perturbation necessary to eliminate the pedestal instabilities come with enhanced transport much further inwards than expected. This has implications for performance predictions of future devices, which typically do not consider these factors due to a lack of information.

Jarred Loughran is a fifth year Ph.D. candidate in the Physics Department at William & Mary. His research focuses on exploring particle flows and fueling in magnetically confined plasma through analysis of experimental data, with a secondary passion for science communication. There's no such thing as a stupid question! 

Doyle S. Weishar, Materials Science & Engineering | University of Virginia

"Tungsten Degradation Mechanisms in Molecular and Atomic Oxygen at Elevated Temperatures"

Advisor: Dr. Elizabeth Opila

Co-Authors: S. Hartman, C. Stephens

Fusion power is of great interest as a solution to ever-increasing power demands. The extreme conditions experienced by plasma facing structural elements preclude many standard materials from use in their design. Tungsten has been identified as a leading candidate for these environments, but the performance of tungsten in potential failure modes in which oxygen is introduced during operation is not well understood. In such an event, the plasma facing elements are expected to remain at elevated temperatures for sufficient time to react with atomic oxygen produced via interactions with the plasma. Understanding the degradation mechanisms in this failure mode are critical to designing safe power generation systems. In this work, tungsten coupons were exposed to both atomic and molecular oxygen at temperatures of 950°C and 1200°C, above and below the oxide volatilization temperature, for times up to 10 minutes. Oxidized samples were characterized by mass change, material recession, X-ray diffraction, scanning electron microscopy, and energy dispersive spectroscopy to probe the impact of atomic oxygen on the degradation mechanisms of tungsten.

Doyle S. Weishar is a third year Ph.D. candidate in Materials Science & Engineering at the University of Virginia. His research areas include materials for extreme environments, ultra-high temperature materials, plasma environments & aerospace structural materials. His current research is exploring the impact of atomic oxygen on tungsten at high temperatures. He holds a B.S. from W&M and an M.M.S.E. from UVA.

Chester Zimmerman, Physics

"Interplay of Thermal and Laser-Driven Effects in Relativistic Transparency"

Advisor: Dr. David Stark

When a strong laser pulse hits a material, it can become a plasma. There exists a density of the material known as the critical density at which the pulse will be totally reflected at the boundary instead of transmitted through. High intensity lasers capable of accelerating electrons to near-light speeds alter the optical properties of a plasma, including the critical density, through the phenomenon known as relativistic transparency. A particle is considered to be relativistic when it is approaching the speed of light, so relativistic transparency refers to the case where the near-light speed of the electrons causes the critical density to increase beyond classical limits, making the plasma transparent. This phenomenon critically influences the laser-plasma energy coupling efficiency and thus must be taken into consideration when choosing parameters for various applications, such as particle acceleration and cancer treatment. This study looks to find these modified critical densities' dependence on temperature and laser intensity using simulation software known as particle-in-cell code and compare them with mathematical predictions. We find that our theory of the critical density's temperature dependence matches simulation at low intensities, but we expect there to be another correction when we consider higher laser intensities, which we hope to model using what is known as relativistic kinetic theory, a theory in plasma physics that predicts the behavior of individual particles in a plasma moving at relativistic speeds.

Chester Zimmerman is a fourth year Ph.D. candidate in Physics at William & Mary. He is interested in studying the physics of relativistic plasmas and applications of laser-plasma interactions to fields like inertial confinement fusion and particle acceleration. He is currently researching the laser intensity dependence on the relativistic transparency threshold in high-temperature plasmas using particle-in-cell code.