In quantitative studies comparing a treatment and a control group, treatment effect is often viewed simply as the difference in group means. However, any treatment can have an impact beyond simply shifting the mean outcome. In this work, we consider a linear treatment effect (LTE) model, meaning we simultaneously consider the difference in means and the ratio of standard deviations between two populations to better characterize the effect of the treatment. Estimation is done using an empirical likelihood (EL) formulation. The EL framework provides a nonparametric approach for conducting inference without making strong assumptions about the underlying population model. Generally, the EL statistic has a limiting chi-square distribution. However, in small sample settings, the EL statistic can exhibit strong deviations from this ideal. To address this issue, we investigate the use of the Bartlett correction, which is a multiplicative adjustment to the EL statistic to improve the chi-square approximation. This correction has been shown to substantially improve confidence region coverage accuracy, especially for small and moderate sample sizes. Through simulation, we examine the performance of the EL statistic in the LTE model, with and without a Bartlett correction applied. Our results demonstrate that the Bartlett-corrected EL approach provides improved performance, yielding confidence regions with coverage closer to desired nominal levels.

In this thesis, we introduce a way to implement Stochastic Processes - particularly Markov chain properties - for analyzing Liar’s Poker, a variant of Poker Texas Hold’Em that incorporates hidden information and a card-switching mechanic. Poker, and in particular Liar's Poker, presents a complex environment in which probabilities evolve as information is revealed and players make sequential decisions under uncertainty, so Markov modeling of this game requires a more flexible state-based representation. The study focuses on two main objectives: first, to construct a state space and transition matrix that are sufficiently compact for analysis while still capturing meaningful changes in hand-strength and game dynamic; and second, to investigate how the game’s exclusive card-switching feature can be incorporated into an optimal decision-making strategy. To address these goals, the thesis models gameplay as a sequence of probabilistic state transitions driven by card draws, hidden information, and strategic actions. By extending Stochastic Process methods to a poker setting with imperfect information and dynamic transition, this thesis aims to provide a structured mathematical framework for evaluating strategy in Liar’s Poker.

Incurred But Not Reported (IBNR) reserves refer to insurance claims that have already taken place, but have not yet been reported to the insurance provider. This presentation formulates a Bayesian modeling framework to estimate the IBNR reserves. The Bayesian framework allows us to incorporate prior knowledge, typically available from historical data and expert opinions, along with the observed claim data, to estimate model parameters and predict future claim liabilities. We emphasize prior models that have heavy tails and therefore can accommodate extreme, rare losses that can be underestimated otherwise. Specifically, we consider Pareto (Type I) and log-t models for the expected ultimate claim amounts for each insurance period. The data generating mechanisms considered are Poisson, negative binomial, and gamma. The analysis of real data also considers model sensitivity to the choice of the prior parameters. In doing so, we aim to produce more robust reserve estimates and better reflect the uncertainty inherent in unpaid claim liabilities. Ultimately, modeling IBNR reserves is important because it ensures insurance companies set aside sufficient funds to cover future claim obligations and avoid unexpected losses that could impact profitability.

Background: Chronic stress among older adults increases risk for depression, anxiety, cardiovascular disease, and cognitive decline. Nature-based interventions may improve psychosocial and physiological stress outcomes, though longitudinal evidence in aging populations remains limited.

Methods: A convenience sample of community-dwelling older adults (N = 21; M age = 74.14 ± 4.59 years; range 65–85) participated in a six-week Nature Rx intervention study conducted in partnership with the Fort Worth Botanic Garden and Texas Christian University. The program included three two-week modules meeting twice weekly: garden yoga, forest bathing, and vegetable gardening. Repeated measures for the Well-Being/Personal Health Index (WPHI), positive and negative affect, nature-relatedness, outdoor activity minutes were assessed at baseline and at three subsequent time points across the program. Measures for handgrip strength (HGS), and salivary cortisol were assessed at the beginning of the first and sixth weeks, respectively. Friedman tests with Bonferroni-adjusted post hoc comparisons evaluated changes in psychosocial outcomes. Paired t-tests and repeated-measures ANOVA assessed physiological outcomes. Analyses were conducted using available cases due to incomplete measurements across time points.

Results: WPHI scores did not significantly change over time, χ²(3, n = 18) = 1.886, p = .596. Positive affect significantly increased, χ²(3, N = 18) = 13.437, p = .004, with higher final scores compared with baseline (padj = .012) and Post 2 (padj = .018). Negative affect also showed a significant overall time effect, χ²(3, N = 18) = 11.131, p = .011, though pairwise differences were not significant after adjustment. Nature-relatedness and outdoor minutes did not change (all p > .05). HGS remained stable, t(18) = −1.08, p = .294, and strength classification did not significantly change (Wilcoxon W = 3.00, p = .157). Salivary cortisol significantly decreased from pre- to post-intervention, t(21) = 7.653, p < .001 (d = 1.63); ANOVA confirmed a significant condition effect, F(1, 12) = 33.09, p < .001, ηp² = .734.

Conclusion: The intervention was associated with increased positive affect and statistically significant reductions in physiological stress, despite minimal changes in global well-being or muscular strength. These findings suggest short-term nature-based programs reduce stress burden in older adults. Larger controlled studies are needed to confirm these preliminary results.

The transcription factor, Nuclear factor erythroid 2-related factor 2 (NRF2), functions by activating genes that help protect the body against oxidative stress, inflammation, and various toxins. Thus, identification of small molecules that can increase NRF2 activity could be helpful to increase the body’s natural defense system against chronic disease. The goal of this interdisciplinary project is to use cell lines generated by the Montgomery lab (Nutrition) that express a fluorescent NRF2 reporter to test a small library of novel compounds generated by the Green lab (Chemistry) for their NRF2 activation capacity. First, our reporter system will be validated with known NRF2 activators. We will then use a luciferase reporter assay to screen 15 novel compounds for their capacity to activate NRF2 compared to the known standards. These data can then be used to inform both labs about their antioxidant capacity and help optimize their furthered development and utility.

Prebiotic sodas are marketed as healthy alternatives to traditional soda, but these claims have not yet been substantiated by research. This study evaluated the effects of fasted consumption of the prebiotic sodas Olipop and Poppi, compared with Diet Coke and Coca-Cola Original, on blood glucose, insulin, glucagon-like-peptide-1 (GLP-1), satiety, gastrointestinal symptoms, and beverage preference. A single-blind, repeated-measures design was employed with 10 participants. Participants completed four randomly assigned trials with a one-week washout period between each. During each visit, blood samples and satiety questionnaires were collected at baseline and throughout a two-hour trial. Beverage preference was assessed post-consumption, and gastrointestinal symptoms were evaluated using a follow-up questionnaire 24h post-intervention. The results from this study are expected to be completed by mid-April (by SRS).

The overall goal of our study is to understand how excess adiposity in women with and without
confounding cardiometabolic risk factors influences breast cancer cell growth and oxidative stress
signaling. I have already collected preliminary data indicating that activity of the antioxidant response
gene, NRF2, and expression of NRF2 targets are decreased in serum from obese subject, regardless of
phenotype. We investigated the functional consequences of these responses
by measuring and quantifying differences in reactive oxygen species (ROS) production. We also
investigated if these changes could lead to changes in breast cancer cell growth. To
investigate this, MCF7 breast cancer cells was grown in 6 distinct treatment groups reflecting varied
human metabolic health: CON (healthy control), NWO (normal weight obese), MUO (metabolically
unhealthy obese), and MHO (metabolically healthy obese), alongside the standard fetal bovine serum-
containing media a negative control. Reactive oxygen species production was assessed using a reagent
that fluoresces when it becomes oxidized by ROS. We expect cells grown in serum from obese subjects
will have higher levels of ROS production and increased invasive capacity. However, the results have yet
to be processed as of Mar 6. This research could demonstrate how total systemic metabolic health
influences oxidative stress responses and invasive potential, linking gene expression to real functional
outcomes. These insights could heavily inform medical assessments.

Oral health is a critical component of overall well-being; however, many children in underserved communities lack access to dental health education and essential hygiene resources. Early oral health education is vital in establishing lifelong preventive habits and reducing the risk of future dental complications. The New Smiles initiative is a student-led outreach program designed to improve oral hygiene awareness and access to dental care resources among elementary school students in the Fort Worth community.
Through interactive presentations delivered to local elementary schools, the program teaches students the importance of proper brushing and flossing techniques, healthy dietary habits, and routine dental care. To reinforce these lessons, hygiene kits containing toothbrushes, toothpaste, floss, and educational materials were assembled in collaboration with CookChildren’s and distributed to participating students. Additionally, a brief survey was administered to assess students’ baseline knowledge of oral hygiene and evaluate the effectiveness of the educational presentation.
By combining hands-on education, community partnerships, and the distribution of essential hygiene supplies, the New Smiles program aims to promote preventive oral health practices at an early age. This initiative seeks to reduce oral health disparities while empowering children with the knowledge and resources needed to maintain lifelong dental health.

Women who are obese have a much higher risk of being diagnosed with breast cancer than women who maintain a healthy body weight. However, excess body fat, even in the absence of excess body weight, a condition referred to as normal weight obesity also increases breast cancer risk. The goal of our study is to determine how serum from human subjects with three distinct obesity phenotypes, metabolically healthy obese, metabolically unhealthy obese, and normal-weight obese, influences breast cancer cell growth and proliferation. We have already collected preliminary data indicating differences in cell viability via NADH measurement, yet metabolic activity alone does not definitively demonstrate growth or vitality because cells may be metabolically active without entering S-phase or replicating. To conclusively show DNA replication (and thus true proliferation/vitality), our plan is to quantitatively measure differences in DNA synthesis using the Click-iT EdU DNA-synthesis assay, which uses a thymidine analog incorporated into newly synthesized DNA which can be detected by the appearance of fluorescent conjugates. Based on our preliminary findings, we expect that the lower rates of metabolic activity in cells grown in serum from obese subjects are not due to reduced rates of cellular proliferation. These findings could be used to inform improved, targeted nutritional and chemotherapeutic strategies for individuals with distinct obesity phenotypes.

Oncolytic viruses, which preferentially target cancer cells and stimulate the immune response, offer a promising avenue for cancer therapy. In the manuscript by Bourgeois et al. (2016), the researchers test a novel oncolytic vesicular stomatitis virus (VSV) strain that induces interferon-γ (IFN-γ) production for its potential as an oncolytic virus. In this study, we utilize an ordinary differential equation model to parameterize the dynamics of viral infection, immune response production (IFN-γ), and tumor growth. Using data extracted from key figures in Bourgeois et al. (2016), we estimate model parameters such as viral titer, IFN-γ levels, and tumor growth by minimizing the sum of squared residuals (SSR). In addition, we compare the dynamics of this novel VSV strain to a control virus, identifying key parameters that maximize tumor elimination. This model provides insights into the therapeutic potential of oncolytic VSV and helps inform strategies for maximizing its efficacy in cancer treatment.

Our Milky Way’s neighbor, the Large Magellanic Cloud (LMC), is a galaxy significantly shaped by powerful explosions from massive, dying stars that drive gas outflows. These explosions release gas and heavy elements, enriching the galaxy's outskirts and contributing to the formation of stars and planets. Understanding these processes is crucial for studying galactic evolution and the mechanisms that drive it. Our research uses observations from the Hubble Space Telescope to characterize the properties of the outflows from the LMC. Our observations are of light from background stars that pass through the LMC’s gas clouds. These clouds block some of the incoming light, and we analyze the missing features to study the physical properties of the outflows. To compare complex stellar spectra on a similar scale, we fit regions of the light that are free from major features blocking it with a best-fit polynomial. This process helps us differentiate components that either belong to the background star or the LMC’s outflowing gas. By examining the missing light, we gain a deeper understanding of how bursts of star formation impact the galactic environment and ultimately connect our existence to the explosive deaths of distant stars.

Zinc oxide (ZnO) is a versatile, inexpensive semiconductor material with unique characteristics. ZnO is particularly known for its inhibitory effects on bacterial growth. ZnO can reduce bacterial growth through mechanisms such as oxidative stress, the deterioration of crucial proteins in the bacterial cell, and the release of Zn²⁺ ions that affect bacterial cell function. The exact mechanism behind ZnO’s antibacterial properties remains unclear. It has been seen that changing the surface and morphology of the particles changes their effectiveness for bacterial inhibition. An additional lesser explored branch of ethanol-based synthesis is solution pH pertaining to ZnO morphology. Our research aims to explore this by doing a wholistic investigation of an ethanol-based synthesis, especially pertaining to how pH affects particle morphology. To produce these materials, we used ethanol-based solvothermal synthesis to create ZnO micro- and nanocrystals. We performed a thorough characterization of these materials to observe changes to the ZnO lattice. This was done by employing scanning electron microscopy (SEM), energy-dispersive X-ray (EDX) spectroscopy, and X-ray diffraction (XRD) spectroscopy.

Graphene quantum dots GQDs possess broad potential in bioimaging and optoelectronics due to their unique optical properties, tunable structure, aqueous solubility, and minimal in vivo and in vitro toxicity. However, despite their solubility, GQD fluorescence may be quenched through interactions with water molecules and aggregation via non radiative decay pathways that reduce emission efficiency. Inspired by the ability of surfactants to prevent quenching interactions for single walled carbon nanotubes, we investigate their utility in preserving GQD fluorescence. Five structurally distinct surfactants, sodium dodecyl sulfate SDS, sodium dodecylbenzene sulfonate SDBS, sodium deoxycholate SDC, sodium cholate SC, and Pluronic F127, are tested across a range of concentrations for preserving fluorescence of top down and bottom up synthesized GQDs to determine optimal conditions. This work reveals that surfactant structure and concentration can non-linearly affect GQD emission in the visible and near-infrared, with SC and SDC providing maximum concentration dependent fluorescence increase. Zeta potential and dynamic light scattering measurements are conducted for each surfactant and GQD system to quantify interfacial charge, colloidal stability, and aggregate size distributions. The present study provides mechanistic understanding of how surfactants influence GQD photophysics, offering strategies to optimize GQD based probes for biomedical imaging and photonic applications establishing a structure-to-function framework that links solution phase organization to fluorescence emission.

Graphene quantum dots (GQDs) have gained significant attention due to their unique optical properties, biocompatibility, and potential applications in bioimaging, biosensing, and optoelectronics. The breakdown of single-walled carbon nanotubes provides an alternative method of producing GQDs that has the potential to be more efficient than current methods. We will investigate the effectiveness of various methods to break down single-walled carbon nanotubes, including through UV-light irradiation. Solutions of carbon nanotubes with sodium hypochlorite are placed under 254nm UV-light for two hours, and fluorescence in the visible spectrum is measured before and after UV-light irradiation to observe the production of GQDs. The use of surfactants in these solutions can affect the resulting fluorescence, so solutions of sodium dodecyl sulfate (SDS) and sodium dodecylbenzene sulfonate (SDBS) are also UV-light irradiated and observed. We will perform transmission electron microscopy (TEM) analysis on the samples to characterize the resulting GQDs and determine their size distribution. The findings from this study will contribute to the broader scientific community by improving an avenue of production for GQDs through conversion of carbon nanotubes into smaller, more functional materials while reducing the toxicity associated with carbon nanotubes.

Defective interfering particles (DIPs) are virions missing the viral genome that allows them to replicate on their own, so they require coinfection with a standard virion to enable replication, interfering with the production of standard virus in the process. DIPs may also stimulate an interferon (IFN) response that further suppresses standard virus replication. Our aim was to evaluate the impact of DIPs and IFN on viral replication. We used Python programming to simulate a mathematical model evaluating the effects of DIPs and IFN on viral replication. Features of the viral titer curve were measured, including peak viral load and area under the viral curve, as functions of IFN parameters and DIP production rates. We examined a range of parameter values for DIP production rate and IFN response strength to assess the effects of DIPs and IFN independently and together. DIP production rate over a range of values resulted in no change in DIP or standard virus population dynamics. However, decreased IFN response resulted in an increase in standard virus and DIP population, while increased IFN response resulted in decreased standard virus and DIP population. DIP production in isolation did not impact viral replication, while IFN demonstrated an inverse relationship to viral replication and DIP production. Increased IFN and DIP production rate led to a reduction in infection intensity. IFN is essential to the antiviral effects of DIPs.

Galaxy simulations are an effective way to study the evolution of galaxies across
cosmic time. They have provided insights into the structural and chemical evolution
of galaxies, gas and star formation, and how LCDM models predict the large scale
structure of universe. Nevertheless, two primary issues have persisted using LCDM -
the core-cusp problem and the diversity of rotation curves for dwarf galaxies of similar
masses. To determine the effect of AGN on these issues, we utilize FIRE-2, which only
includes stellar feedback. We chose this particular galaxy at redshift 0 and compared
the curve to 8 previous observations, and we find that the innermost regions of the
curve are better matched to the data, but diversity still remains a problem. Thus, we
conclude that AGN feedback prescriptions may be removing too much mass from the
center of the galaxy, causing this discrepancy. Hence, more work is necessary to identify
the cause of this issue and potentially resolve it.

Influenza virus causes periodic pandemics and thousands of deaths annually, but many of the details of the viral replication cycle are still poorly understood. This study develops a mathematical model of the dynamic transitions of a virus from the extracellular space through the initial intracellular replication processes. These stages include: binding, endocytosis, HA Acidification, Fusion, and Uncoating. Experimental data from the viral entry phases were fit to a system of differential equations, which represent the biological processes. The model parameters were estimated using optimization techniques that minimize the sum of squared residuals, thereby aligning model predictions with observations. An identifiability analysis was performed to see which parameters can be estimated with the given model and available data. We find that the model fits the experimental data well with identifiable parameters, allowing us to characterize the different stages of viral entry. The model can be used to compare different viral strains or treatment options, in addition to helping explain the kinetics of viral entry.

Syncytia are a type of multinucleated cell that can be formed by virus infection. Quantifying their growth is of particular interest for understanding virus infection within the body. One useful tool we have to understand the growth of these cells is ordinary differential equation (ODE) models. Current models neglect the regeneration of cells that form the syncytia. To account for regeneration, we will discuss a proposed modification of a basic model for cell-cell fusion, which will consider the addition of a logistic growth term. In addition, we will also consider a non-negligible death rate of syncytia. By making these modifications, we can better replicate syncytia dynamics. We present mathematical analysis of this model, which gives insight into the factors that generate long-term syncytia formation as well as the overall biological characteristics of such an infection.

Graphene quantum dots (GQDs) are spherical nanoparticles comprised of stacked layers of graphene known in part for their biocompatibility and fluorescence, which leads to many potential uses in medicine as a diagnostic tool. Solutions of GQDs are known to fluoresce less when the GQDs are allowed to clump together, leading to processes such as sonication being used to break apart these clumps in research environments. Similarly, the addition of surfactants to a solution of GQDs has also been found to modify fluorescence response of the solution. This research explores the effect of introducing four different human blood proteins on the fluorescence response of reduced graphene quantum dots (rGQDs). Fibrinogen, transferrin, gamma globulin, and albumin were added to samples of rGQDs in increments around their respective concentrations in human blood. Generally, we found that the addition of any of the blood proteins lowered fluorescence response in the visible spectrum. In the near-infrared spectrum, smaller concentrations of blood proteins generally increased fluorescence response, while larger concentrations reduced fluorescence response below the control. This has implications for deep-tissue imaging relying on the near-infrared fluorescence of intravenous GQDs.

Pandemics require quick decisions about how to distribute a limited number of vaccines, even when the disease is not fully understood and vaccine delivery is limited. We create a disease model that divides the population into groups based on how likely they are to be hospitalized and how likely they are to get infected, so we can test different group-based vaccination strategies. We compare vaccinating only one group, simple step-by-step priority policies that vaccinate groups for set time periods, and a sensitivity analysis to see which model factors most affect outcomes.

We find that vaccinating people who are both high-risk for hospitalization and highly likely to become infected leads to the biggest reductions in total hospitalized time and deaths, while vaccinating lower-risk groups gives little improvement in severe outcomes. A short step-by-step policy that quickly prioritizes high-risk groups can reduce infections and deaths within about 20 days. The sensitivity analysis shows that the death rate and the rate at which infected people move into hospitalization have the strongest influence on severe outcomes, showing that hospital and clinical processes matter a lot in addition to vaccination. Overall, these results support clear, practical, and easy-to-apply prioritization rules for reducing severe disease when vaccine supply is limited.

Influenza can cause severe respiratory illness and spreads quickly, making prevention especially important for people at higher risk of complications. Because vaccines are not always fully protective, effective antivirals can provide an added layer of defense before infection begins. CD388 is a new antiviral being tested as a preventive treatment for influenza. In this project, we used a mathematical model to better understand how the drug changes the course of infection inside the body. Viral load data from a human challenge study were fit to a target-cell model with an eclipse phase, allowing us to estimate key infection parameters. Compared to placebo, CD388 lowered peak viral load and reduced overall viral burden by about 22%, largely by suppressing viral production. Bootstrap analysis was used to assess uncertainty in the parameter estimates. These results help explain how CD388 limits viral spread and supports its potential as a prophylactic therapy.

Favipiravir, an antiviral that inhibits viral RNA-dependent RNA polymerase, has demonstrated promise as a therapeutic for RNA viral infections such as SARS-CoV-2. Mathematical modeling of viral kinetics provides a tool for analyzing the progression of viral infections and the action of antiviral drugs. In the present investigation, the viral kinetics of SARS-CoV-2 infection in cynomolgus macaques treated with the antiviral drug favipiravir were analyzed using a target cell-limited mathematical model of viral infection. Parameters of the model representing the dynamics of viral infection and replication were estimated by fitting the model to the viral kinetics data. Statistical resampling techniques were applied to analyze the uncertainty of the parameter estimates and to compare viral kinetics between the different treatment regimens. The results demonstrate that antiviral treatment induces measurable effects on viral kinetic parameters, reflecting dose-response effects on viral infection dynamics.

About one quarter of patients hospitalized with influenza-like illnesses are infected with more than one respiratory virus. Coinfections can lead to more serious outcomes for patients and are more complex to treat than infections with single viruses. Mathematical models can be used to help us understand the dynamics of viral coinfections and optimize treatment. Previous mathematical models of viral coinfections assume a cell-free transmission pathway with virus leaving one cell and traveling to nearby uninfected cells. However, viruses can also tunnel directly from one cell to another, which can affect how coinfecting viruses interact. This project analyzes a system of coupled ordinary differential equations that includes both cell-free and cell-to-cell transmission to better replicate actual viral spread. We measure coinfection duration for combinations of five common respiratory viruses as a function of the amount of cell-to-cell transmission. We find that coinfection duration depends nonlinearly on the cell-to-cell transmission rates, with differing patterns for different coinfecting viruses. This study highlights the importance of considering different transmission modes when modeling viral dynamics.

Oncolytic adenoviruses are promising cancer therapies because they can selectively infect and destroy tumor cells, however their replication in cancer cells is sometimes limited leading to incomplete tumor suppression. Recently, researchers have started to modify viruses to enhance their replication in cancer cells. In this study, we use a system of ordinary differential equations (ODEs) to model tumor growth and compare viral treatment dynamics of a modified oncolytic adenovirus ICVB-1042 and a wild-type adenovirus type 5 (Wt Ad5). The model was fit to experimental allowing us to estimate important model parameters for both viruses: infection rate, infected cell death rate, rate of cell protection by the immune response, rate of cell resistance loss, viral production rate, and viral clearance rate. We found differences in the viral production rates and the clearance rates between the two viruses, providing insight into how genetic modifications have altered viral dynamics. These findings highlight how viral properties determine the effectiveness of oncolytic virus therapy.

Title: Evaluation of A Learning Intervention to Support Dementia Care over Two Months
Authors: Emily Q. Anderson, Morgan Shumaker, & Uma Tauber
The Micheal and Sally McCracken Annual Student Research Symposium: April 17, 2026

Abstract:
Research indicates that providing care for individuals with Alzheimer’s Disease and Related Dementia (ADRD) can be highly demanding. Caregivers, who are often relatives or friends, face elevated stress levels and may lack adequate knowledge or resources to effectively support their loved ones (Jorge et al., 2021). This gap in knowledge often leaves caregivers feeling incompetent and may also lead to a lack of confidence in their caregiving skills. We have previously found that our digital health education intervention improves caregivers’ knowledge of behavioral and psychological symptoms of dementia (BPSD). In this research, we extend on our prior work by establishing the impact of our intervention on caregivers’ long-term retention – 2 weeks and 2 months post-intervention. Caregivers were taught 12 categories of BPSD (e.g., anxiety, agitation) by reading information and then either rereading or taking practice tests with detailed, corrective feedback, which has been shown to enhance learning and retention (Ariel et al., 2023; Carpenter et al., 2022; Dunlosky et al., 2013). The study consisted of 3 sessions. Caregivers first completed a screening process to determine their eligibility to participate. Session 1 consisted of teaching caregivers about the BPSD via reading and then rereading or taking tests with feedback. Session 2 consisted of a second round of the learning intervention, as well as taking survey assessments and knowledge tests. Finally, Session 3 consisted of completing final tasks and knowledge assessments. Our goal was to have caregivers learn what physicians would want them to know for how to care for their loved one living with dementia. This study provides caregivers with the knowledge and in turn, confidence to manage BPSD. Data is being collected concurrently, with nearly half of the target participants enrolled so far; thus,, the current presentation reports preliminary observations from this initial cohort.