2022 University of Chicago Undergraduate Research Symposium: Session 2

Cerebral cavernous malfunction (CCM) is a disease of the neurovascular system, characterized by clusters of dilated vessels in the central nervous system. Three genes have been identified in familial cases of CCM: CCM1/KRIT1, CCM2/MGC4607, and CCM3/PDCD10. The role of CCM protein signaling in endothelial cell homeostasis has been established, but their role in epithelial cells has been underexplored. Prior studies have indicated that CCM could act in the gastrointestinal tract. Specific findings demonstrate that Pdcd10 deletion influences the intestine through reduction of the mucus layer, enlargement of the goblet cells, and inflammation of the intestine. Previous findings have indicated that PDCD10 is involved in intestinal homeostasis and therefore intestinal disease development, but further studies were necessary to establish its exact role. It was hypothesized that CCM3/PDCD10 plays a role in intestinal epithelial cell (IEC) homeostasis through modulating intestinal stem cell (ISC) proliferation. Furthermore, the hypothesis indicated that PDCD10-induced disruption of IEC maintenance contributes to Inflammatory Bowel Disease (IBD). The role of CCM proteins was investigated using tamoxifen-inducible mouse knockout models, mouse weight studies, gamma radiation, immunohistochemical staining, and quantitative morphological analysis of intestinal characteristics. Furthermore, PDCD10 and other CCM protein expression levels on the GEO database were compared in control individuals and individuals with IBD. It was found that PDCD10 deletion exacerbated disruption of intestinal homeostasis in 12 Gy and 8 Gy damage-induced conditions, demonstrating the role of PDCD10 in IEC homeostasis. It was further observed that β-catenin activation in intestinal stem cells rescues the irregularities in PDCD10-deleted mice, indicating that β-catenin could act downstream of PDCD10. Finally, PDCD10 was upregulated in certain IBD cases, which could act as a compensatory event and needs to be further explored. Overall, these findings indicate a role of PDCD10 in a signaling pathway for ISC maintenance, and could contribute to development of IBD.

Alex Ellerstein

Christopher Weber, Department of Pathology, University of Chicago; Le Shen, Department of Pathology

Biological & Health Sciences

People with limited English proficiency (LEP) in the US with type 2 diabetes (T2D) face worse health outcomes than native English-speakers. Several interventions have attempted to improve outcomes in individuals with T2D and LEP by offering interventions in study populations’ native languages, including Spanish, Bengali, Korean, Samoan, and Tagalog. This analysis aimed to determine the overall effectiveness of T2D randomized controlled trials (RCTs) at improving HbA1c (glycosylated hemoglobin, a form of hemoglobin associated with sugar) in people with LEP. A literature search, focused on T2D interventions among people with LEP published from 1985 to 2019, was conducted using PubMed, Scopus, PsycInfo, and CINAHL. Broad search terms were utilized for diabetes, race/ethnicity, disparities, and language. Interventions were included if > 20% participants were identified as having LEPhad LEP. Non-RCTs and interventions focused on drug or device efficacy were excluded. A random effects meta-analysis was performed to calculate the pooled effect size to provide summary estimates. The population of interest was adults (age 18+) with T2D living in the US and/or US territories. Initially 111,289 articles were reviewed with 43 meeting the inclusion criteria. Participants were 67% female with a mean age of 56 years, and were racially/ethnically diverse (70% Hispanic, 13% Black, 8% Asian, and 2% Pacific Islander). Most interventions were offered in Spanish (N=39), while others were offered in Bengali (N=1), Korean (N=1), Samoan (N=1), and Tagalog (N=1). Interventions included diabetes educational sessions, lifestyle changes, and telehealth-based support. Overall, patients randomized to the intervention arm had a 0.15% decrease in HbA1c (95% CI -0.29% to -0.02%, p<0.01, I2=85%) compared to the control arm. These finding suggest that non-English T2D RCTs lead to modest improvements in HbA1c. Future research should examine whether characteristics of certain interventions led to larger impacts on HbA1c and other cardiovascular risk factors among people with T2D and LEP.

Alex Rodriguez

Dr. Neda Laiteerapong, UChicago Medicine, Center for Chronic Disease Research and Policy; Amber Deckard, UChicago Medicine

Biological & Health Sciences

During organismal development, embryonic stem cells transition through stable states of gene expression to form numerous tissue-specific cell types. Transcription factors act as critical effectors of signaling pathways driving cell specification. In the developing Drosophila melanogaster eye, the eye imaginal disc, transcription factors Yan and Pointed (Pnt) act antagonistically downstream of Receptor Tyrosine Kinase signaling to regulate the transition of unspecified retinal progenitor cells into specified photoreceptor neurons R1 through R7. Pnt and Yan compete for the same binding sites at target gene regulatory regions. While Pnt generally acts as a transcriptional activator and promotes photoreceptor specification, Yan represses photoreceptor-specific genes to inhibit cell specification. Previous work has shown that increasing the ratio of Pnt to Yan in the developing eye produces ectopic photoreceptors; conversely, decreasing the Pnt:Yan ratio leads to loss of photoreceptor neurons. Altering the level of Pnt expression alone, however, leads to maintenance of the Pnt:Yan ratio by an unknown regulatory mechanism and permits normal photoreceptor cell development. My research addresses whether this regulatory mechanism is bidirectional, meaning robust to changes in both Pnt and Yan, and seeks to define the transcriptional or post-transcriptional mechanisms by which cells sense and adjust absolute transcription factor levels to maintain a normal Pnt:Yan ratio. I will perturb Yan levels using genetic methods (including transgenes and null alleles) and then investigate, using imaging with immunostaining, smFISH, and fluorescent markers, how Pnt levels respond and how these changes impact the initiation and termination of Pnt and Yan target gene transcription. These experiments will produce mechanistic insight into how coordinated expression of the opposing transcription factors Pnt and Yan influences the downstream regulation of key developmental genes to ensure reliable cell fate transitions. More broadly, my work aims to elucidate the complex mechanisms by which transcription factors regulate precise tissue development.

Amanda Hill

Ilaria Rebay, Ben May Department for Cancer Biology, Rebay Lab; Suzy Hur, Rebay Lab

Quad Undergraduate Research Scholar

Biological & Health Sciences

-7/del(7q) Acute Myeloid Leukemia (AML) is a high-risk leukemia that accounts for about 8% of all AML patients, resulting in a poorer prognosis as many generalized therapy methods such as chemotherapy are rendered ineffective. In order to provide more effective and less cytotoxic treatment, a need for targeted therapies which prove lethal to a select group of cells remains a high unsolved priority. To investigate whether any genetic vulnerabilities exist in this subset of patients, 5 previously published genome-wide CRIISPR-CAS9 knockout screens conducted in AML cell lines were used to identify genes on chromosome 7 essential for AML cell proliferation. Out of all genes on chromosome 7, we identified 239 essential chromosome 7 genes in AML cell lines and further narrowed this list to 44 genes which can be targeted by a commercially available drug based on data collected from Drug-gene Interaction Database (DGIdb) and DrugBank. From these potential genes, cyclin dependent kinase 6 (CDK6), important for transition from the G1 phase and an already established therapeutic target in breast cancer, significantly inhibited cell proliferation in AML cell lines such as KG1, K562 and U937 conferring its status as an essential gene in AML through CRISPR knockout of CDK6. CDK6 also demonstrated potential for -7/del(7q) specificity given -7/del(7q) AML cell lines were more sensitive to pharmalogical inhibition with Palbociclib, a potent and specific inhibitor of CDK6. These findings encourage the expansion of research for the inhibition of this gene in relation to AML potentially through AML xenograft mouse models as well as an avenue to continue inhibition of other potential gene candidates to identify a similar trend. Finally, this project also highlights the importance of characterizing genetic vulnerabilities in specific cancer types.

Anjali Kotamarthi

Megan McNerney, Department of Pathology, University of Chicago; Madhavi Senagolage , McNerney Lab

Quad Undergraduate Research Scholar

Biological & Health Sciences

The COVID-19 pandemic has caused an unprecedented strain on hospital systems around the world, including hospital resource shortages and staff burnout. Nations must learn to separate the severe cases from the non-severe cases and ensure hospital preparedness. The purpose of this study is to develop an artificial intelligence-based pipeline to predict a COVID-19 patient’s length of hospitalization and the proper data-driven discharge protocol, as well as to identify risk factors for increased risk of death and prolonged length of hospitalization based on large-scale data analysis. Clinical records and imaging data (thoracic CT scans) from hospitalized COVID-19 patients were retrospectively analyzed. Two random forest machine learning algorithms were created to predict the length of hospitalization and whether a patient can be discharged while they are still improving from COVID-19 or only after full recovery, and AUC analysis was performed. Comorbidities, demographics, and quantitative radiomic imaging data were extracted for use as predictors. Additionally, a Welch two sample t-test was used to determine statistical significance of risk factors for increased length of hospitalization, and a two proportion z-test was used to determine statistical significance of risk factors for increased risk of death. The algorithm predicting length of hospitalization had an AUC value of 0.9933 and predicting discharge condition had an AUC value of 0.9925. Additionally, the risk factors for increased length of hospitalization and risk of death are senior age, high blood pressure, renal disease (only for risk of death), cerebrovascular disease, and neoplastic disease. As hospital systems need more proactive tools to grapple with the COVID-19 pandemic, such AI pipelines as the one described may be effective in ensuring hospital preparedness for years to come. Further validation is needed for clinical use.

Beatrice Katsnelson, Elise Katsnelson

Maryellen Giger, UChicago Medicine Department of Radiology; Jordan Fuhrman, Committee on Medical Physics

Quad Undergraduate Research Scholar

Biological & Health Sciences, Computing Science

The maternal pelvis has been traditionally classified into 4 types: gynecoid, anthropoid, android, and platypelloid. While most women’s pelvises may be intermediate as they do not fit one definition perfectly, this classification helps physicians to better understand the mechanism of labor. As such, different pelvic shapes should be incorporated into simulation models. However, at present, no model exists replicating these different pelvis types, and most of the current birthing simulators do not even contain a bony pelvis as a component. We sought to create different maternal pelvis models using 3D printing technology for educational and training purposes. Utilizing two software, Slicer and MeshMixer, and one database, The Cancer Imaging Archive, four pelvic models representing classical female pelvis types were 3D-printed along with two fetal skulls. The pelvises were also segmented into three sections to illustrate the anatomical areas where there is connective tissue. This particular assembly allows for the models’ pelvic diameters to be adjusted for further procedural individualization. These models replicate the characteristics of the four pelvic shapes and allow the examiners to appreciate their differences during clinical pelvimetry. They also demonstrate their effects on the birthing process. For example, the fetal head is sometimes unable to pass the android pelvis when it is presenting as an occiput posterior, and the platypelloid pelvis prevents internal rotation. Finally, the application of forceps is affected by the pelvic shapes, making it more difficult to insert and articulate the forceps in a non-gynecoid pelvis. Models allow physicians and trainees to have a more realistic learning experience in the clinical assessment of the maternal pelvis, understanding how the mechanism of labor changes in different pelvis types, and becoming more proficient in operative vaginal delivery. These newly constructed pelvis models should be incorporated into future birthing simulators.

Bella Gomez, Nicole Yao, Narvella Sefah

Dr. Yuzuru Anzai, OBGYN, Lenox Hill Hospital

Biological & Health Sciences

In Octopus bimaculoides, arm regeneration requires the ability to regenerate a large variety of unique structures, including the axial nerve cord, which consists of millions of neurons running down each of the octopus’s eight arms. This regenerative process requires the formation of a blastema, a mass of seemingly undifferentiated cells that develop into the regenerating arm. The molecular mechanisms underlying this complex process are not understood, but modern bioinformatics allows for unbiased analysis of the octopus transcriptome at various stages of development and regeneration. Through RNA-sequencing analysis, the expression levels of genes differentially expressed in the regenerating octopus arm, as compared to the developing arm bud and adult arm tips, were characterized. Of note, eight blastema-specific genes were identified based off their relatively high expression levels in the blastema. Of particular interest, in situ hybridization confirmed the differential expression of one of these eight genes, a matrix metalloproteinase (MMP) homolog. This MMP homolog’s expression peaks at the late blastemal stage, with a clear line of expression at the boundary of the resident axial nerve cord and the regenerating blastema. Before this, at the early blastemal stages, and after, at the outgrowth stages, low-level diffuse expression of MMP is present. These results show the utility of utilizing both bioinformatics and wet-lab approaches to understand the molecular mechanisms behind octopus arm regeneration.

Cassie Manrique

Clifton (Cliff) Ragsdale, Neurobiology, University of Chicago; Natalie Grace (Grace) Schulz, University of Chicago

College Research Fellow

Biological & Health Sciences

Cellular senescence, or a state of replicative arrest, occurs when cells undergo major damage. Once cells enter a senescent state, they release inflammatory molecules known as the SASP, which call out to the immune system to destroy the senescent cells. Senescent cells and their SASP have entered centerstage in terms of developing cancer treatments. Because senescent cells straddle the line between healthy and cancerous cells and can be killed using a class of drugs called senolytics, converting cancerous cells to senescent cells could become a viable treatment that sidesteps chemotherapy. The goal of this research project is to gain a better understanding of which elements of the SASP are responsible for senolytic sensitivity in bystander cells. This involves using a quantitative approach that relies on proteomic data analysis to decipher the proteins present in the SASP. We used a protocol pioneered by my mentor Don Wolfgeher using a ProteoMiner Protein Enrichment Kit to detect upregulated elements in the SASP of cancer cells driven to senescence compared to non-senescent controls. Cancer cells grown in media containing the SASP were assayed for their sensitivity to ABT-263, a senolytic known to induce mitochondrial apoptosis. Upon confirmation that the conditioned media containing the SASP actually sensitized bystander cancer cells to senolysis, the conditioned media was filtered to separate proteins in the SASP by size. Cancer cells were grown in the filtered conditioned media and once again assayed for senolytic sensitivity. The range of relevant protein sizes was determined by sensitivity of bystander cancer cells to differently filtered conditioned media. RNAseq was then performed on the bystander cells to determine which pathways and ligand-receptor interactions were activated. From pathway literature research and the protein size range, the proteins in the SASP most responsible for the sensitivity of bystander cells to senolytics could be determined.

Edward Wang

Stephen J. Kron, Molecular Genetics and Cell Biology

Quad Undergraduate Research Scholar

Biological & Health Sciences

COVID-19, caused by the novel 2019 coronavirus, usually presents as a respiratory illness of varying severity, ranging from mild to requiring hospitalization. Thoracic CT scans are commonly used as a diagnostic measure for COVID-19 and to give information regarding the severity of COVID-19 infection in the lungs. The purpose of this study is to understand the extent to which COVID-19 impacts the lungs based on age, gender, and other illnesses. The CT scans were processed via a dual-headed U-Net deep learning model and the ratio of COVID-19 pixels to total lung pixels from a patient’s primary thoracic CT scan upon admission to the hospital was assigned as the patient’s severity score (SevScore). Patients were split up into age groups of <30, 30-65, and 65+; patients in the 65+ group had a significantly greater SevScore as compared to patients in both the <30 and 30-65 age groups, while no statistically significant difference in SevScore was found between the <30 and 30-65 age groups. No statistically significant was found between male and female sexes. SevScores were found to be higher (p < 0.05) in patient groups with the following comorbidities as compared to those without: high blood pressure, renal disease, and congestive heart failure, while the difference in SevScore distribution among patients with and without liver disease, neoplastic disease, cerebrovascular disease, and COPD failed to reach significance (p > 0.05), although these diseases were limited by lower prevalence in this patient population. Additionally, there is a positive correlation between the mean SevScore and number of comorbidities, and patients with no comorbidities exhibited significantly lower average SevScores than those with comorbidities, while patients with 3+ comorbidities had significantly higher average SevScores. Understanding COVID-19 severity can aid in planning and prioritizing care for patients at greater risk to increase survival rates and positive outcomes.

Elise Katsnelson, Beatrice Katsnelson

Maryellen Giger, UChicago Medicine Department of Radiology; Jordan Fuhrman, Committee on Medical Physics

College Global Health Scholar, Quad Undergraduate Research Scholar

Biological & Health Sciences, Computing Science

There are numerous sensory inputs that bacteria can process to elicit physiological change. Two such examples are quorum sensing and nutrient sensing. Bacteria use quorum sensing to gauge cell density, and at high count, express certain genes that are useful as a group. Nutrient sensing is used to detect nutrients in the environment and appropriately regulate gene expression in response. The nutrient-sensing pathway in Pseudomonas aeruginosa involves the sensor histidine kinase CbrA which is known to drive the expression of CrcZ, a small RNA that post-transcriptionally regulates genes involved in carbon metabolism. Furthermore, nutrient-sensing and quorum-sensing pathways intersect to control collective behaviors such as biofilm formation. The goal of this project is to identify and characterize factors that regulate CbrA-CrcZ nutrient-sensing pathway. To do this, we are in the process of building a transcriptional luciferase-based reporter fusion of the crcZ promoter. We will use this reporter to determine levels of CrcZ expression in P. aeruginosa strains that have been mutagenized with transposons. I expect that transposon insertion in activators of this NS pathway will show low levels of luciferase expression when mutagenized. Repressors will have the opposite phenotype. This project will enhance our current understanding of this nutrient-sensing pathway in P. aeruginosa and its ability to potentially cross talk with other systems like quorum sensing or photo sensing. We can potentially use this knowledge to develop therapies that reduce or kill P. aeruginosa during an infection, decreasing its severity in the susceptible.

Ethan Yarberry

Sampriti Mukherjee, Molecular Genetics and Cell Biology

Quad Faculty Research Grant Scholar

Biological & Health Sciences

The random recombination of T-cell receptor specificity allows the immune system to protect against any possible foreign antigen. However, the problem with this random creation of TCRs is that the body will inevitably generate T-cells targeted to tissues of the body, causing autoimmunity. The question is: how can the body distinguish between self and non-self? This distinction is driven in part by a process called negative selection. Medullary thymic epithelial cells (mTECs) display tissue-restricted antigens, and when a TRA is recognized by a TCR, this T-cell is deleted from the repertoire. mTECs, cells of a defined lineage, have a unique ability to express antigens specific to tissues outside of the thymus. For cells to express genes other than those within their defined lineage, expression must be altered at transcriptional and/or epigenetic level. This shaping of chromatin structure and gene expression establishes a permissive epigenetic state in which genes from alternate lineages can be expressed in mTECs. Previous analysis of mTEC development revealed accessibility enrichment in the transition from progenitor to mature cells. Within the regions that become highly accessible during mTEC maturation, the sequence motifs targeted by PPARg are highly enriched. PPARg is a nuclear receptor which acts as both an activator and repressor in many regulatory processes. The hypothesis is that PPARg aids in the promotion of ectopic expression of TRAs through assisting in the establishment of a permissive epigenetic landscape. Bulk genome-wide profiling of chromatin accessibility and gene expression patterns was conducted using a thymic conditional knockout of PPARg. RNA- and ATAC-seq analysis displays those loci altered by PPARg, revealing a possible mechanism of PPARg-induced ectopic expression. This conditional knockout is also used to assess the autoimmune phenotype following the loss of PPARg in mTECs, revealing possible TRAs controlled by PPARg expression.

Gaby Berman

Andrew Koh, Pathology

Quad Faculty Research Grant Scholar

Biological & Health Sciences

A functional immune system relies on T cells to recognize invading pathogens without reacting to self-tissues. This discrimination is largely achieved in the thymus, where developing T cells are selected based on their self-reactivity. Harmful T cells with strong affinity to self are eliminated or inactivated from the repertoire via recognition of their cognate antigen in the thymus. Thus, the scope of this negative selection is defined by the range of self-antigens expressed in the thymus. To extend the diversity of self-representation in the thymus, medullary thymic epithelial cells (mTECs) ectopically express genes usually restricted to unique tissues, such as the pancreas or brain. How cells of a defined epithelial lineage express alternative tissue-restricted genes remains largely unknown. Here, we identify the transcription factor Bhlhe40 as a determinant of this ectopic gene expression. Using genetically engineered mice with Bhlhe40conditionally deleted in mTECs, we demonstrate that Bhlhe40 promotes a chromatin accessibility landscape that is essential for the expression of a subset of tissue-restricted genes. Bhlhe40 deficiency in mTECs leads to the escape of self-reactive T cells to the periphery, causing systemic autoimmunity. Elucidating the mechanistic actions of Bhlhe40 in mTECs will provide novel insights to cellular plasticity and potentially open avenues for therapeutic advances of autoimmune diseases.

Jacob Bernheim

Andrew Koh, Biological Sciences Division, Koh Lab

Quad Faculty Research Grant Scholar

Biological & Health Sciences

Standardized segmentation and subsequent interpretation of MR images is challenging. While clearly defined structures are generally segmented successfully, irregular lesions present difficulties. Deep learning has been applied to segment medical images with varying degrees of success. Cerebral radiation necrosis (RN), a side effect of stereotactic radiosurgery, has diffuse edges and heterogeneous enhancement that have challenged several previous machine learning algorithms. We hypothesized that studying pure RN lesions in an animal model would help us create a more accurate model of RN for automatic segmentation. Healthy mice were irradiated and imaged with contrast-enhanced T1-and T2-weighted MRI. The images were automatically segmented using the deep learning module of MIPAR, a commercially available software, and compared to ground truth manual segmentations created in ITK-SNAP. The program has some difficulty segmenting small regions of RN, as we found a lesion detection rate of 72% with a strongly right skewed distribution of unsegmented lesion areas. In images with large necrotic areas, however, we found a median Dice score of 0.88 (IQR 0.78-0.94), median sensitivity of 0.90 (IQR 0.77-1), median specificity of 0.94 (IQR 0.83-1), and Pearson correlation coefficient of 0.76 (p < 2.2e-16). The data suggests that a commercially available deep learning program could increase objective interpretation of MRI for diffuse and heterogeneously enhancing regions. Our model may serve as the basis of a deep learning program by differentiating RN from GBM recurrence, which could replace invasive methods of differentiation and affect treatment planning. Future work will be directed at improving the data model to better detect small regions of RN.

Jillyn Turunen

Lisa Anne Feldman, MD, PhD, City of Hope National Medical Center Department of Surgery, Division of Neurosurgery; Joel R Garbow, PhD, Washington University School of Medicine, Mallinckrodt Institute of Radiology, Division of Radiological Sciences

Dean's Fund for Undergraduate Research CONFERENCE Awardee, Quad Undergraduate Research Scholar

Biological & Health Sciences, Neuroscience

The COVID-19 pandemic caused by SARS-CoV-2 is a global health emergency. In order to better understand the complications associated with COVID-19, it is critical that research and development teams have access to high-accuracy viral protein structures. Current COVID-19 proteins available on the Protein Data Bank (PDB) contain some moderate- to low-resolution structures. In order to improve the quality of the structures, we use an automated protocol that combines torsional optimization with real-space refinement against the electron density derived from the experimental x-ray data. Our method converts moderate- to low-resolution protein structures at initial (e.g. backbone trace only) or late stages of refinement to structures with increased numbers of hydrogen bonds, improved crystallographic R-factors, and superior backbone geometry. For example, we applied our algorithm to the structure of the SARS-CoV-2 nucleocapsid protein, which is a critical factor in the viral genome packaging process (the N-terminal RNA binding domain, PDB ID: 6M3M). Our protocol improved the R-work by 6.1%, the R-free by 12.9%, MolProbity geometry score by 43.8%, and our own Torsional Statistical Potential (TSP) score by 51.5%. Application of this automated method on COVID-19 proteins has produced higher quality structures that can further aid the counter-pandemic efforts. This method is applicable to proteins of any size at any stage of refinement, and it can be extended to improve NMR and Cryo-EM structures as well.

Joseph Farrell

Dr. Esmael Jafari Haddadian, Biological Sciences Collegiate Division

Dean's Fund for Undergraduate Research CONFERENCE Awardee

Biological & Health Sciences, Computing Science, Engineering, Mathematics, Physics, Statistics

The aorta carries oxygenated blood to all organs and cells in the human body. With age, there is a notable increase in aortic diameter and decrease in aortic distensibility. The aorta also becomes more stiff and anisotropic. Identification of these age-related changes can help inform treatment strategies and timing for intervention for older adults with aortic diseases. Two of the most common diseases involving the aorta are dissections, a tear in the inner layer of the aorta, and aneurysms, an abnormal enlargement of the aortic wall. Because both conditions can occur in aortas below the diameter threshold for elective surgery, methods other than size measurement of the aorta are needed to risk stratify patients. Our research investigates the fragile nature of the aorta by characterizing its geometry and mechanics using mathematical and image-based analysis algorithms. We hypothesize that aortic surface curvature is influenced by each patient’s unique physiological state. By assessing aortic geometry under different physiologic stresses, we can better understand how the surface curvature changes in relation to those stresses. We will investigate the effect of aging on wall stresses and hemodynamics in healthy aortas to better inform treatment protocol in older patients.

Karen Yuan

Luka Pocivavsek, Department of Surgery, University of Chicago

Quad Undergraduate Research Scholar

Biological & Health Sciences

Three-dimensional cell culture methods are gaining importance in the field of tissue engineering, especially spheroids, which are balls of cells that deposit their own ECM to build the structure. The goal is to create a high yield of high-quality spheroids. However, a method for consistent spheroid evaluation has not yet been established. We have been able to establish an index for spheroid evaluation based on grayscale images that allows comparison across different spheroid creation methods. Spheroids composed of mouse mesenchymal stem cells (mMSC) were created using the 96 low-attachment U-bottom plates, large microwell plates, and small microwell plates. The spheroids’ diameter and gray value were evaluated, as well as the ratio of minimum to maximum diameter, indicating spheroid circularity. From these factors, we created a mathematical formula (SQI index) to evaluate the quality of any spheroid regardless of the method used. Finally, histological data was used to confirm that the index effectively ranked the quality of spheroids. From all of these factors, we created an index which positively correlated with spheroid quality. There was a significant difference (p<0.0001) among the three groups using this index. The high quality 33,000 cell mMSC low-attachment plate spheroids had an index of 55.3+/-16.0 which was greater than large dimple spheroids (32.2 +/- 9.3, p<0.0001) and small dimple spheroids (23.5.2 +/- 11.1, p<0.0001). We have established a method for evaluating spheroids using the diameter and gray value of spheroids. This method will contribute to the mass production of spheroids by evaluating and examining spheroids and their production methods.

Katherine Nurminsky, Stephanie Ran

Dr. Narutoshi Hibino, UChicago Medicine Cardiac Surgery, Hibino Lab of Cardiac Tissue Engineering

College Research Fellow, College Summer Research Fellow

Biological & Health Sciences, Engineering

In the Drosophila melanogaster embryo, anterior-to-posterior polarity is established by maternally deposited mRNA of the bicoid gene. After translation early in development, the resulting Bicoid protein acts as a transcription factor, initiating the transcription of segmentation genes required for anterior fates and thereby acting as the Drosophila anterior determinant. At high concentrations, Bicoid shows the pioneer-like ability to induce a transcriptionally primed open chromatin conformation of a subset of target enhancers. Much of the segmentation gene network is conserved across dipteran insects; however, diverse genes encode for anterior determinants in different dipteran species, such as odd-paired in the moth midge Clogmia albipunctata (Cal-Opa). We propose to use dipteran anterior determinants and the segmentation network they act on as a model to study the evolution of developmental gene networks. We hypothesize that anterior determinants are morphogens that show pioneer-like activity in modulating chromatin accessibility of key transcriptional targets necessary for inducing anterior fates. To characterize the putative morphogenetic activity of Cal-Opa, we employ immunological staining across multiple developmental stages in wild type (WT) embryos. We will elucidate the expression of putative Clogmia segmentation genes and test the necessity of Cal-Opa on the expression patterns of these genes via in situ hybridization of WT and Cal-Opa knockdown embryos (via RNA interference). Novel Cal-Opa targets will be determined by identifying regions of Cal-Opa binding through ChIP-seq (Chromatin-Immunoprecipitation Sequencing). These experiments establish the core gene network acted upon by Cal-Opa. Finally, we will define enhancers whose accessibility is dependent on Cal-Opa by performing ATAC-seq (Assay for Transposase-Accessible Chromatin using sequencing) on WT and Cal-Opa knockdown embryos, testing Cal-Opa’s ability to modulate chromatin accessibility. This work sheds light on how the genetic circuit responsible for embryo polarization may accommodate structurally unrelated anterior determinants by establishing the morphogenetic and pioneer-like qualities of these essential genes.

Liviu Megherea

Urs Schmidt-Ott, Department of Organismal Biology & Anatomy, Schmidt-Ott Lab; Ezra Emsy Amiri, Schmidt-Ott Lab

Quad Undergraduate Research Scholar

Biological & Health Sciences

Adenoid cystic carcinoma (ACC) is a relatively low-incidence but aggressive salivary gland cancer. With a persistent and metastatic phenotype and no systemic chemotherapy widely available, recurrence and secondary cancers are common, rendering late-stage ACC functionally incurable. Chemotherapeutic targets for ACC are thus of significant interest to improve prognosis and clinical management. PRMT5 (protein arginine methyltransferase 5) is an epigenetic activator and remodeling protein found to be upregulated in patient-derived ACC samples. Having identified PRMT5 as a potential target, recently discovered selective PRMT5 inhibitor PRT543 has shown promise as a treatment option. In a collaboration with Prelude Pharmaceuticals, we investigated dose-dependent expression of five genes in two PRT543-treated models: human-derived ACC cell lines and NOTCH-mutant ACC organoids. The four genes analyzed were PRMT5, AXL, MYB, and c-MYC, known proto-oncogenes. All four genes are strongly implicated in tumor survival and cell proliferation. Cells in both models were cultured with 7-day courses of PRT543 treatment at six doses and analyzed via immunoprecipitation and Western Blot. Our results demonstrate PRT543’s ability to inhibit all four genes in both models in a dose-dependent manner. PRT543 Phase I clinical trials are currently being conducted; thus, our research provides specific evidence for the regulatory function of PRT543. This in turn supports the prospective translational therapeutic use of PRT543 in patients diagnosed with late-stage ACC, opening potential avenues for treatment.

Manu Sundaresan

Dr. Evgeny Izumchenko, University of Chicago Department of Medicine, Section of Hematology/Oncology

Quad Undergraduate Research Scholar

Biological & Health Sciences

Prior research has shown that arteriovenous fistulas (AVF) in hemodialysis patients commonly fail at the venous cephalic arch, the terminal portion of the cephalic vein located in the shoulder area. The cephalic arch is prone to recurrent obstruction including stenosis, or narrowing of the vein, and thrombosis, or blood clot formation. Maintaining optimal wall shear stress (WSS), the force exerted by blood flow onto the vein walls, is important; otherwise, the vein endothelium can potentially trigger stenosis and thrombosis which leads to loss of dialysis access. Consequently, there is a need for better research tools that allow quantifying hemodynamics, such as WSS, throughout the cephalic arch in order to investigate how these forces influence pathology in hemodialysis patients. This could help improve clinical outcomes of patients once treatment commences. We have developed in vitro cephalic arch models using patient-specific venogram and intravascular ultrasound imaging data to create computational models of their vein geometry. These 3D computational models are then 3D-printed and molded using a soft transparent elastomer to fabricate millifluidic devices. Cephalic vein physiologic blood flow is programmed into pumps, then we imaged flow close to the wall across all models using a blood-mimicking fluid with infused fluorescent particles. Flow imaging data is used to characterize local flow and develop WSS profiles across each cephalic arch geometry. We have successfully generated WSS profiles for both physiologic and pathologic cephalic arch models as well as for two patients at two timepoints, 3 and 12 months after their AVF was created. We aim to increase the achievable flow rates the devices can withstand in order to match patient-specific high pathologic flows. Additionally, we aim to seed the models with endothelial cells to study their biological response to characterized hemodynamics, making our technology imperative to improve clinical care of these patients.

Maren Klineberg

Anidita Basu, Genetic Medicine Department, Basu Lab, UChicago; Andres Moya Rodriguez, Biophysical Sciences Department, Basu Lab, UChicago

Quad Undergraduate Research Scholar

Biological & Health Sciences

The lens of the eye projects an incredibly large and complex 3D world onto the retina, a small piece of tissue that encodes visual inputs to transmit through the optic nerve. In this process, there is the requirement for both efficiency and redundancy. The optic nerve is an information bottleneck - with an upper limit on the amount of information that can be sent at a certain amount of time. In theory, each neuron should send completely unrelated parts of the visual input, but there is also noise in each part of information processing. To prevent loss of important information, there is the requirement for redundancy. The means by which so much information is constantly being compressed and encoded is therefore of significant interest, especially the spatial and temporal structure of the code, with broader implications in theoretical neuroscience and the formation of better retinal prosthetics. Our research has focused on analyzing ganglion cells that tile the retinal sheet, using data produced from a section of larval salamander retina being stimulated with five different movies. The movies are all natural scenes, chosen to investigate how the retina might react to types of motion and scenery that more closely resemble what the salamander would see in its natural environment. We then used this data to calculate spatial and temporal correlation functions, which were compared with those of the natural movies to investigate the implications for neural coding.

Minerva Roscoe

Stephanie Palmer, Departments of Organismal Biology and Anatomy & Physics; Jared Salisbury, University of Chicago

Quad Faculty Research Grant Scholar

Neuroscience

The Nagler lab investigates mechanisms by which the gut microbiome can protect against the development of food allergies. Our lab has shown that butyrate and bacteria of the class Clostridia can strengthen the epithelial barrier and may prevent food allergens from crossing into the intestine and initiating an immune response. We have also shown that germ-free mice colonized with the microbiome of a cow’s milk allergic (CMA) infant have a leaky epithelial barrier, as evidenced through FITC staining experiments, while mice colonized with feces from a healthy infant are protected from food allergies. Some Clostridia, or their metabolic byproducts short chain fatty acids, regulate intestinal barrier function, increase mucus production, and induce regulatory T cells, which express FoxP3 and Rorγt. Previous work has shown that Clostridia induce a barrier-protective phenotype mediated by IL-22, a key cytokine in the gut that regulates mucus production. However, we have not yet examined whether the expression of tight junction proteins by intestinal epithelial cells plays a role in this bacteria-induced barrier-protective response. The focus of my project is to investigate the expression of tight junction proteins by intestinal epithelial cells and the production of IL-22 in the lamina propria. Specifically, I aim to examine expression of the tight junction proteins occludin, claudin-1, claudin-2, claudin-4, and ZO-1 in epithelial cells because we believe that these proteins will strengthen the epithelial barrier and prevent dietary antigens from entering. I will also measure IL-22 in lamina propria tissue as well as downstream antimicrobial peptides. For these experiments we will be using tissue harvested from gnotobiotic mice colonized with the microbiome of a healthy infant or a CMA infant. We hypothesize that mice colonized with the microbiome of the healthy infant will have an increased expression of tight junction proteins and IL-22, protecting the mice from food allergies.

Nidhi Talasani

Cathryn Nagler, Department of Pathology, University of Chicago; Lauren Hesser, Pritzker School of Molecular Engineering

Biological & Health Sciences

The genome of a eukaryotic organism is packaged into strands of DNA-protein complex named chromatin. Recent studies have shown that, in the cell nucleus, chromatin adopts non-random, biologically significant structural organizations, where the expression of certain genes can be turned on or off by bringing specific sites of the chromatin into spatial proximity. This chromatin interaction-based gene regulation plays important roles in critical biological processes such as cell differentiation and cancer development. However, given the significance of chromatin interactions, the field still lacks an effective tool for accurate, surgical perturbation of chromatin organization that enables more controlled studies. To resolve this issue, I plan to construct a chromatin organization remodeler using CRISPR, a highly efficient, programmable technology for targeting specific genomic locations. Specifically, my system features a linker RNA, capped on both ends by CRISPR gRNA sequences, targeting two different locations on the genome, such that when both ends bind to the genome successfully, the linker would bring the two genomic locations into spatial proximity. The first step of my project involves constructing a functional remodeler in budding yeast (S. cerevisiae). Once the system proves successful, I will improve the system’s efficiency by optimizing its components such as the linker sequence. Once optimized, this system will enable controlled and effective programming of chromatin interactions, which will be valuable for both research and practical applications, such as regenerative medicine or cancer treatment.

Paddy Liu

Alexander Ruthenburg, Department of Molecular Genetics and Cell Biology, Ruthenburg Lab, University of Chicago; Gönen Memisoglu, Department of Molecular Genetics and Cell Biology

Biological & Health Sciences

High dose-rate interstitial brachytherapy (HDR-ISBT) is a treatment option for gynecological cancers that involves the insertion of needles into the body to deliver highly localized radiation treatments to a cancerous region. Needle digitization, the process of defining needle locations within the 3D CT frame-of-reference, is a time consuming but critical step in HDR-ISBT as optimized dose distributions are generated based on the digitized needle locations. In fact, the high dose gradients present in HDR-ISBT mean small discrepancies (~1 mm) between the digitized and actual needle locations can produce large dose errors that degrade plan quality. Furthermore, needle digitization for gynecological HDR-ISBT is complicated due to the large number of needles present (15-25) and complex overlapping and abutting needle trajectories. This work aimed to address these challenges by developing and validating an automated needle digitization tool (ANDT) for HDR-ISBT. Training data was acquired over a ten-year period for 57 anonymized patients receiving HDR-ISBT treatment for gynecological cancers. A ResNet-50 deep learning network was trained on over 17302 CT scans (346000 w/ image augmentation) to identify titanium needles. Accuracy measured 94% on the test cases. A 2.5D approach was also implemented where the network learns from an input of three consecutive CT slices to find needles in the middle slice. To define needle trajectories from the segmented image pixels, we proposed a novel clustering algorithm that leveraged the Syed-Neblett template, where the needles are clearly separated, to initialize a 3rd-degree polynomial fit to cluster voxels from the DL network with applicators. The trained ANDT output needle trajectories that were on average within 0.56 mm of manual needle trajectories found by a qualified medical physician. The ANDT typically required 10 minutes to process a single patient’s case. Future work will assess the dosimetric impact from using the ANDT’s automatic trajectories for HDR-ISBT.

Raj Tummala

Jordan M. Slagowski, Department of Radiation and Oncology, University of Chicago and University of Wisconsin-Madison

Quad Undergraduate Research Scholar

Biological & Health Sciences

Homeostasis in the immune system is maintained largely through regulatory T cells (Tregs), which suppress excess or inappropriate effector T cell function in order to prevent autoimmunity. However, in the context of cancer, excess Tregs diminish the ability of effector T cells to battle the tumor. There is the need for the development of therapeutics that prevent Tregs from enacting their immunosuppressive functions in the tumor microenvironment, while still allowing for the maintenance of general-body homeostasis such that any off-target effects are avoided. By targeting the protein-protein interactions (PPIs) between the transcription factor Forkhead Box P3 (FOXP3), the master regulator of Treg function, and its binding partners, we hope to limit suppressive programs in Tregs while allowing for effector T cell expression programs, thus amplifying the anti-tumor immune response. Members of the LaBelle lab have synthesized and tested several stapled alpha-helical peptides (SAHs) that target the FOXP3 homodimer PPI, supporting the hypothesis that the FOXP3 SAHs are effective and have a deleterious impact on Treg function. Here, we investigate the effectiveness of the SAHs in binding FOXP3 recombinant protein through cross-linking and pulldown assays. This contributes to our biochemical characterization of the peptides by allowing us to visualize and evaluate changes in the dimerization of FOXP3 with its binding partners as a consequence of peptide incubation, confirming on-target activity. Based on these assays, the FOXP3 SAHs demonstrate effective disruption of the FOXP3 homodimer interaction. Such research is valuable in paving the path for future studies on specific, effective methods of targeting Tregs for the development of anti-cancer agents. Additionally, by investigating the use of SAHs as a novel therapeutic agent for treating tumors, we aim to contribute to an expanding knowledge base in order to improve upon the design of SAHs in the future and identify additional targets.

Rhea Shah

James LaBelle, MD, PhD, Department of Pediatrics, Section of Hematology/Oncology, LaBelle Lab, University of Chicago

Biological & Health Sciences

Gene expression is tightly controlled by gene regulatory elements, such as transcription factors, promoters, and enhancers. Changes in the level or context of expression of conserved genes can drive the evolution of phenotypes. In this study, we focus on enhancers, which are gene regulatory elements that bind transcription factors, resulting in changes in expression of the target gene. Due to their rapid evolution, poorly conserved sequence homology, and short length, enhancers have been difficult to predict computationally, and only recent advances in high throughput sequencing technology enable us to predict genome-wide enhancer maps. While enhancer maps assembled from chromatin-accessibility and ChIP-seq studies provide excellent predictions for enhancers, they often fail to distinguish active functional elements from merely accessible or accessible/bound sites. For this, technologies utilizing reporter constructs or enhancer traps have been employed. Unfortunately, these systems have limited throughput, constraining studies to a handful of elements. This project aims to develop an assay system that rapidly and efficiently tests the function of a large number of elements in a single species and across species. To understand how the evolution of enhancers has contributed to changes in gene regulation and ultimately to new cell types and body plans, we first identify functional enhancers within a species utilizing an approach known as STARR-seq, in which a large number of elements can be simultaneously tested using a genetic screening approach where an active enhancer drives its own transcription. Differential gene expression in cells can be linked to differences in enhancer activity, providing a genome-wide quantitative enhancer map. Further, we aim to develop this technology for use in single-cell assays wherein the activity of a pool of regulatory elements can be assessed in a restricted tissue or cell-type. In this way, we can efficiently class elements by their tissue or cell-type of action.

Sarah Weber

Heather Marlow, Department of Organismal Biology and Anatomy, University of Chicago

Quad Undergraduate Research Scholar

Biological & Health Sciences

Disparities in cancer incidence and mortality have not been described for the University of Chicago Comprehensive Cancer Center’s patient catchment area (CA), the region from which the center draws its patients. The purpose of this research was to describe incidence and mortality rates (2013-2017) for leading cancers in the CA, which consists of Cook County, IL; DuPage County, IL; Lake County, IL; Will County, IL; and Lake County, IN. Age-adjusted rates (per 100,000) were calculated for leading cancers by sex for each county, the CA, U.S., and IL. Racial disparities were also examined by comparing Black rates to White rates. Incidence rates were higher in the CA compared to the U.S for all cancers data was available for: all sites combined, breast, and prostate. For males, mortality rates of prostate, colorectal, pancreatic, and stomach cancers were higher in the CA compared to the U.S. For females, mortality rates of breast, colorectal, myeloma, uterine, stomach, and bladder cancers, as well as all sites combined were higher in the CA compared to the U.S. By county, males experienced incidence rates 8% to nearly 2.5 times higher and mortality rates 3% to over 1.5 times higher compared to the U.S. for various cancers. By county, females experienced incidence rates 5% to over 4 times higher and mortality rates 4% to over 1.5 times higher compared to the U.S. for various cancers. Lake County, IN generally had the highest incidence and mortality rates. Racial disparities in incidence and mortality were present for most cancers in the CA overall and by county. For males, racial disparities in mortality for prostate, stomach, and larynx are of greatest concern. For females, racial disparities in mortality for myeloma, cervical, and uterine are of greatest concern. This research highlights important health inequities and priorities for the UChicago Cancer Center.

Stephanie Zhang, Joel Ssepuuya

Brandon Pierce, Public Health Sciences

Quad Faculty Research Grant Scholar

Biological & Health Sciences

Hypertensive disorders of pregnancy (HDP) are associated with maternal and neonatal morbidity. Readmissions are more common among women who have HDP. We sought to characterize racial differences in readmissions among women with HDP. This is a retrospective study of patients with HDP admitted at an urban tertiary care center from January 2019 through November 2019. Medical information was collected by chart review as part of a bundled quality improvement initiative for women with HDP. Readmission data was collected up to 6-week postpartum period. 729 patients were included with HDP over this time period and 80 patients were readmitted within 6 weeks. 60 of the 80 patients were Black. Indication for readmission differed by race with Black patients more likely to be readmitted for preeclampsia with severe features (43.3 % vs 10.0%). Black patients had higher first systolic blood pressures on admission, in the first 24 hours, and were more likely to be discharged on oral anti-hypertensive medication. Despite an established protocol for patients with HDP to be evaluated in Labor and Delivery, Black patients were more likely to be evaluated in the emergency room initially compared to non-Black patients (43.3% vs 15%). Our results suggest differences in readmission characteristics by race among women with HDP. At the patient level, Black patients had more severe blood pressures and were more likely to have preeclampsia with severe features. At the system level, Black patients were more likely to be evaluated in the incorrect location per established hospital protocol.

Tinyan Dada

Sarosh Rana, Obstetrics and Gynecology, University of Chicago; Victoria Oladipo, Pritzker School of Medicine; Sunitha Suresh, Obstetrics and Gynecology

College Global Health Scholar, Dean's Fund for Undergraduate Research Awardee, Dean's Fund for Undergraduate Research CONFERENCE Awardee

Biological & Health Sciences

Inorganic arsenic (iAs) is a carcinogen, and chronic exposure is associated with adverse health outcomes, including cancer and cardiovascular disease. In Bangladesh, ~50 million individuals are chronically exposed to iAs through drinking water. Consumed iAs can undergo sequential reduction and methylation reactions catalyzed by gluthathione and arsenic methyltransferase (AS3MT), respectively, producing monomethylated and dimethylated forms of arsenic (DMA and MMA). Methylation of iAs helps facilitate excretion of arsenic in urine, with DMA making up the majority of arsenic species excreted. The rate at which individuals undergo this process is known as their arsenic metabolism efficiency (AME), which can be estimated through the measurement of arsenic species (iAs%, MMA%, and DMA%). Past studies have identified genetic variation in the AS3MT (10q24.32) and FTCD (21q22.3) regions through their associations with the proportions of each species present in urine, but any such association with arsenic species present in blood has not been examined. Here, we use data from the Health Effects and Longitudinal Study (HEALS) to both compare AME measurements in blood and urine through OLS regression and examine the association of previously-identified genetic variants on arsenic species in both urine and blood of 334 individuals in a GWA (genome-wide association) study. We confirm our hypothesis that all genetic variants known to affect arsenic species composition in urine (an excreted by-product of metabolism) are also present in blood (a tissue type that directly interacts with many organs, including those prone to arsenic toxicity). This consistency provides further support for the broader hypothesis that AME SNPs identified to date impact the efficiency of arsenic metabolism and elimination, thereby influencing internal dose of arsenic (and circulating metabolite levels), potentially impacting exposure across multiple toxicity prone organs.

Yohhan Kumarasinghe

Brandon Pierce, Biological Sciences, Pierce Lab; Lizeth Tamayo, Pierce Lab

Quad Faculty Research Grant Scholar

Biological & Health Sciences, Statistics

For video games, death is often what players use to measure difficulty. This paradigm—failure through death—works to incentivize the player to optimize their performance and privilege game completion. Despite these given conventions, some games have mitigated the disruption of “flow” brought about by death, from relaxed penalty (LIMBO) to non-fatal scenarios (the Pokémon franchise). My research explores a tension between these branching developments in death mechanics, specifically in games narratively aboutdeath that remove the mechanicalpossibility of the avatar’s demise. How does this design allow players to resonate with themes of failure, loss, and dying without offering its highest form—the avatar’s death? To investigate this tension, I exploit digital media theories of gaming difficulty, the paradox of failure, and the avatar as a player’s corporeal double. Then, I employ Stephen Curtis’s theory of deathseticsto outline a framework on how gamers confront game death narratives. Finally, I analyze existing archival records of gamers’ experiences with indie puzzle games Journey (2012) and Monument Valley (2014), investigating the ways in which deathseticspermeates their testimonies. I argue that these “deathless” systems allow players to understand death through its abstract narrativity, allowing players to reinscribe punitive measures into their gameplay. In turn, this evokes a hyper-awareness of avatar vulnerability and the real-life counterpart of their own human mortality. Beyond an alternative explanation for the relationship between game mechanics, narrative, and player agency, my research suggests the role of indie games as a space of “refuge” to contemplate anxieties over suffering, failure, and the inevitability of death—perhaps, in a way, that the corporate expectations of AAA games rarely offer.

Alina Kim

Jon Satrom, Cinema and Media Studies, Media Arts and Design

Media Arts and Design

Since its founding in 1833, The University of Chicago advanced a commitment to freedom of expression. Our team is producing freedom of expression case studies for high school and college audiences in collaboration with Professor Leila Brammer and the Parrhesia Program for Public Discourse. Focused on instances in which UChicago’s stance on freedom of expression was challenged, we analyzed historical documents from UChicago’s archives and relevant literature on UChicago’s history. We also studied archival records of student newspapers like The Maroon and national newspapers like The Chicago Tribune and The New York Times. In this specific case study, our team focused on controversy surrounding the 1949 Broyles Commission of the Illinois State Legislature, which targeted UChicago students, administration, and professors after members of a communist student group protested an anti-communist bill in the state legislature. The case examines questions about how best to protect student speech, as well as overarching themes of student censorship within academic institutions and the role of university officials not just to allow student speech but to encourage it. With the case studies produced by this research, audiences will examine the historical conflict itself as well as the response of the UChicago students, administration, faculty, and the community at large. Working through the case studies, students will analyze historical documents and relevant literature on the topics and deliberate about how they, in the role of students, faculty, or administrators, would respond in similar situations today. Through these instances in UChicago’s history, we hope to expose students to the principles of freedom of expression, work them through contestations, and develop capacities that foster productive discourse. In addition to the case study itself, we will craft an accompanying pedagogical guide.

Elisabeth Snyder, Anna Guzman

Leila Brammer, Parrhesia Program for Public Discourse

Quad Undergraduate Research Scholar

History, Humanities, Interdisciplinary Humanities

Three of the eleven extant plays we have from Aristophanes—the main playwright of the Attic Old Comic tradition whose work remains available to us—have women as their central characters: Lysistrata, Ecclesiazusae, and Thesmophoriazusae. Most well-known among them is his Lysistrata, but all three comedies contain rather transgressive portrayals of women, who are shown acting democratically: coordinating successful coups, holding meetings like politicians, and completely taking over the running of the state. In a preliminary exploration of the role of these portrayals, I engage largely in textual analysis, comparing the political competence of the plays’ women with the obvious incompetence of the men, and the latter’s superficial feminization with the ways the women are able to effect change through the preservation of their actual societal roles, leading to a view that these portrayals endorse female involvement in the political sphere. Straying somewhat from Aristophanes’ scripts, I then look at a broader range of sources, including tragedy, pottery, and philosophy (Plato’s Republic specifically) to explore the possibility that an overarching cultural shift regarding attitudes toward women was taking place around this time. However, I then pivot to a situation of these plays within their performed context: dramatic festivals of the 4th–5th centuries BCE. Sue Blundell argues that gender relations were commonly used as a means of exploring other divisions in Athenian drama, and Ober proposes that these are being used in these works to explore the validity of “social facts,” entailing a questioning of the structure of the polis and its separation from the oikos. I thus conclude that the inverted gender relations in these plays serve to highlight the polar nature of contemporary Athens’ city structure, condemning the state and management of the Athenian polis, and advocating for internal harmony between the interconnected political and domestic spheres.

Hannah Halpern

Ada Palmer, History

Quad Faculty Research Grant Scholar

History, Humanities, Classical Studies

Jacques Derrida famously defines “the movements of deconstruction” as a form of criticism that “do[es] not destroy structures from the outside” and “[is] not possible and effective, nor can [it] take accurate aim, except by inhabiting those structures.” This essay turns this nondestructive deconstruction into the lens for probing nonsovereign play, a nonnormative mode of gaming that challenges skill-based and control-centered conceptualizations of games, as well as for examining how the 2019 role-playing game Disco Elysium (ZA/UM) compels the player to confront nonsovereignty through both mechanics and narrative. Specifically, I compare the main features of Disco Elysium to those of resource-allocation action games, with the intention of demonstrating that Disco Elysium employs game features that often uphold player agency, only to encourage the player to embrace a future that is rarely under her control and to engage with a story that foregrounds nonsovereign experiences. With this case study, I argue that the countermovement against control-centered understandings of games could be a movement of deconstruction, destabilizing the traditional emphasis on player sovereignty by inhabiting the very structure of normative games. Ultimately, I propose that such a deconstructive approach may allow independent games to challenge normative modes of play without sacrificing the specificities of games as a rich, interactive medium.

Irene Li

Patrick Jagoda, Cinema Media Studies

Interdisciplinary Humanities

On subscription-based pornography platforms such as OnlyFans, performers become at once marketers of a product and the product itself. With private chatting, custom video requests, and the like available to anyone willing to pay, that product comes wrapped in a promise of digital intimacy, compelling performers to provide increasingly greater access to themselves in the pursuit of subscriber pleasure and financial security. The challenges of sex work are heightened for black women and trans people, who have been systemically devalued under racial capitalism. This work inquires into the professional branding strategies of queer, Black, dark-skinned creators of digital sexual content, from whom the desire economy mandates a performance of utmost authenticity. I seek to answer, “How do these performers find sanctuary amidst the violent visual schemas of gender and race upon which their desirability is negotiated?” What kinds of pleasure do they exact for themselves within the labor relation of consumer and consumed? Combining digital ethnographic methods with textual analysis, I elucidate the modes of resistance taken up by creators to the power dynamics of commodified erotic exchange. I posit mischief at the heart of it all; performers employ strategies of disidentification and humor in order to ‘play games’ with audience expectations as a means of self-preservation. This research was undertaken in service of my broader Fundamentals question, “What is sexy?” To that end, this cohort holds a wealth of embodied knowledge on the function of desire as an agent of power.

Jo Blankson

Eva Pensis, Music, Theater and Performance Studies; Malynne Sternstein, Fundamentals (Chair), Slavic Languages and Literatures

College Summer Institute Scholar

Gender and Sexuality Studies, Visual & Performing Arts, Digital Ethnography

Boasting one of the oldest maps in world history, Nuzi or modern Yorghan Tepe, Iraqi Kurdistan, appears an apt archaeological site for analysis of emic spatial concepts, sensory perception of space, and spatialized approaches to social interaction given that its ancient inhabitants, at least of the Old Akkadian period, clearly contemplated these variables themselves. Owing to its bountiful cuneiform corpus and widely exposed strata of the 15th to 14th centuries BCE, this later period of Nuzi’s history has indeed been the focus of studies in genealogy and geography as extracted from family archives’ economic and legal documents. However, the potential of this data, both archaeological and textual, to inform more complex spatial analyses of interaction and perception has not yet been realized. In my thesis research, I reconstruct and interpret the spatialpatterns of a unique form of social interaction – witnessing and being witnessed in the Nuzi legal system – by intersecting statistical social network analysis with the built environment and its “space syntax,” focusing on the archival and architectural record of the “Eastern Area.” It is widely assumed that the identity of witnesses is a simple function of convenient proximity to the witnessed – nearest neighbors. However, this assumption has not been questioned, much less proven, for Nuzi or more broadly. Who are the witnesses, and where do they come from? Were there strategic concerns in the selection of one’s witnesses which might override the (assumed) friction of distance? How and why do these actors diverge from simple least-cost expectations in the navigation of their social worlds? In addition, understanding how space intervenes in the socio-legal transactions and ties of inter-observer obligation, liability, and (in)security forged in such constructed spaces should allow for further inferences regarding the emic interpretation of spatially-dependent sensory and affective experiences of these contexts and their boundaries.

Maddie Ouimet

Susanne Paulus, Near Eastern Languages and Civilizations, The Oriental Institute

College Research Fellow, College Summer Research Fellow, Hoeft Undergraduate Research Awardee, Quad Faculty Research Grant Scholar, Beinecke Scholar, Neubauer Undergraduate Research Fellow

History, Humanities, Interdisciplinary Humanities, Digital Humanities, Archaeology, Assyriology

Independent documentary distribution is a complicated process with surprisingly high barriers to entry. While submissions to festivals themselves are inexpensive, many aspects, from the specific make of camera that was used to the subject itself, are fair game for dismissal when it comes to landing substantial deals with streaming services. In this work, we analyzed the tensions between production processes and distribution goals for a documentary series entitled “Curiosity: The Making of a Scientist.” The primary distribution goal of this series was to expose it to as broad an audience as possible. The series sought to expose non-scientists to the culture of science, so screenings in universities were insufficient. Due to this, it was important to research distribution possibilities such as local theatrical screenings, applications to film festivals, and conversations with small producers and distributors that could help the films slowly build up acclaim. The range of “out-of-box” solutions investigated will be discussed, particularly in reference to the tension between reaching a large audience and ensuring that the project remains a student-made enterprise.

Meira Chasman

Sunanda Prabhu-Gaunkar, Pritzker School of Molecular Engineering, University of Chicago

Interdisciplinary Humanities, Physics, Visual & Performing Arts

A visitor’s experience in a museum gallery is determined by the artworks on display just as much as how and why they are displayed. These considerations constitute the basis of curatorial praxis, broadly defined as 1) the scholarship, ethics, beliefs, and critiques that shape and inform decisions to exhibit, acquire, or deaccession objects and 2) the practical labor of researching, interpreting, displaying, and assembling objects. My research as one of the 2020-2021 Curatorial Research Associates at the Smart Museum of Art culminated with my first curatorial project, Ambiguous Beauty, a thematic grouping of fifty-six artworks from the Smart’s permanent collection that engaged with queer aesthetics, relation, and embodiment. Central to this project was the notion of queerness as that which refuses meaning, especially with regard to beauty, aesthetics, and relations of becoming. As I return to the same artworks in preparation for my thesis, I am developing a critical framework in response to this notion, provisionally named “the tactile gaze,” as a way of seeing that is based on Levinas’ metaphor of the caress and phenomenologically closer to touch. My presentation will offer a description of the gaze’s characteristics, its conditions, and its ethical stakes. I will then provide a case study of how the gaze may function as curatorial praxis by revisiting a number of artworks in Ambiguous Beauty, critiquing their existing wall labels, as well as the introductory label of the whole grouping, and offering alternative interpretations based on this critical framework. This presentation is part of my thinking towards a queer ethics of seeing and an ongoing inquiry into the relation between trans bodies and meaning.

Seth Nguyen

Orianna Cacchione, Department of Art History, Smart Museum of Art at the University of Chicago

Gender and Sexuality Studies, Humanities, Curatorial Studies

Digital capitalism posits us as users embedded in everchanging realities. Our research traces the aesthetic genealogy of what we call Z Revival, an attempt to reconnect with a lost future promised to us by the technology of our youth. Today, young artists are revisiting what dominates our memory of the years 2008-2014, the futurepast of Miis, Cable TVs, iPhone 3Gs, and Black-Eyed Peas. With dead malls and the subsequent loss of public space as our origin point, we survey vaporwave, the hauntology of liminal spaces, and cottage core, as precursors to this burgeoning aesthetic movement. Nostalgia for a time when our interactions with technology were seemingly benign and explicable is to be expected when in the present, we know they’re not. The cultural weight placed on individual aesthetic choices reveals something pathetically misled about our search for agency and simultaneous abundance of such online. Our research aims to answer the questions: Can collective nostalgia be forged as a weapon against capital? And what is so political about rejecting high definition? Internet aesthetics are cathartic expressions of the cultural residue sticking to the walls of our social psyche; their deconstruction is a growing area in media research. Just because it’s happening online doesn’t make it any less real.

k80 ambrose; Navid Mazidabadifarahani

Nick Briz, Media Arts and Design

Humanities, Media Studies

Earth’s core is mostly made of iron, but the seismologically determined densities of the solid inner core and liquid outer core do not match that of iron under core conditions. Lighter alloying elements are, therefore, likely present. One possible candidate is sulfur, since it readily alloys with iron at high temperatures and pressures and is found in iron-rich meteorites thought to originate from the cores of planetesimal building blocks of terrestrial planets like Earth. Previous experimental investigations indicate that sulfur likely exists within the core as Fe3S. It is critical to understand the density-pressure-temperature relationships of Fe3S at the pressure-temperature conditions of Earth’s core to constrain a possible composition that resolves the density deficit. In this study, we determined the room temperature equation of state (EoS) of Fe3S using single-crystal X-Ray diffraction techniques. We then constructed a thermal EoS using the room temperature EoS combined with high P-T data collected usingpowder X-ray diffraction techniques. The Fe3S thermal EoS, and that of iron, were compared to the density of Earth’s liquid outer core. Our results suggest that 16 wt% S is required to match the density profile of Earth’s outer core, given sulfur as the sole alloying light element.

Abigail Case

Andrew Campbell, Geophysical Sciences

College Summer Research Fellow, DAAD RISE Scholar

Geophysical Sciences

Quantum sensing promises to create new methodologies to uncover previously inaccessible information in many scientific fields. One interest in cellular biology is the heat shock response, where a cell can adapt to proteins denaturing in a cell due to temperature and restore their function. The components of the process are not well understood, and current methods that heat clumps of cells cannot distinguish between extracellular and intracellular protein regulation. To better describe this biological phenomenon that occurs in all known organisms, we propose quantum sensors, specifically nanodiamonds, to sense temperature change at a small spatial resolution, smaller than any cell. A small quantum sensor, nanodiamonds read local temperature fluctuations well at room temperature and in a cellular matrix. However, these nanoparticles drift rapidly and randomly within a living cell, making them challenging to help with biological applications. This research aims to create a consistent temperature modulator that withstands the drift of nanodiamonds. By synthesizing novel temperature reading schemes and spatial tracking algorithms, temperature fluctuations will be read on a fast timescale to allow real-time quantitative measurement of the temperature to understand the intracellular response to protein degradation.

Aidan Jones

Peter Maurer, Pritzker School of Molecular Engineering, Maurer Lab; Uri Zvi, Maurer Lab

Quad Undergraduate Research Scholar

Engineering, Physics

Phase stability between femtosecond pulses is critical for extracting accurate spectral lineshapes to then derive short-time dynamical information of chemical and biological systems. Current methods for measuring our pulses use an autocorrelation technique called Transient Grating - Frequency Resolved Optical Gating (TG-FROG) that lacks the ability to retrieve complete and correct phase information of the pulses. Recently, we have encountered spectral artifacts that usually result from non-optimal pulse compression and TG-FROG, but have not yielded satisfactory results in detecting the compression problem. I will apply image deconvolution algorithms developed by the Trebino Group and determine the parameters needed to measure the pulse intensity, phase, and stability of the pulse train of our sub-10 fs pulses. FROG works by measuring the signal spectrum as a function of the time delay. This spectrogram is created by gating the pulse field with a clone pulse, and the pulse shape can then be mathematically retrieved from the spectrogram. Applying this algorithm will enable us to understand our pulse completely and subsequently dictate ideas to solve the compression problem. Beyond using FROG to understand the complexities in our pulse shape, I also built a box for our two-dimensional electronic spectroscopy (2DES) setup as well as an aluminum plate for the addition of a pair of chirped mirrors. The box will aid in reducing phase drift that results from environmental instabilities and the chirped mirrors are used for compressing the pulse in time. Ultimately, combining the FROG algorithm with these physical alterations allows us to generate more stable pulses as well as accurately measure their phase shape.

Ainsley Iwanicki

Greg Engel, Chemistry; Indranil Ghosh, Chemistry

Quad Undergraduate Research Scholar supported by the NK Cheung Chemistry Research Fellowship

Chemistry

Binary systems are a common byproduct of the stellar evolution process, but much about their formation remains poorly understood. Of particular interest is the relationship between the close binary fraction and metallicity, which offers insight into a variety of important processes, including accretion, proto-binary fragmentation, type Ia supernovae progenitor rates, and the chemical evolution of galaxies. This relationship has been continuously debated over the last three decades, with several independent studies reaching conflicting results. However, recent work by Moe et al. (2019) appears to have made significant strides towards resolving this discrepancy by re-examining five prior studies on this topic. By applying additional corrections for incompleteness, Moe et al. showed that all five studies could be made to agree on a single conclusion: the close binary fraction is strongly anti-correlated with metallicity. In this work, we present a test of Moe et al.’s hypothesis based on data from the Magellanic Clouds. As the average metallicity of the Large Magellanic Cloud (LMC) is roughly two-and-a-half times the average metallicity of the Small Magellanic Cloud (SMC), our test consists of a straightforward comparison of the close binary fractions in each galaxy. Approximately 380,000 light curves from each galaxy were transformed into two-dimensional images that represent information about the shape of the light curves via “dm-dt” mappings as described in Mahabal et al. (2017). These images were then fed into a convolutional neural network (CNN) that had been trained on a combination of simulated images and a small set of real, identified images. The CNN was able to identify binary objects with an accuracy rate of approximately 94%. After additional visual verification of the final binary sets, the binary fractions in both galaxies were computed and compared. We will report preliminary results on the relative LMC/SMC binary fraction.

Alexandra Masegian

Michael Shara, Astrophysics, American Museum of Natural History; David Zurek, Astrophysics, American Museum of Natural History; Eve Armstrong, Astrophysics, American Museum of Natural History

Astronaut Scholar, Goldwater Scholar

Astronomy and Astrophysics

In this project, we demonstrate the ability of a commercially available graphic arts scanner to produce scientifically useful scans of astronomical photographic plates. We describe a method to extract accurate magnitude measurements from the star images on sky survey plates, which are stored in observatory archives around the world. The affordability of our approach increases the potential of astrophysical research using measurements from up to over a century ago. We detail the use of a star-focused method of measurement as well as a novel approach to measurement calibration utilizing galaxies on plates taken by E. E. Barnard in 1905 and 1911 at Yerkes Observatory. Additionally, we examine the effects of our scanning method on our magnitude measurements and present a case study of nine red supergiant stars appearing on a chosen Barnard plate. Finally, we discuss our approach to future projects, including our investigation of digitization differences in commercial scanners and digital cameras, as well as the methodology behind our ongoing analysis of the Nova Herculis photographic plates.

Audrey Scott, Rowen Glusman, Isaiah Escapa

Dr. Richard G. Kron, Department of Astronomy and Astrophysics

Astronomy and Astrophysics

In efforts to create a sustainable economy and mitigate the effects of climate change, electrocatalytic reduction of CO2 offers a potential tool for the production of high-demand chemicals such as carbon monoxide and ethylene. CO2 catalysis in non-aqueous media is an interesting alternative to aqueous catalysis, as it minimizes the competing hydrogen evolution reaction (HER). To date, the state-of-the-art alkali salts used as electrolytes for selective CO2 reduction in aqueous media are unusable in non-aqueous systems due to the inactivation of the electrode surface. This is attributed to the carbonate production side reaction which causes most modern electrolysis methods to suffer from low energy and carbon efficiencies. It was found that carbonate passivation through precipitation with alkali metals has been preventing the use of these key electrolytes and expanding the non-aqueous electrocatalysis space. Thus, solving the issue of carbonate passivation is necessary to expand the number of available electrolytes and continue to study the system dynamics for further optimization. In this work, we demonstrate how an acidic non-aqueous media enables the sustained electroreduction of CO2 with common alkali salts in DMSO. This method was applied to both a Cu and an Au working electrode demonstrating its wider applicability. Electrochemical experiments have shown that conducting CO2reduction in acidic media allows carbonate to be reformed to CO2 preventing any build-up of nonvaluable side products, while reaching faradaic efficiencies up to 80% for formic acid and carbon monoxide. Thus, an acidic, non-aqueous electrolyte could increase energy and carbon efficiencies compared to previously reported setups, while maintaining high selectivity toward valorized products.

Benjamin Kash

Dr. Chibueze Amanchukwu, Pritzker School of Molecular Engineering

Astronaut Scholar

Chemistry, Engineering

Usage-Based Insurance (UBI) is a form of vehicle insurance in which costs are determined based on the type of vehicle, measured against time, distance, driving behavior and location. The data for UBI is collected by the driver’s vehicle and analyzed by the insurer, who then uses it to determine how much of a discount the driver is eligible to receive. The goals of UBI are to enable insurers to promote safer driving behavior, reduce the frequency and magnitude of claims and help improve costs and revenues. In order to reach these goals, the idea is to leverage Machine Learning (ML) methods to classify driving behaviors at different risk levels. To classify driving behavior, we first apply a Convolutional Neural Network (CNN) model to a time series of simulated GPS data for different drivers. Under the assumption that high-risk/low-risk drivers always exhibit high-risk/low-risk driving behavior, we observed a 96% classification accuracy. In order to account for a mixed distribution of driving behavior, we leverage a Gaussian mixture model within the Variational AutoEncoder framework (GMVAE), the results of which are still on-going research.

Blaise Munyampirwa

Rebecca Willett, Computer Science; Willem Marais, UW-Madison

Quad Undergraduate Research Scholar supported by the Liew Family Research Fund

Computing Science, Statistics

Thermobaric convection, or convection that results from the increasing thermal expansivity as pressure increases, likely contributes substantially to high-latitude ocean convection and associated deep-water formation. However, the effects of thermobaric convection remain poorly constrained. To evaluate the role of thermobaric convection in the Southern Ocean (and, therefore, the role of thermobaracity in the formation of Antarctic Bottom Water), Ocean Convective Available Potential Energy (OCAPE), the potential energy arising from thermobaracity within a conditionally stable ocean column, is diagnosed in simulations using the MITgcm coupled ice-ocean model. As found in other models, and consistent with some observational evidence, Antarctic Bottom Water formation in these simulations is highly variable, showing oscillatory behavior at multiple frequencies. We find that OCAPE in the Southern Ocean is high while convective activity is lowest while being depleted rapidly once deep convection sets in. The build-up of OCAPE seems to arise due to the advection of anomalously salty and warm water at depth. Hence, OCAPE accumulation appears to play an important role in the variability of bottom water formation and provides a potential predictor of when deep convection will occur.

Cal LeDoux

Malte Jansen, Geophysical Sciences

Quad Faculty Research Grant Scholar

Geophysical Sciences

Competitive systems can exhibit both hierarchical (transitive) and cyclic (intransitive) structures. Despite theoretical interest in cyclic competition, which offers richer dynamics, and occupies a larger subset of the space of possible competitive systems, most real-world systems are predominantly transitive. Why? Here, we introduce a generic mechanism which promotes transitivity, even when there is ample room for cyclicity. Consider a competitive system where outcomes are mediated by competitor attributes via a performance function. We demonstrate that, if competitive outcomes depend smoothly on competitor attributes, then similar competitors compete transitively. We quantify the rate of convergence to transitivity given the similarity of the competitors and the smoothness of the performance function. To test that theory, we ran a series of evolution experiments designed to mimic genetic training algorithms. We considered a series of canonical bimatrix games and an ensemble of random performance functions that demonstrate the generality of our mechanism, even when faced with highly cyclic games. We varied the training parameters controlling the evolution process, and the shape parameters controlling the performance function, to evaluate the robustness of our results. These experiments illustrate that, if competitors evolve to optimize performance, then their traits may converge, leading to transitivity.

Christopher Cebra

Alexander Strang, Statistics

Statistics

Pooling and sharing data increases and proliferates its value, but often data organizations are hesitant to share for regulatory, privacy, and legal reasons and because data cannot be revoked once shared. Take, for example, hospitals: improved care and immense research value is gained by pooling patient data, yet it is even more crucial to control its release. Therefore many organizations that could benefit from sharing do not, and the few exceptional scenarios are built around agreements resulting from long and tedious one-off negotiations. Worse, because the organization cannot know the exact contents of the data beforehand, after the exchange the data is not guaranteed to be useful. This arrangement thus induces a significant upfront risk and cost of gaining value from others’ shared data. We introduce the Data Station, a system designed to enable the storing, sharing, and executing of computation on data. Data owners send data to the Data Station knowing it will not be released without their consent. Data usersassign computations to run on the Data Station, which potentially leverages multiple datasets from various owners. Computation is done within the Data Station privately and securely, allowing data processing before any release of data. This scheme bypasses the upfront cost of sharing data until the user knows the computed products are useful; afterwards, the sharing agreement can be decided. The Data Station leverages hardware enclaves, a new and promising technology, to establish trust that owners’ data and computed products are private unless shared correctly and explicitly.

Christopher Zhu

Raul Castro Fernandez, Computer Science, ChiData

Computing Science

NOTE: This abstract contains mathematical notation which may not render correctly below. To view the properly formatted abstract as a PDF, click here.

Cosmological analyses with type Ia supernovae (SNe Ia) typically assume a single empirical relation between color and luminosity (β) and do not account for varying host-galaxy dust properties. However, from studies of dust in large samples of galaxies, it is known that dust attenuation can vary significantly. Here we take advantage of state-of-the-art modeling of galaxy properties to characterize dust parameters (dust attenuation Av, and a parameter describing the dust law slope Rv) for the Dark Energy Survey (DES) SN Ia host galaxies. Utilizing optical and infrared data of the hosts alone, we find three key aspects of host dust that impact SN Ia cosmology: 1) there exists a large range (from ~ 1 - 6) of host Rv; 2) high stellar mass hosts have Rv on average ~ 0.7 lower than that of low-mass hosts; 3) there is a significant (> 3σ) correlation between the Hubble diagram residuals of red SNe Ia that, when corrected for, reduces scatter by ~ 13% and the significance of “the mass step” to ~ 1σ. These represent independent confirmations of recent predictions based on dust that attempted to explain the puzzling ‘mass step’ and intrinsic scatter in SN Ia analyses. We also find that red-sequence galaxies have both much lower and more peaked dust law slope distributions (Rv = 1.8, σRv = 0.5) in comparison to non red-sequence galaxies (Rv = 3.2, σRv = 0.9) and that when analyzing the DES sample of spectroscopically confirmed SN Ia light curves separately for those in red-sequence hosts versus all other hosts, we find that the SN Ia β and σint both differ by > 3σ (red sequence: β = 2.09 ± 0.15, σint = 0.05 ± 0.015; non red-sequence: β = 2.71 ± 0.11, σint = 0.10 ± 0.007). This agreement between fitted host-Rv and SN Ia β & σint suggests that SN Ia color-luminosity standardization is driven by host dust properties and supports the claim that SN Ia intrinsic scatter is driven by Rv variation.

Cole Meldorf

Alex Drlica-Wagner, Department of Astronomy and Astrophysics; Antonella Palemse, Department of Astronomy and Astrophysics

Astronomy and Astrophysics

Heat shock proteins (Hsps) are molecular chaperones that maintain cellular protein homeostasis by promoting protein folding and dissolving aggregated proteins during cellular stress, such as heat shock. The interaction of Hsp104 with protein aggregates has been previously demonstrated at the single-molecule level using a nonnative test system, firefly luciferase. A more recent study by our collaborators has shown that heat-induced Pab1 assemblies, which form in direct response to heat shock, are much more efficient substrates for Hsps than nonnative luciferase. Therefore, we expect to gain novel insight into physiological Hsp function through Pab1. To investigate Hsp104-Pab1 interactions at the single-molecule level, we aim to develop a single-molecule assay that enables us to (1) immobilize Pab1 aggregates on a glass slide and observe them over extended durations (seconds-minutes) with high fidelity, and (2) observe Hsp104 interacting with Pab1 in the absence of confounding factors, such as nonspecific protein binding to imaging chamber surfaces. We aim to combine specific anchoring of a target molecule through a biotin-avidin interaction with nonspecific repulsion of the interacting species by a polyethylene glycol (PEG) layer. So far, I have demonstrated that treatment of PEG coating on glass slides is able to consistently repel binding of heat shock proteins. Repeated washes of the protein solution in the imaging chamber further reduce protein binding. The next step is to treat the slides with PEG-biotin and the target molecules with avidin, to explore the effect of the biotin-avidin interaction on specific binding of the target molecules onto the glass slide. The assay will ultimately enable us to discover how Hsps disperse a native, near-physiological substrate, heat shocked Pab1.

Dana Lin

Allison Squires, Pritzker School of Molecular Engineering, University of Chicago; Kyle Lin, Graduate Program in Biophysical Sciences

Quad Undergraduate Research Scholar

Biological & Health Sciences, Chemistry, Physics

Storage, the home of Big Data, has grown enormously over the past decade. As data is continuously generated and persisted at large scales, storage performance is critical for many applications. However, evidence from industry reveals that storage systems still suffer from high latency problems–above 50 milliseconds per 4 KB I/O over long durations–and drive failures, many of whose causes remain inexplicable due to the complexities of modern, multi-leveled storage architectures. In this project, we employ a white-box approach, using large data sets, comprising approximately 1,000,000 disks, and data science methodology to study high latency and performance failures in storage systems. We are interested in answering the question, “Is performance failure highly correlated with any anomaly definitions or data features from a given system?” Anomaly definitions refer to values of combinations of performance features that are likely to appear when a disk becomes anomalous with respect to itself and others in a RAID group. We develop two approaches to define anomalies, namely “Pairwise Correlations” and “High-Dimensional Anomaly Detection.” In the former, we examine pairwise correlations of the 14 available data features, determine the anomaly values ranges, and apply a clustering or outlier detection model–K-means or DBSCAN–to automatically identify the anomalous region. In the latter, we use a dimensionality reduction method–PCA or t-SNE–and input all 14 dimensions of data to separate outlier values and identify them using K-means or DBSCAN. We anticipate that, by correlating these anomaly definitions with disk failures, we will be able to provide statistics illuminating certain behaviors of a drive before it fails. Not only will this knowledge be useful for potential future studies on disk failure prediction, it can also enable storage experts to pre-emptively handle anomalous drives and avoid system unavailability.

Dang Nguyen

Haryadi Gunawi, Computer Science, University of Chicago UCARE Lab

Quad Faculty Research Grant Scholar

Computing Science

Within the framework of Einstein’s theory of gravity, known as general relativity, there is a notion of angular momentum of gravitational radiation for asymptotically flat spacetimes. Such a quantity is of interest for the study of gravitational waves and black hole mergers; however, it is not obvious how to define the angular momentum of a system. There is a widely-accepted definition known as the Dray-Streubel angular momentum, first proposed more than 30 years ago. Recently, though, due to this ambiguity, there have been numerous alternative definitions put forth and used in the literature. Hence, the goal of this research was to analyze these new definitions to better understand if they are well-defined. When a definition for a quantity, such as the angular momentum of a system, is proposed, it should fulfill several desirable properties to be meaningful. One criterion is that the difference in the angular momentum at different times should be given by the integral of a well-defined flux. We showed that this is not possible for one of the new definitions. Current work is ongoing to show that a second recently proposed definition also does not admit a well-defined flux. Therefore, this research shows that the Dray-Streubel angular momentum is the most useful definition for the angular momentum of gravitational radiation to be used for other studies.

Daniel Paraizo

Robert Wald, Physics, Enrico Fermi Institute

Quad Undergraduate Research Scholar

Physics

The structure and evolution of dark matter halos encodes important information about the fundamental questions in cosmology, like the nature of dark matter, gravity and the connection between observed galaxies and their halos. Analytic, idealized simulations of self-similar halo collapse predict caustic-like, localized enhanced density features through the dark matter distribution in these objects. I study the distribution of dark matter and its evolution in the full phase space structure of halos, to see how transient features like caustics, which are predicted in these analytical models, evolve in N-body simulations and if they can be observed in data. While the outermost caustic associated with the splashback radius and its effect on the halo profile has been studied in some detail over the last few years, there is evidence for a second feature arising at a smaller halo-centric radius. The focus will be on this inner, second caustic-like feature and on when they occur in realistic halos. Correlations will be studied in the phase space structure with the accretion histories of dark matter halos to see how different phases of slow (fast) accretion and mergers affect the evolution of this feature.

Diego Garza

Andrey Kravtsov, Astronomy & Astrophysics, KICP

Astronomy and Astrophysics

Machine learning (ML) has been identified as a powerful tool in a variety of fields from trend discoveries in large databases to selective processing of changing patterns. It is the development and utilization of computational algorithms that improve and adapt spontaneously by increased experience based on data. Battery demands have increased with the development of sustainable energy, yet traditional experimental methods are insufficient to meet such demands. Thus, ML aided predictions are critical for the accelerated development of novel batteries, while a systematic review of ML approaches is necessary to facilitate its applications. Herein, two functioning schemes (first-principle oriented and system-modeling oriented) of ML are reviewed, and common ML models utilized in each scheme are presented, including artificial neural networks (ANN), kernel methods, and linear fitting. Additionally, ensemble ML models that incorporate multiple learner models are introduced and categorized as bagging, boosting, and stacked generalization, each with enhanced adaptations. Applications of multiple ensemble ML models in battery research are discussed on the state of health and residual capacity predictions for Li-ion batteries. Furthermore, metal-organic framework (MOF) materials have been found to improve electrochemical properties in batteries, but discreet MOF selection is required for optimization, which could be aided by ML. Yet challenges for ML applications are imposed by MOF due to the structural complexity, limited data availability, and lack of appropriate ML models. Thus, novel ML solutions for MOF materials are proposed and categorized with regards to the choice of descriptors, learning mechanisms, and ML model reconstruction. Applications of ML models with distinctly combined solutions are discussed including composition optimization for Li-MOF and metallicity and potential energy approximation for MOFs. Such ML functioning schemes, ensemble ML models, and novel solutions for ML challenges provide valuable insights into future ML development and potential ML applications in expanded fields.

Eric Jiahan Zhao

Professor Junhong Chen, Pritzker School of Molecular Engineering, The University of Chicago; Dr. Haihui Pu, Pritzker School of Molecular Engineering, The University of Chicago

Chemistry, Computing Science, Engineering

Hybrid quantum networks are an area of great interest. In the Bernien Group’s nanophotonic quantum network design, single atoms are optically coupled to nanophotonic crystal cavities, combining the advantages of atomic high coherence and fiberoptic-based telecom networks. At one network node, atoms are excited with precise laser pulses and emit photons in the telecom band. This platform can be used to create and distribute entanglement via optical fiber over long distances across the network. On another node of the quantum network, memory nodes are required to store quantum information. In this poster, I will report on two homebuilt cat-eye style external cavity diode lasers and their application in our quantum network project. The cat-eye style laser offers single-mode selection via an interference filter and “cat-eye” retroreflector. One laser, “BrightPheonix” at 935nm is used to make optical tweezers for “magic” trapping of Caesium atoms near nanophotonic cavities in a vacuum chamber; optical tweezing is an exciting phenomenon in which single atoms are held in place by focused laser beams and is a key ingredient in the experimental setup. The second laser, “TeleRare” at 1530nm, is intended for spectroscopy on 166Er3+:7LiYF4, a rare-earth quantum memory candidate in the telecom band. I will present the laser design, performance with robust single-mode behavior, tunability over several wavelengths, and successful trapping of atoms, as well as further directions for the quantum network project.

Haley Nguyen

Hannes Bernien, Pritzker School of Molecular Engineering; Yuzhou Chai, Pritzker School of Molecular Engineering; Noah Glachman, Pritzker School of Molecular Engineering

Quad Undergraduate Research Scholar

Physics

Perovskite quantum dots (QDs) are a promising area of study for their various potential optoelectronic applications, including photovoltaics for solar energy. Previously, a reliable method to synthesize stable luminescent cesium bismuth iodide QDs has been obtained. The goal of this project is to show the robustness of the synthesis method by expanding it to cesium bismuth bromide QDs, so the established synthesis is being optimized to similarly structured quantum dots with bromide now as the halide. The synthesis method involves mixing the salts CsBr and BiBr3 in the non-polar solvent toluene under room or elevated temperature conditions in the presence of the ligands oleylamine and oleic acid. The optimization is being accomplished by varying the concentration of the salts and ligands, the temperature of the reaction, and the stirring time. To characterize or analyze the results of the synthesis, linear absorption techniques such as UV-Vis and Photoluminescence spectroscopy are being used. Conventional perovskite quantum dots contain lead, which is toxic and not environmentally friendly, so there has been significant research done to replace lead with nontoxic metal elements. However, when lead is substituted in these perovskites by some non-toxic element, the photoluminescence quantum yield is decreased. The goal of this project is to create samples for probing with nonlinear spectroscopy techniques such as Two-Dimensional Electronic Spectroscopy (2DES), which is the focus of the Engel Group. This method will allow us to directly measure hole transport and optical response from these novel materials.

Hugh Cairney

Greg Engel, Chemistry, Engel Group; Coco Li, Engel Group

Quad Undergraduate Research Scholar

Chemistry

Laser cooling and imaging of cold atoms rely on the absorption of photons whose frequency is in resonance with the atoms. The efficiency of cooling and imaging is often limited by the coherence and linewidth of the light used to excite the atoms. Developing ultra-stable, narrow-linewidth laser sources is therefore important for high-fidelity experiments. In this poster, we present narrow-linewidth 852 nm external cavity diode lasers (ECDLs) in the cat-eye configuration. The cat-eye geometry uses the idea of optical feedback to give narrower laser linewidth and higher intensity. It also has advantages over other ECDL configurations due to its increased mechanical stability. We observe a strong correlation between laser feedback and spectral linewidth and report narrow linewidths of less than 70 kHz and power outputs of up to 70 mW. In addition, we present a feedback scheme based on a phased-locked loop circuit to optimize the coherence of the cat-eye lasers. Using this scheme, we achieved a locking speed of less than 100 μs. Upon better characterizing the lasers’ linewidths in real experiments, the cat-eye lasers will be integrated into the Quantum Matter Synthesizer experiment for imaging Caesium atom clouds at single-atom resolution and in quantum simulation experiments.

Huiting Liu

Cheng Chin, Department of Physics, James Franck Institute, Enrico Fermi Institute; Jonathan Trisnadi, Chin Lab; Mingjiamei Zhang, Chin Lab

Quad Undergraduate Research Scholar supported by the Liew Family Research Fund

Physics

The development of large-area micro-channel-plate-based photodetectors (MCP-PMTs) opens opportunities for many applications that would benefit from high-precision 4D imaging. In such systems, the electronic channel count is a major cost driver. Incorporating a capacitively coupled anode attenuates the cost problem through external pickup electrodes with patterns of individual channels optimized for occupancy, rate, and time/space resolution for each application while using a shared photodetector design. The implementation is both economical and flexible in its applications, where the signal pickup antenna can be economically implemented as a printed circuit card with a 2D array of pads for high-occupancy/high-rate applications such as in particle colliders and medical imaging, or a 1-dimensional array of strips for a lower channel count in low-occupancy/low-rate applications such as large neutrino detectors. In 2021, we proposed pad patterns that enhanced signal-sharing between pads to substantially lower the channel count per unit area in large-area systems, while maintaining sub-millimeter spatial resolutions. Patterns that group non-contiguous pads using multiple signal layers can lower the channel count even further and thus further reducing cost by moving the scaling behavior in the number of pads from quadratic to linear. Several anode designs using strips and pads have been implemented as multi-layer PC boards and are under testing.

Jacky Li

Henry J. Frisch, Physics, Enrico Fermi Institute

Quad Undergraduate Research Scholar

Engineering, Physics

Traditional methods of disseminating scientific results are inherently opaque and inaccessible to the general public. A visually stimulating medium, such as documentaries, could close the gap between the public and provide higher accessibility to science as well as scientists. Documentary films can incorporate several techniques such as color, light, and sound to enhance the content that is being presented. We explore several methods for color representation in documentaries, especially related to scientific environments, such as laboratories. Representing colors in laboratories presents several challenges for filmmaking, due to lighting that is built into those locations. In particular, lab lightings are shadowless, meaning the amount of luminance is uniform over all surfaces. While such standards ensure safety, the resulting look lacks dynamic range and contrast, making the environment less appealing on camera. We present several methods that we employ in the production and post-production phases of filmmaking to obtain and enhance color representation to help the audience visualize the scientific settings and to elicit emotions. We also provide recommendations for the early stages of the filmmaking process, such as cinematography, to assist with the later stages of the filmmaking process and to provide ample room for experimentation during the post-production editing. Although initially developed for lab settings specifically, the color grading process has implications for a wider range of settings outside the lab as well. With this research, we want to reveal scientific research and the lives of scientists in a relatable manner for the general public.

Jason Chen, Maia Driggers

Nancy Kawałek, Pritzker School of Molecular Engineering, STAGE Lab; Sunanda Prabhu-Gaunkar, STAGE Lab

Fine & Performing Arts, Visual & Performing Arts

Advancements in silicon solar cell efficiency have plateaued in the last two decades. This is primarily due to poor external quantum efficiency in the blue and ultraviolet (UV) region. Ytterbium-doped inorganic halide perovskites have garnered attention for applications in high-efficiency solar cells, as they exhibit quantum cutting (QC), generating two near-infrared (NIR) photons (1.25 eV) from each blue or UV photon absorbed at energies >2.5 eV. Such materials address the current deficiencies in solar cells by shifting blue and UV photons to the NIR region, where silicon solar cells are more efficient. Current leading QC materials utilize lead, which is toxic, and the photoluminescence quantum yield (PLQY) of these materials decreases at high photon fluence, thus developing lead-free QC materials is of interest. We deposited ytterbium-doped Cs2AgBiX6 (X= Cl, Br) thin films using physical vapor deposition. We investigated the effects of precursors’ fluxes, substrate temperature, and post-deposition annealing conditions on the films’ optical properties. Structural and optical properties of perovskite films were characterized with X-ray diffraction, Raman spectroscopy, absorption, and photoluminescence. We synthesized ytterbium-doped Cs2AgBiBr6 thin films with a PLQY of 47.5%, which is a significant improvement over the highest previously reported PLQY of 28% from lead-free materials.

Joseph Geniesse

Eray Aydil, Chemical and Biomolecular Engineering, NYU Tandon; Minh Tran, NYU Tandon

Dean's Fund for Undergraduate Research CONFERENCE Awardee

Chemistry, Engineering

We have proposed the use of PET scanners using low atomic number media that undergo a persistent local change of state along the paths of the Compton recoil electrons, which significantly lowers the required dose, while providing higher resolution, by identifying the gamma trajectories with higher precision than traditional detectors. We propose an implementation using photoswitchable molecules of the diarylethene family. One possible mechanism for enhanced imaging of Compton-scattered electron tracks is that ionization-driven excited states in the solvent are transferred to a solute triplet sensitizer (TS). In this model, the excited TS converts short-lived singlet excitations into long-lived triplet excitations with a high quantum yield via intersystem-crossing. The triplet excitations are sufficiently long-lived for the TS to find and transfer the excitation to a dilute photoswitchable molecule. The molecule uses the excitation to switch to its fluorescent form, it and can be optically excited multiple times to recover its location. The molecule can then be switched back with visible light. We are investigating this excitation pathway with X-rays and optical excitation and have proposed experiments with ionizing radiation at Fermilab.

João Shida

Henry Frisch, Physics; Allison Squires, PME; Patrick La Riviere, Radiology

College Research Fellow, College Summer Research Fellow, Dean's Fund for Undergraduate Research Awardee, Liew Family Research Fellow, Quad Undergraduate Research Scholar

Physics, Photochemistry

In the realm of total synthesis, indole alkaloid natural products are a well-known and long studied class of targets, due to their proven bioactivity and interesting structural features. Picrinine is a highly complex alkaloid, and its cage-like structure makes it a formidable synthetic challenge. Herein, we present our progress for the total synthesis of picrinine, an indole alkaloid isolated from the tree alstonia scholaris. Using a chiral pool strategy, a Pictet-Spengler reaction is first used to furnish a 4H-β-carboline. This is followed by a gold-mediated 6-endo-dig cyclisation, and an oxidative Barton decarboxylation which establishes the core of the molecule. Finally, the last cyclisation is achieved using tin radical conditions, thus obtaining the desired caged scaffold. Although this molecule has been synthesised twice before, the route proposed herein would be both more efficient, as well as potentially allow divergent access to various other indole alkaloids not yet synthesised.This synthesis can be particularly useful for medicinal purposes, since picrinine has exhibited activity as a moderate central nervous system depressant. Additionally, synthetic analogues accessible from this route can be developed to modify a compound’s functionality and provide further medicinal use.

Julien Lee Heberling

Scott A. Snyder, Chemistry Department; Russell Kielawa, Chemistry

Quad Undergraduate Research Scholar

Chemistry

Future improvements in particle accelerator performance are predicated on increasingly accurate, online modeling of accelerators. Hysteresis effects in magnetic, mechanical, and material components of accelerators are often neglected in online accelerator models used to inform control algorithms, even though reproducibility errors from systems exhibiting hysteresis are not negligible in high precision accelerators. In this work, we combine the classical Preisach model of hysteresis with machine learning techniques to efficiently create non-parametric, high-fidelity models of arbitrary systems exhibiting hysteresis. We demonstrate that our technique accurately predicts hysteresis effects in physical accelerator magnets. We also experimentally demonstrate how these methods can be used in situ, where the hysteresis model is combined with a Bayesian statistical model of the beam response, allowing characterization of hysteresis in accelerator magnets solely from measurements of the beam. Furthermore, we explore how using these joint hysteresis-beam models allows us to overcome optimization performance limitations when hysteresis effects are ignored.

Kabir Dubey

Young-Kee Kim, Physics

Computing Science, Physics, Statistics

Transportation and mobility contribute, on average, nearly one-third of the annual greenhouse gas (GHG) emissions. To address detrimental health side effects of GHG emissions and climate change, many cities promote focus on reducing emissions from transportation using shared micromobility, like rental services for compact lightweight vehicles with low maximum speeds (e.g. bicycles, mopeds, and scooters). This research project consists of two parts: a literature review and data analysis. The literature review compares the methodologies and findings of existing studies on the usage patterns and environmental effects of different modes of micromobility, notably CO2 emissions and the traffic effects of substituting other forms of transportation with micromobility. For clarification, usage patterns encompass user socio-demographics, times of use trip distances, and vehicle lifetimes, while traffic effects concern congestion, travel time, and the breakdown of transportation modes in a city. The data analysis portion examines 2+ million scraped e-scooter trip data points from 5 major German cities for usage patterns, a process that entails transforming datasets into heat maps overlayed on city maps in R and identifying differences in the trip starting and end point densities in relation to local establishments and public transportation systems. One major finding is that riders largely appear to use e-scooters to connect to other modes of transportation, like buses, subways, and trains. The findings from this project will help design more rigorous future studies on micromobility and shape urban road planning decisions to better accommodate micromobility, which is becoming increasingly popular for short-distance urban transportation.

Kelly Zhang

Nora Schelte, Environmental Technology and Resource Management, Bochum University of Applied Sciences Sustainable Technologies Laboratory

DAAD RISE Scholar

Computing Science, Environmental & Urban Studies

Two-dimensional, jammed packings undergoing cyclic shear can be trained to reach periodic orbits in which the particles return to the same positions in each subsequent cycle. These packings deform via abrupt rearrangements, which are relatively large shifts in the positions of a cluster of particles. A model that captures important characteristics of the periodic response describes these rearrangements as interacting, hysteretic, and two-state systems (hysterons) that “flip” between states as the packing is sheared beyond certain strain values. In this work, we study such rearrangements in a simulation of sheared packings of particles. After the system has reached a periodic orbit, we analyze to what extent the rearrangements can be described as a set of interacting hysterons. By shearing very small packings of particles to different strain amplitudes, we explore how rearrangements may be paired up to form “candidate hysterons” and directly measure how the hysterons interact with one another. We find that changing the state of one candidate hysteron can significantly alter the behavior of another candidate hysteron. Examining these rearrangement pairs in detail will help us better understand the kinds of information that can be trained into jammed systems.

Keyer Thyme

Sidney R. Nagel, Physics, UChicago (James Franck Institute, Enrico Fermi Institute); Chloe Lindeman, Physics Department, UChicago

Physics

WACCM (Whole Atmosphere Community Climate Model) is an atmospheric model that has found many applications in a plethora of different scientific fields. After testing, this model has been found consistent in many different scenarios on the earth; however, its consistency and accuracy have not been fully tested in a chemically-dynamic exoplanetary setting. Thus, our focus is to simulate the spectroscopic observation of a WACCM-modeled exoplanet using NASA’s Planetary Spectrum Generator (PSG). In the future, this data could be compared to a similar experimental observation to verify the predictions of the WACCM. To create the simulated data, we used previously modeled WACCM data describing the atmospheric response of an exoplanet whose host star was undergoing solar flaring with four planetary configurations that describe an O2 rich or O2 poor atmosphere under either a K-star or a G-star. We then evaluated this data using the PSG to simulate observation of such a planet by NASA’s proposed IR/O/UV space telescope with a large (~6m) aperture. We will use this spectroscopic output to investigate the detectability of flare-driven atmospheric chemical changes and constrain the planetary and instrumental conditions needed to attain a detectable signal-to-noise ratio. Current findings are that NO2 seems to be the most promising atmospheric gas for detectability, as other molecules are insufficiently flare-driven or difficult to detect.

Kyla Mullaney, George Afentakis

Dorian Abbot, University of Chicago Geophysical Sciences

Quad Faculty Research Grant Scholar

Astronomy and Astrophysics, Geophysical Sciences

The study of hydrodynamic instabilities in viscous flows allows scientists to describe a range of real-world phenomena from bands in planetary atmospheres to vortex formation in oceans to volcanic eruptions. Traditionally, the study of such instabilities relies on the dissipation of mechanical energy in the system. However, this is not the case in certain types of systems that can be modeled as fluids with odd viscosity. These include suspensions of moving cells, chemically driven colloids, or magnetized fluids where mechanical energy is externally pumped into the system. In this presentation, I examine the effect of adding odd viscosity into a fluid on the properties of hydrodynamic instabilities in that fluid. Specifically, I focus on the Raleigh-Bernard convection instability, which arises from an imposed temperature gradient at the boundaries of the fluid, and which is well understood both analytically and numerically in the traditional case. Using numerical methods, I find that the addition of odd viscosity suppresses the appearance of the instability as well as alters the functional form of the relationship between the imposed and emergent properties of the fluid.

Lara Braverman

Vincenzo Vitelli, Physics, James Franck Institute; Tali Khain, Physics, James Franck Institute

Quad Undergraduate Research Scholar

Physics

Diffusive interactions between biomolecules are ubiquitous in the biochemical processes of life, yet the kinetics and dynamics of molecular recognition and binding remain unclear. We report a novel infrared (IR) compatible CaF2-based microfluidic mixer that enables the direct observation of bimolecular association processes, such as DNA hybridization by structure-sensitive IR spectroscopy. Nucleic acids are an excellent target for this technique due to their distinct mid-IR spectrum, arising from in-plane vibrational stretching motions of nuclear bases that are reshaped in response to changes in hydrogen bonding and base stacking environments. We integrated the mixer with FTIR microscopy and performed a time-resolved measurement of DNA hybridization of a 10-base oligo. The hybridization timescale is 2.5 ms and agrees with results from temperature-jump experiments. The mixer can easily be integrated with a laser-based IR source to yield a time-resolution in the order of tens of microseconds, which will be instrumental in understanding microsecond dynamics of many biomolecular processes involving proteins and nucleic acids.

Max Moncada Cohen

Andrei Tokmakoff, Chemistry, James Franck Institute; Ram Itani, Tokmakoff Group

Quad Undergraduate Research Scholar

Chemistry

Next generation lithium-ion batteries require electrolytes that have high ionic conductivity, Coulombic efficiency, and electrochemical stability for effective energy storage. As lithium-ion batteries near their maximum performance potential, new electrolyte chemistries are needed to support high energy density alternatives like lithium-metal batteries. Ether-based electrolytes have been a promising field of battery research due to their strong cycling performance, relatively high conductivities, and excellent reductive stability. Current literature also reports a class of fluorinated silanes, which demonstrate both immiscibility and high oxidative stability for use in dual-phase electrolytes. However, fluorinated compounds are expensive, and their fluorinated chains pose a challenge for high ionic conductivity. In this work, we study three novel electrolytes, comprised of a hydrogenated silane solvent and lithium bis(fluorosulfonyl)amide salt, inspired by previously reported ether and fluorosilane electrolytes. These novel electrolytes exhibit iconic conductivities as high as 1 mS/cm at room temperature (30 °C, 1M salt) and full-cell Li/LiFePO4 cycling performance comparable to tetraglyme. We propose ion solvation structure as a potential explanation for the higher conductivity observed in the hydrogenated compounds versus the fluorinated compounds through nuclear magnetic resonance spectroscopy and other characterization techniques. Finally, we use density functional theory computations and relevant electrochemical experiments to study the oxidative and reductive stability of these hydrogenated silanes. By studying these novel electrolytes, this study contributes promising candidates for silane electrolyte development and concludes that fluorinated silanes are not necessarily better than hydrogenated silanes for lithium-based battery electrolyte performance.

Michael Han

Chibueze Amanchukwu, Pritzker School of Molecular Engineering, Argonne National Laboratory; Peiyuan Ma, Pritzker School of Molecular Engineering

Dean's Fund for Undergraduate Research Awardee

Chemistry, Engineering, Environmental & Urban Studies

Cancer is a major global health issue whose drug treatment has been limited by inherent lack of target specificity and harmfulness of side effects. Nanoscale coordination polymer (NCPs) nanoparticle delivery systems have proven to be a particularly useful approach in chemotherapy because of their ability to stabilize, circulate, and effectively deliver anticancer therapeutic molecules to tumor cells. NCPs bypass the current setbacks of chemotherapy in that the NCP system protects chemo-drugs from premature degradation and demonstrates spatiotemporally controlled uptake and release in tumor cells. However, efficacious accumulation of NCPs in the tumor microenvironment is severely inhibited by nonspecific binding to in vivo off-target components and clearance by the mononuclear phagocytic system. Here, NCPs carrying cytotoxic oxaliplatin and SN38 prodrugs were used to demonstrate an active transport strategy via targeting of the low-density lipoprotein receptor (LDLR). Low-density lipoproteins play a dominant role in the transfer of cholesterol and cholesteryl esters to peripheral cells by LDLR-mediated endocytosis. In tumor cells, LDLR is overexpressed and highly active. By loading cholesterol into the shell of the NCP in the form of a stabilized cholesterol-conjugated SN38 prodrug, these NCPs exhibited preferential adsorption by low-density lipoproteins and enriched in tumors specifically via LDLR endocytosis, allowing for targeting of tumor cells for effective delivery and release of the anticancer payloads. This effect was observed in vitro with immunofluorescence analysis and isothermal titration calorimetry as well as in vivo with pharmacokinetic LDLR knockout studies. This NCP platform increased tumor deposition of OxPt and SN38 by 4.9 and 6.0 times, respectively, compared to free drug administration, thereby achieving 92-98% tumor growth inhibition without causing serious systemic toxicity. This paradigm of low-density lipoprotein mediated transport and delivery offers a promising method for targeted nanoparticle therapy.

Morten Lee

Wenbin Lin, Chemistry

Quad Undergraduate Research Scholar

Chemistry

According to the standard model of cosmology (standard model), the structure of the universe originated from quantum-mechanical fluctuations, during the Big Bang, that were then dramatically amplified by a process of exponential expansion known as inflation. Light emitted during the universe's earliest moments, known as the cosmic microwave background (CMB), is still visible today, and CMB temperature fluctuations provide the empirical foundation for the standard model. There are several known statistical anomalies, in the CMB, unexplained by the standard model. An alternative holographic cosmological model creates an opportunity to connect and explain these anomalies via new, fundamental physical symmetries that result from quantum-gravitational effects not included in the standard model. This motivates the following research question: Are there statistical signatures in the CMB that indicate the presence of symmetries predicted by this new model, based on holographic quantum gravity, but not by the standard model? To answer this question, we are examining a class of functions that measure the correlation between the temperature of CMB light from each individual direction and the temperature of CMB light from all other directions. The standard model predicts that the average correlation between the temperature of light from each direction with the temperature of light from all other directions should lie within a relatively broad range. Our model posits a new, exact symmetry that follows from effects due to holographic quantum gravity: for every direction, the temperature of light from that direction should not be correlated with the temperature of light from any direction within a certain range of angular distance away from the original direction. The significance of our research is as follows: if we find evidence for these new symmetries, it will explain several long-standing anomalies in the CMB, suggesting that the standard model must be radically altered to accommodate them.

Nathaniel Selub

Craig Hogan, Astronomy and Astrophysics, Kavli Institute for Cosmological Physics; Stephan Meyer, Physics, Astronomy and Astrophysics

Quad Undergraduate Research Scholar

Astronomy and Astrophysics, Physics

Computational models in the natural sciences depend on the data they are given and the approximations they employ to succeed in simulating physical systems. This is no less true in computational chemistry, an area of research that uses computing resources to model the observable properties of atomic and molecular systems. In quantum chemistry, the Hartree-Fock (HF), or self-consistent field (SCF) method, approximates the ground state solution to the Schrödinger equation of an electron system. However, one must improve upon this “single-reference” method to obtain chemically accurate results in many cases, an important subset therein are “multireference” systems, which must be constructed as a linear combination of electronic configurations. The present work I am undertaking is to help with the development of additional functionals to be used with multiconfiguration pair-density functional theory (MC-PDFT), a modern formulation which depends on electron pair density in addition to multireference wave functions to compute electronic energy. It is important to note that this method requires its own “translated” functionals that differ from the standard, widely available functionals found in the literature. A promising technique to develop these functionals is to use multiconfiguration data-driven functional methods (MC-DDFMs), whereby machine learning is used to train new functionals based on the differences between less-expensive calculations and highly accurate, computationally expensive reference results. We are generating complete active space self-consistent field (CASSCF) results on the singlet-triplet gaps of carbene molecules, chosen for their inherent, multireference character of their electronic structure. These results are then used for the parameterization of machine-learned MC-PDFT functionals, with the use of an artificial neural-network (ANN), to minimize the error on corrected energies. The parameterization of new MC-PDFT specific functionals is an important step in its long-term development, and it is a critical area of research in lower-cost, highly accurate multireference methods.

Noah Dohrmann

Laura Gagliardi, Chemistry, Pritzker School of Molecular Engineering, James Franck Institute, & Argonne National Laboratory; Daniel King, Chemistry

Quad Undergraduate Research Scholar supported by the NK Cheung Chemistry Research Fellowship

Chemistry

Prior research has explored techniques for impacting taste perception through electrical stimulation of the tongue or added inhaled odors to food items. However, these devices can neither be used during eating nor elicit specific changes in taste profiles. We propose a novel method for altering our perception of tastes by delivering chemical compounds to the user’s mouth; we call this process taste retargeting. We identified eight such taste modulators that produce targeted changes in our perception of basic tastes: amiloride (salt suppression), bortezomib (sour enhancement), clofibric acid (umami suppression), gymnemic acid (sweet suppression), lactisole (umami and sweet suppression), miraculin (sour-to-sweet modulation), neohesperidin dihydrochalcone (sweet enhancement), and zinc sulfate (sweet and bitter suppression). These compounds bind to taste receptors, altering taste transduction and producing short-term, yet fully reversible, changes in the user's subsequent taste experiences. We engineered a delivery system that hooks at the mouth corner (oral commissure) and pumps small amounts of modulators around the gums, changing the user's taste perception while still enabling them to eat freely. We will evaluate our approach through three studies. First, we are currently measuring the changes in taste perception elicited by each modulator. We designed two virtual reality (VR) experiences that leverage taste retargeting to map one food item to several virtual foods, with different taste profiles, using these ongoing insights and literature. In our second study, participants will evaluate the realism of these experiences. Finally, we will explore potential implications of this device as an intervention experience, for which the user will receive modulators based on the approximated nutritional attributes of their chosen food item.

Noor Amin

Pedro Lopes, Department of Computer Science, Human-Computer Integration Lab; Jas Brooks, Human-Computer Integration Lab, Department of Computer Science

College Summer Research Fellow, Quad Faculty Research Grant Scholar

Computing Science, Engineering

A technique often used to obtain the redshifts of objects in large astronomical surveys is the estimation of Photometric Redshifts (Photo-z). This is the case for the Dark Energy Camera Local Exploration (DELVE) Survey. The estimation is done due to the usefulness of having a large dataset of redshift for millions of galaxies. This dataset enables further studies, as redshifts are frequently used to calculate the distance to galaxies, among other useful quantities. Photo-z, as opposed to spec-z (Spectroscopic Redshift), are a much cheaper alternative to the use of spec-z, as the latter involves the measurement of the spectrum of all the objects in the survey, which is, although precise, financially costly. On the other hand, the estimation of photo-z does not involve the direct measurement of spectrum in any way. Instead, this technique employs the already existent flux or magnitude measurements to calculate the redshift based on an interpolation of the flux data obtained and what the spectral energy distribution of the particular object being analyzed should be. In this work, we will estimate the redshifts of around 100 million objects in the DELVE catalog, across approximately 15,000 square degrees. This is to be done in three steps. First, we will obtain a matched catalog of DELVE and another smaller survey’s data, BOSS, which contains a large number of spectroscopic redshifts. This matched catalog is an intersection of the objects present in both catalogs. Second, we will use this smaller matched catalog to calibrate our photometric redshift estimation software by comparing it to the spectroscopic redshifts to better produce more precise redshifts. Third, with the calibration completed, we will estimate redshifts for the entirety of the DELVE catalog.

Raul Gomez del Estal Teixeira

Chihway Chang, Astronomy and Astrophysics, Kavli Institute for Cosmological Physics; Francisco Javier Sanchez Lopez, Kavli Institute for Cosmological Physics/Fermilab

Quad Undergraduate Research Scholar

Astronomy and Astrophysics

Organic-inorganic heterostructures combine the favorable optoelectronic properties of atomically thin inorganic materials with a wide variety of organic semiconductors to generate new electronic structure and functionality at the interface. Transition metal dichalcogenides (TMDs) are a candidate 2D material of particular interest for their semiconducting properties, strong light-matter interactions, high charge transport mobility, and tunable bandgap. Combining the high absorption efficiencies characteristic of organic semiconductors with the unique properties of TMDs may give rise to a new generation of optoelectronic devices that derive their novelty from the electronic structure and dynamics of the organic-TMD interface. However, the interfacial electronic structure and dynamics that result in the altered behavior of the heterostructure are not fully understood. The King Lab is currently building an interferometer for heterodyne-detected electronic sum-frequency generation (eSFG) to directly interrogate the electronic energy landscape and excited state dynamics at the interface of materials. As a first step to selectively probing the interfacial electronic structure of the PTCDA-MoS2 interface using eSFG, we have characterized the dynamics of the PTCDA-MoS2 heterostructure. Photoluminesence (PL) spectroscopy is used to examine the modification in the luminescence properties of PTCDA-MoS2 and comparing the transient absorption spectra of bare MoS2 with that of PTCDA-MoS2 elucidates the effects of PTCDA on MoS2 excited state dynamics. Forthcoming eSFG investigations will lend insight into the novel properties of organic-TMD heterostructures and clarify the interfacial electronic structure and processes that result in their altered behavior and thereby inform the design of other organic-TMD heterostructures.

Sarah Melton

Sarah King, Chemistry, King Lab; Clare Keenan, King Lab; Nasim Mirzajani, King Lab

Quad Undergraduate Research Scholar supported by the NK Cheung Chemistry Research Fellowship

Chemistry

Wearable and implantable biosensors that can seamlessly adhere to the human body for long-term monitoring of health conditions are rapidly attracting interest. To achieve this, the biosensor needs to provide high sensitivity, for which a transistor-based biosensor with its amplification capabilities is a good option. Alongside, it needs to have tissue-like softness and stretchability that is highly compatible with the human body for mitigating the foreign body response. However, there is currently a substantial modulus mismatch between the rigid biosensor (>1 GPa) and human tissue (<10 kPa). One promising way to overcome this is to create a sensing layer that has tissue-like mechanical properties, and directly interface it with the human body. The semiconductors used as organic electrochemical transistor (OECT) central materials are hypothesized to decrease modulus through water uptake and crystallinity expansion. This softening process can be further enhanced by incorporating a secondary hydrogel network with a polymer semiconductor. The target of this project is to construct a soft hydrogel-like semiconductor for OECTs with an elastic modulus of less than 10 kPa. Here, to develop an ultra-soft OECT, a suitable semiconductor/hydrogel system of low modulus (<1 kPa) and great electrical performance (transconductance >20 S/cm) was identified by combining the structural, mechanical, and electrical characterizations.

Shivani Chatterji

Sihong Wang, Pritzker School of Molecular Engineering, Wang Research Group; Yahao Dai, Wang Research Group, PME

Quad Undergraduate Research Scholar

Chemistry, Engineering

Our predictions of how the universe should behave are not consistent with observed gravitational effects. Measurements of the cosmic microwave background and the motion of galaxies indicate that there is far more matter in the universe than we observe. We call this unobserved matter dark matter. One potential explanation for dark matter is a yet-undiscovered elementary particle that does not interact normally with other matter. The XENONnT dark matter experiment attempts to test this model by looking for these particles. The experiment uses a dual phase time projection chamber (TPC) to detect potential interactions of dark matter particles with the xenon target. An interaction in the liquid xenon causes the production of excitation and ionization. The excitation leads to the emission of prompt scintillation light, which is recorded by arrays of photo multiplier tubes on the top and bottom of the TPC. This first light signal is referred as S1. At the same time the free ionization electrons are drifted to the top of the detector, under the effect of an applied electric field. Once reaching the gas phase of the TPC, the stronger electric field accelerates them, producing a delayed scintillation signal called the S2. This signal is also detected by the photomultiplier tubes. This method allows for the localization of interactions within the chamber. Based on the nature of the observed interactions with xenon atoms in the detector target, we are able to distinguish potential dark matter interactions from background interactions. In this project, I analyze S1 and S2 data from the experiment in order to discard events for which we cannot uniquely match a single S1 signal with an S2.

Tori Ankel

Luca Grandi, Physics, Enrico Fermi Institute, Kavli Institute for Cosmological Physics

Quad Faculty Research Grant Scholar

Physics

Lithium metal batteries are widely regarded as the next generation battery systems, offering higher energy densities than traditional lithium-ion systems. Solid state electrolytes, like Li7La3Zr2O12 (LLZO), can enable lithium metal batteries because of their nonflammability. However, LLZO is plagued by poor interfacial contact. To improve the stability during battery cycling, a popular approach is to mix LLZO with a polymer electrolyte such as polyethylene oxide (PEO) with lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) to make a hybrid electrolyte. However, transport mechanisms through hybrid solid-state electrolytes remain an open question, especially since the pure inorganic phase is typically more conductive than the polymer phase. This study seeks to resolve these conduction mechanisms, specifically in high weight LLZO content hybrids. To this end, we observed that even at 90 wt% LLZO, the Li-ion conductivity mimicked that of pure PEO, in direct contrast to some literature claims. Additionally, when analogous LLZO hybrids were made with ion blocking polyethylene (PE), the conductivity plummeted. All of this is strong evidence to suggest that even with “PEO in LLZO” hybrids, the primary effect of the LLZO is to plasticize the polymer—but it does notprovide favorable lithium transport routes. Differential scanning calorimetry and x-ray diffraction can provide further evidence of PEO plasticization and help confirm that PEO is the strongly favored conducting phase. With this surprising conclusion, it is clear that hybrid electrolyte designs must consider ways to force conduction through the highly conductive inorganic phases instead of the poorly conductive polymer phase. The insights from this study will greatly inform the design of solid-state electrolytes, which could enable lithium metal batteries.

Walker Gillett

Chibueze Amanchukwu, Pritzker School of Molecular Engineering, Amanchukwu Lab; Priya Mirmira, Pritzker School of Molecular Engineering

Dean's Fund for Undergraduate Research Awardee, Quad Undergraduate Research Scholar

Chemistry, Engineering

Does gentrification pave the way for more diverse social mixing? Previous research has come to differing conclusions around gentrification’s ability to reduce racial and class-based segregation. While gentrification can potentially bring benefits to under-served communities such as better resources, a stronger local economy, and potential economic opportunity, actual evidence of increased social mixing as a result of these opportunities is unclear. In general, census data does not offer enough granularity to see social mixing in gentrifying neighborhoods, making this a difficult question to answer. However, using granular, anonymized mobile app GPS data, we analyze the social diversity of neighborhoods at different stages of gentrification and compare the degree of racial and/or socioeconomic diversity at each stage. Built on these findings is a web-based platform to visualize the relationship between gentrification and neighborhood diversity across different cities, designed to make the information both understandable and accessible.

Allie Chu

Wenfei Xu, Center for Spatial Data Science and Mansueto Institute for Urban Innovation

Quad Undergraduate Research Scholar supported by the Liew Family Research Fund

Environmental & Urban Studies

In order to develop effective mental health interventions for Muslim students in the United States, who represent a rapidly increasing portion of the population, research must be done on how these individuals understand their mental health in the context of their lived religious and cultural identities. This study seeks to understand how Muslim students construct a framework to conceptualize mental health while growing up, how this framework evolves when they attend school at the University of Chicago, and barriers that still exist, preventing access to mental health resources. Twenty Muslim students at the University of Chicago participated in hour-long semi-structured interviews during the summer of 2021. These interviews were then qualitatively coded and analyzed to find: (1) participants’ understanding of mental health was largely shaped by family stigma towards mental illness in childhood, (2) attending college created distance that allowed students to rework and expand their existing framework to understand mental health, (3) students struggled cultivating support systems in college due to difficulty making friends and/or finding mental health professionals who shared their religious and cultural identity. These results suggest that in order to support mental health outcomes in Muslim college students, investment is needed in resources that can provide social support (religious and cultural) to these individuals.

Areeha Khalid

Eman Abdelhadi, Comparative Human Development

Earl R. Franklin Research Fellowship

Comparative Human Development

Much of the recent growth in the Charter School sector in New York City has come through charter school network expansion. In such networks, regulators face a potential tension between economies of scale and market power when deciding whether to approve the opening of more schools in a particular network. We propose 5 channels from industrial organization theory to explain network growth: economies of scale, increased input market power, rent-seeking, replicating successful results, and branding. We then attempt to quantify the contribution of each of these channels to network growth, giving insights on the degree of positive or negative impact on student welfare by each. Due to a recent 2010 legal policy change in New York State which enabled schools to form merged corporations constituting a single legal status, the identification challenge for schools in versus out of networks can be studied using a difference-in-differences methodology. We retrieve geographical data, State University of New York (SUNY) charter school performance reports, Common Core of Data (CCD) education data, and New York State Education Department (NYSED) data in our study.

Bruce Ke Zhen Wen

Michael Dinerstein, Kenneth C. Griffin Department of Economics, University of Chicago

Quad Undergraduate Research Scholar supported by the Liew Family Research Fund

Economics & Business

This winter and spring quarter I assisted Dr. Brammer with the development of a historical deliberation guide for the National Portrait Gallery examining women’s suffrage at the turn of the 20th century. This guide was created from historical and archival research focused on the expansion of suffrage in the United States to include women and Black Americans, and the implication of these changes for the goals and motivations of broader civil and social rights movements. As per the National Portrait Gallery’s interest, this curriculum utilizes the public debates between Ida B. Wells and Frances Willard which presented fundamental arguments about suffrage that existed in American social movements for temperance, abolition, and civil equality. This curriculum will expand on the arguments found in the Wells/Willard debates, utilizing arguments made by both women in their published writing, and the changes instituted by Constitutional amendments and legislation. The curriculum will focus specifically on the accessibility, security, and legitimacy of the ballot while considering the implications of race, gender, class, education, and other forms of sociopolitical stratification on political engagement. The developed curriculum will be structured as a deliberative framework, with a focus on encouraging public discourse much like that between Wells, Willard, and their supporters.

Emma Kugelmass

Dr. Leila Brammer, Parrhesia Program for Public Discourse

Quad Undergraduate Research Scholar

History

From Florida’s ‘Don’t Say Gay’ bills to Russia’s ‘Anti-Gay Propaganda’ laws, the legislative processes of various cities, states, and countries have been used to silence the LGBTQ+ population with the stated goal of protecting the moral values of families. In Brazil, for instance, the Nonpartisan School Program (ESP) is a set of federal bills that seek to prohibit discussions about gender and sexual orientation in schools to preserve the right of families to teach their own ‘moral, sexual, and religious’ values to their children. Since 2014, 15 ESP bills have been introduced in Brazil’s Chamber of Deputies. Together, these bills produce a narrative of moral crisis, point out the ‘indoctrinating’ agents in education, and indicate and justify actions to mitigate the alleged indoctrination of children. Given the proliferation of ESP bills in Congress, the erosion of Brazilian democracy in the last decade, and the history of moral education in Brazil, this research project analyzes how the conservative Christian Right has used the legislative process to advance its ideals of morality in Brazil. Through a historically situated discourse analysis of the 15 pro-ESP bills, I demonstrate that the conservative Christian Right first used the bills to portray a multifaceted moral enemy to then focus on the mitigation of such an enemy within the principles of the Brazilian Constitution. As such, the conservative Christian Right has reshaped its national project of moral education - last seen during the 1964-1985 civil-military dictatorship - to promote a populist hegemony of the traditional patriarchal family under the guise of democracy. Therefore, the Nonpartisan School Program is part of the larger process of democratic erosion happening in Brazil. As similar bills sprout in other countries, it is imperative to analyze them to better know how to fight for the preservation of a pro-democratic education.

Giácomo Rabaiolli Ramos

Dr. Brodwyn M. Fischer, History

Education & Scholarship of Teaching, Gender and Sexuality Studies, Political Science

A common way to find solutions is to begin with a known problem. However, recent research suggests that people can also spontaneously generate solutions to a previously unrecognized problem after encountering an object, a process known as need-solution pair recognition. Little is known about the mechanisms that support this recognition-based solution-finding. One possibility is that the occurrence of need-solution pair recognition is precipitated by analogical reasoning about (an) encountered object(s). If this is the case, then there should be a positive relationship between an individual’s analogic transfer abilities and their need-solution pair recognition abilities. Additionally, successful analogical transfer is putatively reliant on both the search and retrieval of an appropriate object from long-term memory (LTM) as well as some mapping process between the current object and the analog object from LTM. Given that both of these cognitive tasks are considered active processes that draw on working memory resources, this leads to the additional hypothesis that working memory capacity may mediate the positive relationship we see between analogical reasoning performance and need-solution pair recognition performance. To test these hypotheses, I have been working on constructing and organizing tasks that measure these three types of abilities. In the future, if my hypothesis is supported, I would be interested in training analogical reasoning and observing whether this training improves need-solution pair recognition.

Joshua Kim

Shannon Heald, Psychology, APEX Lab; Yena Kim, APEX Lab

Quad Undergraduate Research Scholar

Social and Behavioral Sciences

This research project aimed to contextualize quantitative homicide patterns in San Francisco between 1960 and 2003. By reviewing analyses of historical homicide data and studying the history of San Francisco’s neighborhoods from books and newspaper archives, I hoped to identify the areas of the city with the highest homicide rates and hypothesize which historical patterns or modern developments may have caused these dynamics. In my research, I identified six “high-activity” neighborhoods containing blocks with consistently high homicide rates: Western Addition, Hayes Valley, Japantown, The Tenderloin, South of Market, and Bayview Hunters Point. These neighborhoods have historically been some of the city’s most impoverished areas, being home to most of the city's single-occupancy room (SRO) residential buildings, as well as having the highest rates of preventable hospitalizations. The highest-activity blocks in these neighborhoods have also been the site of city-led revitalization efforts, such as the Yerba Buena Center and Hunters Point Redevelopment Area, though homicide dynamics have remained largely unchanged since their implementation. Future research would map locations of SROs in San Francisco onto yearly geospatial estimates of homicide rates and use historical news archives to better quantify and categorize homicides in these areas.

Juan-Pablo Armes

Robert Vargas, Division of Social Sciences; Caitlin Loftus, University of Chicago

Quad Faculty Research Grant Scholar

Environmental & Urban Studies

As essential workers during the COVID-19 pandemic, American nurses and retail workers have been frequently deemed “heroes” by their employers and the public. This study first introduces what hero rhetoric is before investigating its implications, as a shift in how these workers are framed culturally and in their occupational prestige, for workers’ experiences and occupational fulfillment. It draws on a mix of qualitative data, including examples of hero rhetoric from a variety of sources, pre-pandemic impressions of workers, forum posts and interviews on workers’ pandemic experiences, and pandemic-era work policies. This study examines how workers engage in mechanisms like role distancing to manage the importance of their occupations in their personal identities when placed in undesirable working conditions. It further considers workers’ occupational fulfillment, arguing that cultural frame shifts must coincide with structural changes for it to be meaningfully improved; otherwise, workers, regarding their occupations as undesirable, are motivated to strategically distance themselves.

Julia Du

Kristen Schilt, Sociology

Economics & Business, Public Policy, Social and Behavioral Sciences, Sociology

Understanding the neural mechanisms that mediate speech processing is critical in the broader project of understanding some aspects of perception and cognition that are often viewed as unique to humans. One key focus in the field of speech processing is the forward model: a mechanism where motor movements required for speech are predicted by the brain before they are executed. These movements may then be simultaneously compared with sensory information so that future motor movements may be executed properly. However, even with decades of studies reinforcing the existence of the forward model for speech, the physiological mechanisms behind such a model remain poorly understood. How does the brain predict the auditory feedback produced during speech to manage speech motor control? We may begin to answer this question by analyzing zebra finch birdsong, which anatomically mirrors human speech as both coordinate lung pressure and vocal tension to produce speech sounds. Prior research has found that zebra finches have neurons which fire at such a precise time during speech that they are too late to be motor commands but too early to be associated with sensory feedback – these neurons are thought to represent the zebra finch’s forward model. If humans produce speech in a physiologically similar way to zebra finches, it is possible that humans could also have a similar mechanism for the forward model for speech. To investigate this, we have devised an experiment where human speech production mechanisms, laryngeal tension and lung pressure, may be respectively analyzed with an electroglottogram (EGG) and a subglottal pressure monitor. Concurrently, neural recordings may be taken using an electroencephalogram (EEG). If neurons fire in that similar unique way during human speech, this would provide strong evidence that the human and zebra finch forward models for speech operate similarly.

Koby Rosen

Howard Nusbaum, Psychology, APEX Lab; John Veillette, APEX Lab

Quad Undergraduate Research Scholar

Neuroscience, Social and Behavioral Sciences

Midwives have been proposed as a potential avenue for combating climbing maternal mortality rates and severe racial disparities in the US. Midwifery is often thought to attenuate some systemic inequalities and aspects of the prevailing physician model of care that disproportionately harm mothers of color. However, policies that define a midwife’s role in maternal care currently vary widely state by state. Yet, few studies have evaluated if midwives lead to improved outcomes, or how midwifery policies across states influence outcomes. US natality data (2014) is merged with state-level midwifery integration policy data to test the association between birth attendant type (physician or midwife) and birth outcomes, evaluate whether these relationships vary by maternal race/ethnicity, and explore if state-level midwifery policies impact the aforementioned associations. Even after controlling for the different risk profiles of physician- and midwife-attended births, midwife attendance is associated with lower odds of cesarean section and other adverse outcomes such as preterm birth, low/very low birth weight, and low Apgar score. Midwife attendance is particularly consequential for improving some of these outcomes for Black mothers relative to White mothers. Moreover, states with more supportive and integrative midwifery policy tend to have higher average birth weights than states with more restrictive policies in place. The findings from this work are poised to contribute to emerging research and policy interventions that improve maternal health in the US.

Laura Chen

Aresha Martinez-Cardoso, Department of Public Health Sciences

College Global Health Scholar, Dean's Fund for Undergraduate Research CONFERENCE Awardee

Biological & Health Sciences, Public Policy, Social and Behavioral Sciences

It requires a certain level of cognitive control to monitor one’s responses in social situations, so as to not appear biased. To see if this requirement of active monitoring has a measurable effect on day-to-day social interaction, we measured participants’ response latency to questions comparing society’s perception of different pairs of social groups. Some pairs involved groups that are targets of common discrimination in the United States (Black people and White people, disabled people and non-disabled people, etc.), and other pairs did not (parents and teachers, accountants and engineers, etc.). Data from 20 adult participants in the United States indicate that individuals do not tend to begin speaking later in response to questions regarding discriminated groups but take longer to verbally reach the point of discussion regarding society’s judgment of them. In regards to certain discriminated groups, marginally significant results suggest that individuals take longer to verbally reach the point of discussion regarding society’s judgment of them. This finding shows that measuring “relevant utterance latency” could be used as a future method of measuring social discomfort or cognitive load in social speech.

Lev Copelan

Susan Goldin-Meadow, Psychology, Goldin-Meadow Lab; Anjana Lakshmi, Goldin-Meadow Lab

Quad Undergraduate Research Scholar

Social and Behavioral Sciences

Differences in childhood attention and working memory performance may lead to disparities in future learning outcomes. Because previous research suggests that aspects of the physical environment affect cognitive development, understanding which physical environmental factors influence children’s attention and working memory abilities could inform effective interventions to minimize such differences. We asked how levels of neighborhood fine particulate matter (PM2.5), which signal worse air quality, predict improvement in preadolescents’ attention and working memory performance from age 9-10 to age 11-12. To address this question, we analyzed data from the Adolescent Brain Cognitive Development (ABCD) Study, which is the largest longitudinal study of child brain development in the United States with over 11,000 participants. In this study, demographic, environmental, behavioral, and brain data are collected from the participants biannually from when they are aged 9-10 to 19-20. As we were interested in attention and working memory, we examined participants’ performance on the n-back task, which provides behavioral measures of these cognitive functions. We performed multiple regression to predict changes in n-back task performance from the baseline study visit (when participants were 9-10 years old) to the 2-year follow-up study visit (when they were 11-12) from neighborhood PM2.5 levels (n = 3344). To control for effects of general cognitive performance and socioeconomic factors, we included measures of cognitive performance at the baseline study visit along with other physical and social environmental and demographic measures in the regression model. We found that PM2.5 negatively predicted changes in n-back performance above and beyond effects of baseline n-back scores, general neurocognitive abilities, and social and demographic factors. This suggests that air quality may be worth considering for interventions designed to improve children’s cognitive development. Future directions include examining whether air quality modulates brain connectivity patterns, which may in turn affect developmental cognitive performance.

Mudmee Sereeyothin

Prof. Monica Rosenberg, Psychology, Cognition, Attention, and Brain Lab; Dr. Omid Kardan, Cognition, Attention, and Brain Lab

Hoeft Undergraduate Research Awardee, Quad Faculty Research Grant Scholar

Social and Behavioral Sciences

When we encode a visual memory, the information retained can be categorized into detailed and general information, or the gist. In previous research, we have considered the effect of time on drawings representing visual memory and the retention of the gist of the image. In this experiment, we used virtually created scenes to study the effect of grammatically inconsistent objects within a scene, and how exposure duration affects gist retention and visual memory. Scene grammar, the heuristic which guides expectations of what will be present within a scene, is also understood to be a highly relevant heuristic for visual processing and attention. Participants drew scenes, both with consistent and inconsistent scene grammar, from memory across a variety of exposures to the stimulus images. Consistency was manipulated by replacing an object within the stimulus with an item considered irregular to the stimulus description; for example, replacing a car in a garage with a horse. A separate group of participants scored the presence of objects in those drawings in order to quantify visual memory content. Statistical analysis revealed no significant relationship between consistency, time, and memory content. However, the utilization of entirely virtually simulated scenes to create the stimuli is a part of the experimental design which should be considered. Previous research shows that memory content increases over time when real-world scenes are used as stimuli. The fact that the data became insignificant in indicating changes in memory indicates that virtually created scenes may impact or interact with visual memory in a yet untested way that differs from a real-world stimulus. This will likely lead to lowered significance for variables established to influence real-world scenes. Virtual reality may be analyzed in a different way than real-world stimuli, thus relying on slightly altered heuristics or mechanisms for memory.

Rebecca Greenberg

Wilma Bainbridge, Psychology, Brain Bridge Lab

College Summer Research Fellow, PRISM Scholar

Neuroscience, Social and Behavioral Sciences

The gender disparity in many prestigious careers (e.g., the STEM domain) manifests to its full form in adulthood, yet it takes root in early childhood. By the age of 6, children begin to endorse the stereotype that brilliance is a trait of men and boys (the gender brilliance stereotype), and girls show less interest than boys in participating in games described as requiring brilliance. The aim of this project is to uncover the psychological mechanisms through which the gender brilliance stereotype undermines girls’ interest. We propose that self-efficacy is a key contributor to this gender disparity. Specifically, given the negative stereotype against females’ high intelligence, girls may perceive themselves as incapable of performing activities said to require brilliance and thus become less interested in these activities. To test this, we presented 4- to 9-year-old children (N = 81, planned = 120) with two novel games: the “Smart” game, described as a game only for really, really smart children, and the “Hardworking” game, described as a game only for really, really hard-working children. After hearing about each game, children were asked two questions assessing their self-efficacy (e.g., “Do you think you would do well at the Smart game?”). Each question was rated on a 6-point scale (1= really not well, 6 = really well). Preliminary results suggested that with age, boys (but not girls) show increased self-efficacy for the Smart game, while girls (but not boys) show increased self-efficacy for the Hardworking game. These results provide initial evidence that self-efficacy is a powerful mechanism explaining why girls are less interested in activities portrayed as requiring brilliance. This project adds to our knowledge on the psychological barriers blocking girls from entering certain domains and sheds light on potential ways to address the gender gap at its developmental roots.

Sophie Barth

Lin Bian, Department of Psychology, Early Social Thinkers Lab, University of Chicago; Molly Tallberg, Department of Psychology, Early Social Thinkers Lab, University of Chicago

Quad Undergraduate Research Scholar

Social and Behavioral Sciences, Developmental Psychology

Girls are discouraged from pursuing science around the time when formal schooling begins (e.g., Bian et al., 2017; Miller et al., 2018). To tackle the gender gap at its developmental roots, the present research focuses on 6- to 7-year-olds and explores potential interventions to strengthen girls’ resilience in science fields. Specifically, past work suggests that asking girls to pretend to be a competent character improves their persistence in challenging tasks (White et al., 2017). Building on this work, we ask whether pretending to be a scientist would enhance girls’ resilience in science tasks. In addition, we examine the key factors that may influence the effectiveness of this strategy, that is, the characteristics of the role model. Children participated in a science game in one of three conditions: they pretended to be a dedicated (dedication) or brilliant science role model (brilliance) or did not engage in pretend play (baseline). In the science game, they were asked to guess whether various objects would sink or float when dropped in water and were told whether their answers were correct or incorrect on each trial. Resilience was measured by the number of trials completed following an initial incorrect prediction. Preliminary results (N = 75) suggest that boys show higher resilience in the baseline and brilliance conditions than the dedication condition. However, girls were more resilient in the dedication condition compared to the baseline and brilliance conditions. These results provide supporting evidence that asking girls to pretend to be a dedicated scientist strengthens their resilience in science through continued persistence. Using pretend play as a tool can help to promote equal gender representation in science fields.

Tatiana Rachlin

Lin Bian, Psychology; Gabrielle Montiel, Early Social Thinkers (EAST) Lab

Quad Undergraduate Research Scholar

Psychology

Every piece of artwork is unique, and viewing art is often seen as a subjective experience. Additionally, many works of art are made to last in people's memories and leave an impact on the viewer. However, what makes a work of art memorable? We determined and studied the memorability of 4,021 paintings at the Art Institute of Chicago (AIC) both in an online and in-person task. In an online task, over 3,200 people participated in a continuous recognition task to determine the memorability scores of the paintings. Participants were significantly consistent in the pieces they remembered and forgot, and that memory performance was predictable by a neural network (ResMem; Needell & Bainbridge, 2021), showing that an objective memorability score can be quantified for a piece of art. We identified properties that influence memorability scores: paintings with text or unusual content tended to have higher memorability scores, and paintings with the lowest memorability scores were usually scenes or landscapes with little visual information and a darker color palette. Additionally, the paintings that caused the most false alarms tended to be beige landscapes or scenes. Importantly, famous pieces were judged as significantly more memorable by ResMem, suggesting that certain perceptual features of a painting can influence its success. For the in-person task, a separate set of participants walked through and observed 162 paintings on both the first and second floors of the American Art wing of the AIC. Participants answered a mobile experiment indicating which paintings they remembered seeing, intermixed with foils. Foil images were from the same region and time period as the targets, and many were by the same artists. We found that ResMem memorability scores were also able to predict in-person memory behavior, suggesting a consistent influence of images on our memories.

Trent Davis

Wilma Bainbridge, Psychology, Brain Bridge Lab

Social and Behavioral Sciences, Visual & Performing Arts

Non-fungible tokens (NFTs) are “digital collectibles”: unique, indivisible, durable digital assets on blockchains, often used to represent works of visual art. The NFT market has recently experienced explosive growth, driven by and attracting many investors hoping to achieve impressive returns. Because of the opaque nature of NFT as an asset class, a commonly held view is that experienced investors can profit much more than inexperienced investors from their superior information about the quality and potential of NFT projects. Our research addresses this issue using a comprehensive dataset of NFT transactions on the Ethereum blockchain. We find that experienced investors systematically outperform: they attain roughly 10 percentage points higher returns on each trade than inexperienced investors for trades with similar holding periods. NFT collections purchased by experienced investors are more likely to sell out in primary markets, sell out faster, and experience subsequent higher price growth in secondary markets. Moreover, changes in the share of collection items owned by experienced investors also predict changes in collection prices. Our results suggest that NFT markets are characterized by high degrees of informational inefficiency, allowing investors with informational advantages to extract profits systematically.

Wanran Zhao

Anthony Lee Zhang, University of Chicago Booth School of Business

Quad Undergraduate Research Scholar

Economics & Business