Project in Neuroscience
NEURONAL ACTIVITY IN OLFACTORY PROCESSING: A META-ANALYSIS AND EXPERIMENTAL STUDY
Ethan Abramovitz, Grace Calhoun, Emily Chen, Maanav Chittireddy, Sapna Patel, Survani Sinha, Titus Solomon, Joshitha Sripada, Arjun Subramanian, Hillary Xie, Jessica Yao
Advisor: Graham Cousens, Ph.D.
Assistant: Harris Naqvi, B.S.
ABSTRACT
When the brain processes odorant molecules, olfactory receptor neurons (ORNs) in the olfactory epithelia project cilia into the nasal mucosa, providing our sense of smell. However, research shows that when ORNs process odorants, inhibition occurs. To address this phenomenon, we wanted to find: 1) differences in inhibition/excitation between brain regions, 2) how inhibition varies for different functional groups, and 3) if inhibition occurs simultaneously between neurons in oscillatory patterns. By examining mouse olfactory systems through single-unit electrophysiological recordings from our lab and other studies, we generated plots using MATLAB to analyze the firing rates of odor-cell pairs when exposed to different scents. We hypothesize that odor-invoked neuronal responses in the olfactory bulb will show a higher proportion of excitatory responses than primary cortical regions; simultaneous responses will be detectable between neurons of the same cortical region; neurons will show oscillatory spiking patterns with dominant frequency components in the beta and gamma bands; and the degree of mixture suppression would directly correlate to the similarity between functional groups. Through our meta-analysis and experiment, we found that inhibition was most prevalent in early olfactory regions such as the olfactory bulb (OB) and piriform cortex (PC), but substantially lower in higher-order areas like the posterolateral cortical amygdala (plCoA) and dorsal tenia tecta (TT). This supports a hierarchical model in which inhibition has a key role in odor filtering but has a lesser role in downstream regions involving higher-order functions.
Project in Archaeology
ARCHAEOLOGY AND MATERIAL SCIENCE: Sacred Stones or Clever Fakes?
Natalie Chen, Zachary Ciappa, Jack Ecklund, Abigail Frey, Lucas Horowitz-Kurtzberg, Darby Ko, Serena Lorenzo, Julianne Ma, Andie Park, Alexica Perez, Luke Shao
Advisor: Dr. Maria Masucci
Assistants: Ece Onatca and Katelyn Rohlf
ABSTRACT
Around the globe, greenstone beads have been found in burial and ritual contexts, suggesting their cultural significance as symbols of wealth and status. Our study analyzed the composition of greenstone beads from suspected production sites in Guangala and Manteño settlements in the southwestern coastal region of Ecuador. Understanding the origin and composition of these beads is essential to understanding the history of pre-Columbian societal structure and cross-cultural trade, as well as the value of greenstone beads in the coastal Guangala and Manteño cultures. In this study, we aim to determine the raw materials used to manufacture the greenstone beads, and determine the location and rarity of the minerals. We performed a multidisciplinary analysis including mineralogical analysis with optical petrography, chemical analyses through a hydrochloric acid test, and Fourier Transform Infrared Spectroscopy with Attenuated Total Reflectance (FT-IR-ATR) examined with Pearson’s and Spearman’s correlation rank order analysis of the spectra to identify and compare the samples. Through the analyses, we determined that the artifacts were composed of different minerals, predominantly present in local and common rock formations. Our results suggested that the greenstone beads served as an essential domestic export for societies in a greater coastal trade network. Our diverse methodology further provides a novel approach to archaeological research, demonstrating the utility of combining cultural, chemical, geological, and statistical analyses.
Project in Psychology
COMPARATIVE ANALYSIS OF COGNITIVE ILLUSION SUSCEPTIBILITY IN HIGH-PERFORMING STUDENTS VS. THE GENERAL POPULATION
Mishi Chaturvedi, Tanvi Daita, Francesca DiMarcello, Glen Finnie, Mahek Khan, Bryanna Liu, Aidan Ryu, Arlind Sinani, Audrey Wang
Advisor: Patrick Dolan
Assistant: Brenna Hezel
ABSTRACT
Cognitive illusions are common and predictable lapses in thinking, decision making, and other mental processes. This study aimed to explore the connection between intellectual ability and susceptibility to cognitive errors. To do so, a survey framed as a personality assessment was distributed to two groups: 46 high-achieving high school students attending the New Jersey Governor’s School in the Sciences and 46 online participants representing the general population with varying ages and academic backgrounds. The survey contained personality-related questions as well as various standard cognitive illusions that tested memory, judgement, and attention. The participants’ responses were anonymously collected and compared by group. Results varied based on the specific illusion, which may be attributed to the nature of each task. One consistent trend was that Governor’s School scholars were more susceptible to the illusions related to self-evaluation. Overall, the results show that while academic intelligence may correlate with susceptibility to cognitive illusions, even high-performing individuals are ultimately vulnerable to them due to the inherent tendencies of the brain to make assumptions and have selective attention. This research highlights the importance of studying these illusions in order to gain further insight into the mechanisms behind human psychological behavior.
T12 Final Paper >
Project in Computer Science
AI-ASSISTED DISCOVERY OF PHYSICS LAWS
Harinarayan Asoori Sriram, Ishani Bakshi, Katherine Estevanell, Henry Gaus, Oliver Kahng, Aarav Khatri, Chloe Krawczak, Logan Miller, Joshua Moore, Prasham Shah, Kaitlin Zhang
Advisor: Dr. Minjoon Kouh
Assistant: David Hoyt
ABSTRACT
The discovery of fundamental physics laws has traditionally required extensive human analysis of experimental data. However, as the scope of the field grows, so does the complexity and volume of the data, and in turn, the difficulty of finding underlying physics laws and their mathematical formulations. In this study, we took advantage of advances in artificial intelligence, namely core components of the “AI Feynman” algorithm (Udrescu and Tegmark, 2020), to explore how algorithms inspired by physics properties can discover underlying equations from data. Our approach combines linear and polynomial regression with symmetry detection, powered by artificial neural networks, to sequentially simplify and fit data. We tested our system on synthetic datasets representing various physics equations, evaluating its ability to identify translational, sum, and scale symmetries, as well as its performance under noise. Our results show that the algorithm performs well on low-noise datasets and that the range of data should be considered for symmetry detection. However, our system is built to discover a limited class of algebraic equations. Future work may include an expansion of this algorithm to deal with differential equations or higher noise levels. This work confirms the physics-based approach of the AI Feynman in equation discovery, highlights its current limitations, and suggests potential avenues for future improvement of such methods.
T22 Final Paper >
Project in Molecular Biology
HOST-PATHOGEN INTERACTIONS: USING VESICULAR STOMATITIS VIRUS TO DISCOVER THE RELATION BETWEEN UNTRANSLATED REGIONS OF MRNA AND HUMAN IMMUNITY
Addisyn Fisher, Ananya Gupta, Jessica Hong, Chloe Koo, Claire Lee, Joonwoo Lee, Meredith Levy, Michelle Lin, Ronak Pathak, Luc Reyes, Michael Royzman, Hrishi Shah, Shaunak Soni, Isabelle Wang
Advisor: Dr. Tyler Dorrity
Assistant: Christiana Perez
ABSTRACT
Vesicular stomatitis virus (VSV) is a negative-sense RNA pathogen that infects livestock and humans with major implications in modern-day vaccines. Recently, VSV has been found to decrease the length of the human 3’ untranslated regions (3’UTR) of messenger RNA in cells. Prior research suggests that 3’UTR length contributes to innate immune system regulation. Double-stranded RNA (dsRNA) is often used as a viral genome, so when long 3’UTRs form dsRNA, PRRs mistake “self” dsRNAs as foreign from viral pathogens, thus activating the immune response. This indicates that VSV has a selective advantage by shortening 3’UTR length to avoid detection and reduce host innate immunity. A more novel UL (ultra long) to ORF (open reading frame) ratio technique, derived from qPCR fold change, was normalized to a negative control to ascertain which gene (and associated protein) in VSV was responsible for this phenomenon. We found that the VSV glycoprotein (G protein) is most likely to be responsible. Compared to VSV’s other protein-coding genes, only the qPCR fold change ratios for the VSV-G gene indicate 3’UTR shortening of both the PKR and SMC1A genes’ mRNAs in THP-1 cells, being 0.349 and 0.254, respectively. Additionally, qPCR fold change values of IFN production with VSV-G demonstrate a decreased abundance of these innate immunity signalling molecules. Such data imply VSV-G’s immense role in host cellular immune dysregulation amidst viral infection, bringing into question the safety of recombinant VSV vaccines and calling for future studies that investigate the exact mechanisms utilized by the VSV-G protein for pseudotyping purposes.