Dissertation: Flexible, Real-Time Shaping of Zebra Finch Vocal Learning

Advisor: John Pearson, Ph.D.

• Built adaptive, closed-loop real-time pipelines for streaming analysis and reinforcement of zebra finch vocalizations using Python, LabVIEW, and \0{}MQ
• Developed and deployed deep learning models (VAEs, CNNs) and ML classifiers (KNN, RF, SVM) for high-dimensional spectral and temporal song feature extraction
• Designed and conducted closed-loop behavioral reinforcement experiments; analyzed high-dimensional song features in response to real-time auditory feedback
• Supervised an undergraduate research assistant

Behavior and neural dynamics of mouse models of neuro-developmental and psychiatric disorders

Advisors: David Carlson, Ph.D. & Kafui Dzirasa, M.D., Ph.D.

• Developed pipelines for fine-grained behavioral and neural analysis across multiple assays using deep learning, ML, statistical topic models, and image/video processing
• Characterized behavioral and population-level LFP dynamics in genetic mouse models of ASD and chronic stress via spectral and time-series analysis
• Supervised an undergraduate CS independent study in ML for quantitative behavioral analysis (2019–2020)

Honors Thesis: Low-Dimensional Mapping of Corticostriatal Circuitry Dynamics Underlying Pair Bonding

Advisor: Gordon Berman, Ph.D.

• Developed pipelines for high-dimensional neural data analysis using dimensionality reduction, clustering, and statistical topic models
• Characterized corticostriatal circuit dynamics underlying social behaviors in mouse models of ASD via multi-electrode array recordings and optogenetic manipulations

Unpublished Honors Thesis/NSF GRFP Research Proposal: The Effects of Metabolic Regulation on Local Field Potentials in the Basolateral Amygdala during Emotional Behaviors

Advisors: Donald Rainnie, Ph.D., Chia-Chun Hsu, Ph.D. & Teresa Madsen, Ph.D.

Funded by the Emory University’s S.I.R.E. Research Partners Program (2016–2017)

• Assessed the role of cAMP metabolism in anxiety-like and social behaviors via intracerebroventricular infusion
• Developed automated pipelines to quantify social and non-social behaviors in rat models of ASD, reducing analysis time and human bias
• Assessed fear learning, recall, and extinction via Pavlovian fear conditioning and extinction paradigms

PEDOT:PSS microwires for selective stimulation of single cells

Advisors: Craig Forest, Ph.D., Christine Payne, Ph.D., & Scott Thourson, Ph.D.

Funded by the Georgia Institute of Technology’s S.U.R.E. Program

• Characterized spatially selective electric fields of polymer-coated microwires using ex vivo patch-clamp electrophysiology and impedance spectroscopy
• Validated experimental observations against COMSOL physics simulations