University of Pennsylvania, Department of Neuroscience, Philadelphia, PA
This summer, I had the pleasure of working at the Luo Lab in the neuroscience department at the University of Pennsylvania. The broader focus of the lab is to understand how the somatosensory system functions.
The objective of my project was to automate scratching behavior analysis in mice and guide future experiments for studying acute and chronic itch pain. The end goal of the project is to connect the mouse scratching patterns to human conditions in order to suggest possible treatments. I handled mice, including perfusions and brain dissections, and performed behavioral experiments and analysis. I utilized Python to analyze data and deep learning to generate accurate models for research. By being exposed to coding and AI, I learned how to take advantage of cutting-edge technologies to solve specific problems in research and automate manual work. For example, labeling 36,000 picture frames (20 videos, 20 min/video, 30 frames/second) manually is truly laborious and time consuming; this same task can be accomplished by a computer in seconds.
The specific mechanism I studied was the “itch” sensation, which can be acute or chronic. A portion of my time was dedicated to watching, annotating and analyzing recorded videos of mice. I then modified the existing Python programs to convert from video to picture frame formats, and went through each frame individually to precisely annotate any scratching behavior. After tracking the pattern and analyzing it, I created a model using deep learning that will predict future mouse scratching trains. This model was then checked by comparing it with the manually annotated scratching trains. Prior to this project, I was unaware of how intertwined AI and medicine could be and how AI can help in research by eliminating manual tasks. This will surely impact the selection of my future courses.
Once a week, we had a lab group meeting during which each member shared their findings of that week. These meetings helped me improve my communication skills with scientific professionals. Moreover, the overall project helped me develop independence and become self-directed in learning—essential skills for research.
I entered this internship with very little experience in the wet lab and none in deep learning/coding. I started as an aspiring pre-med student but left with an even stronger desire to become a doctor. Being in a position to help people is what motivates me. Before this internship, I had a very broad background in biology and none in neuroscience and computer science. Yet, this internship has sparked in me a huge interest in both of these fields.
I consider myself privileged to have worked with such an incredible team. I truly had a remarkable experience and would like to extend my greatest gratitude to the Class of 1951 and the ’68 Center for Career Exploration for their support.