Claire Lee: using machine learning to study human cognition

Wednesday, May 13, 2020
by Sharon Adarlo

Claire Lee, 23, Class of 2020


Lee’s bachelor’s degree is in neuroscience, and she is set to complete requirements for the Certificate Program in Statistics and Machine Learning at the Center for Statistics and Machine Learning (CSML) and the Certificate Program in Cognitive Science.


After her sophomore year, Lee joined the lab group of Yael Niv, professor of neuroscience and psychology, who studies the neural and computational processes underlying human learning and decision-making. In this setting, Lee was introduced to computational psychiatry while working with Daniel Bennett, a postdoc in the lab.

In the Niv lab, I learned interesting quantitative approaches to studying human cognition and behavior. We recruited participants and had them play games. We then linked their patterns of behavior to various computational models using methods from neuroscience, behavioral economics and machine learning,” said Lee.

Lee delved further into computational psychiatry during a gap year after her junior year with the Computational Psychiatry group at IBM Research.

“Studies by members of the group at IBM have shown that automated analysis of natural speech can help diagnose and even predict the onset of certain psychiatric conditions,” said Lee. My mentors – Sara Berger, Guillermo Cecchi, and the rest of the group – exposed me to the data-driven approaches to understanding cognitive disturbances using AI.”  

In her senior year, Lee continued her interest in data-driven psychiatry through her project for SML 310 - Research Projects in Data Science, taught by Michael Guerzhoy, lecturer at CSML and advisor of Lee’s project following the course’s completion.

Unlike most other illnesses, many psychiatric disorders like depression and bipolar disorder are largely based on subjective measures like patients’ self-reports and responses to questionnaires. As a result, an alarming number of individuals with bipolar disorder are initially misdiagnosed with unipolar depression, which poses serious health concerns,” she said.

Lee’s research over the years has culminated in two projects: her senior thesis project whichuses computational modeling to examine mood-driven risk-taking behavior and her independent project in which she harnesses artificial intelligence to classify and predict mood episodes (mania and depression) for YouTubers with bipolar disorder.

After graduation, Lee plans to continue research and pursue COVID-19 related volunteer work during her glide year before medical school.

By equipping me with both foundational and comprehensive tools, I’m thankful that CSML has prepared me well for a career in the increasingly quantitative, objective and data-driven field of medicine,” she said.

Extracurricular activities:

Lee is a former violist in the Princeton University Orchestra, former managing editor for The Daily Princetonian, and founder of Princeton MediHack. The latter is a student organization that hosts 36-hour tournaments for participants to hack promising medical solutions, from computational neuropsychiatry to machine learning, for drug discovery.

For fun:

Lee unwinds by watching movies and visiting cafes with her friends to talk and hang out.