Using machine learning to analyze social behavior


Portrait of Megan Nelson, an undergraduate student in the ASU Department of Psychology. Nelson stands on a sidewalk under a tree. She has shoulder-length brown hair and wears a burgandy blouse while smiling at the camera.

Megan Nelson, a third-year undergraduate in the Department of Psychology and a student in Barrett, The Honors College at Arizona State University, is changing how artificial intelligence and computational modeling are used to understand behavior.

|

Megan Nelson, a third-year undergraduate in the Department of Psychology and a student in Barrett, The Honors College at Arizona State University, is changing how artificial intelligence and computational modeling are used to understand behavior. 

She was recently selected as a SOLUR Scholar by the School of Life Sciences for her work. The SOLUR Scholar is a paid position where students work with a faculty member on a research project and also participate in a weekly seminar with other students in the cohort. Throughout the year, the scholars attend research symposiums and present their research findings at conferences. 

As a member of the SOCIAL (Study of Circuits in Adolescent Life) Neurobiology lab with Assistant Professor Jessica Verpeut, Nelson helps study the development of neural circuits and structure in early life to understand behavior. She is majoring in psychology and biological sciences with a concentration in neurobiology, physiology and behavior. Additionally, she is minoring in math and earning a certificate in computational life sciences.

“Currently, I’m interested in cerebellar perturbations on social behavior in adolescent mice,” Nelson said. “Previously, the effect of cerebellar injury has been linked to autism in clinical studies. Some other studies have tested the effects of cerebellar injury on a wide variety of behaviors, like associative learning, anxiety-like behaviors and even social behavior.”

One of the major challenges of studying social behavior in animal models is that it is so complex and there are so many contextual variables and body parts involved. 

“Social behavior hasn't been able to be tracked super well until extremely recently with the development of machine learning algorithms,” Nelson said. “The rise of these machine learning algorithms also allows us to use different tests to measure social behaviors that were previously impossible.”

Nelson presented her initial research findings last year at the Arizona Psychology Undergraduate Research Conference and has since expanded her research into computational modeling.

She is hoping to use this year as a SOLUR Scholar to gather more research and develop new models and algorithms to better understand the social behavior of adolescent rats. At the end of the year, she will present her findings at the School of Life Sciences Undergraduate Research Symposium. 

Long term, her goal is to go to graduate and earn a PhD in neuroscience to further investigate the origins of social behavior. 

“If I was given unlimited resources to investigate a question further, I would be interested in studying the role of genetics or specific genes on the development of the brain. I’d like to understand the genetic connections and see how developmental changes in behavior start on a genetic level,” Nelson said.

More Science and technology

 

Food insecurity dashboard SGSUP

ASU, St. Mary’s Food Bank partner to tackle food insecurity in Arizona

Arizona State University and St. Mary’s Food Bank (SMFB) have joined forces to create an interactive data dashboard that tracks…

Female Mojave desert tortoise

Study reveals genetic insight into desert survival

The deserts of the American Southwest are home to the Mojave and Sonoran desert tortoises, two seemingly similar yet genetically…

Image of a human brain and a tractor.

Study reveals lasting effects of common weed killer on brain health

Environmental exposure to toxins in the air, water or certain chemicals can increase the risk of ill health effects, including to…