Alex Hsu Earns Biological Sciences Fellowship
By Kirsten HeuringMedia Inquiries
- Associate Dean for Communications, MCS
Alexander (Alex) Hsu, a doctoral candidate in Carnegie Mellon University's Department of Biological Sciences, was awarded the Glen de Vries Fellowship for the Biological Sciences for his research using computational methods to analyze neuroscience data.
"It has always been my plan to learn how to analyze big data better and establish where machine learning can benefit neuroscientists," Hsu said. "I wanted to basically learn how to analyze big data with machine learning."
Hsu has plenty of experience with big data. Between finishing his master's of science degree in anatomy and neurobiology from Boston University and starting his Ph.D. at Carnegie Mellon, he worked on speech recognition at IBM Watson Research Center, the same artificial intelligence that was featured on Jeopardy.
As part of the Yttri lab, Hsu has created software to process large behavioral data to find correlations in neuroscience data. He records multiple points of data from mice, monitoring neural activity in their motor cortex while recording their actions 24 hours a day. The data is fed into a computer application he created, which makes connections to determine which neural impulses cause which movements.
"For decades, we've been training subjects to do very specific, very reduced tasks," said Eric Yttri, an assistant professor of biological sciences. "We use these in order to have just one behavior to study, but that's not what our brains are designed to do. In an ideal situation, you'd be able to just watch a subject do whatever it's going to do and then extract the neural correlates of those behaviors. We designed a tool to actually do just that."
Hsu and Yttri want to apply their research to Parkinson's disease patients. By recognizing what impulses cause which movements, they could find a way to target and limit those impulses.
"Parkinson's patients right now have DBS or deep brain stimulation devices, an implant in their brain, and currently, the technology is constantly stimulating regardless of what the patients are doing," Hsu said. "If we understand that a certain brain circuit has gone awry for whatever reason during a proposed movement, we can use our software to selectively stimulate when the action is detected."
Hsu and Yttri share the program with other labs across the country, including at Columbia and Princeton. Designed for Parkinson's, the program can be used to study other neurological conditions, such as chronic pain and OCD. While conducting research, Hsu has taken on a mentorship role for other members of the Yttri lab. He has advised four masters' students and five undergraduate students. He aims to make them feel valued in the lab, ensuring their contributions are noted, especially on any code he creates.
"If they had helped me with a certain part of a code, I would create a GitHub account that properly credits them," Hsu said. "It sounds really small, but that makes people want to continue because it feels like it's their own property. I think mentoring is really fun when you allow them to flourish."
Hsu's research is a prime example of what an interdisciplinary mindset can create. He has taken classes on the School of Computer Sciences on machine learning to further develop his computational skills and apply what he's learned to neuroscience.
"Alex's ability to think those within the realm of computer science as well as neuroscience and to make those bridges to make meaningful steps forward, it shows a lot of his maturity, and bridging fields is always one of the most difficult things we can do," Yttri said.