College: Brooklyn College
Awards: National Science Foundation Graduate Research Fellowship, 2013
Making Autism Compute
With an estimated 1 million to 1.5 million Americans having an autism spectrum disorder and the prevalence of autism believed to have risen to 1 in 88 births, there needs to be a reliable method of diagnosis so that young children can get the early intervention that will help them to live full lives.
Sarah Ita Levitan (Brooklyn College, B.S. in computer science, 2013) hopes to develop an objective, computer-based system that would analyze children’s speech, looking for patterns that would identify those with autism spectrum disorders.
Her idea helped her win a 2013 National Science Foundation Graduate Research Fellowship. This three-year, $126,000 grant is the most prestigious award for graduate study in the science, technology, engineering and mathematics (STEM) disciplines, and it will support her research in Columbia University’s computer science doctoral program.
“As of now, there isn’t a simple diagnostic test for autism,” Levitan says. “It is done by a set of subjective assessments to evaluate if children have behavioral disorders. There have been a lot of studies in psychology on different speech patterns. They have looked at things like echolalia, where children repeat things back to you, or turn-taking in conversation. Whereas most people try not to cut each other off, some individuals with autism spectrum disorders seem to have trouble communicating and turn-taking; they tend to be a lot slower in responding or cut people off.”
Her interest in autism spectrum disorders dates to high school and college, where she volunteered and then worked for the Hebrew Academy for Special Children in Brooklyn. “While working with these children, I observed first-hand how early detection could make a world of a difference,” says Levitan.
At Brooklyn College, she worked on a computational biology research project with Dina Sokol, an associate professor of computer and information science. “She studies tandem repeats in DNA, which are used to diagnose diseases and in human identity testing. She developed an algorithm to find the repeats, but there are so many, it’s hard to know what to do with them. I’ve been working with her on clustering the data into groups of similar repeats,” Levitan says.
She credits Sokol with being “an incredible mentor. She introduced me to research as an undergraduate, encouraged me to apply for an undergraduate research grant, and she is the faculty advisor for the Women in Computer Science Club.”
At Sokol’s suggestion, Levitan applied for a Distributed Research Experiences for Undergraduates award from the Computer Research Association’s Committee on the Status of Women in Computing Research. That led to spending the summer after her junior year conducting research in the laboratory of Julia Hirschberg, director of Columbia’s Spoken Language Processing Group.
She studied entrainment, the common phenomenon of people unconsciously sounding like one another during a conversation, through a computer analysis of supreme court discussions. “In conversation, people tend to adapt similar speech patterns, such as adopting a higher pitch or speaking louder. There’s a lot of fascinating psychology involved as well as computer science,” she says.
The experience encouraged Levitan to start a chapter of the women’s computing organization, the Association for Computing Machinery-Women, on campus. “As an undergraduate, I’ve seen a major gender imbalance in computer science. I’d like to encourage more women to consider majoring in computer science and provide a forum for networking and support. For many women, computer science isn’t on their radar when they consider what fields to go into. That is unfortunate because there are so many interesting and exciting research areas where a degree in computer science can take you.”