Leading the Vanguard: Maithilee Kunda

Maithilee Kunda smiling

Vanderbilt Kennedy Center (VKC) investigator Maithilee Kunda, Ph.D., is an assistant professor of Computer Science and Computer Engineering. Dr. Kunda’s research focuses on the role of visuospatial cognition in learning and intelligent behavior, especially in developmental disabilities such as autism and various learning disabilities. She uses methods from artificial intelligence (AI) to build computational models of reasoning that can be used as testbeds to better understand how differences in basic cognitive processes such as memory, attention, etc., affect behavior.

In the interview below, Kunda shares what inspires her research in developmental disabilities, what she’s learned through her work, and how membership with the Vanderbilt Kennedy Center helps her achieve her goals.

Tell me about your attraction to developmental disabilities research. Do you have a personal connection to disability?

I first became interested in studying autism when I read the autobiography “Thinking in Pictures” by Temple Grandin, who is a professor of Animal Science at Colorado State who also happens to be on the autism spectrum.  Dr. Grandin writes about how her propensity for visual thinking gives her great benefits in her job as an equipment designer for the livestock industry, a role in which she has gained national prominence.  She also writes about how her visual thinking style has led to challenges throughout her life in areas like communication and understanding abstract concepts.

My research aims to understand human visual cognition in all its myriad forms, including how and why different people show different patterns of strengths and weaknesses in visual cognition.  Looking into atypical forms of cognition, like those found in persons with developmental disabilities, gives us a window into the development and workings of the human mind.  In addition to advancing the basic science of cognition, I also aim to use our research findings to improve methods for assessment and education for people with a diversity of cognitive backgrounds.

What are your current research interests and what problem(s) or challenge(s) does it address?

First, several collaborators and I have a special education research grant from the Institute of Education Sciences to develop a new, technology-based intervention for helping adolescents on the autism spectrum improve their social reasoning skills.  The original idea for this project came from conversations with Gautam Biswas, Ph.D., another computer science faculty member and VKC Member who specializes in educational technologies, particularly in the area of science education.  We had this idea that, perhaps, an effective strategy for teaching social reasoning to young people on the autism spectrum could emulate current practices for teaching scientific reasoning in general education, especially those that lean heavily on visual supports for students.  Both scientific reasoning and social reasoning often involve thinking about complex situations, with multiple interacting parts, some of which are observable (like pH levels in a lake, or people’s facial expressions) and some of which are unobservable (like the health of fish living in a lake, or people’s internal beliefs).  Our project essentially involves adapting Betty’s Brain, a science-centered educational technology platform developed by Professor Biswas in prior research, into a new, social-reasoning-centered educational game called Film Detective.

A second project involves using AI to improve standard visuospatial cognitive assessments, in order to better identify the source of individual differences on visual cognitive tasks — to understand not just how well a person can perform a particular visual task, but also how they are performing it. We have been studying the block design test, which is a very well-known test of visuospatial reasoning, and using a combination of machine learning, computer vision, data mining, and computational cognitive modeling to get a more complete picture of a person’s block design task performance.

Do you have a story about a research participant or a breakthrough that illustrates the impact of your work?

Part of our research involves building computational cognitive models that essentially use “visual thinking” to solve problems from visuospatial intelligence tests.  Once, I was presenting this work to a group of high school students who were on the autism spectrum. I was explaining to them how psychology and neuroscience researchers had found that most typically developing individuals solve many of these test problems verbally, even though the problems themselves are visual; there is evidence that many people convert the visual problems into something like mental words, and then use their language processing skills to reason through the problem. And, along these lines, most prior AI models had done a sort of “verbal reasoning” to solve the test problems.  Our model provided a contrasting perspective, in which the reasoning was all done visually.

Before I could even get to the end of this explanation, the students stopped me in bewilderment.  “Wait, wait,” they said, “are you telling us that it’s possible to solve these problems VERBALLY?!”  They were so flabbergasted by this idea, and so fascinated to hear about these “other minds” (of typically developing individuals) that worked so differently from their own.  We were all laughing, because the students thought it was a hilarious idea that anyone would think in this very odd way.  I had to convince them; I remember laughing and saying, “No, no, I promise this is what the research really does say,” as they all looked at me sort of skeptically.  It was especially funny to me because, this was the exact same look of skepticism I used to get when presenting my research at AI conferences, and many of the researchers would say, “Are you telling us that it’s possible to solve these problems VISUALLY?!”

To me, this was a great reminder of how, sometimes, the things that we think are obvious and certain are, in fact, not at all obvious and certain to other people, especially (but not only) when people are coming from different cognitive backgrounds.  And it’s why I think cognitive science research is such a critical part of how we understand developmental disabilities.  No two people think exactly alike, and the better we can understand how individuals think, the better we can design supports and tools and assessments to help people overcome their individual challenges and really make the most of their individual potential.

What are your reasons for becoming a Vanderbilt Kennedy Center (VKC) Investigator? How does the VKC enhance the work you do? 

The VKC offers such an amazing community of expertise and collaboration.  I can’t say how many times I have come up against a tough question in my research, only to realize that some of the world’s leading experts are right there in the VKC for me to ask!  I have been able to talk to, and work with, experts on everything from neuroscience and neuropsychology to special education and statistics. And all just a short 10-minute walk from my office!

In addition to its people, the quality of my work has also just been utterly transformed by VKC research resources, like for example the participant and family recruitment databases maintained by TRIAD.  There is no way that we could be doing the kinds of research we are doing without this level of infrastructure and support that the VKC provides. For me, especially coming from my computer science background, the VKC has really served as a springboard to enable me to translate my ideas into robust and high-quality research studies.

Elizabeth Turner is associate director of VKC Communications.

Giving Banner

This is a monthly email of Vanderbilt Kennedy Center Notables published by the Communications staff of the Vanderbilt Kennedy Center. Between issues of Notables, you can stay up to date on the latest Vanderbilt Kennedy Center news, information, and resources via the Vanderbilt Kennedy Center’s Facebook page.