Researchers use artificial intelligence to help early detection of autism spectrum disorder | Colleges and Universities

Can AI be used to help early detection of autism spectrum disorder? That’s a question researchers at the Arkansas Agricultural Experiment Station and the University of Arkansas are trying to answer. But they are taking an unusual path.

Han Seok Seo, associate professor of food sciences at the Agricultural Experiment Station, research arm of the U of A System Division of Agriculture, and Khoa Luu, associate professor of computer science and computer engineering at U of A, will identify sensory cues from different foods. In both children with the neurotypical pattern and those known to be on the spectrum. Machine learning technology will then be used to analyze biometric data and behavioral responses to those smells and tastes as a way to detect indicators of autism.

There are many behaviors associated with autism, including difficulties with communication and social interaction or repetitive behaviors. Autistic people are also known to display some abnormal eating behaviors, such as avoiding certain foods, specific mealtime requirements, and antisocial eating. Avoiding food is of particular concern, as it can lead to malnutrition, including vitamin and mineral deficiencies. With this in mind, Seo and Luu intend to identify sensory cues from food items that lead to atypical perceptions or behaviors while eating. For example, scents such as mint, lemon, and cloves are known to elicit stronger reactions in people with autism than in those without them, which can lead to increased levels of anger, surprise, or disgust.

Seo is an expert in the fields of sensory science, behavioral neuroscience, biodata, and eating behavior. He is organizing and leading this project, including examining and identifying specific sensory cues that can distinguish children with autism from children without autism with regard to cognition and behaviour.

Luu is an AI expert with specializations in biometric signal processing, machine learning, deep learning, and computer vision. It will develop machine learning algorithms to detect ASD in children based on unique patterns of cognition and behavior in response to specific test samples.

This is the second year of a three-year, $150,000 grant from the Arkansas Institute of Biological Sciences.

Their goal is to create an algorithm that displays equal or better performance in early detection of autism in children when compared to traditional diagnostic methods, which require health and psychiatric professionals to do assessments, longer assessment periods, provider-provided questionnaires and additional medical costs. Ideally, they would be able to validate a low-cost mechanism to help diagnose autism.

While their system likely won’t be the last word in diagnosis, it can provide parents with an initial screening tool. Ideally, children who are not candidates for autism spectrum disorder will be screened while ensuring that the most likely candidates are followed up for a more comprehensive evaluation.

Siu said he became interested in the possibility of using multisensory processing to assess autism spectrum disorder when two things happened: he began working with graduate student Asmita Singh, who had a background in working with students with autism, and the birth of his daughter.

Like many first-time parents, Seo paid close attention to his newborn baby, worried that she would be healthy. When he noticed she wouldn’t make eye contact, he did what most nervous parents do – he turned to the internet for an explanation. He learned that avoiding eye contact was a well-known feature of autism spectrum disorder.

While his child has not been diagnosed with autism, it has sparked his curiosity, particularly about the role of smell and taste sensitivities in autism spectrum disorder. Additional conversations with Singh led him to believe that other anxious parents might benefit from an early detection tool — perhaps inexpensively to alleviate fears at first.

Subsequent conversations with Luu led the couple to believe that if machine learning, developed by Luu graduate student Xuan-Bac Nguyen, could be used to identify normal reactions to food, it could be taught to recognize atypical responses as well.

Seo is looking for volunteers between the ages of 5 and 14 to participate in the study. Both children with the neurotypical pattern and children already diagnosed with ASD are needed for the study. Participants receive a $150 eGift card to participate and are encouraged to contact Seo at hanseok@uark.edu.

Leave a Comment