DURHAM, N.C. - New technology being developed at Duke University could help identify symptoms of autism in toddlers.
Virginia Beach resident Chris Mahon recalls when he and his wife, Carolynn, found out that his son, Luke, has autism.
“It was really around 18 months where we finally had a concrete diagnosis,” Mahon told News 3. “When he was younger, if there were loud noises, he would cover his ears to try and block out that sensory.
“My son does not hold eye contact with people very well. If he does, he’s very limited,” Mahon added. “To be able to see him grow and overcome any adversity is so heartfelt.”
Mahon said early detection and intervention was crucial for his family.
“As the mind is developing, as the child is developing, it’s kind of a way to help steer them in the right path,” he said.
Dr. Geraldine Dawson, Director for the Duke Center for Autism and Brain Development, and other researchers have been studying an app they've created as a potential early autism screening tool.
“What we wanted to do is to develop a tool that could be feasibly used in a pediatrician’s office or even at a parent’s home,” Dawson told News 3. “It’s remarkable, to me, that we’re able to do this on an iPhone or an iPad.”
The app measures eye gaze patterns while kids watch short videos on a phone or tablet, using technology to find out whether they're looking more at people in the video or objects.
“One of the early signs of autism is that these young children are not paying attention to social information in their environments, and actually they’re much more drawn to paying attention to toys and other objects,” Dawson said.
“We’re recording, with the camera that’s embedded in the device, the child’s gaze behavior and then the videotapes are uploaded, and then using a technique called computer vision analysis, the engineers are able to determine where that baby was looking,” Dawson added. “We found that by tracking where the babies are looking, we could then predict which one of these toddlers would go on to be diagnosed with autism.”
Dawson said the results in the study, so far, are promising.
“Using just these three measures, two measures of social attention, one measure of tracking conversation, we were able to predict with 90% accuracy whether this child would be diagnosed with autism,” Dawson said.
While Duke researchers continue studying the app, including automated feedback, Mahon believes, if the app is released, it could be an asset for families.
“It may be a great tool to kind of take to their doctor, or their pediatrician,” Mahon said.