A recent study published in Nature Medicine examines a novel AI-driven app called SenseToKnow that can test to see if a child is on the autism spectrum. This app was created by researchers at Duke University and holds the potential to help parents and health care providers identify specific behavioral indicators that could help assess and treat children who are on the autism spectrum.
“Autism is characterized by many different behaviors, and not all children on the spectrum display all of them equally, or at all,” said Dr. Geraldine Dawson, who is the director of the Duke Center for Autism and Brain Development and a co-author on the study. “This screening tool captures a wide range of behaviors that more accurately reflect the complexity and variability found in autism.”
SenseToKnow is built on previous research that examined how specially tailored movies could help diagnose autism by monitoring a child’s eye movements. Now, researchers have designed SenseToKnow to conduct a larger selection of behaviors, including blink rate, head movements, gaze patterns, and facial expressions. Not only is SenseToKnow easy to use, but it has also exhibited efficiency race, ethnicity, and sex, and given it’s an app, that means SenseToKnow could be used to evaluate potential autism behaviors anywhere, including a patient’s home.
“The AI we’ve built compares each child’s biomarkers to how indicative they are of autism at a population level,” said Sam Perochon, who is the study’s lead author and a PhD student in the lab of study co-author, Dr. Guillermo Sapiro. “This allows the tool to capture behaviors other screening tests might miss and also report on which biomarkers were of the most interest and most predictive for that particular child.”
As part of the study, the researchers administered SenseToKnow to 475 children, 49 of which were later diagnosed with autism and 98 exhibiting what’s known as developmental delay but without autism. SenseToKnow successfully recognized 87.8 percent of the children who had autism while recognizing 80.8 percent of the children who did not have autism. Additionally, 40.6 percent of children who were identified by SenseToKnow as exhibiting autism traits were later medically diagnosed with autism, as well.
A toddler playing a bubble-popping game which is part of a 10-minute tablet app that can assist in vetting children for autism behavioral symptoms. (Credit: Duke University)
For next steps, the team is presently in the middle of a study where parents are using the app at home with the goal of further establishing the success of SenseToKnow while monitoring how a child progresses when part of an early intervention program.
How will SenseToKnow help diagnose and treat autism in children in the coming years and decades? Only time will tell, and this is why we science!
As always, keep doing science & keep looking up!