Short home videos may provide enough information to determine whether or not a child is on the autism spectrum, a new study suggests.
By analyzing one- to five-minute videos of young children, researchers say that they were able to distinguish those with autism from typically-developing kids with nearly 90 percent accuracy.
For the study, a group of non-expert raters were asked to watch videos of 162 children, 116 of whom had autism and the remainder who had no diagnosis. All of the videos showed the child’s face and hands, the opportunity to use toys or other objects as well as the option for social engagement. The raters — who did not know about the children’s diagnostic histories — were asked to answer 30 questions with a simple “yes” or “no” about whether each child made eye contact, expressed emotion and other behaviors.
Information from the raters was then input into eight different algorithms. The most successful model relied on ratings for five behaviors and was able to accurately label 94.5 percent of children with autism and 77.4 percent of those without.
A subsequent test of the five-question model involving an additional 66 videos correctly flagged 87.8 percent of those with autism and 72.7 percent of the kids without.
“We showed that we can identify a small set of behavioral features that have high alignment with the clinical outcome, that non-experts can rapidly and independently score these features in a virtual environment online in minutes, and that the model we used to combine those features is effective in producing a score that matches the clinical outcome,” said Dennis Wall, an associate professor of pediatrics and biomedical data science at the Stanford University School of Medicine who worked on the study published recently in the journal PLOS Medicine.
Currently, many children face long waits to be evaluated for autism. Though the developmental disorder can be diagnosed at age 2, most children are not identified until after age 4, according to the Centers for Disease Control and Prevention.
Those behind the study say they hope their model can one day be used to help speed the process of getting an autism diagnosis, allowing children to enter treatment sooner when it’s considered most effective.
“Our long-term dream is that a tool like this will give general pediatricians more confidence in making diagnostic decisions about autism and other developmental disorders,” Wall said.