An app that can study people's facial expressions and emotional responses could one day be helpful in detecting autism signs in children, new research found.
The iPhone app, called "Autism & Beyond," was developed by scientists and software developers at Duke University in North Carolina and uses mathematical algorithms to automatically detect people's expressions and emotional cues, based on muscle movements in the face. The app is currently available as a free download from Apple's App Store, for families who are interested in participating in a six-month medical research study, the researchers said.
Children in the study will be presented with a short video clip designed to elicit emotional responses and social interactions. Using the iPhone's front-facing camera, the app will then measure the children's responses to the videos through video analysis and machine learning, the researchers said. [11 Facts Every Parent Should Know About Their Baby's Brain]
"A core component of the app is emotion," said Dr. Guillermo Sapiro, a professor of electrical and computer engineering at Duke University, who developed the algorithm.
The app works by "following facial landmarks that are automatically detected and from classifying emotions as well as head position," Sapiro told Live Science.
The app can be downloaded for children up to 6 years old, with components of the app differing based on the child's age, Sapiro said. Autism spectrum disorders often appear in infancy and early childhood, characterized by signs such as failing to make eye contact, not responding to his or her name being called, or playing with toys in unusual and repetitive ways, according to Autism Speaks, an advocacy organization in the United States.
Sapiro said that the app is not designed to be a self-diagnosis resource but could be used as a potential screening tool for autism. In the future, the app, which was developed with funding from the U.S. Office of Naval Research, could also be used to screen for other developmental health issues, including post-traumatic stress disorder (PTSD), traumatic brain injury and depression.
But it may be too soon to know if this app will be effective in the real world, said Dan Smith, head of innovative technologies at Autism Speaks, who is not involved with the current research. For autism in particular, there is no specific medical "test" to diagnose the health issue, which is estimated to affect about 1 percent of the population worldwide and more than 3.5 million Americans, according to the Autism Society.
"It's way too early to know whether [the app] picks up differences in people with autism or PTSD, and it's too early to say if those differences could predict or identify people with a condition better than [can] current methods," Smith told Live Science. "We just don't know yet if this type of information they're gathering is useful for improving diagnostics or assessing the risk."
However, Smith said he thinks that, at some point, "there will be some reliable markers to track certain stimuli with certain facial expressions or response to emotional stimuli."