Google Glass System Assessed for Autism Therapy

Alex, one of the participants in a study of a Google Glass system to treat autism spectrum disorder (Stanford University)

3 August 2018. A pilot study evaluating an autism spectrum disorder therapy that combines Google Glass with a smartphone app using machine learning shows a decrease in autism-related behaviors among the children in the test. The system was developed at Stanford University medical school, with a description and test results appearing in yesterday’s issue of the journal npj Digital Medicine.

Autism spectrum disorder is a collection of neurodevelopmental conditions marked by communication difficulties and impaired social interaction, as well as repetitive and stereotyped patterns of behavior. Some 1 in 59 children in the U.S. have autism spectrum disorder, according to Centers for Disease Control and Prevention, with males 4 times more likely to have the disorder than females. Classic autism is considered the most severe form of the syndrome.

Medical and behavioral researchers, along with computer scientists at Stanford are seeking to make therapies for autism spectrum disorder more widely available to children and their parents. The authors led by autism researcher Dennis Wall note that families are advised to start therapy as early as possible, but face waiting lists as long as 18 months for specialized treatments for children with the disorder. Thus their system, called Superpower Glass, is designed to provide home-based alternatives to teach social skills to children with the condition.

“We have too few autism practitioners,” says Wall in a university statement. “The only way to break through the problem is to create reliable, home-based treatment systems. It’s a really important unmet need.”

Superpower Glass is built around Google Glass, a wearable interactive technology platform with plastic display lenses worn like eyeglasses. The Superpower glass software is designed to operate in the child’s daily environment during social interactions with family members and friends. The software, using Glass’s outward facing camera, employs machine learning to recognize 8 universal facial emotions, such as happiness, sadness, and anger, with images displayed on the Google Glass lenses providing feedback to the child. The system can also be operated as more structured games where the wearers guess the emotions of people in their field of view or the children act out emotions for others to guess.

The study is the Stanford team’s first assessment of the system’s feasibility among families with children having autism spectrum disorder. Wall and colleagues recruited 14 families to try out Superpower Glass with their children, 11 of whom were male and an average age just over 9 years. The families were asked to use Superpower Glass in 3 sessions per week, with each session lasting about 20 minutes. The families reported using the system in their homes on average for 72 days.

The study assessed participating children on a standard index of social impairment associated with autism spectrum disorder, known as Social Responsiveness Scale, or SRS-2. The results show an improvement in SRS-2 scores among the children, meaning lower scores since higher scores indicate more severe autism spectrum disorder behaviors. The team also evaluated the children’s abilities to recognize facial emotions, which also increased after the Superpower glass sessions. And in follow-up interviews, 12 of the 14 families reported that the children were exhibiting more eye contact than before, as well as finding the system fun and engaging for the children.

The researchers are now conducting a larger study that includes a control group, which was missing from this first effort.

More from Science & Enterprise:

• Robotics, A.I. Employed for Personal Autism Therapy
• 115M Euro Grant for Precision Medicine, Trials in Autism
• Cancer Drug Seen to Reverse Autism Spectrum Symptoms
• Umbilical Cord Blood Evaluated for Autism Treatment
• Autism App Adapts Facial Analysis Software

*     *     *