Skip navigation

Calls to Action

Robots as Autism Co-therapists

Study suggests that including interactive robot in therapy helps some children with autism learn social skills that hold up in real life
May 02, 2013


It’s become a truism that many individuals with autism gravitate toward gadgets –even preferring them to people. This has inspired great interest in developing devices – including robots – that can interact with individuals on the autism spectrum.

However, little of this research has involved evaluating how effectively robots might be used to help children with autism overcome social-communication disabilities, says University of Notre Dame psychologist Joshua Diehl, Ph.D.

Dr. Diehl presented the promising results of just such a study at this year’s International Meeting for Autism Research (IMFAR), in San Sebastian, Spain.

Dr. Diehl’s team found that a robot “co-therapist” can, in fact, increase gains in social skills among children with autism spectrum disorder (ASD). However, the benefits varied widely from child to child.

The investigators enrolled 19 children – 17 boys and 2 girls, ages 6 to 13. Their communication abilities varied widely. All the children completed 12 one-hour sessions of Applied Behavioral Analysis therapy with a trained human therapist. In six of these sessions, a robot “co-therapist” joined the mix. Hidden from view, a second human therapist operated Kelly. The robot had a distinctly “computerized” voice and limited mechanical movements (reaching, fist pumping, etc.).

Modeling communication
Through the robot, the hidden therapist modeled socially appropriate questions and responses. (“How was your day?” “That is so cool.” etc.) It also encouraged the child to interact with the human therapist in the room.

Roughly half the children received the robot-assisted therapy for their first six sessions. The others interacted with the robot during their final six sessions. 

The researchers measured social behaviors with standardized checklists during the sessions and parents filled out similar checklists outside of the sessions. They also recorded the children’s emotional levels using wristband biosensors.

“Overall we found that including the robot co-therapist increased gains in social skills,” Dr. Diehl says. “Importantly, we found that children who learned skills while interacting with the robot would then use these skills with their human therapist and at home.”

Robot boosted gains for most

Across all the sessions, 17 of the 19 children showed an overall improvement in the social skills targeted by the therapy. The majority showed greater gains during the six sessions that included the robot.

However, the children showed great variability in their responsiveness to the robot. For example, Dr. Diehl recalls one child who politely set the robot aside so he could focus on his human therapist.

Importantly, the researchers acknowledge that the cost of two therapists plus a robot would make such an approach impractical for broad use. Instead, the aim was to determine whether or not such an approach could produce real benefits.

Encouraged by the results, the team hopes to adapt “robot co-therapy” into a more practical form. One possibility is for a therapist to use video clips of the robot that could be cued from, say, a laptop in the therapy room, Dr. Diehl says. Another possibility might be to use robots to allow consultants to join therapy sessions from remote locations.

“As parents and therapists know well, the most effective interventions build on a child’s strengths and interests,” comments Lauren Elder, Ph.D., Autism Speaks assistant director of dissemination science. “As they continue their research, we hope these investigators find ways to adapt their findings into practical innovations that make therapy more effective.”

More IMFAR 2013 coverage here.