Using Robots to Diagnose Autism in Children
Croatian researchers introduce Robot Rene to observe and diagnose autism in young children.
ZAGREB—Four-year-old Luka is enthusiastic about his new friend.
“He can write. He can sit down. And he can drink water,” he says.
The two also like to dance together.
But it is an unusual friendship. Luka is a child with autism spectrum disorder (ASD), a developmental disability that inhibits social, behavioral, and communications skills. And his friend, Rene, is a robot.
They met thanks to a joint project of the University of Zagreb’s Faculty of Education and Rehabilitation Sciences and its Faculty of Electronics and Computer Sciences. The initiative aims to use robots to improve the diagnosis and assessment of children with the disorder, a process that until now has been highly complex and subjective.
The robot is intended to assist, not replace the clinician.
“For children with autism, the robot is a stimulus that is very simple and always the same,” says researcher Jasmina Stosic. “Its eyes are always in the same place. Its mouth is always in the same place. People are rather complicated for such children because when we talk we make various gestures. And one day we’ll wear a red t-shirt and the next day, a blue one. The robot is one constant stimulus, and the children don’t need to think about so much different information and instead can concentrate on the essence.”
Because of the difficulties of assessment, ASD children are often diagnosed quite late. In Croatia, most ASD children are only diagnosed and begin receiving specialized education and therapy at around 5 or 6 years of age.
Marija Cukelj, whose 4-year-old son Filip is part of the robot project, has experienced the frustration of trying to get specialist assistance for children who are younger than this.
“Filip started to close down when he was 14 or 15 months of age,” he says. “He stopped looking people in the eye—even us, his parents. Within about two months we started looking for help. Unfortunately, it took time because he was 1 1/2 [years old] and we were told: ‘The child is too small, anything could happen. We need to wait and see.’ It took a year between the time we told doctors something was wrong and when our son got a reliable diagnosis. We lost that time.”
According to Cukelj, meeting Rene was a ground-breaking moment for Filip.
“The first time he saw the robot, he simply sat on a chair and watched it,” Cukelj says. “For Filip, who is so energetic and who is calmed down by very few things, it was a great success to see him sit down and carefully watch something.”
Researcher Maja Cepanec says that Rene has elicited positive reactions in trials with children so far.
“Children with attention-deficit issues, who have trouble making eye contact, react relatively well to the robot,” she says. “They watch it and they are excited about it. So far, our experiences have been relatively positive.”
Rene was made in France and several research institutions around the world have been using similar robots to work with ASD children. The researchers in Croatia, however, are focusing on using the robots to develop a standardized diagnostic protocol.
Cepanec says the goal of the project is to use the robot to collect data on, for instance, recurrences of repetitive behavior, and to conduct uniform testing of particular behaviors such as drinking from a cup or addressing the robot by name.
“We believe that ASD assessment, empowered by advanced behavioral and social-signal processing, might become more objective and reliable,” says Cepanec. This improved process would also include using objective quantitative metrics.
“The robot is equipped with a camera, microphones, speakers, and it can record things we might miss,” says Cepanec. “It can code a child’s vocalizations, his or her closeness to the parent, how many times the child initiates communication, how much eye contact the child makes, and so on.”
In order to build on the project’s promising start, the researchers are working with the Croatian National Science Fund and European sources in hopes of expanding the number of children they can include in their research.