Children with autism spectrum disorder are unable to sustain play, fantasy, and fluid social interaction. At least with real people. But psychologist and linguist Justine Cassell of Northwestern University in Evanston says that interaction with virtual peers releases hidden social skills in these kids.
A virtual child is a cartoonish-looking, gender-neutral 8-year-old that appears on a TV or projection screen. When it interacts with a real child, half of the action takes place in the real world, and half in the virtual world. For example, the virtual child "watches" the real child as he or she plays with dolls, thanks to sensors on the toys. The virtual child can also talk to the real child in the pre-recorded voice of a real child, and even uses lifelike expressions and gestures.
Cassell and colleagues originally developed virtual children ten years ago, in part to study how literacy and conversation develop in normal children. But "every time I presented the work," Cassell said at a press briefing yesterday, "a parent of an autistic child would come to me and ask, 'Please, can I get a copy of this software for my child?'"
Indeed, the group's newest research, presented at a session this morning, demonstrates that interaction with virtual playmates may unlock social aptitude in autistic children. During unsupervised play with typical children, autistic children don't fill in pauses in conversation, nor do they ask or answer questions in a natural flow. But with a virtual playmate, autistic children begin to do all these things after as little as 20 minutes. In another experiment, the autistic children were given the opportunity to "become" the virtual child. When they hid behind a curtain and manipulated the digital kid via a control panel, their virtual stand-ins interacted with typical kids in socially sensitive ways.
Cassell says autistic children may be more at ease with virtual playmates because the virtual kids are more predictable, which could make them seem less threatening. Preliminary brain scans show that typical people have to think harder to relate to a virtual human than to a real one, and Cassell speculates that the reverse may be true in autistic children. But will the autistic children's' new-found social fluidity translate to subsequent interactions with real children? "That's the million dollar question," Cassell told Science.
Cynthia Zocca, a graduate student at the University of Connecticut in Storrs, who attended Cassell's presentation, says that as a theoretical linguist, it's nice to see applications of linguistics helping people in the real world.
--Elsa Youngsteadt
