Penn State Researchers Develop Computer Game To Help Autistic Teens Read Body Language

A team of researchers at Penn State will soon pilot a computer game they designed to help adolescents on the autism spectrum improve their ability to understand body language.

People on the autism spectrum often have difficulty picking up on nonverbal social cues, such as facial expressions, eye gaze and subtle body language. Since people typically convey a great deal of information through body language, those with autism can be at a disadvantage in schools, the workplace and social situations. People with autism can learn to read body language, but some families don't have access to qualified therapists who can teach those skills. A team of Penn State researchers in psychology and biobehavioral health and medicine are developing the game with the goal of helping children on the spectrum learn those critical skills, and also to research the effectiveness of using computer games to help autistic kids.

The game is designed for kids aged 11 to 18. In the game, they play as detectives chasing a criminal through an underground maze below a city. To find the criminal, "players must rely solely on nonverbal cues from bystanders to lead them through the hazy, low-lit maze to catch the bad guy: if they make a wrong turn, the criminal might escape," according to information from the university. Players must actively seek out social interactions with the silent in-game characters and interpret their finger pointing, head turns and eye gaze to navigate through the maze to the criminal.

To prevent players from memorizing the game's mazes as a way of avoiding character interactions, the researchers randomized the combinations of mazes, and the mazes become more complex as players progress through the four levels. The easiest level requires only five turns through the maze, while the highest level requires 20 turns.

The researchers deliberately designed the game without a point system. "We want the kids to be immersed and feel like it's actually them in the game — that way when they catch the criminal the reward is intrinsic and not based on an external point system," said Elisabeth Whyte, a postdoctoral research assistant, in a prepared statement. "If they can internalize that they are good at making correct social decisions and not that they scored 1,000 points, their abilities and confidence will grow."

The initial funding for the project was crowdsourced, and subsequent funding was provided by the university's Center for Online Innovation in Learning (COIL). The researchers collaborated with students in Penn State Erie, the Behrend College's game design certificate program to build the game template, and with undergraduate students at the University Park campus to complete and develop the game and art design.

The first four levels of the game are currently in testing and will soon enter the pilot phase, at which time the researchers can begin collecting data. The team said it hopes to add more game levels in the future, and if the clinical trials are effective, the researchers said they hope to make the game available on a larger scale.

About the Author

Leila Meyer is a technology writer based in British Columbia. She can be reached at [email protected].

Featured