Game tests theory of mind in autism, intellectual disability

A simple game that requires few language skills may help researchers assess theory of mind in children who have autism and intellectual disability, a new study finds.

By Nicholette Zeliadt
7 October 2014 | 4 min read
This article is more than five years old.
Neuroscience—and science in general—is constantly evolving, so older articles may contain information or theories that have been reevaluated since their original publication date.


Children with autism often have difficulty understanding the beliefs of others. Traditional tests of this ability, called theory of mind, require children to have good enough language skills to understand what they’re being asked to do.

That means the tests can’t reliably assess theory of mind in children with language impairments or intellectual disability, both of which are common among individuals with autism. A new study, published 24 September in Autism Research, suggests that swapping these language-based tasks for a ‘penny-hiding game’ can help researchers assess theory of mind in these children.

The traditional test for theory of mind is a ‘false-belief task.’ This task often involves telling a child a story about two characters named Sally and Ann who put a toy into a basket. When Sally leaves the room, Ann hides the toy in a box. The child passes the test by reasoning that Sally will look for the toy in the basket when she returns.

Most typically developing children pass this test by age 5. Children with autism, however, fail the test into adolescence. Those with language impairment or intellectual disability also perform poorly on the task. Many children who have severe intellectual disability simply cannot participate at all, and are thus excluded from experimental studies.

Child’s play:

In the new study, the researchers played the penny-hiding game with 132 children who have moderate to severe intellectual disability, including 56 who also have a diagnosis of autism. All of the participants, who ranged in age from 5 to 19 years, have language impairments.

The game starts with an experimenter showing the child a penny in one hand and then putting both hands behind his or her back. The experimenter hides the penny in one hand, then brings both hands forward again and invites the child to point to the fist that holds the penny.

About 91 percent of the children grasped the guessing part of the game, despite their language impairments. Children who learned the game after three attempts then swapped roles with the experimenter and hid the penny themselves during six additional trials.

The researchers recorded the number of times the child tried to trick the experimenter into choosing the wrong hand. They also tallied the number of errors each child made, such as not fully hiding the penny, using the same hand every time or transferring the penny from one hand to the other in plain sight. These errors suggest that the child is unaware of what the experimenter might glean about the penny’s location.

Children with autism made more errors and used fewer tricks in the experimenter role than those without the disorder, the researchers found. The children with autism were also more likely to fail the test, defined as making two or more errors. Overall, 88 percent of the children who have both autism and intellectual disability failed the test, compared with 70 percent of the children with intellectual disability alone.

Overall, children with higher scores on measures of verbal ability were more likely to pass the test. But the children with autism needed more advanced verbal abilities to pass the test than those without the disorder. This suggests that some basic language skills are still necessary to succeed in the game.

Higher scores on measures of social skills also predicted fewer errors on the test. This link is weaker for children with autism when the researchers factor in differences in language ability. The results suggest that children with autism rely primarily on language skills to pass the test, whereas those without the disorder use both language and social skills.

This game alone doesn’t seem to distinguish between the children with and without autism, but demonstrating that most children with intellectual disability can complete the test is a step in the right direction. Hopefully, this and other alternatives to traditional tests will help include this poorly understood group in research studies.