By imaging the brains of adolescents with a high-functioning form of autism as they played a social-interaction game, scientists have identified a physiological deficit specific to the disorder. The researchers believe that the change is linked to a diminished sense of self. The findings, recently published in the journal Neuron, could help guide future research into the nature of autism and potentially lead to new ways to diagnose and treat the disorder.

“I think this is an exciting advance,” says Uta Frith, a professor at University College London, in England, who wrote a preview of the paper for Neuron. Most studies find only subtle differences in people with high-functioning autism, “so it’s quite impressive to find such a big difference,” she says.

Autism is a complicated and heterogeneous developmental disorder marked by problems in language and social behavior. No medical tests exist to detect the disorder, so children are typically diagnosed based on doctors’ observations. Scientists are avidly searching for more objective markers of autism, but identifying specific brain abnormalities has been a challenge.

Researchers at Baylor College of Medicine, in Houston, believe that they have now identified a specific physiological marker of the disorder. Read Montague, Pearl Chiu, and their colleagues scanned the brains of adolescents with Asperger’s syndrome, a high-functioning form of autism, while they played an interactive trust game.

In the game, one person, designated the investor, chooses an amount of money to send to a second player, the trustee. The money is tripled en route, and the trustee must then decide how much to give back to the investor. When played by normal volunteers, the game unfolds in a very characteristic fashion: generous gestures are met with generous responses, while selfish ones inspire selfishness in return.

Brain activity also follows a stereotyped pattern. A study by Montague and his colleagues. published in 2006, imaged the brains of both the investor and the trustee as they played the game. The researchers discovered a specific signal in the cingulate cortex, part of the brain that integrates information from both the cortex and the body, that was detected only when the investor thought about how much money to give the trustee. A second signal was seen only when the investor received his or her return from the trustee. “We see a ‘self, other, self, other’ pattern,” says Montague, director of the Human Neuroimaging Lab at Baylor. “We think that’s an unconscious assessment of who the actions should be attributed to.”

According to the new findings, people with Asperger’s play the game just as a nonautistic person would, but they lack the characteristic “self” signal in the brain. Normal people lack the signal only when they think that they are playing against a computer, suggesting that autistic people view interactions with other people similarly to the way that normal people think about interacting with a computer. “This approach allows a somewhat objective look at something hopelessly subjective–sense of self,” says John Gabrieli, a neuroscientist at MIT.

While the findings are clearly intriguing, it’s not yet clear what they mean. One popular theory of autism is that people with the disorder lack a normal theory of mind–the ability to imagine the thoughts and actions of others. Identifying a specific deficit linked to thoughts of self could help narrow down what has gone wrong in that process. “People think autism is linked to a lack of understanding of what a partner is doing,” says Chiu. “But maybe they don’t understand their own role in the social exchange.”

Other scientists interpret the results further, suggesting that this signal is linked to a sort of internal reputation assessment in the brain. “If you are a normal person, when you invest money in the game, you are thinking about how you will look in the eyes of your partner,” says Frith. “That’s precisely what the theory of mind hypothesis would project is wrong with people with autism.”

Other autism experts are unwilling to make such a leap. “I’m skeptical about how much [the Baylor College study] tells us about which capacities are intact and engaged in autism,” says Matthew Belmonte, a scientist at Cornell University, in Ithaca, NY. “I’m not convinced they have a deficit at all. Maybe they have adopted a different cognitive strategy.”

Regardless of the deeper meaning of their findings, Montague and his team ultimately hope to develop the brain-imaging results into a diagnostic test. They have converted the activity signal from the cingulate cortex into a simple numerical score, which correlates well with a clinical test for the severity of autism. Eventually, they hope to be able to show, for example, “that if you get a 3 rather than a 14, you are 80 percent more likely to be a high-functioning autistic,” says Montague. Such a tool could potentially also be used to test the effectiveness of new behavioral treatments, he says.

However, much work remains to be done before such a test could be used in a doctor’s office. “We need to make it simpler and test people with a wider range of IQs,” says Montague. The Asperger’s volunteers in the current study had an above-average IQ.