Autism diagnosis currently relies on clinical evaluation, but a new study suggests it may be possible to detect the disorder with near perfect accuracy using brain scans.

Researchers say they were able to identify “thought-markers” — or differences in the way the brain responds to certain thoughts — specific to those with autism. The method was successful in identifying whether or not a person had autism with 97 percent accuracy, according to findings published Tuesday in the journal PLOS ONE.

For the study, researchers used functional magnetic resonance imaging to perform brain scans on 17 adults with high-functioning autism and 17 typically-developing controls. During the scans, the individuals were asked to think about various social interactions like “persuade,” “adore” and “hug.”

Advertisement - Continue Reading Below

The researchers then used a machine-learning technique to assess patterns in brain activity and decode how the thoughts were processed. They found a clear distinction between those who were and were not on the spectrum.

Among the typically-developing controls, thoughts of a hug, for example, activated an area of the brain associated with a representation of one’s self, but this was largely absent in individuals with autism.

“We found that we could tell whether a person has autism or not by their brain activation patterns when they think about social concepts,” said Marcel Just of Carnegie Mellon University who led the study. “We’ve shown not just that the brains of people with autism may be different, or that their activation is different, but that the way social thoughts are formed is different. We have discovered a biological thought-marker for autism.”

Researchers said their approach could lead to quicker and more certain diagnosis and allow for therapies that are more targeted to specific areas of the brain.