In a small new study, researchers report that they can use brain scans and a computer formula to identify individuals who have autism with 97 percent accuracy. They propose that their technique could be used in combination with traditional methods of diagnosing autism based on behavior checklists. However, others caution against over-interpreting the results.
The scientists, from Pittsburgh’s Carnegie Mellon University, used functional magnetic resonance imaging (fMRI) to track brain activity while they asked 34 study participants to think about words associated with social interactions. Examples of these social cue words included “persuade,” “adore” and “hug.” Half the participants were on the less-severely affected end of the autism spectrum. The others did not have autism.
The brain images from participants not affected by autism showed clear activation in parts of the brain associated with self-awareness. By contrast, the brain scans of those in the autism group showed little activation in these areas.
The researchers then used a mathematical computer formula to accurately identify “autism” or “not autism” in 33 of the 34 participants.
“When asked to think about persuading, hugging or adoring, the neurotypical participants put themselves into the thoughts; they were part of the interaction, says senior study author Marcel Just. “For those with autism, the thought was more like considering a dictionary definition or watching a play – without self-involvement.”
The use of fMRI to track brain responses to social cues is not new to autism research. Several recent studies have used fMRI scans to measure changes in the brain’s response to social cues following behavioral therapy or virtual reality training.
“This study is consistent with a growing body of research showing that some individuals with autism have different brain activity patterns than do those without autism,” comments Dan Smith, Autism Speaks senior director for discovery neuroscience. “It takes us a step further by looking at the brain activity patterns caused by specific social-emotional cues.”
However, the Carnegie-Mellon researchers go further in proposing their method as an objective tool to help diagnose autism.
“Unfortunately, we can’t determine whether biomarkers like brain activity patterns can predict and diagnose autism by conducting studies on people after they have already received a diagnosis,” comments Dr. Smith. “Predicting whether an individual goes on to develop autism and identifying a marker that correlates with an existing autism diagnosis are apples and oranges. We need to see more research engaging in the very challenging task of determining whether any markers can truly predict autism. Only then will we have the objective biological measures we need to complement and improve our current behavioral diagnostic process.”