By Serena Gordon
Known as functional connectivity MRI (fcMRI), the scan gives a peek at how different regions of the brain work together. As it turns out, certain areas that are connected also seem linked to autism risk, the researchers said.
The fcMRI allowed the researchers to accurately predict 9 out of 11 high-risk babies who later showed behavioral signs of autism.
"We used functional brain imaging information at 6 months and clinical information from 24 months to figure out if we could identify which high-risk infants would go on to develop autism," said study author Robert Emerson, a postdoctoral fellow at the University of North Carolina at Chapel Hill.
The hope is that such a prediction tool could one day be used to identify babies in need of early intervention.
"What we found is exciting, but our findings have to be replicated," said co-senior study author Dr. John Pruett Jr. He's an associate professor of psychiatry, radiology and psychological and brain sciences at the Washington University School of Medicine in St. Louis.
Autism spectrum disorders affect about 1 in 68 children in the United States, according to the U.S. Centers for Disease Control and Prevention. Currently, children can't be diagnosed until they're around 2 years old. At that time, behavioral symptoms of the disorder -- such as difficulty behaving, communicating and interacting with others, repetitive behaviors and obsessions -- begin to show.
The earlier a child gets behavioral intervention services, the better the outcome generally, according to the CDC. And a recent study even found that when an intervention was started in babies before symptoms first appeared, those babies had better attention, language, communication and social skills at age 3. The study was presented recently at the International Meeting for Autism Research, in San Francisco.
"There's very little behaviorally that tells us about autism in the first year of life," Pruett said. "It would be very important if we could identify brain-based features early in life. We could identify infants at an even higher risk and get them into studies of infant adaptations of current toddler-age interventions."
The new study included 59 infants considered at high risk of autism because they had a sibling with autism.
"In this high-risk group, there's a 20 percent conversion rate to autism," Emerson said.
The babies underwent fcMRI when they were 6 months old. They were sleeping during the test.
The fcMRI viewed neural activity across 230 different regions in the brain. The researchers looked for areas with coordinated activity, and focused on those connections known to be tied to features of autism -- language skills, repetitive behaviors and social behavior.
The researchers then developed a computer program to help them sort through this information and identify which babies were likely to develop autism, and which would probably not.
Eleven of the 59 babies developed autism. The test and program were able to predict 82 percent of those cases. All of the children who didn't develop autism were correctly identified as unlikely to get the disorder in the toddler years, the researchers said.
Thomas Frazier is the chief science officer of Autism Speaks. He wasn't involved with the study, but reviewed the findings.
"Autism has been thought to be a disorder of connections in the brain, and the fact that the function connectivity MRI is a good predictor of autism helps confirm those suspicions," Frazier said.
"Another interesting thing is that it shows how early in life we can see differences in autism. It's remarkable that they could capture brain changes related to autism at 6 months," he said.
Like the researchers, Frazier said this study needs to be replicated. And he wondered, "Can it be replicated in a way to expand its use beyond high-risk siblings?"
The findings were published June 7 in the journal Science Translational Medicine.