One of the great mysteries of autism is predicting which children will develop the condition, Gizmodo reports. But scientists at UNC Chapel Hill have now used brain scans to predict with near pinpoint accuracy whether babies as young as 6 months would develop autism by age 2. Two is the age when telltale repetitive behavior and other signs of autism usually emerge; diagnosis often doesn't happen before age 4. But the new findings raise the possibility that doctors may one day be able to use a tool to identify babies likely to develop autism. That would allow treatment to begin far earlier, when brains are most malleable, Scientific American reports. "This is a game-changer for the field,” autism expert Kevin Pelphrey tells the mag.
Reporting in Science Translational Medicine, researchers used a special MRI to analyze how the regions of the brains interact in 59 babies at age 6 months. They repeated the scans at 24 months, and developed an algorithm that spotted patterns in the brain connectivity of autistic toddlers, per Gizmodo. From there, using artificial intelligence, the team devised a model that correctly predicted 9 out of the 11 infants who developed autism. Although researchers caution the findings will need to be replicated, outside autism experts were optimistic there may be hope for earlier treatment of condition that afflicts one out of 68 US kids. "This is a really, really important advance," one Yale professor tells Scientific American. (An earlier study by the same team noted excessive brain growth as a sign of autism.)