A team from the University of North Carolina has developed a method to diagnose autism in the first year of life. The researchers predicted the onset of the disease in high-risk 8 out of 10 children, all under the age of 2 years. The first results are exceptional, but further studies will be needed to confirm the value.

It is estimated that autism is diagnosed in 1 child out of 160 in the world. Most at risk are children with a close relative suffering from autism. It is estimated that a child with an autistic brother will have greater probability of 1 to 5, in turn, be autistic. Despite the increased attention placed on high-risk subjects, seldom can be diagnosed before 2 years of the child's life. It is only at this age that any behavioral abnormalities begin to become apparent. Moreover, when there is neither a cause unique and specific genetic biomarkers for the disorder. The study aims to anticipate the diagnosis, so as to put in place immediately the most effective treatments.

Joseph Piven, who heads the team, studied since the 90 MRIs of the brains of high-risk children. Children with autism have a larger brain than the average, but it is not known when begin the abnormal growth. The researchers then periodically scanned the brains of 106 children at high risk and 42 low-risk children. To compare the two groups of images, they used a machine learning algorithm. The method allowed them to diagnose autism in 8 cases out of 10.
Resonances showed abnormalities in the cortical surface of the brain of autistic children. These are extras anomalies between 6 and 12 months of age, before the onset of symptoms. The discovery could have important clinical implications, even if you need more confirmations. We should also see if the new diagnostic method is applicable only to those at high risk or to all children.

Source: wired.it