New research from the University of Minnesota could lead to early autism detection in children at high risk of developing the disorder.
Using MRI brain scans, researchers across the country, including the U of M, were able to pinpoint changes in the brains of children who later developed autism. And they were able to predict that diagnosis with 80 percent accuracy.
Jed Elison, assistant professor of child development at the U of M, co-authored a paper on the research published Wednesday in the scientific journal Nature, and said the results shed new light on autism.
"Traditionally, autism diagnoses are made based upon behavior signs," he said, "but we'd like to understand the underlying biology. People have been searching for these biomarkers for a long time."
Studies from the early '90s showed that children with autism often have larger brain volume. Elison said that link wasn't well understood, but it was a good place to start.
Over the last 12 years, more than 100 children were scanned at 6, 12 and 24 months old. Elison said researchers concentrated on children with older siblings already diagnosed with autism — which puts them at a much higher risk of developing the disorder themselves.
They noticed that some children who later developed autism had a rapid expansion of brain surface area from six to 12 months of age. Researchers built a computer program using the brain scans from those children to search for the same changes in other children.
That program was able to predict autism diagnoses with 80 percent accuracy, Elison said.
"This is a learning program," he said. "The more data we build in, the more accurate it will get."
Most of the time, it takes a few years for parents and doctors to notice the behavior changes in a child with autism.
Diagnosing the disorder a full year early could make a huge difference, Elison said. The earlier behavioral treatment starts, the better it works. But Elison added it will be a long time before the research is ready for clinical use.
Researchers will have to replicate their results in larger, more complex studies. Even then, Elison said, MRI scans for infants will likely never be common.
MRIs are very expensive, and it's difficult to convince a 12 month old child to lay still long enough to get a good brain image. During the study, researchers had to wait for each child to fall asleep. It wasn't easy.
"We're looking at this as an option for high-risk children," Elison said. "This isn't something for the general population yet."