U of M research may lead to earlier detection of autism

A computer program based on the research was 80 percent accurate in predicting who'd be diagnosed with autism.
Author:
Updated:
Original:

University of Minnesota researchers are excited about a study they were part of that could lead to much earlier detection of autism.

The breakthrough research published in a scientific journal Wednesday involves biological signals that a child is at risk of being on the autism spectrum. Those signs could be spotted even before symptoms of the brain disorder show up in a child's behavior. In the study they predicted with 80 percent accuracy who would be diagnosed with autism.

Growth of the brain is the key

The U of M researchers were part of a big study across the U.S. and Canada that's been going on for years. It focuses on babies who have older siblings on the autism spectrum, which increases their chance of developing autism themselves.

Scientists used MRI scans to measure the size of a baby's brain at various ages – 6 months, 12 months, and 24 months, a statement from the U of M explains. The researchers noticed that babies whose brains expanded more than usual between 6 months and 12 months were more likely to be diagnosed with autism later in childhood.

They used those brain scans to build a computer program that searches for the same changes in the brains of other infants.

Currently, an autism diagnosis is based on a child's behavior and the earliest it happens is normally at age 2 or 3.

Jeb Elison of the U of M's Institute of Child Development is a co-author of the new study. He says it's possible that in the future doctors could see in the first year of a child's life who is most likely to get an autism diagnosis.

Parents and doctors could then start some early intervention "before the symptoms that define autism consolidate into a diagnosis,” Elison says.

Money for the study came from the National Institutes for Health. The advocacy group Autism Speaks also helped and has more on the study here.

Next Up

Related