Modern Medicine

Scientists uncover an algorithm that predicts autism in high-risk babies

Megan Sheehey, left, works with one student (name not used-school policy) inside the Red Cloud Peak classroom.The Englewood-based Joshua School, highly regarded for its programs for elementary school age kids with autism, recently opened a second campus in Centennial called Joshua Early Childhood Center.
Kathryn Scott | The Denver Post | Getty Images

A team of researchers has shown that measuring the growth of brains in babies can predict the onset of autism later in childhood.

Predicting the disorder early could allow doctors to begin treating the condition earlier.

About 1 out of every 68 children in the United States has some form autism, according to the Centers for Disease Control and Prevention. It is a developmental disorder that varies in severity from person to person.

As the name autism spectrum disorder suggests, autism can manifest in many different ways among patients, but in general it can be characterized by certain social difficulties and a tendency toward highly repetitive and ritualistic behaviors.

A team from several leading institutions in the U.S. and Canada published a paper Wednesday in the journal Nature, demonstrating an algorithm they created that improved early diagnosis of the condition among several children known to be at high risk.

"In the field we are always trying to detect autism at younger ages, so we can start treatment earlier, but we hit a wall around 2 to 3 years of age, because the symptoms don't start showing up until around then," said the study's senior author, Joseph Piven, a professor of psychiatry, psychology and pediatrics at the University of North Carolina, Chapel Hill, in an interview with CNBC.

Piven said the research can be likened to similar efforts to detect other brain disorders earlier in life, such as Parkinson's or Alzheimer's, before they begin to impair patients.

In the study, the researchers scanned the brains of three different groups of subjects with magnetic resonance imaging machines. They looked at infants with a high family risk of autism who were later diagnosed with the disorder, high-risk infants who did not develop autism and low-risk infants who did not develop the condition.

The team scanned the brains of all three groups three times — at 6, 12 and 24 months of age.

The brains of the autistic children had faster rates of growth on the surface of their brains from 6 months of age to 1 year, and faster overall brain size from 1 to 2 years of age.

For the second part of the experiment, the team created an algorithm that used the scans to correctly predict the onset of autism in 8 out of 10 high-risk infants.

The researchers cautioned in their report that more research is needed, but that the results suggest machine learning could help doctors identify the disorder early, and perhaps develop therapies or treatments that could improve the well-being of patients, or, perhaps one day, even stop the progression of the disorder.