New scan could predict autism before a baby’s first birthday

A new scan may allow children at risk of autism to receive help from an early age.
A new scan may allow children at risk of autism to receive help from an early age. Photo: Getty Images

A team of researchers from the US say that it may be possible to use magnetic resonance imaging (MRIs) to predict which babies will develop autism. The new research is considered to be a huge breakthrough because the scans might be able to predict autism before a child's first birthday.

At the moment, the earliest that a child can be diagnosed is age two, which is when hallmark behaviours and communication problems, such as avoiding eye contact, emerge.

Speaking to the Huffington Post, Dr Joseph Piven from the University of North Carolina said that researchers have struggled to predict autism earlier than age two.

"We've kind of reached a wall around 2 years of age. Prior to that, behavioural markers just don't seem to help in detecting kids that end up with autism," he explained.

In the study, which was published in the journal Nature, researchers at autism centres around the US ran MRI scans of babies' brains when they were 6 months old, 1 year old and then again when they turned 2.

The study included 160 babies who were considered to have a high risk of developing autism because they had an older sibling with the disorder. Researchers also looked at the brains of 40 babies who were considered low-risk.

The MRIs showed that the babies who went on to develop autism experienced a more rapid growth of the brains surface area (the folds of the brains surface) than the babies that didn't develop autism.

In addition to this, researchers also found that between ages 1 and 2, the rate of brain volume growth they experienced was also highly accelerated.

Researchers believe that the brain volume "overgrowth" is linked to the emergence of social symptoms related to autism in the children's second year. This can include delayed speech and language and behaviour such as not being able to engage in role-play or let's pretend games.


Scientists have created a computer algorithm using all of the data about brain changes. They were then able to correctly identify 80 percent of high-risk babies using the algorithm.

"We view this is, particularly in this high-familial risk sample, as a very real possibility of pre-symptomatic detection," said Dr Piven.

"So, detecting autism before it really appears. Before the consolidation of symptoms and brain deficits and at a time when the brain is most malleable, giving us the greatest chance of having an impact with early intervention."

The new findings are particularly promising for families of children who are at high-risk for the disorder.

"There's a developmental sequence," Dr Piven notes, "and it raises the possibility that we could sort of disrupt that sequence early on."