Brain Scans and Artificial Intelligence Could Predict Which Infants Will Develop Autism

Autism Spectrum Disorder (ASD) tends to manifest at different times and in different ways in each child, but in most cases, symptoms will begin to appear around the age of two.

Now, researchers may be getting closer to pinning down a way to detect autism much earlier. In a new study, which appears in the journal Science Translational Medicine, a team of scientists was able to predict, with 96 percent accuracy, which 6-month old infants would go on to be diagnosed with autism as toddlers. They were able to do so with brain scans and artificial intelligence.

The team used MRI (magnetic resonance imaging) machines to capture brain scans of the neural activity of 59 infants in 230 different areas across the brain. Each of the infants had at least one older sibling on the autism spectrum.

Instead of focusing on differences in brain anatomy, the researchers analyzed how the different brain areas connected with one another. The synchronization of these different regions is known to play a critical role in language, repetitive behaviors, and social skills – areas which often pose challenges to those with autism.

Through this analysis, the researchers identified 974 pairs of connections that were linked with autism. These classifiers were then entered into an artificially intelligent computer program which could almost always accurately predict which six-month-olds would go on to be diagnosed with autism at the age of two.

“When the classifier determined a child had autism, it was always right,” researcher Robert Emerson from the University of Carolina said in a press statement. “But it missed two children. They developed autism but the computer program did not predict it correctly, according to the data we obtained at six months of age.”

However, it’s not likely that single brain scans will be used as a basis to diagnose autism in the future. Presumably, a combination of pre-tested evaluations would be used to confirm results and better predict autism.

“I think the most exciting work is yet to come, when instead of using one piece of information to make these predictions, we use all the information together,” Emerson said. “I think that will be the future of using biological diagnostics for autism during infancy.”

Being able to better predict the onset of autism could help families provide therapy for their children at earlier ages. Plenty of research shows that the earlier a child receives ABA (Applied Behavior Analysis) therapy, the better chance there is to improve in critical developmental areas like language, social skills, motor skills, and day-to-day living skills.

“The more we understand about the brain before symptoms appear, the better prepared we will be to help children and their families,” researcher Joseph Piven concluded.