Using Brain Activity Patterns to Identify Autism in Kids as Young as 2

In a large new study, researchers at Boston Children's Hospital used EEG to identify specific patterns of brain activity that can distinguish children with autism


In one of the largest studies to date, researchers at Boston Children’s Hospital used electroencephalography (EEG) to measure brain activity in children and determined 33 specific patterns that could differentiate an autistic child from a neurologically typical peer.

The researchers measured EEG patterns in 430 children with autism and 554 control subjects ages 2 to 12. Those with autism had activity patterns that consistently showed reduced connectivity between brain regions, especially in areas associated with language on the left side of the brain.

“The brain works like a series of computers and they have to hook to one another through nerves in the brain in order to connect and function together,” says study author Dr. Frank H. Duffy of the department of psychiatry at Boston Children’s Hospital. “We can estimate from EEGs how well regions connect to one another. If there is high coherence between different regions of the brain, this indicates the brain is well connected.”



The researchers eliminated children with high-functioning autism or Asperger’s syndrome, focusing instead on those with ”classic” autism symptoms who had been referred for EEG by neurologists, psychiatrists or developmental pediatricians to rule out seizure disorders. “We studied the typical autistic child who is seeing a behavioral specialist. These children are hard to study and it is usually difficult to get EEG recordings from them,” says Duffy.

To get an EEG reading, children must wear a cap of electrodes that record electrical signals signifying brain activity. The research team used certain techniques to allow clean readings from the participants, such as letting the children take breaks and adjusting for behaviors like body and eye movement and muscle activity that can throw off recordings.

The researchers compared EEG signals from multiple electrodes placed on the children’s scalps; if two or more signals, or waves, rose and fell in unison, it indicated that the brain regions were tightly connected. The team identified synchrony, or coherence factors, among 4,000 unique combinations of EEG signals; among those, they found 33 such factors that consistently differentiated autistic children from typical controls. The coherence factors held up in all age groups: 2 to 4, 4 to 6, and 6 to 12.

The researchers repeated their experiment 10 times, each time splitting their study population in half different ways. They used half the participants to identify the coherence factors and the other half to confirm them. They were able to identify autistic children all 10 times.


“The findings are remarkable, real and they work,” says Duffy. “What is surprising is that this pattern of 33 factors was so useful from age 2 to 12. It seems to establish that no matter how you come to have autism, you have a distinct pattern of abnormality.”

Duffy says EEG, a widely used and affordable test, could be used to identify children on the autism spectrum, but there isn’t necessarily a need for it in the U.S. “In the U.S., well-trained child neurologists, pediatricians and psychologists can tell you when your child is on the autisitic spectrum within a minute or so of meeting your child,” he says. “I can pretty much guess from the noise in the hall before the kid comes into the room if the child is on the spectrum. If I don’t know in 30 seconds, I won’t know at the end of the hour and will need to do some tests.”

In developing countries, however, where specialists are in short supply and autism screening is less commonplace, Duffy suggests EEG could help identify kids with the disorder. If doctors are not automatically looking out for autism, EEG could be a useful diagnostic tool.


The researchers are hopeful they can now apply their research to high-functioning children with Asperger’s syndrome in order to confirm whether or not these kids should be categorized on the same spectrum as autistic children.