Discover SPARK

How Baby Siblings Can Bring Insight to Autism

Emily Singer

Date Published: May 11, 2020

Children with an older brother or sister who has autism, also known as ‘baby sibs’, have a 10- to 20-fold greater chance of being diagnosed with autism than the general population. This is because siblings are likely to share inherited genetic risk factors for autism.

This illustration shows how children with an older brother or sister who has autism have a 10- to 20-fold greater chance of being  diagnosed with autism than the general population.

Scientists are using this fact to better understand the earliest stages of autism. Many of the classic symptoms become clear around 2 years old. But research hints that in some children, some subtle differences appear earlier.

By studying babies that will go on to develop autism, scientists hope to discover changes that happen in the brain before babies show outward symptoms. This may eventually lead to earlier diagnosis and treatment.

In baby sib studies, scientists collect information on brain development and behavior in the same at-risk child over time, starting early in infancy. Once these children are old enough to be assessed for autism, researchers can look back at the data to search for signs that predict who would go on to receive an autism diagnosis and who would not.

Joe Piven, director of the Carolina Institute for Developmental Disabilities at the University of North Carolina, runs one such study, called the Infant Brain Imaging Study, or IBIS. Piven’s team has found clear differences in brain development in children who are later diagnosed with autism. But the nature of the differences depended on the child’s age.

At 6 months old, brain structure in babies who will later be diagnosed with autism was organized differently compared with other infants in the study, even though they lacked obvious signs of autism at that age. But at 12 months, the two groups looked roughly the same. At 24 months, a new set of differences emerged.

Piven says that the findings highlight how important it is to study how the brain develops in the same person over time. “We talk about autism as a disorder of development, but we don’t study it that way,” he says. For example, if the researchers had studied 12 month-olds only at that age, they would have missed these differences all together.

Piven’s team is now following up on an earlier study to develop a brain imaging test that, if repeated, will reliably predict which baby sibs of children who have autism will go on to develop the condition. In a study published in 2017, they showed that a type of brain scan called a magnetic resonance imaging scan, or MRI, in the first year of life can predict 8 out 10 babies that will go on to develop autism. The test isn’t yet ready to use in general medical practice. They are now trying to repeat the findings in another group of children.

If the findings are confirmed, Piven says that the test would be aimed at families who already have one autistic child. That’s because brain scans are expensive and time consuming, so it makes sense to focus on the most at-risk infants. “It’s not cost effective to do brain scans on every 6-month old,” Piven says.

Researchers are working on other measures, such as questionnaires, that can predict autism risk in 1-year olds who don’t have an older autistic sibling. Babies who show a higher risk could then get brain scans.

No widespread treatments are available yet for babies who are at high risk. If doctors can predict who will go on to develop autism, they can more efficiently develop interventions that might help this group. The only way to test these kinds of treatments is to focus on at-risk infants. “If we can identify kids at 6 months who are highly likely to develop autism at two years of age, we can then start studying which treatments… and that’s a big deal,” he says.

SPARK participants with baby sibs can update their status on their dashboard under ‘sibling update’. Knowing whether or not a baby sib develops symptoms can help doctors better understand families’ genetic results.



  • Emerson R.W. et al. Sci. Transl. Med. 9, (2017) PubMed
  • Hazlett H.C. et al. Nature 542, 348-351 (2017) PubMed