A gaze into your eyes. A smile back. A gurgle and babble. These are treasured moments between parents and their babies. Yet, for families with a high familial risk of autism spectrum disorder (ASD), these precious moments may be elusive. Researchers have long suspected that autism emerges in gradual ways during the first year of life, but it has been impossible to predict with confidence which high-risk infants are likely to be diagnosed with this complex developmental disability.
In a novel study conducted by the multi-center Infant Brain Imaging Study (IBIS) network that includes the Center for Autism Research (CAR) at Children’s Hospital of Philadelphia, scientists found a brain biomarker that could help to identify children with ASD earlier in life, with the help of a computer-generated algorithm. Their findings, reported in the journal Nature, suggest that brain changes may precede behavioral manifestations of ASD.
“The results of this study are a real breakthrough for early diagnosis of autism,” said Robert Schultz, PhD, who directs CAR and led the CHOP study site. “Currently, the earliest clinicians can diagnose a child with autism is around 2 or 3 years old, because that is when clinicians can observe the earliest behavioral presentations of autism. This research opens the door to be able to identify those needing intervention during the first year of life, before the full emergence of autism. Delivering early intervention offers hope that we can blunt the development of autism and set the stage to dramatically improve long-term outcomes.”
Earliest Detection of Enlarged Brain Size May Define the Core of Autism
Behavioral symptoms of ASD are characterized by difficulties in social interaction, verbal and nonverbal communication, repetitive behaviors, and restricted interests. By the time autism is diagnosed in the preschool years — on average around age 4 in the U.S. — their brains have already changed substantially and tend to be enlarged. Increased brain size was one of the earliest brain markers discovered in autism, but until now, brain size has only been studied in children and adults after the full onset of autism. The development of differences in brain anatomy before the first outward manifestations of autism appear has been a mystery.
“Brain imaging studies in older children have taught us a great deal about structural and functional differences in the way a child on the autism spectrum’s brain develops, but a key question has always been: What are the first changes in the brain, and how do neuroanatomical differences unfold during early development?” Dr. Schultz said. “These earliest developmental differences in the brain are likely the most important for understanding autism, for defining the core of autism.”
In the current study, the IBIS investigators from four sites in the U.S. used magnetic resonance imaging (MRI) to look for anatomical differences in brain development at 6 months, 12 months, and 24 months of age in 106 children at high risk for developing autism by virtue of having an older sibling with ASD. At each age, they measured brain volume and specific structural characteristics of the brain, in particular the thickness and surface area of the cerebral cortex in each of its many subregions. The researchers compared the measurements to a group of 42 low-risk infants without any family history of autism. The results showed that the babies who developed autism experienced much more rapid growth of the brain’s surface area between the ages of 6 and 12 months compared to babies who did not develop autism by 24 months of age. These findings suggest that the cerebral cortex expands first, and then enlarged brain size follows.
“The rate of cortical surface area expansion in the second half of the first year of life predicted with high accuracy which babies would later develop autism and which would not,” Dr. Schultz said. “This surface area expansion also helped us to understand why, on average, children with autism have bigger brains. Although brain size was not yet enlarged by age 1, the cortical surface growth rate from 6 to 12 months accurately predicted which children would show significantly larger brains at 2 years of age among the children who did develop autism.”
Honing in on Differences in ‘Social’ Areas of the Brain
While the entire cortical sheet expanded between 6 and 12 months, the greatest enlargement was confined to specific parts of the temporal and frontal lobes of the brain for children who were later diagnosed with autism. These brain areas are involved in language and nonverbal social understanding like facial expressions and complex social thinking. The researchers noted that these are the same brain regions that always show the greatest structural and functional differences in older children with autism.
Scientists do not fully understand why bigger is not better for the brains of children with autism. During normal development, there is significant pruning of neuronal connections called synapses that presumably enables improved cognition and social functioning, Dr. Schultz explained. It has been hypothesized that inefficient culling of connections that are not essential and perhaps not needed at all is the reason why brain size is enlarged in autism.
Applying ‘Smart’ Data to Neuroscience
The study team used a type of machine learning algorithm known as deep learning to establish their prediction. This process enabled more than 80 percent accuracy in predicting which babies would develop autism by age 2 and more than 95 percent accuracy in predicting which babies at high risk for developing autism would not develop autism.
“While the machine learning algorithm that predicted autism at age 2 drew upon many brain features, the prediction was largely carried by the growth rate of the surface area at 6 to 12 months,” Dr. Schultz said. “This was the single biggest part of our predictor for who is going to have autism at 2 years of age.”
On the Horizon: Translating Earlier Diagnosis to Earlier Intervention
The researchers envision that this kind of procedure might one day be part of routine clinical care for high-risk infants. However, the research results will first need to be reproduced in additional studies in order to be ready for clinical uses.
The community of researchers, clinicians, and families are excited by these findings because the ability to identify autism risk during infancy could lead to the development of valuable early interventions. Research has shown that children with autism who receive the earliest treatment tend to reap the most benefits.
For example, a specific social behavior called joint attention is impaired in children with ASD. Joint attention is the coordination of attention to objects between two people for the purpose of sharing. It’s motivated by the desire to share. Children with autism don’t readily initiate joint attention; they don’t use nonverbal cues to seek for others to attend to what is capturing their attention.
Joint attention emerges toward the end of the first year of life, and these interactions foster a common frame of reference critical for learning language. It is also a core building block for social development. Joint attention training, which has been very successful in preschoolers and school age children, is one example of the kind of intervention that now might be used before autism emerges, Dr. Schultz explained.
In a related IBIS study that CAR investigators also were involved in, the study team used functional MRI scans to identify brain networks involved in initiation of joint attention. They evaluated the strength of connections between the brain’s vision, attention, and default modes to identify patterns of neural activity that also may allow for earlier diagnosis. Those findings appeared in the January issue of the journal Cerebral Cortex.
The IBIS study has been underway for a decade, and Dr. Schultz is thrilled that the study teams are on the cusp of reporting many more significant findings that also may have predictive value for earlier diagnosis of autism.
“We’ve known for a long time that ASD symptoms emerge over the first two years, but it has been difficult to find evidence of those symptoms in the first year of life,” said Dr. Schultz, who also is R.A.C. Endowed Professor, Department of Pediatrics, Perelman School of Medicine at the University of Pennsylvania. “These studies are hard to do because they take a long time, but now the results are coming in spades, and it’s wonderful.”
Researchers at the Carolina Institute for Developmental Disabilities at the University of North Carolina, which is directed by the Nature study’s senior author, Joseph Piven, MD, worked on the study. Other data collection sites included CHOP, the University of Washington, and Washington University in St. Louis. Investigators at McGill University, the University of Alberta, the College of Charleston, and New York University were involved in data analyses and interpretation.