Eye-tracking device predicts autism in young children

An eye-tracking device provides measures of social visual engagement that are predictive of autism in young children as shown in a new study.  

In all children, sensitivity was 71 percent for the social visual engagement and specificity was 80.7 percent relative to expert clinical diagnosis.

In 335 children whose autism diagnosis was certain, the EarliPoint Evaluation device was 78 percent sensitive and 85.4 percent specific. Moreover, device accuracy in predicting variance in social disability, verbal ability, and nonverbal cognitive ability was 74.1 percent, 88.8 percent, and 77.9 percent, respectively. [JAMA 2023;330:854-865]

“The findings provide an opportunity to shorten the time to autism diagnosis in this patient population and support early treatment,” said the researchers led by Dr Warren Jones from the Marcus Autism Center at Children’s Healthcare of Atlanta in Atlanta, Georgia, US.

Late diagnosis

Getting social skills training as toddlers can make a big difference for children with autism. But often, they do not get diagnosed until after they have started school.

Jones said guidelines call for the assessment of verbal and cognitive abilities, on top of the detection of social disability, to diagnose autism in children. “However, there’s a big challenge in the community as there are not enough clinicians to meet the needs of all the children who have concerns about their early development. There are long waitlists at almost all the major autism treatment centres and specialty diagnostic centres in the US.”

A delay in diagnosis can result in delayed language and social skill development, which can have long-term consequences for children’s quality of life.

The concept of eye-tracking tools came about as high-risk children show fundamental social disability and reduced attention to social cues, Jones explained. He added that they did not intend to replace clinicians’ assessments and merely wanted to quantify qualitative insights in their study.

Potential biomarker

The study included 475 children aged 16–30 months, assessed for autism in six specialty clinics in the US from April 2018  to May 2019. Care staff were blinded to clinical diagnoses of autism and were asked to use the automated device to measure the eye-tracked-based social visual engagement of children who were prompted with social cues, such as the sound of their own name or children playing together.

Overall, 221 children were diagnosed with autism by expert clinical diagnosis. The eye-tracking test results correlated with the clinical assessments for individual levels of social disability, verbal ability, and nonverbal cognitive ability.

“Our results suggest that measurement of social visual engagement, for example, how children look at and learn from their social environment, may offer such a biomarker for autism,” Jones said.

However, further evaluation of the role of this eye-tracking test on early diagnosis and assessment of autism in routine specialty clinic practice may be necessary, he added.