Diagnoses of autism spectrum disorder have skyrocketed in the U.S. since the turn of the millennium, when one in every 150 children were diagnosed with the condition. Two decades later, according to the CDC, that now includes one in every 44 children, measured among the nation’s 8-year-olds.
Despite the growing recognition of autism’s prevalence in young children, its diagnosis can still be severely delayed. The median diagnosis age sits around four or five years old—even though symptoms can be observed within the first 18 months of life—stretching out the time it takes for kids to get the support and treatment they need.
The rise of digital diagnostics may be the key to speeding up those diagnoses, which can be even further delayed for girls and for children from low-income and nonwhite families.
Last year, the FDA handed down its first clearance for an app from Cognoa that uses machine learning to analyze videos of kids interacting with others and performing tasks, as well as surveys from caregivers and doctors, to help diagnose autism spectrum disorder.
A year later, the agency has cleared another digital tool to detect signs of autism, this one from EarliTec Diagnostics.
The EarliPoint Evaluation uses eye-tracking technology to monitor a child’s focus and responsiveness while viewing short videos of social interactions between other kids. According to EarliTec, the “looking behavior” of young children with autism can deviate from that of typically developing children as many as 1,000 times during a single testing session.
Artificial intelligence analyzes the child’s eye movements, then compares those findings to benchmarks for the child’s age to spot any disparities between the social learning expectations and reality. The results of EarliTec’s test distinguish between different levels of social, verbal and non-verbal abilities.
The technology was developed by researchers from Children’s Healthcare of Atlanta, the Emory University School of Medicine and Yale University.
“How we quantify moment-by-moment behavior of a child not only provides objective measures of each child’s strengths and weaknesses today—measures that can be universally available and accessible to all families—it provides a digital health platform that can support care in the future, so that all individuals affected by autism receive timely, individualized care,” said Sreeni Narayanan, EarliTec’s chief technology officer.
Compared to the Cognoa app—which is indicated as a diagnostic aid for children between 18 months and six years old—the EarliTec tool covers a narrower and slightly younger age group. The FDA cleared it for use with children as young as 16 months, and up to those aged two and a half years.
The FDA clearance comes shortly after another win for EarliTec that arrived in the form of an eight-figure funding round. The Atlanta-based startup announced the $19.5 million financing in February, courtesy of Home Depot co-founder Bernie Marcus and the Georgia Research Alliance.
At the time, EarliTec said the money would help support its FDA submission and subsequent commercialization of the technology, as well as the expansion of its digital biomarkers and AI platform.