AI Test Examing Retinas Correctly Identifies 100% of Cases of Autism

January 08, 2024

Application Also Accurately Determined the Severity of ASD Symptoms between 48% to 66% of the Time

A Korean study has discovered that an artificial intelligence (AI) system can correctly identify children with autism by examining images of their retinas. This recent research found that the optic disc area is a significant region for ASD screening and that retinal alterations could be a potential biomarker for the disorder. The study’s design involved testing an AI application on the retinas of 958 children, age four and older, who were either on the spectrum or were typically developing. The findings showed that the technology could identify children with an autism diagnosis perfectly based on the images. Interestingly, the AI application was also able to determine the severity of ASD symptoms in children with the disorder correctly between 48% and 66% of the time. The authors suggest that this AI system could provide an additional way to measure or assess if an individual may have autism by relying on quantitative methods based on biology. They also indicate that AI could be used as a pre-screening method, alongside regular psychometric testing, to speed up diagnosis and detect it earlier, providing interventions earlier.

Original Article

Original Study 

Show Buttons
Hide Buttons