Teen Innovator Creates AI Tool for Autism and ADHD Diagnosis
Nadia Petrov
A high school student developed RetinaMind, an AI tool that uses retinal images to diagnose autism and ADHD, aiming to enable earlier interventions and improve patient outcomes.
Edward Kang, a 17-year-old high school senior from Bergen County Academies in Hackensack, New Jersey, has developed RetinaMind, an AI tool that diagnoses autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD) using retinal images. His invention earned second place and a $175,000 prize at the 2026 Regeneron Science Talent Search, a renowned STEM competition for high school students in the United States. The competition recognized Kang’s work for its innovative combination of artificial intelligence and biological research.
Kang's journey began three years ago when he encountered a study from the Chinese University of Hong Kong that used retinal images to diagnose autism. Inspired by the study, he taught himself coding and machine learning to refine the model, ultimately creating a tool that achieves an accuracy rate of 89 percent. RetinaMind analyzes retinal images to determine the likelihood of autism or ADHD, providing a heat map to highlight the key areas influencing its diagnosis.
Advancing Neurodevelopmental Diagnostics
Autism and ADHD are among the most common neurodevelopmental disorders in the United States, affecting millions of children. Current diagnostic methods rely on behavioral and developmental assessments, which can delay early intervention. Kang’s RetinaMind aims to bridge this gap by offering a faster and more objective diagnostic tool. Using advanced computational techniques like ensemble learning, the model combines predictions from multiple algorithms to improve reliability and accuracy.
Kang also explored the biological mechanisms behind retinal differences in individuals with autism and ADHD. His research identified potential genetic links, such as the ABCA4 gene, which may play a role in retinal toxicity and degradation. These findings could pave the way for deeper understanding and improved diagnostic methods for neurodevelopmental disorders.
Implications for Urban Health Systems
For city leaders and municipal health systems, innovations like RetinaMind highlight the potential of AI to transform public health diagnostics. Early and accurate detection of neurodevelopmental disorders could reduce the burden on urban healthcare providers by enabling timely interventions and tailored support for affected families. This is particularly relevant in cities where access to specialized diagnostic services is limited.
Municipal health departments may consider integrating AI-driven tools into community clinics and public health programs to enhance early screening efforts. Additionally, city leaders could explore partnerships with technology developers and educational institutions to pilot such innovations locally. Addressing equity in healthcare access will be critical to ensure that underserved populations benefit from these advancements. By leveraging AI tools like RetinaMind, cities can improve health outcomes and reduce long-term costs associated with delayed diagnoses.
Questions That Matter
How did Edward Kang develop RetinaMind in Hackensack?
Edward Kang, a high school senior from Bergen County Academies in Hackensack, was inspired by a study from the Chinese University of Hong Kong using retinal images for autism diagnosis. He taught himself coding and machine learning to refine the model, resulting in RetinaMind, which achieves an 89% accuracy rate.
What prize did RetinaMind win at the 2026 Regeneron competition?
RetinaMind, developed by Edward Kang, earned second place at the 2026 Regeneron Science Talent Search. The award came with a $175,000 prize, recognizing the innovative use of AI and biological research in diagnosing autism and ADHD.