Early identification of autism spectrum disorder remains one of child psychiatry's hardest logistical problems: specialist assessment is slow and scarce, and waiting lists mean children often miss the window when intervention helps most. In May 2026 Yonsei University College of Medicine announced the start of a project to build an AI screening platform that approaches the problem through an unexpected channel, the eye.
The platform, led by Severance Hospital child psychiatrist Cheon Keun-ah as principal investigator with Professor Park Yurang and the digital healthcare company Huraypositive, is a multimodal system combining analysis of eye fundus images, non-invasive photographs of the back of the eye that carry signatures of the central nervous system, with developmental behavioural indicators. The project was selected for a public research outcomes demonstration pilot run by the Ministry of Science and ICT and the Korea Technology Commercialisation Promotion Agency.
Backed by roughly 1.32 billion won over 33 months, from April 2026 to December 2028, the team aims to refine existing algorithms with clinical data, test usefulness in real patients, and reach technology readiness level 7 or higher, the threshold associated with commercialisation. The ambition is a tool quick and simple enough for primary care clinics and home use, moving autism screening out of specialist centres and towards the point where families first raise concerns.