Photographs of the retina may aid in diagnosing ASD in children, adolescents

An artificial intelligence program designed to find patterns in photographs of the retina of the eye can identify children with autism spectrum disorders (ASD) with a high degree of accuracy, according to a new study from South Korea.

Jae Han Kim and colleagues used artificial intelligence to detect patterns in the retinas of children with ASD. They then tested their model on 958 children and adolescents between four and 19 years of age. Half of the children had ASD, and half were neurotypical.

The researchers report that their model identified every one of the autistic participants accurately and generated no false positives, although the model was less accurate at measuring the severity of autistic symptoms. They say their findings indicate that “retinal photographs may serve as an objective screening tool starting at least at age four years… which is earlier than the average age of 60.48 months at ASD diagnosis.” However, they say it is unknown whether the technique will work for children younger than four years of age, as the retina is still developing in this group.

The researchers caution that their study had some limitations. For example, they did not control for medication use in participants, and they excluded children with co-existing conditions. However, they say, “Our findings suggest that the optic disc area is crucial for differentiating between individuals with ASD and [those with] typical development.”

They conclude, “Although future studies are required to establish generalizability, our study represents a notable step toward developing objective screening tools for ASD, which may help address urgent issues such as the inaccessibility of specialized child psychiatry assessments due to limited resources.”

Citations

Development of deep ensemble to screen for autism and symptom severity using retinal photographs,” Jae Han Kim, JaeSeong Hong, Hangnyoung Choi, Hyun Goo Kang, Sangchul Yoon, Jung Yeon Hwang, Yu Rang Park, and Keun-Ah Cheon, JAMA Network Open, December 15, 2023 (free online). Address: Yu Rang Park, Department of Biomedical Systems Informatics, Yonsei University College of Medicine, Yonsei-ro 50-1, Seodaemun-gu, Seoul 03722, Republic of Korea, [email protected]).

—and—

“Using AI to diagnose autism in children,” Bob Yirka, Medical Xpress, December 18, 2023.