HOME Science & Technology

Artificial Intelligence detects childhood autism through advanced eye photo analysis

2024.01.12 07:20:47 Eileen Yoon

[A photo of an eye. Photo Credit to Pixabay ]

Recently, researchers have disclosed a unique approach using artificial intelligence (AI) to identify childhood autism with unparalleled accuracy of 100%, a development that could transform the field of autism diagnosis.

Autism spectrum disorder, abbreviated as ASD, is a neurodevelopmental disorder marked by difficulties with social interaction, communication, and repetitive activities.


The difficulty in diagnosing autism spectrum disorder (ASD) in its early stages has prevented affected children from receiving the assistance and intervention they need until now.


However, this new AI-driven method uses highly sophisticated algorithms to examine images of the eyes and spot subtle patterns that could correspond to signs of autism spectrum

disorder (ASD).


The study, conducted by a team of researchers from Yonsei University, examined the retinas of 958 young children, ages 7 to 8 on average, half of whom had already received an autism diagnosis.


Although the training data in this case was restricted to a certain age group, the researchers believe that it will still be effective for even younger children in a future study.


To build models for diagnosis, the AI was trained using 85% of the retinal images and scores on the severity of the symptoms.


The remaining 15% of the eye pictures were saved to evaluate the AI's capacity for diagnosis.


The screening process was based on a cutting-edge procedure that examined the junction of the optic nerve and retina with the optic disc in the back of the eye to provide information about the brain.


The structure, an outgrowth of the central nervous system, opens the brain.


Scientists have begun to take advantage of this by readily and non-invasively accessing body parts to gather crucial data about the brain.


As a result of the study, the AI system achieved 100% accuracy in diagnosing individuals with autism spectrum disorder as children.


The key to the success of the system lies in the AI’s ability to detect patterns imperceptible to the human eye.


The AI is able to identify typical patterns linked to autism by examining parameters like pupil size and shape, eye spacing, and other minute details.


One of the major implications of this accomplishment is the possibility of early intervention and assistance for autistic children since early diagnosis is critical for implementing effective therapy procedures and providing required support to those on the autism spectrum.


The findings underscore the potential of AI technology as an unbiased screening tool for the early identification of autism, especially in scenarios where access to pediatric psychiatrists with specialized training may be restricted.


Remarkably, one in 36 people is thought to be on the autistic spectrum, illustrating the

significance of early identification for enhanced self-awareness and overcoming everyday challenges.


The researchers conclude that retinal photos are a viable option for an objective means of evaluating the severity of symptoms and screening for autism, opening up new possibilities for early intervention and assistance in neurodevelopmental diseases.


It is clearly shown that AI has the potential to revolutionize the healthcare field.

Eileen Yoon / Grade 11
Fayston Preparatory of Suji