Adelaide, Australia :A groundbreaking study from the University of South Australia and Flinders University suggests that artificial intelligence (AI) could revolutionize Autism Spectrum Disorder (ASD) diagnosis in children. Using a single flash of light and AI analysis of retinal responses, researchers may have unlocked a faster and more accurate method for identifying ASD.
The study employed electroretinography (ERG), a test that measures electrical activity in the retina following light stimulation. Analyzing data from 217 children aged 5-16 (71 with ASD and 146 without), researchers observed distinct differences in retinal responses between those with and without ASD. Notably, the strongest indicator emerged from a single bright flash to the right eye, processed by AI to significantly reduce test time.
Currently, diagnoses require lengthy psychological assessments
This research, conducted in collaboration with the University of Connecticut and University College London, holds immense promise for earlier and more accessible ASD diagnosis. Currently, diagnoses require lengthy psychological assessments, often leaving children and families waiting for vital support.
“This test offers a much faster alternative,” explains Dr. Fernando Marmolejo-Ramos, a UniSA researcher. Within 10 minutes it, RETeval ERG allows data collection and rapid screening.This not only saves time and resources but also reduces stress for families.”
The non-invasive nature of the test further adds to its potential. “The procedure is well-tolerated by children, making it a more comfortable experience for everyone involved,” Dr. Marmolejo-Ramos emphasizes.
The connection between the eye and brain provides a unique window into neural development in individuals with ASD, as explained by Dr. Paul Constable, Flinders University researcher and project lead.
While further research is required to assess the test’s applicability to younger children and those with other conditions, this initial study marks a significant milestone. Dr. Hugo Posada-Quintero, co-researcher and Assistant Professor at the University of Connecticut, highlights the future potential: “Our findings support the promising application of advanced signal processing and machine learning in analyzing retinal responses. With further development, these methods could become valuable tools for clinicians to effectively screen and diagnose ASD, and potentially other neurodevelopmental disorders.”
This groundbreaking research paves the way for a future where rapid and accurate ASD diagnosis becomes a reality, paving the path for earlier intervention and improved quality of life for countless children and their families.
