|
Getting your Trinity Audio player ready...
|
In a study conducted by the University of South Australia and Flinders University, researchers have unveiled a potential breakthrough in the diagnosis of autism spectrum disorder (ASD) in children using artificial intelligence (AI). Employing an electroretinogram (ERG), a diagnostic test that gauges the electrical activity of the retina in response to light stimuli, the researchers utilised AI to identify distinctive features for classifying ASD.
The study, conducted in collaboration with the University of Connecticut and University College London, involved measuring retinal responses in 217 children aged 5-16 years, comprising 71 with diagnosed ASD and 146 without an ASD diagnosis. The researchers observed distinct retinal responses in children with ASD compared to their neurotypical counterparts.
Remarkably, a single bright flash of light to the right eye yielded the strongest biomarker, and AI processing significantly reduced the testing time. The study revealed that higher frequency components of the retinal signal were diminished in ASD.
Dr Fernando Marmolejo-Ramos, a researcher at the University of South Australia, highlighted the potential of this test to revolutionise ASD diagnoses. According to him, the test, conducted with the RETeval electroretinogram testing unit, can collect data and complete a screening for autism within as little as 10 minutes. This rapid and non-invasive method could offer an improved approach to diagnosing ASD, potentially streamlining support for thousands of children on the spectrum.
Dr. Marmolejo-Ramos emphasised the significance of early interventions and appropriate support for children with ASD, acknowledging the current lack of a simple diagnostic test. He emphasised that individuals often undergo lengthy psychological assessments and reports to obtain a diagnosis, and the proposed test could significantly alleviate the time, stress, and financial burden on parents and their children.
The prevalence of ASD in Australia stands at one in 70 people, with approximately 353,880 individuals on the autism spectrum. Notably, ASD is four times more common in boys than girls. Globally, the rates of autism vary, with the World Health Organization estimating a prevalence of one in every 100 children.
Dr Paul Constable, a researcher at Flinders University and the project lead highlighted the exciting prospect of using the electroretinogram with signal analysis and machine learning to classify ASD with greater accuracy. He emphasised the connection between the eye and the brain, underscoring the potential of exploring new ways to understand the brain’s development in individuals with ASD.
Acknowledging the need for further research, Dr Constable emphasised the importance of studying younger children and those with other conditions, such as attention deficit hyperactivity disorder, to ascertain the specificity of the test. He deemed this initial step crucial in exploring innovative approaches to employ the electroretinogram alongside signal analysis and machine learning for enhanced ASD classification.
Co-researcher Dr. Hugo Posada-Quintero, Assistant Professor in the Department of Biomedical Engineering at the University of Connecticut, echoed the promising potential of analysing retinal responses using advanced signal processing and machine learning techniques.
He emphasised the importance of further research and technological development to transform these analytic methods into practical tools for clinicians. These tools could aid in accurately and efficiently screening for and diagnosing ASD and related neurodevelopmental disorders.
The collaborative research effort between the University of South Australia, Flinders University, the University of Connecticut, and University College London presents a significant advancement in the quest for a faster and more accurate method of diagnosing ASD in children. The integration of AI with the electroretinogram opens new avenues for research and technological development, offering hope for improved outcomes in the early identification and support for children with ASD.
