ADHD Diagnosis Could Be Revolutionized by AI
An accurate diagnosis of ADHD is essential in providing the right support to individuals who need it. However, current diagnosis methods are often time-consuming and inconsistent. A recent study suggests that artificial intelligence (AI) could potentially revolutionize the way ADHD is diagnosed.
Researchers in South Korea conducted a study where they trained machine learning models to analyze characteristics in photos of the fundus at the back of the eye and connect them to a professional diagnosis of ADHD. The study tested four machine learning models, with the best achieving an impressive 96.9 percent accuracy in predicting ADHD based solely on image analysis.
The team discovered that higher blood vessel density, shape and width of vessels, and specific changes in the optic disc of the eye were key indicators of the presence of ADHD. For years, it has been hypothesized that changes in brain connectivity associated with ADHD could also manifest in the eyes. By identifying these visual markers, a faster and more reliable method for detecting the disorder could be developed.
The researchers, led by a team from Yonsei University College of Medicine, published their findings, highlighting the potential of retinal fundus photographs as a noninvasive biomarker for screening ADHD and stratifying executive function deficits in the visual attention domain.
The AI system was tested on 323 children and adolescents diagnosed with ADHD and 323 individuals without an ADHD diagnosis. The results showed that the AI system excelled in predicting ADHD and identifying characteristics of the disorder, such as impairments in visual selective attention.
While various machine learning techniques have been explored for ADHD screening, this approach stands out for its speed, simplicity, and scalability. Unlike other methods that rely on a diverse set of variables, this model focuses solely on retinal photographs, enhancing clarity and utility.
Moving forward, the researchers aim to expand their tests to larger and more diverse groups of people, including different age ranges. They also acknowledge the need to refine the system further, particularly in distinguishing ADHD from other conditions like autism spectrum disorder.
With approximately 1 in 20 individuals estimated to have ADHD, the potential impact of a quicker and more accurate diagnosis is significant. Early screening and intervention can greatly improve social, familial, and academic functioning in individuals with ADHD.
The research has been published in npj Digital Medicine, offering a promising glimpse into the future of ADHD diagnosis with the help of AI technology.
