Autism spectrum disorder is most often diagnosed only at the age of 3-5 years, although often later. In recent years, scientists have increasingly turned to the capabilities of artificial intelligence and computer models that should help the doctor pay attention to certain symptoms in a child in time. Researchers from the United States decided to conduct diagnostics in a game format and developed a technology for assessing the child's movements, which can be used to judge the presence of ASD.
During the diagnostics, the child is asked to repeat the dance movements of a video avatar for one minute. During this time, a computer assessment of the imitation of movements is carried out. The accuracy of the new system was assessed in 183 children aged 7 to 13 years. The results are published on the Kennedy Krieger Institute website.
The test showed 80% accuracy in identifying ASD and 70% accuracy in distinguishing ASD from attention deficit hyperactivity disorder. “This is a very important advantage, since ADHD and autism often occur simultaneously and even experienced doctors have difficulty making a diagnosis,” the authors emphasize.
The researchers now plan to adapt the diagnostics for young children, as well as expand testing for more serious developmental problems. The goal is to make it a simple, accurate, and patient- and physician-friendly tool that can be used in any medical setting.
Recently, other scientists have found signs of autism in disrupted gut microbiota. This discovery also suggests the possibility of creating a simple and accurate test for autism, regardless of gender, age, and other factors.