A new AI system can predict autism in young children with limited information, according to a study published in JAMA Network Open. The team from Karolinska Institute, Sweden, suggests this can be used to provide parents with the proper support.
“With an accuracy of almost 80 percent for children under the age of two, we hope that this will be a valuable tool for healthcare,” said Kristiina Tammimies from the Department of Women’s and Children’s Health, Karolinska Institutet.
Using an extensive database with information on approximately 30,000 individuals, the team analysed a combination of 28 different parameters, which were used to create four distinct machine-learning models to identify patterns in the data. Crucially, all the information was obtained before the participants were 24 months old.
Of the four models tested, the best-performing one was named ‘AutMedAI.’ This model was able to identify 80% of children with autism. Some factors used in the model included the age of first smile, the first short sentence, and the eating difficulties. “The study results are significant because they show that it is possible to identify individuals who are likely to have autism from relatively limited and readily available information,” said Shyam Rajagopalan, an affiliated researcher at Karolinska Institutet.
Early diagnosis is critical to implement effective interventions that can help children with autism and their parents. “This can drastically change the conditions for early diagnosis and interventions and ultimately improve the quality of life for many individuals and their families,” said Shyam Rajagopalan.
The authors plan how to implement and validate the model in clinical settings. They’re also working on ways to include genetic information in the model, which may lead to even more specific and accurate predictions. “To ensure that the model is reliable enough to be implemented in clinical contexts, rigorous work and careful validation are required. I want to emphasize that our goal is for the model to become a valuable tool for health care, and it is not intended to replace a clinical assessment of autism,” said Kristiina Tammimies.
Shyam Sundar Rajagopalan, Yali Zhang, Ashraf Yahia, Kristiina Tammimies. Machine Learning Prediction of Autism Spectrum Disorder from a Minimal Set of Medical and Background Information. Jama Network Open, 2024, doi: 10.1001/jamanetworkopen.2024.29229