AI Predicts Autism in Toddlers with High Accuracy

Researchers at Karolinska Institutet have developed a machine learning model, named AutMedAI, capable of predicting early autism in children under two years old with nearly 80% accuracy. The model was created using data from the SPARK study, which included information on approximately 30,000 individuals, both with and without autism spectrum disorders. AutMedAI identified about 80% of children with autism among 12,000 participants, focusing on patterns related to social communication difficulties, cognitive abilities, and developmental delays.

The research team, led by Shyam Rajagopalan and Kristiina Tammimies, plans to further improve and validate the model in clinical settings. They are also working to incorporate genetic information, which could enhance the model's accuracy. This study was funded by the Swedish Foundation for Strategic Research, Hjärnfonden, and Stratneuro. The early identification facilitated by AutMedAI could potentially enable timely interventions, improving the quality of life for affected individuals and their families.