AI Model Shows Promise in Early Autism Detection for Toddlers

Karolinska Institutet's machine learning tool predicts autism with 80% accuracy, potentially transforming early diagnosis and intervention.

The AI model, AutMedAI, analyzes 28 developmental parameters to predict autism in children under two years old.
Developed using a large US database, the model identified about 80% of children with autism in validation tests.
Key predictors include age of first smile, speech development, and eating difficulties.
Early diagnosis through this tool can lead to timely interventions, improving outcomes for children and families.
Further validation and improvements are planned to ensure clinical reliability and incorporate genetic information.