AI Model Successfully Predicts Autism in Young Children

A new AI model predicts autism in children under two years old with 80% accuracy by analyzing behavioral features. Developed by researchers at Karolinska Institutet, it utilizes readily available information, eliminating exhaustive assessments. The findings, published in JAMA Network Open, aim for early, efficient autism diagnosis.

A groundbreaking AI model has shown remarkable promise in predicting autism in children under two years of age with an accuracy of nearly 80%, according to researchers at Karolinska Institutet in Sweden.

This innovative model analyzes 28 behavioral features of a child, using readily available information, thereby bypassing the need for extensive assessments and medical tests. The model was trained on a U.S. database comprising approximately 30,000 individuals, both with and without autism.

Early autism diagnosis is notoriously challenging due to a general lack of awareness among parents regarding early signs such as not responding to their name or not smiling. This study's significant findings highlight the feasibility of using limited information for early identification, potentially transforming autism diagnostics.

(With inputs from agencies.)