Artificial Intelligence Methods in Early Detection of Autism Spectrum Disorder: A DSM-5 Criterion-Based Systematic Review

This systematic review evaluates AI-based approaches (machine learning, deep learning, and hybrid methods) for early detection of autism in children aged 0-18, assessing their alignment with DSM-5 diagnostic criteria and methodological quality across 43 included studies. The analysis found classical machine learning and deep learning approaches were equally prevalent, with convolutional neural networks and support vector machines being the most commonly used techniques for identifying behavioral markers of autism.