Artificial intelligence in ADHD assessment: a comprehensive review of research progress from early screening to precise differential diagnosis.

Attention deficit hyperactivity disorder (ADHD) diagnosis traditionally relies on subjective assessments, which lead to challenges like symptom overlap, heterogeneity, and misdiagnosis risk. Artificial intelligence (AI), especially machine learning (ML) and deep learning (DL), offers objective assessment opportunities by processing complex multimodal data (behavioral, neurophysiological, neuroimaging, genetic). This paper reviews AI's current applications in objective ADHD assessment, covering early screening, risk prediction, diagnostic assistance, classification, assistance in precise differential diagnosis, symptom quantification, and heterogeneous subtype identification. While AI models show significant potential in extracting objective biomarkers and improving assessment efficiency, the field faces challenges: insufficient standardized data, limited generalization, interpretability issues, potential biases, and lack of rigorous clinical validation. Future research must establish large-scale, standardized multimodal databases, develop robust, interpretable, and fair AI models, and conduct rigorous clinical translation validation to achieve responsible, precise, objective, and personalized ADHD assessment and management.