Artificial Intelligence for Mental Health Care: A Systematic Review of Diagnostic Applications

Abstract

Mental health conditions pose a critical health challenge globally, yet the diagnosis is complex and often hindered by multiple factors such as overlapping symptoms and subjective assessments. Adopting Artificial Intelligence (AI) in mental health care is groundbreaking and has presented ways to handle issues faced in traditional approaches. This paper aims to analyse the pertinence and effectiveness of AI in mental health for diagnostic purposes. This systematic review presents the applications of AI in the evaluation of mental health conditions. A database search (Google Scholar, Web of Science (WoS), Scopus, IEEE Xplore, PubMed, ScienceDirect) was conducted for the years 2010 – 2025 and a total of 32 studies relevant to this study were included based on a defined inclusion criterion. Findings reveal that the models in machine learning and deep learning, mainly Support Vector Machines (SVM), Random Forests, and Convolutional Neural Networks (CNN), have demonstrated strong potential in diagnosing mental health disorders with approximately 80% accuracy on controlled datasets. This review study highlights the present state of AI-based diagnosis of mental health disorders, identifies key limitations and outlines future directions for robust AI-based diagnostic systems.

Ähnliche Arbeiten