Artificial intelligence in psychiatry: Current and emerging trends, clinical applications, and research gaps explored through a bibliometric analysis.

OBJECTIVE: The aim of this study is to analyze AI-based scientific publications in psychiatry using bibliometric methods to identify prominent themes, research trends, and future opportunities.

METHODS: Original articles published between 1980 and 2025 and indexed in the "Psychiatry" category of the Web of Science were screened. The search strategy included keywords such as "artificial intelligence," "machine learning," and "deep learning." A total of 2328 original research articles were included in the analysis. Bibliometric analyses were conducted using the Bibliometrix package in RStudio. Publication trends over the years, keyword analysis, trend analysis, and factor analysis were applied.

RESULTS: Keyword analyses revealed that themes such as depressive disorders, schizophrenia, suicide, anxiety, bipolar disorder, psychosis, and digital health were central. Trend analyses showed that neurobiological markers dominated between 2014 and 2017; early psychosis, symptom severity, and fMRI were prominent between 2018 and 2020; and neurodegenerative disorders and connectivity concepts gained importance after 2021. The rise of themes such as "mental health," "digital health," and "interpretability" in 2024-2025 reflects both the digitalization of healthcare delivery and the growing demand for algorithmic transparency.

CONCLUSION: The findings indicate that AI in psychiatry holds transformative potential not only for diagnosis and prediction but also for digital health, personalized treatment, ethical governance, and community mental health. Moreover, the intensive examination of focus areas such as depression, schizophrenia, and suicide highlights the strong alignment of the field with clinical priorities. However, external validity, data heterogeneity, explainability, and ethical integration remain critical research gaps to be addressed.