Social Media Use and Mental Health: Insights from a Targeted Thematic Review

Abstract

This thematic review examines eighteen studies published between 2017 and 2025. It focuses on using social media data to detect and predict mental health conditions, particularly depression and related disorders. The review synthesizes current trends in computational methods, psychological and demographic factors, and ethical issues across major platforms such as Facebook, Instagram, Reddit, and Twitter. The findings show that advanced machine learning models, especially transformer-based approaches, improve the accuracy of mental health detection. Psychological mediators such as social comparison, body image, and loneliness impact the link between social media use and mental health outcomes, especially for adolescents and females. Ethical concerns related to privacy, informed consent, and bias reporting vary greatly among studies, stressing the need for standardized guidelines. While there is promise in this area but challenges remain. These include a lack of diversity in methods, difficulties in generalizing findings beyond Western, English-speaking populations, and integrating results into clinical practice. This review serves as a valuable reference for researchers, clinicians, and policymakers, helping them guide future efforts and ensure the responsible use of social media analytics in mental health surveillance.