Mapping the Digital Mind: A Meta-Analysis of EEG Biomarkers in Cognition, Emotion, and Mental Health

Background: Electroencephalography (EEG) provides millisecond-resolution measurements of neural activity, offering a unique potential to identify biomarkers of cognition, emotion, and mental health. However, the proliferation of methodologically diverse studies necessitates systematic synthesis to establish the reliability and clinical utility of proposed EEG biomarkers. Methods: Following PRISMA 2020 guidelines, we systematically searched PubMed, PsycINFO, Web of Science, and Scopus for studies published 2015-2025 examining EEG correlates of cognitive control, learning, emotion regulation, and mental health. From 3847 initial records, k = 210 unique studies (estimated n ≈ 9935 participants across 38 countries; see Methods for sample size derivation) met the inclusion criteria. Random-effects meta-analyses estimated pooled effect sizes for primary EEG markers across five research domains. Results: Frontal-midline theta demonstrated robust effects for cognitive control (k = 12; d = 0.89, 95% CI [0.72, 1.07]; I² = 0.0%) and learning/memory (k = 10; d = 0.70, 95% CI [0.50, 0.89]). The late positive potential indexed emotional processing (k = 18; d = 0.87, 95% CI [0.75, 1.00]) and regulation success (k = 14; d = -0.65, 95% CI [-0.79, -0.51]). Neurofeedback showed very large effects for PTSD (k = 2; d = -1.98, 95% CI [-2.50, -1.47]) and moderate effects for anxiety (d = -0.62), ADHD (d = -0.60), and depression (d = -0.42). Alpha event-related desynchronization marked cognitive engagement (k = 18; d = -0.70, 95% CI [-0.85, -0.55]). Heterogeneity was negligible (I² = 0.0%) in most analyses, except for clinical interventions, which showed condition-explained heterogeneity (I² = 75.4%). Conclusions: EEG biomarkers demonstrate substantial effect sizes and a notable consistency across cognitive and clinical domains, supporting their potential as candidate neurophysiological indicators for diagnostic research, the investigation of treatment response, and intervention monitoring. Causal claims are not warranted from this evidence base alone. A four-phase implementation framework is proposed to facilitate clinical translation. Future research should prioritize methodological standardization, diverse samples, and real-world validation.