Normative age-related changes in resting-state EEG alpha power and theta/alpha ratio across eyes-open and eyes-closed conditions.

Resting-state electroencephalography (EEG) provides a sensitive measure of large-scale neural dynamics and has been widely used to characterize age-related changes in brain function. However, many prior studies have relied on relatively small samples or single-condition recordings, limiting the generalizability of findings. In this study, we investigated normative age-related trajectories of EEG spectral features across eyes-closed (EC) and eyes-open (EO) conditions in a large cohort of healthy adults. After reproducible preprocessing and artifact rejection, a total of 3,229 resting-state EEG recordings from 608 participants aged 20-70 years were included. EEG data were analyzed across eyes-closed (EC) and eyes-open (EO) conditions, pre- and post-task timepoints, and baseline and follow-up sessions when available. We focused on occipital alpha relative power, global theta/alpha ratio, and occipital alpha peak frequency as spectral markers. Linear mixed-effects models with participant-specific random intercepts were used to account for repeated measurements, and model estimates were reported with standard errors, confidence intervals, test statistics, and false discovery rate-adjusted p-values. Results showed a robust age-associated decrease in occipital alpha relative power, with lower alpha expression in EO than EC. In contrast, the global theta/alpha ratio increased with age, suggesting an age-associated shift in conventional spectral balance toward relatively greater low-frequency predominance. Occipital alpha peak frequency showed weaker and more variable patterns and was therefore interpreted as a supplementary marker. These findings provide normative descriptive information on age-associated resting-state EEG spectral variation in healthy adults and suggest that EC and EO recordings may be more informatively analyzed as distinct recording contexts rather than assumed to provide fully interchangeable baseline estimates.