Normative characterization of age-related periodic and aperiodic activity in resting-state real-world clinical EEG recordings.

INTRODUCTION

The relevance of electroencephalographic (EEG) biomarkers is increasing, as advancements in spectral analysis enable computational decomposition of complex neural signals into quantitative EEG (qEEG) parameters. Especially the differentiation of periodic and aperiodic components can reveal insights into neural function, disease biomarkers, and therapeutic efficacy. The aim of these analyses from real-world clinical routine EEG recordings was to provide normative values of physiological age-related oscillatory (periodic) and non-rhythmic (aperiodic) activity.

METHODS

We analyzed 532 physiological EEGs of patients between 8 and 92 years of age. EEG segments were preprocessed, and the power spectrum was computed using a multitaper method. We decomposed the power spectrum into periodic (peak power, frequency, and bandwidth) and aperiodic (intercept and exponent) components. Linear regression models were used to investigate age-related changes in these parameters.

RESULTS

We observed significant global age-related changes in the periodic alpha (-0.015 Hz/year) and gamma (+0.013 to +0.031 Hz/year) peak frequency as well as in the aperiodic exponent (-0.003 to -0.004 μV2/Hz/year). In the other parameters there were solely regional or no significant age-related changes.

CONCLUSION

Decomposing the power spectrum into periodic and aperiodic components allows for the characterization of age-related changes.

SIGNIFICANCE: This study provides the first spectrum-wide normative characterization of age-related changes in periodic and aperiodic activity, relevant for non-invasive brain stimulation with alternating current targeting ongoing oscillatory activity.