In 261 RAM SEEG adults aged 18–64, beta peak height rose and peak frequency fell with age, while relative beta power did not change and the height effect was not beta-specific

We tested a pre-registered prediction that the 13–30 Hz (beta) peak above the aperiodic line increases with age. We analyzed unchanged 0.5 Hz per-pair spectra from 261 RAM epilepsy patients aged about 18–64 years, recorded during 60 s at the start of a memory task. The spectra came from OpenNeuro ds004789 v3.1.0 and ds004809 v2.2.0. Using the inherited aperiodic fit and peak finder, we estimated partial Spearman correlations adjusted for site, sampling rate, sex, log number of accepted pairs, and dataset. Beta peak height increased with age (rho +0.270, 95% CI +0.156 to +0.381, permutation p 0.0001); the association remained after adjustment for implant location (rho +0.312). Beta peak frequency decreased with age (rho −0.277, Holm p 0.0008; slope −0.57 Hz per decade). The age difference was in the 14–22 Hz range; there was no difference at 26–30 Hz. Relative 13–30 Hz power did not change (rho +0.039, p 0.52). After adjustment for the 4–13 Hz peak height, the beta height association was small and not significant (rho +0.106, Holm p 0.28). The data therefore do not show an age effect specific to beta. A larger shared periodic component in 4–22 Hz, or a property of the aperiodic fit, could explain both peak-height effects. Results are limited to epilepsy patients aged about 18–64, task-state recordings, a permissive peak threshold, and location control using median MNI coordinates. The low-beta residual may include the upper flank of the alpha peak, and post hoc tests are exploratory.

stereo-EEG · SEEG · aging · beta oscillations · spectral peaks · aperiodic activity · memory task · epilepsy