With a 250 microvolt range limit, stage 2 gives a less negative linear age slope for frontal 0.5–4 Hz EEG power in 78 Sleep-EDF adults

Sleep stage selection changed the cross-sectional age slope of frontal slow EEG power in 153 nights from 78 adults in Sleep-EDF Expanded 1.0.0. The primary measure was mean epoch log power from 0.5 to 4 Hz in Fpz-Cz. With a 250 microvolt range limit, weighted linear age slopes were −0.368 dB per decade for stage 2 and −0.522 for stages 2–4, a paired difference of +0.154 (95% person bootstrap interval +0.095 to +0.217). At a 150 microvolt limit, the difference was +0.027 (−0.013 to +0.065); that limit removed 88.2% of stage 4 epochs accepted at 250 microvolts. The stage contrast therefore depends on the voltage rule, and these results do not determine which rule gives a more valid neural measure. The original local package generated 146,880 epochs with two EEG channels and seven power bands. SCI omits that source-derived cache and source labels and metadata under their ODC Attribution terms; it retains analysis code, numerical results, figures, provenance, and repeat instructions. Independent signal and model checks passed. An offline repeat reproduced all 17 numerical result files. These are cross-sectional associations under a common linear age assumption, not within-person age changes.

sleep EEG · age association · sleep stage · slow-wave power · spectral power · Sleep-EDF · bootstrap · signal quality · cross-sectional analysis