SEEG sleep data show no pre-registered age effect in preserved contacts, but exploratory slopes steepen with age

We tested if age changes the aperiodic spectral slope (20-45 Hz) and the 70-110 Hz power of SEEG contacts that clinicians labeled normal. We used OpenNeuro ds007118, ds007119 and ds007120 (NREM sleep). We read 60 s from the middle of the first run of each patient and made bipolar pairs of adjacent contacts on the same shaft. The plan was registered before we loaded signals. In the primary set (normal and preserved contacts, 20 patients, 330 pairs), age did not change the slope (rho -0.15, Holm p 0.58) or the 70-110 Hz power (rho -0.25, Holm p 0.58). Both predictions failed. In a declared second set (normal contacts with all resection states, 49 patients, 725 pairs), the slope became steeper with age (rho -0.38, Holm p 0.018). The 70-110 Hz power relative to 4-30 Hz power decreased (rho -0.36). Absolute 70-110 Hz power did not change. A post hoc check showed that the effect came mainly from 14 patients with a 60 Hz line peak (slope rho -0.69). In 35 patients with a 50 Hz line peak, the effect was weak (rho -0.15). Thus site or hardware can change the age effect. Use these results only as exploratory. The package has code, metadata provenance, per-pair features, statistics and figures.

SEEG · NREM sleep · aperiodic slope · high-frequency activity · age · bipolar recording · intracranial EEG