In 13 SEEG patients, SOZ high-frequency activity showed no significant coupling or lag difference from a cross-shaft up-state reference

We tested a pre-registered question with OpenNeuro ds007118, ds007119 and ds007120 (NREM sleep, 90 s per patient). Is high-frequency activity (HFA, 80-300 Hz) at seizure-onset-zone (SOZ) bipolar pairs less coupled, or differently lagged, relative to the brain-wide up state? For each pair, we correlated the slow HFA envelope with a reference G. G is the mean envelope of normal pairs on other shafts. The positive control passed: the coupling of normal pairs was above a circular-shift null in 10 of 10 patients (mean r 0.22 vs 0.00, p = 0.0017). H1: SOZ pairs had lower coupling in 10 of 13 patients (mean AUC 0.376, 95% CI 0.278-0.474, raw p = 0.036). This did not pass Holm correction (p = 0.072), so H1 is not supported. H2: lag did not differ (AUC 0.52, Holm p = 0.80). In an exploratory check, G used all non-SOZ pairs. Then the coupling AUC was 0.51 (p = 0.88, 16 patients), but HFA power still separated SOZ pairs (AUC 0.65, p = 0.033). Coupling did not correlate with the local phase-amplitude index (mean rho 0.04). The main limit: 28 of 43 patients had usable pairs on only one shaft, so only 13-16 patients entered. The study has low power. The package has the plan, code, per-run byte-range provenance with SHA-256, per-pair features, HFA envelopes and the reproduction steps.

intracranial EEG · non-REM sleep · seizure onset zone · high-frequency activity · up state · coupling · epilepsy