Two-second delta power ranked seizure-onset-zone SEEG channels above chance in 11 patients; spike rate needed about 60 seconds

In 11 patients with SEEG from one center, we compared five fixed features for ranking clinically labeled seizure-onset-zone (SOZ) bipolar channels using non-overlapping windows from 2 to 120 s of one interictal run. Delta power (1–4 Hz) ranked SOZ channels above chance in all 11 patients at every window length (exact two-sided sign-flip p = 0.00098). A 2 s window retained 84% of the 120 s gain over chance (mean AUC 0.715 versus 0.756); the small losses at 2, 5, and 10 s were consistent, while differences from 120 s were not significant from 20 s onward. Spike rate was at chance at 2 s and reached 94% of its 120 s gain only at 60 s. Line length and 70–115 Hz power produced stable channel maps (median disjoint-window rho 0.96), but neither SOZ effect reached p < 0.05. Combining delta power and spike rate did not improve on delta power alone. The analysis used the first 120 s of one interictal run in 11 analyzed patients selected from the first 12 SEEG patients in file order; one patient without an SOZ label was excluded. The sample is small, from one center, and not random. Labels are clinical and partly based on interictal abnormalities, sleep state is unknown, the spike detector was not validated against expert marks, and no correction was applied across 35 tests. AUC therefore does not measure surgical outcome. The package preserves the original analysis code, tables, figure, source URLs, ETags, and SHA-256 hashes of fetched EDF prefixes; the raw EDF data are not included.

SEEG · epilepsy · seizure-onset zone · interictal · delta power · spike rate · AUC · window length · channel ranking