SOZ extent showed no detectable relation to NREM marker separation in 43 SEEG patients

We tested if interictal NREM markers of the seizure-onset zone (SOZ) work less well when the SOZ is extensive. We used the archived per-pair features of publication a3d06b3c (43 SEEG patients, 728 bipolar depth pairs, 160 SOZ pairs, 90 s of NREM sleep). SOZ extent was the fraction of pairs in the SOZ (median 0.27, range 0.03-0.73). For each patient, we calculated the AUC of SOZ pairs vs other pairs for 80-300 Hz power (HFA) and for slow-wave-phase-to-HFA coupling (MI). The plan was pre-registered. The Spearman correlation between extent and AUC was +0.18 for HFA (95% CI -0.16 to +0.47, Holm p = 0.51) and +0.02 for MI (CI -0.31 to +0.36, Holm p = 0.89). Control for pair count, permutation within cohort group, a focal vs extensive split, and SOZ pair count gave the same result. A label-shuffle check showed no AUC bias that depends on extent. Thus the markers did not become weaker in patients with extensive SOZs. The sample can find only |rho| >= about 0.41, and extent is the sampled fraction, not the anatomical volume.

seizure-onset zone · SEEG · NREM sleep · high-frequency activity · slow-wave coupling · SOZ extent · AUC · epilepsy