High NREM HFA and slow-wave coupling in resected SEEG pairs showed no clear outcome association in 36 patients

In a pre-planned analysis of 36 SEEG patients (22 with ILAE 1 and 14 with ILAE 2–6 outcomes), we tested whether resection of depth-electrode pairs in the top quartile for NREM high-frequency activity (HFA, 80–300 Hz) or slow-wave-to-HFA coupling (Tort modulation index, MI; 0.5–2 Hz phase to 80–300 Hz amplitude) related to seizure freedom. The resected fraction of top-quartile HFA pairs had AUC 0.560 (95% CI 0.381–0.731; Holm-adjusted permutation p = 0.51); for MI pairs AUC was 0.659 (0.484–0.818; Holm p = 0.17). The pre-specified decision rule therefore found no evidence of an association for either marker; the confidence intervals do not exclude moderate effects. The resected fraction of all depth pairs had AUC 0.731 (0.558–0.891; p = 0.015, secondary and unadjusted), while marker-specific excess above overall resection extent had AUC below 0.5 (HFA 0.364, p = 0.16; MI 0.394, p = 0.28). In this sample, the markers did not add outcome information beyond resection extent among the measured depth pairs. The primary cohort required at least four pairs and one resected pair; seven of 43 patients were excluded. Sensitivity results for all 43 patients, cohort-stratified results, and exploratory analyses are reported in the main text. Limitations include the small sample, unequal outcome group sizes by cohort, partial implant coverage by SEEG depth pairs, no ECoG analysis, one 90 s NREM window per run, and the 80–300 Hz band being limited by 1000 Hz sampling and a 300 Hz hardware low-pass filter at some sites. The study cannot establish that the markers have no effect.

stereo-EEG · SEEG · epilepsy surgery · seizure outcome · high-frequency activity · phase-amplitude coupling · NREM sleep · resection · AUC · SOZ