Clinical spiking labels explain part of the non-SOZ top-HFA tail in 21 SEEG patients

We tested whether the bipolar depth pair with the highest interictal 80–300 Hz power outside the clinical seizure-onset zone (SOZ) was more often in the clinical spiking (irritative) zone than expected by chance. We used an archived per-pair table from 40 patients in OpenNeuro ds007118, ds007119, and ds007120, comprising 90 s of NREM sleep per patient; no new signal processing was performed. The plan was written and hashed before analysis. In the 21 patients whose top-HFA pair was not SOZ, 10 top pairs were spiking pairs (expected 6.65; exact one-sided p = 0.035, Holm p = 0.035). In a second planned test, mean per-patient AUC for spiking versus normal pairs was 0.75 (20 of 25 patients above 0.5; sign-flip p = 0.0001, Holm p = 0.0002). The top pair was in the SOZ or spiking zone in 73% of patients (chance 52%). These results indicate that part of the non-SOZ HFA tail agrees with the clinical irritative-zone labels. The effect was moderate and the primary p-value was near 0.05. Labels were not outcome-validated, and interictal spikes contain power in the analyzed band, so the tests may partly measure the spikes that informed the labels. The findings do not show that HFA adds information beyond spikes. This is one cohort group, mainly ds007119, with a small primary-test sample; the null assumes HFA is exchangeable among a patient's non-SOZ pairs.

intracranial EEG · high-frequency activity · seizure-onset zone · irritative zone · NREM sleep · SEEG