Spike-rate asymmetry was positive but did not pass correction in two independent SEEG cohorts

This preregistered replication tested whether interictal spike rate is higher in the seizure-onset-zone (SOZ) hemisphere, as reported in one center, using archived per-pair features from two independent cohorts. OpenNeuro ds003876 included 15 patients from four centers; ds005398 included 9 UCLA pediatric patients with sleep clips. In ds003876, the primary all-pairs lateralization index (mean within-patient z score on the SOZ side minus the other side) for spike rate was +0.21 z (95% CI 0.02 to 0.39; 11/15 positive; one-sided p = 0.026; Holm p = 0.106), so the preregistered corrected threshold was not met. Without SOZ pairs, the index was +0.08 z (p = 0.19). In a secondary name-matched mirror-pair analysis, spike rate was +0.30 z (Holm p = 0.14), while broadband 1–40 Hz log power was +0.59 z (Holm p = 0.020); this power result was not part of the original test. Delta fraction, high-gamma power, and line length did not lateralize in the all-pairs contrast. In ds005398, no feature lateralized, and all mean indices were negative for both clips. The hemisphere with more implanted pairs matched the SOZ side in 13/15 and 7/9 patients, more often than any interictal feature. The spike-rate result has the same sign as the earlier HUP result (+0.36 z) but is smaller and does not meet the corrected threshold. Limits include small cohorts, hemisphere assignment from shaft names in ds003876, clinical SOZ labels, and different spike detectors. The analysis used archived features and did not reprocess signals.

stereo-EEG · epilepsy · seizure onset zone · interictal spike rate · lateralization · replication · OpenNeuro