Sigma peak was not lower, and detected sigma events were more frequent, at seizure-onset-zone SEEG pairs

In 43 patients with stereo-EEG (SEEG) recordings labelled as non-REM sleep, we compared 11–16 Hz activity at clinical seizure-onset-zone (SOZ) bipolar depth pairs with other pairs in the same patient. The analysis used 240 s per patient from OpenNeuro dataset versions 1.0.0 and included 728 pairs (160 SOZ pairs). The mean per-patient AUC for sigma prominence above a fitted 1/f background was 0.497 (95% CI 0.399–0.594; Holm-adjusted p=0.95), providing no evidence of a reduction at SOZ pairs; the confidence interval does not exclude a small effect in either direction. Envelope-detected sigma event density was higher at SOZ pairs (AUC 0.608, 95% CI 0.526–0.687; Holm-adjusted p=0.027), opposite to the prediction. Raw sigma power and broadband 2–40 Hz power were also higher (AUCs 0.709 and 0.708). In post-hoc checks, the density difference was absent after fitting for broadband power (AUC 0.512, p=0.76), and SOZ pairs had more rejected large-amplitude events. These checks suggest, but do not establish, that broadband activity and epileptiform transients may affect the detector. The recordings were not scored for N2 versus N3, the detector was not validated against expert marks, and each patient contributed one 240-s window that overlaps windows in related analyses. The findings do not support reduced sigma prominence at SOZ pairs; interpretation of detected events is limited by the detector and recording window.

sleep spindles · sigma activity · stereo-EEG · seizure-onset zone · epilepsy · intracranial EEG · signal processing