Longer HUP SEEG seizures show a trend toward broadband postictal suppression, but the primary gamma result does not pass Holm correction

We tested if seizure duration predicts the postictal change in SEEG band power in the same patient. We used archived per-pair features from ce52bbb6 (OpenNeuro ds004100, HUP; 120 clips, 33 patients). For each seizure, R = mean over bipolar pairs of log10(post/pre power). Post = offset+2 s to +20 s. Pre = onset-30 s to -5 s. For each patient with 3 or more seizures (26 patients, 109 seizures), we calculated the Spearman rho between log duration and R. Plan was hashed before analysis. Primary gamma (30-55 Hz): mean rho -0.27 [-0.49, -0.04], sign-flip p = 0.031, Holm p = 0.062. Primary delta (1-4 Hz): rho -0.23, p = 0.084. The sign is opposite to the slowing prediction. All bands and line length have negative mean rho (-0.23 to -0.33). Pooled within-patient slopes per doubling of duration are negative, and their CIs do not include 0. Pre-ictal level does not change with duration (all p >= 0.25). Postictal level alone decreases (alpha-beta rho -0.47, p = 0.0003). With seizures >= 20 s only, the effect is weaker. The effect is clearer at non-SOZ pairs than at SOZ pairs. Conclusion: a consistent trend to broadband postictal suppression after longer seizures, not a confirmed effect. Seizure type was not available and can confound the result.

seizure duration · postictal suppression · SEEG · intracranial EEG · band power · epilepsy