In 27 HUP SEEG patients, absolute power but not the 30–55 Hz exponent separated clinical seizure-onset channels
In 27 patients with stereo-EEG (SEEG) and at least one clinically marked seizure-onset-zone (SOZ) bipolar channel and three non-SOZ channels, the pre-specified 30–55 Hz aperiodic exponent did not separate SOZ from non-SOZ channels (mean patient-level AUC 0.553, 95% CI 0.462–0.643; sign-flip p = 0.26). Results were also null for each run, same-shaft controls, a strict SOZ label, a 2–40 Hz exponent, and the median-PSD exponent. Absolute log power was higher in SOZ channels across bands; total 1–175 Hz power had AUC 0.696 (95% CI 0.627–0.764; p = 5×10⁻⁵; q = 2×10⁻⁴), with a same-shaft AUC of 0.685 (n = 20; p = 0.002). Relative 125–175 Hz power was lower in SOZ channels (AUC 0.320, 95% CI 0.250–0.393; p = 5×10⁻⁵). No feature differed significantly by surgical outcome (15 good, 12 poor; all permutation p > 0.17). Features were stable between two runs (median channel Spearman ρ 0.86 for the primary exponent and 0.97 for log total power; per-patient AUC correlations 0.85 and 0.90). The analysis used only the first 120 seconds of each five-minute interictal clip and lacked anatomical controls; power differences may reflect tissue type, contact location, or interictal spikes rather than epileptogenicity. Clinical SOZ labels are imperfect.
stereo-EEG · SEEG · epilepsy · seizure onset zone · aperiodic exponent · spectral power · interictal · OpenNeuro · HUP iEEG
References
1- usesdoi:10.18112/openneuro.ds004100.v1.1.3
The analysis used interictal SEEG EDF prefixes, channels.tsv SOZ and resection labels, and participants.tsv outcome data from OpenNeuro ds004100 snapshot 1.1.3. Per-run hashes of downloaded byte prefixes are recorded in results/provenance.tsv.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.