In 18 pediatric SEEG patients, 10-s sleep segments showed higher 1–250 Hz power in SOZ contacts but not fast-ripple power or spike rate
We analyzed all 18 SEEG patients marked as methods == 0 in OpenNeuro ds005398 version 1.1.3. For each patient, we used one 10-s sleep segment starting at 600 s and analyzed the bipolar channels as stored. Among 1,285 usable channels, 181 were clinically marked seizure-onset-zone (SOZ) contacts. Within-patient median AUCs for SOZ versus non-SOZ contacts were 0.70–0.79 for power from 1 to 250 Hz (16–17 of 18 patients had AUC > 0.5; sign-flip p ≤ 0.0001). Fast-ripple power (250–500 Hz; AUC 0.53, p = 0.92) and spike rate (AUC 0.51, p = 0.18) did not separate the groups in this short window; line length had AUC 0.61 (p = 0.015). The pattern remained in grey-matter contacts only. A leave-one-patient-out logistic model had median AUC 0.75; its 30-shuffle null had a 95th percentile of 0.56 and permutation p = 0.032 (the minimum possible with 30 shuffles). The analysis used one window per patient, with unknown sleep stage. Clinical SOZ labels may be partly circular, bipolar power depends on contact distance and tissue, and only three SEEG patients had resection, so surgical outcome validation was not available. Null fast-ripple and spike-rate findings do not establish absence of these signals.
stereo-EEG · seizure-onset zone · sleep · intracranial EEG · power spectrum · pediatric epilepsy · AUC
References
1- usesdoi:10.18112/openneuro.ds005398.v1.1.3
This version of OpenNeuro ds005398 supplied the SEEG recordings, channel labels, and participant methods field used in the analysis.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- supportsAlpha-power maps were stable across six 2-s windows and matched full-night STE HFO maps in 18 pediatric SEEG recordings
The power-based SOZ separation is reproduced in every one of six windows across the night (median AUC 0.73-0.78), so it does not depend on one sample time.
- usesAlpha-power maps were stable across six 2-s windows and matched full-night STE HFO maps in 18 pediatric SEEG recordings
We use the same features and method and extend them from one 10-s window to six 2-s windows across the night.
- supportsIn 18 pediatric SEEG patients, 10-minute STE HFO maps approached full-recording SOZ separation, while MNI maps did not
The same 18 ds005398 SEEG patients. That result found no SOZ separation by spike rate in 10 s and said that event measures need longer data. Here, HFO-event-rate maps (STE) separate the SOZ with 5-10 min windows (AUC 0.73-0.77), but not with 1 min (0.54).
- usesIn 18 pediatric SEEG patients, HFO-rich contacts had less stable 80–250 Hz sleep power
Same cohort and channel selection; it showed higher 80-250 Hz power on SOZ contacts, which we extend to the selection of BCI channels.