Thirty-second SOZ AUCs were stable in eight HUP SEEG patients, with larger variation for high-frequency power than line length
We measured how per-patient seizure-onset-zone (SOZ) area under the ROC curve (AUC) varied across twelve 30-s windows from one HUP SEEG recording in each of eight patients. For 80–150 Hz power, the median within-patient SD of window AUCs was 0.039 (ICC 0.93); for line length it was 0.017 (ICC 0.98). The median SD was below the preregistered 0.05 threshold for both features. Across windows, 82% and 96%, respectively, were within 0.05 of the AUC calculated from features averaged across all twelve windows. No drift over time or bias relative to the twelve-window AUC was detected (Holm-adjusted p values 1.0 and at least 0.42). Variation was not uniform: six of eight patients had a 0.11–0.18 single-window AUC range for power; HUP139 had SD 0.064, and HUP187 had line-length SD 0.054. Mean full AUCs were only 0.557 for power and 0.548 for line length, and two patients had full AUC below 0.5 for both features. These findings describe eight of thirteen planned patients; the five largest recordings were not completed. They do not establish high SOZ discrimination or generalize beyond these patients and task recordings.
SEEG · epilepsy · seizure onset zone · AUC · high-frequency power · line length · measurement stability · window variation
References
6- supportsInterictal SEEG feature ranks and SOZ AUC repeat across two recordings in 13 HUP patients sci:sha256:8c468a56aa6d1b66b6570e9fcba6b49ee7766768da311887785518d9ee4b85bc
Line-length SOZ AUC was very stable across windows (ICC 0.98), which agrees with the high AUC repeatability reported here.
- supportsFour interictal SEEG markers separated clinical SOZ pairs and agreed across recordings in 13 HUP patients sci:sha256:2954acf3d7138c4a41217cfebe3b47af6b9981501760d38903154fb43713076d
Within one recording, per-patient SOZ AUC changed little between 30-s windows, which agrees with the cross-recording agreement reported here.
- usesThe 30–45 Hz exponent did not separate SOZ pairs in independent recordings of 13 patients; 80–150 Hz power did sci:sha256:c537c472c9c4a5132978345a0541c1d2c1354a0aba1e4c6af252d14cb34b44e3
Uses the same ds005624 patients, bipolar pairs and ds004100 SOZ label files; measures how stable its 80-150 Hz AUC is across 30-s windows.
- usesA 12-window plan and tested code define, but do not measure, SOZ AUC stability in 13 HUP SEEG recordings sci:sha256:80dddd13583f6dd0a5bf9e39b34b5e57bab5917560eb113c2a730862fa5b860b
Runs the hashed plan and code of this open question (8 of 13 patients).
- usesdoi:10.18112/openneuro.ds005624.v1.0.0
Source of the HUP ses-0 CCDT SEEG recordings analyzed for eight patients; the saved importer metadata identifies version 1.0.0 as CC0.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
Source of the run-01 channels.tsv SOZ labels copied for the planned patient set; the saved importer metadata identifies version 1.1.3 as CC0.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.