In 13 HUP patients, bipolar SEEG markers showed similar SOZ separation in awake task and interictal recordings, with pair ranks agreeing across datasets
We compared bipolar stereo-EEG (SEEG) amplitude markers in two public datasets for the 13 HUP patients present in both: OpenNeuro ds005624 v1.0.0, which contains awake Color Change Detection Task recordings, and ds004100 v1.1.3, which contains clinical interictal recordings and seizure-onset-zone (SOZ) labels. Labels from ds004100 were transferred by channel name. Using the same analysis code on both datasets, delta power (1–4 Hz), broad power (1–45 Hz), and line length separated SOZ from other bipolar pairs in the task data (mean AUC 0.733, 0.724, and 0.681; Holm-adjusted p = 0.0011, 0.0011, and 0.0044). Paired task-minus-interictal AUC differences were −0.017 (95% CI −0.043 to 0.010), −0.003 (−0.028 to 0.023), and +0.008 (−0.013 to 0.028), respectively; none was significant. Pair-value ranks agreed strongly between datasets (mean Spearman rho 0.95–0.97 across features; sign-flip p = 0.00012). Exploratory comparisons found a larger task AUC for beta power (+0.037, Holm-adjusted p = 0.014 across eight bands); gamma did not remain significant after that correction. These findings apply to one center and 13 patients with the same implants and labels. Fixed contact properties may explain much of the rank agreement; recording interval, sleep or wake state of the interictal clips, and task timing are unknown.
stereo-EEG · epilepsy · seizure-onset zone · signal features · intracranial EEG · OpenNeuro · AUC · rank agreement
References
4- In 13 HUP patients, interictal task SEEG recordings separated clinical SOZ pairs, with rankings stable across recordings sci:sha256:e06dfd0cbb1b4c65fa9ff725d4e98f660204040b688b3236dbb844bdd901aae0
[Legacy relation: reproduces] Same 13 patients and label transfer; independent segments and features give the same SOZ separation in ds005624 (mean AUC 0.68-0.73).
- Interictal SEEG power separated SOZ pairs in 13 HUP patients, but cross-session reliability was underpowered sci:sha256:c037a20dfc1e5bde5033cc9dca5400f54a3883a2eb5a8602b22443ad4156965b
[Legacy relation: reproduces] Replicates the 1-45 Hz and 1-4 Hz power SOZ AUC in ds005624 with 60-s segments at 1/3 and 2/3 of ses-0; adds a between-dataset rank test for all 13 patients.
- usesdoi:10.18112/openneuro.ds005624.v1.0.0
Source of the awake Color Change Detection Task SEEG recordings analyzed in this study.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
Source of the interictal SEEG recordings and channel-level SOZ labels analyzed in this study.
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesIn 13 HUP SEEG patients, a recording and state change added no large error to SOZ AUCs, while differences between patients were larger
Per-patient AUCs of ds005624 task and ds004100 interictal recordings with the same pairs (E_rec).