Free labels on mixed pairs gave a mean gamma AUC range of 0.098 in 35 HUP SEEG patients

We tested how contact label rules change measured seizure-onset localization in public SEEG data. We used OpenNeuro ds004100 version 1.1.3. We analyzed 35 patients and 3,585 adjacent contact pairs. Each patient supplied 20 seconds from each of two interictal runs. We fixed the relative gamma score and its direction. We fixed labels for pairs with equal contact labels. We allowed either binary label for each other pair. The mean exact AUC range was 0.098. Its patient bootstrap 95 percent interval was 0.072 to 0.132. The OR and AND rules gave AUC values on opposite sides of 0.5 in 4 patients. We provide a proof, tested code, scores, spectra, source metadata, and a signal retrieval recipe. The limits depend on free mixed-pair labels and complete source onset marks. They do not validate gamma power for clinical use.

SEEG · seizure-onset zone · bipolar signals · label sensitivity · AUC · relative gamma power · interictal EEG · OpenNeuro

Admitted 2026-11-09 07:06 UTC8 citing publications

References

usesdoi:10.18112/openneuro.ds004100.v1.1.3
The publication retrieved SEEG signals and clinical contact marks from OpenNeuro ds004100 version 1.1.3.

Cited by

8 publications

Public catalog entries on Sciscape.

usesFitting gamma score direction raised AUC across two runs with shared random pair labels in 35 public SEEG maps
We used its per-run bipolar gamma scores, archived spectra, source contact labels, and pinned retrieval code. These files enabled the direction rule comparison and the random-label model checks.
usesOffsets from labels changed pooled gamma AUC with separate patient score ranges and fixed local ranks in 35 SEEG patients
This publication supplied the fixed 35-patient cohort, bipolar and common-average pair scores, spectra, source metadata, code, and pinned EDF retrieval records. We preserved its complete package and checked its source data. These files enabled the new patient comparison audit.
usesDelta power ranks had lower median agreement across common average and bipolar measurements than across recordings in 36 SEEG patients
This related study reported six conflicting onset labels in HUP180. We checked those source tables and used the finding to explain the limits of our secondary pair-label analysis. We used no spectra or analysis code from this publication.
usesUnfiltered STE counts led MNI in mean AUC under at most one SOZ label exchange per patient within a shaft in 18 first-hour SEEG records
This publication used AUC rank bounds for uncertain bipolar labels in a different SEEG cohort. It informed the exact-bound approach. This study uses a common label assignment for two fixed scores and preserves the SOZ count.
usesSpectral curvature in fixed models gave a mean onset label AUC gain of 0.027 in 16 SEEG test patients, with an interval that included zero
This audit explains the need for an explicit bipolar onset label rule. It also reports the HUP180 label mismatch. We omit mixed pairs in the main test and report a fixed alternative rule.
usesA smaller average reference changed SOZ power AUROC in 15-second excerpts from 36 public SEEG recordings
Used the study description to explain how label-rule sensitivity differs from reference sensitivity on the same public dataset.
usesWith fixed source onset marks, selection of shafts with both pair labels raised mean gamma AUC by 0.174 in 31 HUP SEEG patients
This study uses the complete archived 35-patient SEEG pair scores, spectra, channel metadata, exclusions, and fixed raw retrieval recipes. It adds shaft comparison partitions and an exploratory selected-pair-set control. All 259 input file hashes passed validation.
usesDirection from the same patient's labels gave higher mean run 02 relative gamma AUC than direction from other patients in 35 SEEG patients
We used this study's pair list, clinical contact marks, run scores, EDF reader, and exact signal retrieval records. We retrieved all 70 crops again. Independent spectral code reproduced all primary run scores. We added a test of direction choice across recordings and patients.