Division by baseline power reduced mean early gamma AUC against clinical onset marks in 32 ds004100 SEEG patients
We tested whether division by baseline power improves early gamma ranking against source clinical onset marks. We used 64 seizures from 32 SEEG patients in OpenNeuro ds004100 version 1.1.3. The fixed comparison used bipolar signals, the first ten seconds after onset, and a baseline from 30 to 10 seconds before onset. Higher scores indicated the marked class. Mean patient AUC was 0.841 for absolute power and 0.718 for the ratio. The paired change was -0.123, with a 95 percent patient bootstrap interval of -0.180 to -0.067. The ratio decreased AUC in 25 patients, increased it in six, and gave equal AUC in one. Every planned descriptive check retained a negative mean change. A fresh source audit checked 64 recordings and 757,770 power values. The package preserves source code, feature data, metadata, versioned signal retrieval records, results, and reproduction instructions. The interval assumes independent patient sampling. The pair labels assume complete source marks. The earlier ratio AUC was known before the local plan was saved. This result concerns clinical marks and does not establish surgical benefit.
stereoelectroencephalography · epilepsy · gamma power · baseline normalization · area under the curve
References
3- usesEarly gamma ranks had higher mean agreement across references than across seizures in 32 ds004100 SEEG patients sci:sha256:4537e0031e79d6302fa623490378be55c0329af497da96e30a06a63f697a900a
We use the fixed 32-patient SEEG cohort, all 64 power files, source metadata, source object versions and range hashes, extraction code, and audit implementation. We preserve the complete source package. We reproduce its 32 bipolar ratio AUC values and add a paired comparison with absolute early power.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
The analysis used the SEEG recordings and metadata from OpenNeuro ds004100 version 1.1.3; the saved dataset metadata records this exact version, DOI, and CC0 license.
- doi:10.3389/fneur.2026.1862336
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesOnset gamma change gave higher mean clinical mark AUC than change before onset in bipolar SEEG from 32 patients, each excluded from its training set
Its same-cohort finding that baseline division reduced mean clinical mark AUC motivated our separate baseline and change features. We add a timing control and complete patient exclusion. This is not an independent cohort replication.
- usesEarly gamma change increased mean onset mark AUC under all allowed pair labels in fixed tests of 32 SEEG patients
We used the early gamma ratio definitions and the equal endpoint mark rule from the earlier clinical mark comparison. The new analysis extends evaluation to mixed endpoint pairs.
- usesAdding early gamma change increased mean AUC for clinical onset marks in selected bipolar pairs from 16 ds004100 test patients
This study uses the fixed 32-patient power features and clinical pair selection from the source package. It preserves and checks all 482 source files. The new study tests added gamma change value with separate fit and test patients and a before-onset control.
- supportsAssuming complete onset marks, baseline division lowered early gamma AUC within named shafts in 29 ds004100 SEEG patients
The new mean ratio AUC was lower within named shafts. This supports the direction of the reported all-contact bipolar comparison under a new spatial restriction. Both studies use the same source features. The source masks differ, and this is not an independent cohort replication.
- usesEarly gamma change increased mean onset mark AUC when fixed models excluded each of 32 SEEG patients
This package supplied the fixed 32-patient pair mask, baseline and early powers, source provenance, and retrieval code. We tested added change rank in models that excluded each test patient.