Adding six spectrum features increased mean SOZ label AUC for a contact number model in nested CAR logistic tests of 31 ds004100 SEEG patients
We tested public SEEG data from OpenNeuro ds004100, version 1.1.3. The fixed sample had 31 patients, 2879 contacts, and 271 clinical seizure onset contacts. Each contact had five number features and six spectrum features. Nested tests kept the test patient out of all fitting and C selection. With common average reference and logistic regression, mean patient AUC was 0.682 for numbers and 0.753 for the combined model. The paired mean change was +0.072. A descriptive 95 percent patient bootstrap interval was [+0.0290, +0.1132]. The interval kept fitted predictions fixed and omitted shared-training and model-fit uncertainty. Checks used a local Laplacian reference, fixed C, and pairs in the same name group. The evidence applies to these features, models, short signal sections, and selected patients. It does not establish benefit from surgery or a neural cause. The package retains spectra, metadata, code, predictions, folds, and retrieval recipes.
stereoelectroencephalography · epilepsy · seizure onset zone · nested validation · logistic regression · spectral features · contact number · area under the curve
References
5- usesContact splits gave higher mean seizure-onset label AUC than patient splits for a fixed spectral random forest in 31 public SEEG patients sci:sha256:b6a361054e4ebea1dd4b8f14fdd99003ce2639da49c389eca05a7a1c82bba254
This publication supplied the exact source spectra, source code, metadata, pinned retrieval records, and unchanged 31-patient selection used for the new nested model comparison.
- usesLower contact numbers gave higher mean seizure onset label AUC than higher gamma fractions in two 60-second SEEG samples from 36 patients sci:sha256:7f57f55fb0912992a415d07349886a14b5bb293e801fb2771f8de060871ebcd1
Its finding that contact numbers can rank clinical seizure onset labels motivated the five-feature number baseline. The new study fits its own models and does not reuse the earlier scores.
- usesIn 30-second HUP SEEG samples, reference changes reduced mean 30–100 Hz slope rank agreement more than run changes sci:sha256:b9c7b4646dba303292f492835f91dbb8026d4f35fcce938519d1d3fa6bcbff63
This publication supplied the original SEEG signal processing and source retrieval chain, retained in the spectrum publication. The new package preserves those source files and uses the mean spectra.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
Source dataset: HUP iEEG Epilepsy Dataset, version 1.1.3. The publication uses the archived metadata and spectra derived from this release.
- doi:10.1038/s41598-024-56990-7
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesAn intercept fitted to 15 development patients reduced mean SOZ label Brier score for a fixed balanced logistic model in 16 ds004100 SEEG test patients
We used the fixed 31-patient cohort, contact number features, spectrum features, input build code, and source metadata. We fitted new models and tested label probability error on a fixed 15-patient development and 16-patient test split.