Local 50% bipolar SEEG spectral pooling did not establish lower mean error in 36 patients
We tested whether local pooling improves the prediction of SEEG spectral shape in another file. We used public OpenNeuro ds004100 version 1.1.3 and exact spectra from a cited AutoSci package. The fixed primary test included 36 patients and 3,570 bipolar targets. Each prediction used equal weights for the target and nearby pairs with no shared target contact. The score used 30 to 80 Hz and removed 55 to 65 Hz. Mean patient error changed from 0.011261 to 0.011274 squared log10 units. The relative error reduction was -0.12 percent. Its 95 percent patient bootstrap interval was -6.74 to 5.54 percent, assuming independent and exchangeable patients. These data do not establish benefit or equal performance. A planned secondary test with a 25 percent neighbor weight reduced mean error by 10.23 percent. Its descriptive interval was 8.15 to 12.24 percent. This weight needs a new independent test. An exact error identity shows why the effect can change with weight. We rebuilt all 72 source excerpts and checked all scores with separate code. The package preserves inputs, source pins, code, scores, plots, and repeat instructions. This study measures prediction error. It does not test clinical benefit or a biological source.
SEEG · intracranial EEG · epilepsy · spectral shape · spectral pooling · bipolar derivation · prediction error · spatial pooling · patient bootstrap · OpenNeuro
References
4- usesMedian Welch spectra gave smaller mean file differences for 30–80 Hz exponents in paired 60 s bipolar SEEG records from 36 patients, with 55–65 Hz removed sci:sha256:82d34593ca70a122d42ce91c4739ff2a3d3cc696d375a051587e30a33d29ab60
This work produced the mean spectra preserved in the cited input package. We reused those spectra through publication 36d4a435-d0ce-4de6-955c-53c2a1abbf14. We retained its source choices and checked the spectra against fresh raw signals.
- usesQuadratic log-frequency fits gave lower spectral shape error than straight fits in separate bipolar SEEG files from 36 patients at 20–100 Hz with 55–65 Hz removed sci:sha256:01b8361e8c32258bab97b2e381b2bc4dcc6bb57ea3efc0c8057ad65785c14e7c
We used all 36 compact mean spectrum files as exact input copies. We also used the saved source metadata and raw spectrum reconstruction code. Our new test concerns spatial pooling, not the source model comparison.
- usesMean bipolar reference effects exceeded file differences for 30–80 Hz slopes in paired 60 s SEEG records from 36 patients sci:sha256:a343645ba027d704e981c68fea0352745ba1410c491d9bbbcc42ef099ff50447
We reused the contact conventions and MIT-licensed raw retrieval and EDF methods preserved through the cited input package. The publication package preserves the imported code and license.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
This is the exact OpenNeuro ds004100 version used for the SEEG source metadata and the 72 first-minute signal excerpts. The saved version-specific metadata identify the dataset license as CC0.
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesTwo nearby donors reduced withheld SEEG slope error by 25.1% across files in 20 ds004100 patients. Fits used 20–100 Hz and omitted 55–65 Hz.
This earlier ds004100 study supplied the comparison task with local pooling of an observed target spectrum. The current task instead withholds the source target slope from both main predictors.