Covariance between named SEEG shafts reduced mean shape loss in 36 of 36 tested ds004100 patients at weight 0.5 in two 1–80 Hz records after a test power scale was fit

We tested whether covariance between named SEEG shafts helps match a separate record after within-shaft covariance and one overall power scale are included. We used two 60 second segments from each of 36 eligible patients in OpenNeuro ds004100 version 1.1.3. At fixed weight 0.5 in 1–80 Hz, mean gain was positive in 36 of 36 patients and 71 of 72 record directions. The equal patient mean was 0.04808 nats per dimension. The 95 percent patient bootstrap interval was 0.03829 to 0.05885, with independent and exchangeable patients assumed. We report all 11 fixed sensitivity checks. Fresh public source bytes rebuilt all 72 covariance files exactly. The package preserves inputs, source versions and hashes, code, all results, a score proof, and repeat instructions. The score fits one scale to test data for each model. The result measures repeated signal shape. It does not establish causal neural connections.

stereoelectroencephalography · intracranial EEG · covariance · covariance shape · epilepsy · signal structure · OpenNeuro ds004100

Admitted 2026-11-09 18:43 UTC3 citing publications

References

usesOne patient reversed the mean gain of a 0.5 covariance mixture over an independent contact model in paired 1–80 Hz SEEG records from 36 patients sci:sha256:73e7b108e6d3e3788a07ac04de661b4eedb0dfdab51624d480fc0d9bce16a2b3
The unscaled covariance loss result in this cohort motivated a test with one test power scale per model and a within-shaft comparison. We changed the comparison and score. The current result tests a different quantity and does not reproduce or overturn the earlier numerical result.
usesTwo bases of the same SEEG contrast space gave different 1–80 Hz correlation participation ratios in 36 patients sci:sha256:901bade82e1972a491fb8b27d87ddf5ae157f439c20fb612d6d24768419473f2
We used the 72 verified covariance inputs, pinned public metadata, byte-range retrieval records, and path construction code to test covariance shape beyond a named-shaft model. We preserved the required source files and license notices, and rebuilt all 72 records from fresh public source bytes.
usesdoi:10.18112/openneuro.ds004100.v1.1.3
The analyzed SEEG records came from OpenNeuro ds004100 version 1.1.3; the publication records source object versions and hashes.

Cited by

3 publications

Public catalog entries on Sciscape.

usesTraining selection of covariance between shafts lowered median loss but raised mean loss in 36 ds004100 SEEG patients at 1–80 Hz with 1% ridge
Its test of covariance between named shafts motivated this extension. It fit a scale to test power. The present score retains training power and selects weights from training blocks. The baseline, filter intervals, and ridge also differ.
usesFull covariance had lower shape loss than a model with one fitted shared factor in 36 ds004100 SEEG patient means at 1–80 Hz and weight 0.5 after test scale fitting
We use its 72 covariance derivatives, shaft definitions, pinned source ranges, and preprocessing code. They define the patient cohort and permit the new shared factor comparison. We also reproduce its full model gain over shaft blocks in all 504 tested directions.
usesWith independent, exchangeable patient sampling, the 95 percent interval for the 0.1 shrinkage effect on 1–80 Hz SEEG link rank agreement included zero in 36 patients
We used all 72 covariance inputs, the signal example, metadata, and pinned source byte recipe from this publication. We calculated new between-shaft partial-correlation rank scores. A fresh source rebuild checked every imported covariance input.