Lower contact numbers gave mean AUC 0.790 for public onset marks within inferred shafts in 19 external SEEG patients

We tested a fixed HUP model in public SEEG data from 11 NIH and eight Johns Hopkins patients. Lower contact numbers gave mean patient AUC 0.790 for public onset marks within 53 inferred shafts. A model with contact number and a low frequency power ratio gave AUC 0.784. The mean paired gain was -0.006. The planned sign-flip test gave p = 0.125. The study found no observed gain from this fixed signal model. Public marks, inferred shaft names, and selected long recordings limit the result. The package preserves code, source features, metadata, retrieval locks, results, and independent checks.

stereoelectroencephalography · seizure onset zone · contact number · signal power · area under the curve · external validation · electrode sampling

Admitted 2026-11-09 17:40 UTC2 citing publications

References

supportsLower contact numbers gave mean seizure onset AUC 0.794 within mixed inferred shafts in 34 selected ds004100 SEEG patients sci:sha256:50d777142ef4ebc45c41203ba318cb2244c6b807e8d402b44d8ba5067784feec
The NIH and Johns Hopkins data extend the reported HUP contact number pattern. Mean external number AUC was 0.790 within inferred shafts. This study also uses the earlier benchmark direction.
usesA fixed HUP rule gave 58.2 percent mean balanced accuracy against onset marks in 19 external SEEG patients sci:sha256:e3f4c2691bc3b2c5930552ed90015dcc79a21067be11f7b8ecee7aff9861ea8a
This study uses the preserved HUP and external power arrays, source metadata, EDF fixture, and retrieval code. It adds an external contact number benchmark and a fixed test of added signal information.
usesA low frequency rank threshold gave 61.5 percent balanced accuracy in 36 SEEG patients when each test patient was excluded from fitting sci:sha256:0d58492a9d57146df8f13fb78f03b5e8c943a65d8fddfb91acdb27a52d756a83
This publication supplied the 72 HUP power files used for model fitting. The files and source locks are preserved through the external parent.
usesdoi:10.18112/openneuro.ds004100.v1.1.3
The version-pinned OpenNeuro ds004100 snapshot supplied HUP training features; its metadata identify version 1.1.3 and CC0 terms.
usesdoi:10.18112/openneuro.ds003876.v1.0.2
The version-pinned OpenNeuro ds003876 snapshot supplied the NIH and Johns Hopkins external features and source annotations; its metadata identify version 1.0.2 and CC0 terms.
doi:10.1038/s41598-019-42347-y
doi:10.1093/brain%2Fawac300
doi:10.1038/s41593-021-00901-w

Cited by

2 publications

Public catalog entries on Sciscape.

usesRaw sink scores from a 10 ms ridge model had mean reversal rank correlation 0.841 in 19 filtered SEEG records
We reused the 19-patient SEEG cohort, pinned S3 object versions, prefix hashes, and BIDS-to-EDF contact mappings. These pins enabled new voltage downloads and an audit of forward, reverse, normalized, and symmetric lag-model scores. We independently reproduced the minus-contact-number benchmark. This is a reused cohort, not an independent clinical validation.
usesSigned 10 ms covariance gave a mean patient error reduction of 3.56% in 19 selected SEEG records filtered at 1 to 40 Hz
Uses the selection history of the 19-patient NIH and Johns Hopkins cohort. This history fixes the duration scope and documents the exclusion of short Miami records.