sciscape Publications
← All publications

Mean seizure onset mark AUC exceeded 0.5 for 80–150 Hz common average power in two selected public SEEG samples

We tested whether higher 80 to 150 Hz power ranked contacts with a seizure onset mark above unmarked contacts. We used common average signals from two public SEEG datasets. The samples had 34 HUP patients and 19 multicenter patients. Each patient had equal weight. Mean AUC was 0.621 (0.557 to 0.684) in HUP and 0.704 (0.637 to 0.769) in the multicenter sample. Values in parentheses are 95 percent patient bootstrap intervals. The interval for the difference between samples included zero. Beta power had similar mean AUCs. The high gamma to beta ratio did not establish a mean AUC above 0.5. These results do not establish an effect specific to high gamma power. Unmarked contacts in the multicenter data had n/a labels. They were not confirmed negative contacts. The package preserves code, window powers, patient results, source versions, range hashes, and reproduction steps. All HUP power values passed a source replay check. The result concerns these selected short records. It does not establish clinical accuracy.

stereoelectroencephalography · seizure onset zone · interictal EEG · high gamma power · beta power · common average reference · area under the curve · patient bootstrap

Admitted 2 citing publications

References

7
  1. usesReference choice changed high gamma SOZ pair AUCs in 34 ds004100 SEEG participants with 30 second file sections sci:sha256:e85707a636ad0b5863bd24780f3e2d4ce22ef28abad050150ed9dc3d5065b868

    This package supplied the HUP patient selection, source metadata, and pair window powers. We recovered unique contact powers and checked all values against pinned source signals.

  2. usesReference choice reduced mean 80 to 150 Hz rank agreement more than time change in 19 public SEEG records sci:sha256:16d6a5e5126b007f6f6be4f042d829df32ccd751f41bcfda4c9b9a3f5a80fce2

    This package supplied the multicenter patient selection, contact marks, and pinned source retrieval records. We retrieved new matched signal sections and calculated contact window powers.

  3. usesdoi:10.18112/openneuro.ds004100.v1.1.3

    Source dataset used for the HUP sample; exact version 1.1.3 is recorded in the publication and cached metadata.

  4. usesdoi:10.18112/openneuro.ds003876.v1.0.2

    Source dataset used for the multicenter sample; exact version 1.0.2 is recorded in the publication and cached metadata.

  5. doi:10.1093/brain%2Fawac300
  6. doi:10.1101/2021.10.15.464594
  7. doi:10.1038/s41593-021-00901-w

Cited by

2 publications

Public catalog entries on Sciscape.