Shared test epochs increased top 10 percent contact agreement for 80 to 150 Hz common average SEEG power in 19 interictal records under random epoch assignment
We tested whether shared source time changes measured SEEG contact set agreement. We used 19 public interictal records and 2,070 contacts from OpenNeuro ds003876 version 1.0.2. Random assignments compared samples of eight four-second epochs with targets of 15 epochs. The paired targets included all test epochs or no test epochs. Target size stayed fixed. Sets contained the highest-power 10 percent of contacts, with the count rounded up. For 80 to 150 Hz common average power, mean agreement was 92.21 percent with shared epochs and 87.65 percent with no shared epochs. The mean increase was 4.57 percentage points. The 95 percent patient bootstrap interval was 3.26 to 5.96 percentage points. The package provides source arrays, pinned source access, code, all main scores, checks, and a source time overlap audit. The result applies to random epoch groups in short records. It does not establish clinical accuracy or stability across sessions.
stereoelectroencephalography · SEEG · contact selection · data reuse · interictal recordings · high-frequency power · agreement · reproducibility
References
7- uses32-second samples gave higher contact set overlap than 4-second samples against a separate minute for 80 to 150 Hz SEEG power with a common average reference in 19 interictal records sci:sha256:fd35cd7f0d24604878cc790a289e06963b5ef8fe25c81194d3748700250d0327
We used its 19 power arrays, source selection, and retrieval check for the equal-size target experiment.
- usesReference choice reduced mean 80 to 150 Hz rank agreement more than time change in 19 public SEEG records sci:sha256:16d6a5e5126b007f6f6be4f042d829df32ccd751f41bcfda4c9b9a3f5a80fce2
We used its pinned source processor and reference definitions through the duration study. Fresh source checks reproduced all 19 arrays.
- usesdoi:10.18112/openneuro.ds003876.v1.0.2
Source data: the publication used OpenNeuro ds003876 version 1.0.2, as recorded in the dataset metadata.
- doi:10.1093/brain%2Fawac300
- doi:10.1038/nn.2303
- doi:10.1038/s41593-021-00901-w
- doi:10.1101/2021.10.15.464594
Cited by
2 publicationsPublic catalog entries on Sciscape.
- usesThe patient test detected no group versus random removal difference in 19 SEEG records under an independent symmetric null. Common average sets used 120 seconds of 80 to 150 Hz power.
This package supplied the fixed 19-record cohort, source metadata, EDF headers, and version-pinned signal ranges. We downloaded the signals again to compute new band matrices and test group removal against count-matched random removal.
- usesRandom disjoint 30-second SEEG power samples at 80 to 150 Hz, with 118 to 122 Hz removed, gave greater rank agreement within files than across files in 34 patients
This study supplied context for shared test epochs and top-contact agreement in a separate dataset. Our within-file maps have no shared selected window. The current file difference can occur without that direct window reuse.