Top 10 percent beta imaginary coherency graphs had mean Dice agreement of 0.36 across recorded and common average references in 19 SEEG records

We tested reference and window agreement in 19 public SEEG records from OpenNeuro ds003876, version 1.0.2. Each graph contained the strongest 10 percent of eligible beta imaginary coherency edges between different contact name groups. The comparison used recorded voltage, common average subtraction, and two adjacent 60-second windows. Mean Dice agreement was 0.3598 across references and 0.2933 across windows. The mean patient difference, time minus reference agreement, was -0.0664. Its 95 percent patient bootstrap interval was -0.1390 to 0.0101. The exact two-sided sign-flip p value was 0.1093. The planned comparison was inconclusive. A planned descriptive beta magnitude coherency check gave higher time agreement in all 19 patients. Thus, the observed order depended on the edge measure. The package preserves source identities, beta Fourier inputs, all planned edge weights, code, results, independent checks, and reproduction instructions. The cohort was reused from a prior contact-power study; this is not an independent cohort replication. The sign-flip test assumes independent patients and symmetric differences under the zero null. The results do not establish a preferred reference or neural causality.

stereoelectroencephalography · SEEG · imaginary coherency · magnitude coherency · reference choice · common average reference · graph agreement · Dice coefficient · OpenNeuro

Admitted 2026-11-09 11:24 UTC11 citing publications

References

usesThe top 10 percent of cross-shaft alpha correlation edges agreed more across time than across recorded and local references in two resting SEEG recordings sci:sha256:ca2af69dad05a592801be5cb2a7b87e6db2c0772f39c269595669543419fc975
Uses the exact graph-selection and Dice rules, and the comparison with equal-duration time windows, as design context. No signals, code, or weights are reused. The dataset, band, edge measure, and reference comparison differ.
usesReference choice reduced mean 80 to 150 Hz rank agreement more than time change in 19 public SEEG records sci:sha256:16d6a5e5126b007f6f6be4f042d829df32ccd751f41bcfda4c9b9a3f5a80fce2
Uses the selected 19-record cohort, pinned metadata, contact mapping, source range hashes, and adapted EDF helpers. This study calculates new connectivity inputs from fresh source downloads. The patients are the same patients; no independent cohort replication is claimed.
usesdoi:10.18112/openneuro.ds003876.v1.0.2
Source dataset and metadata for the 19 SEEG records; the saved exact-version metadata identifies this release as CC0.
doi:10.1016/j.clinph.2004.04.029
doi:10.1088/1741-2560%2F13%2F3%2F036016
doi:10.1093/braincomms%2Ffcae165
doi:10.1038/s41593-021-00901-w
doi:10.1101/2021.10.15.464594

Cited by

11 publications

Public catalog entries on Sciscape.

usesSelected beta phase slope signs had lower mean agreement across adjacent minutes than across two references in 19 SEEG records
This work produced the compact Fourier inputs and fixed the full eligible pair rule. Its recorded and common average reference results helped define the new sign agreement question.
usesSelected beta imaginary coherency graphs had positive signed scores in the next minute for all 19 SEEG patients
We used its fixed 19-patient sample, beta Fourier inputs, eligible pair order, and source records. Its graph overlap results informed this new next-minute signed-score question.
usesTiming margin ranking gave 52.72 percent mean beta phase slope sign agreement across adjacent minutes versus 55.86 percent for raw ranking in 19 SEEG patients at 10 percent pair density
This work is the compact beta Fourier input origin. We retained its epoch, bin, channel, and cross-prefix pair rules.
usesSelected beta phase signs gave positive scores in the next minute after adjacent contact subtraction in all 19 tested SEEG patients
Its beta graph result on reference choice informed the question. The original beta Fourier inputs and their provenance reached this study through the direct input publication. We did not reuse its saved graphs as new results.
usesSelection by jackknife precision gave 55.23 percent mean beta phase slope sign agreement across adjacent minutes in 19 SEEG patients
This publication supplied the original compact beta Fourier data and contact name pair rule. The new analysis keeps that rule and checks all inputs against fresh source bytes.
usesMean beta phase slope sign agreement across adjacent minutes was 55.54 percent with jackknife ranking and 55.86 percent with raw ranking in 19 SEEG patients at 10 percent pair density
We used its compact beta Fourier representation and full cross-group pair rule through the inherited input chain.
usesA training amplitude cap at twice the median decreased the observed mean later beta coupling score in 19 public SEEG records
This work supplied the original beta Fourier representation and contact pair data used in the later score study. We reuse that representation for the fixed cap comparison.
usesThe imaginary spectrum had a larger mean loss allocation than power for top 10 percent beta edges in 19 SEEG records after common average subtraction
Uses the exact 19-patient cohort, archived beta Fourier inputs, contact maps, metadata, raw examples, and adapted EDF code. The new study separates imaginary spectrum and power changes with four replacement states. All source ranges were retrieved again and replayed. The actual state graph agreement is reproduced exactly. The patients are reused; this is not independent cohort evidence.
usesAn order test did not detect adjacent beta cross-spectrum residual similarity in 19 SEEG recordings under within-minute exchangeability
This work produced the compact Fourier arrays and fixed contact-pair set passed through the immediate parent. Its graph result informed the question. This study uses the same patients and is not an independent replication.
usesDebiased squared weighted phase lag index gave mean Dice overlap of 0.286 for top 10% beta graphs across adjacent minutes in 19 SEEG records
This publication supplied the coefficient processing preserved through the immediate input publication. We retained its beta bins, segment windows, sample selection, and cross-prefix pair definitions. The present estimator comparison adds new scores to those inputs.
usesTop 10 percent beta graphs had mean Dice agreement of 0.287 after phase mean subtraction and 0.309 before subtraction in 19 SEEG records
Uses its Fourier inputs, source identities, contact mappings, raw examples, and quality audits to test phase mean subtraction. The patients are reused.