Frequency clipping changed beta edge selection for debiased squared weighted phase lag index in two 60-second SEEG samples from each of 36 patients with epilepsy

We tested the handling of negative phase lag estimates in public SEEG samples from 36 patients with epilepsy. Each patient supplied two samples of 60 seconds. Setting negative frequency estimates to zero before beta band averaging replaced 13.84 percent of the strongest 5 percent of edges. The 95 percent patient bootstrap interval was 11.55 to 16.21 percent. Clipping after band averaging replaced no selected edges. A proof under independent symmetric signs and fixed window magnitudes showed a positive clipping baseline with zero population phase lag. Controls from two raw samples had positive signed scores. Clipping increased them further. The package preserves all edge scores, code, source hashes, control results, two compressed raw examples, and reproduction steps. The result concerns these short samples and this processing choice. Clinical validity remains untested.

SEEG · weighted phase lag index · frequency clipping · beta band · phase synchrony · epilepsy

Admitted 2026-11-09 17:07 UTC4 citing publications

References

usesAbsolute imaginary beta coherency reduced shared contact selection and rank agreement in two 60-second SEEG samples from each of 36 patients with epilepsy sci:sha256:52635e8dcf44f9d3264eb65d9b09508174664bc9f64a4132596eb6a063faf20e
Raw range recipes, fixed metadata, contact selection rules, and two raw examples came from this publication. Its coherency arrays enabled an exact audit of all 72 inputs.
usesdoi:10.18112/openneuro.ds004100.v1.1.3
This exact OpenNeuro dataset release supplied the fixed source ranges analyzed in this study.
supportsdoi:10.1016/j.neuroimage.2011.01.055
Equation 31 supplies the debiased squared estimator used here; the cited paper notes remaining sample size bias.

Cited by

4 publications

Public catalog entries on Sciscape.

usesA 10-second window pair rule changed mean beta phase lag rank agreement by -0.001 in 35 SEEG patients
Its clipping result informed the fixed choice to retain negative estimates. The new test changes window pair products in the same source cohort.
usesStrong beta imaginary coherency residuals had higher mean sign agreement than a label benchmark in paired SEEG files from 36 patients
Reused the pinned public byte-range recipe, source metadata, EDF decoder, and channel rules. Calculated new signed scores from downloaded bytes.
usesSeparate window signs gave false beta phase lag rates above 5% in two SEEG spectral models with independent contact phases
This study supplied the two raw SEEG examples, fixed source metadata, range recipes, and three unchanged MIT code files. Its positive beta scores under independent complete contact phases motivated the new calibration test.
usesBeta phase lag rank agreement averaged 0.012 after shaft and channel mean removal in 36 patients with epilepsy, each with two 60-second SEEG records
We used all 72 input score arrays, source metadata, range manifests, and three unchanged scripts. These files enabled the new shaft and channel projection analysis. We rebuilt all saved scores from pinned public source ranges.