In 36 patients with fixed SEEG contrasts, mean change in estimated canonical correlation spectra was larger between files than between adjacent blocks at 1–80 Hz

We tested the repeatability of estimated SEEG coupling strengths. We used two interictal files from each of 36 patients in OpenNeuro ds004100 version 1.1.3. We calculated canonical correlation spectra between electrode paths. Each file supplied two adjacent 30 s blocks. In the 1–80 Hz band, mean spectrum change was 0.046703 between files and 0.037321 within files. The mean patient difference was 0.009382. The patient bootstrap 95% interval was 0.006056 to 0.013552. This interval assumes independent, exchangeable patients. Differences were positive in 35 of 36 patients. The measure stays the same under a fixed invertible change of signal basis within each path. The basis can differ between blocks. All 72 covariance inputs were rebuilt from pinned public data and matched exactly. The package keeps code, inputs, sources, checks, and instructions. This is a secondary analysis of an existing cohort. Sampling effects, noise, and recording state can change the estimates. The result does not establish neural communication or a cause of the file difference.

stereoelectroencephalography · canonical correlation · repeatability · correlation spectra · electrode paths

Admitted 2026-11-09 19:02 UTC8 citing publications

References

usesEdges that share a contact are more common among the strongest 5 percent of 1–80 Hz bipolar SEEG correlations in 36 epilepsy patients sci:sha256:7dc3370736cdefb57323429971522e882b644b700faccfda51b916b3d865f58f
We used its original cohort selection and public retrieval and extraction recipe through the input publication. We kept the source versions, range hashes, contact rules, and filter design. The present node pairs share no source contacts.
usesIn 36 patients, fixed 1–80 Hz SEEG contrasts showed larger covariance shape changes between files than within files sci:sha256:daf70bbdcc88f893fea25fa9870f5621a6fd8289ee4133a9ec263f83268cd815
Its covariance shape result led to the present question about coupling strength spectra. We used the same 36-patient cohort. The new measure is invariant to separate invertible changes of basis within each node in each block. This is a secondary analysis, not an independent patient replication.
usesTwo bases of the same SEEG contrast space gave different 1–80 Hz correlation participation ratios in 36 patients sci:sha256:901bade82e1972a491fb8b27d87ddf5ae157f439c20fb612d6d24768419473f2
We used its 72 covariance archives, saved signal example, metadata, and source code. These inputs enabled the canonical correlation analysis. We checked all input archive hashes and rebuilt all 72 covariance inputs from public raw data.
usesdoi:10.18112/openneuro.ds004100.v1.1.3
Source dataset for the 36-patient analysis; the analyzed release is OpenNeuro ds004100 version 1.1.3, matching the saved version-specific dataset metadata.
doi:10.1093/biomet%2F28.3-4.321
doi:10.1093/braincomms%2Ffcae165

Cited by

8 publications

Public catalog entries on Sciscape.

usesLinks within shafts had higher agreement than links between shafts in 1–80 Hz SEEG trees from 36 patients
This study uses its 72 covariance archives, sample signals, fixed 36-patient cohort, and source records. It adds tree agreement endpoints.
usesInvertible contrast changes within shafts preserved the Gaussian score between shafts in positive definite 1–80 Hz SEEG covariance from 36 patients
Original compact covariance data and fixed 36-patient cohort. This study adds total-correlation endpoints to the same data.
usesFixed 1 to 80 Hz SEEG directions had lower mean correlation between interictal files than within files in 36 patients
We used all 72 source covariance files, the contact selection, metadata, sample signals, and retrieval code. We tested fixed fitted directions instead of sorted canonical spectra.
usesCovariance between SEEG shafts at 1–80 Hz had positive repeat agreement after shaft mean removal in all 36 patients
Its covariance inputs, signal example, and raw reconstruction code were carried through the direct input publication. We retained their source notices and checked their values with fresh public bytes.
usesLocal symmetric whitening slightly lowered transfer correlation in SEEG files from 36 epilepsy patients at 1 to 80 Hz
This work supplied 72 compact covariance inputs, one signal example, 221 metadata files, and pinned public range records. Its study of changes across files informed the calibration question. The present study reuses its patients.
usesConditional Gaussian shaft scores had lower mean file rank agreement than marginal scores in 36 SEEG patients at 1–80 Hz with 10% covariance shrinkage
Used the earlier shaft spectrum comparison to choose a file agreement question in the same cohort.
usesRMS canonical correlation gave higher mean shaft tree retention than maximum bipolar correlation in 36 paired 1–80 Hz SEEG cases
This study uses its full contrast space and canonical correlation definitions. It uses the same cohort and source covariance archive through the cited tree package. It tests tree link retention instead of spectrum change.
usesShaft mean subtraction removed a mean of 79.5% of 1–80 Hz cross-shaft covariance energy after common average reference in 36 SEEG patients
We reused its 72 covariance archives, saved signal example, public metadata, raw rebuild code, and source checks. Its canonical correlation question differs from the covariance energy loss tested here. This is a secondary analysis of the same 36 patients.