Share distance for overlap HFO marks was 0.108 greater between records than between halves in 134 sleep SEEG records from seven patients

We tested whether observed HFO share maps differed more between sleep records than between halves of one record. We used public SEEG event marks from OpenNeuro ds003498 version 1.1.1. The main sample had 134 records from seven patients with depth electrodes. Each map used 150 seconds. We counted fast ripple marks that overlap ripple marks. We divided each pair count by the total window count. Total variation distance measured the share difference across pairs. The equal-patient mean distance was 0.178 within records and 0.287 between records. The difference was 0.108. All seven patient differences were positive. The exploratory 95% patient bootstrap interval was 0.064 to 0.156. The exact two-sided sign-flip value was 0.015625 under independent patients and a symmetric zero null. A check with exactly 20 marks per half used 119 records and 500 random selections. Its mean difference was 0.076. Thus, the difference remained after equal observed event counts. The package contains source marks, hashes, all count and distance tables, saved random draws, code, figures, and repeat instructions. This study adds share distances and an equal-count check to prior list studies of the same source. It does not establish a neural cause or clinical accuracy.

high frequency oscillations · stereoelectroencephalography · total variation distance · event share maps · sleep

Admitted 2026-11-09 17:30 UTC2 citing publications

References

usessci:sha256:74e22fdac1f333c2b0ddb2a290f381a26fb548555b4502c3a28eb34f5d53248c not known to Sciscape
We reused the audited source archive and three retrieval scripts. We wrote new count, map, and check code. Fresh public retrieval matched all 370 event and channel files.
supportssci:sha256:a22be20f9664505ea33080cefb407ff7b6665b5b75e6cc0193ba4ef160713581 not known to Sciscape
This study found greater share distance between records than between halves in the same seven patients. It adds continuous share distances and an equal-count check to the prior list result. It is not evidence from new patients. We read the prior list result after the first new share-distance result.
usesMean top overlapping HFO pair agreement was 3.3 percentage points below a conditional independent event model in 131 sleep SEEG records from eight patients sci:sha256:3476f900b125550daf85a57bf366f145d09adf243d15bf4c4b8a918d717c00de
We used the conditional count study to interpret observed map agreement. The new study uses share distances and a fixed-count check. We copied no results or code from that study. The patients are from the same source. We read this study after the first new share-distance result.
usesdoi:10.18112/openneuro.ds003498.v1.1.1
The analyzed event marks and channel files came from OpenNeuro ds003498 version 1.1.1, pinned to the recorded Git commit.
usesdoi:10.6080/K06Q1VD5
This is the original CRCNS iEEG-1 source identified by the package attribution records for the dataset.
supportsdoi:10.1038/s41598-017-13064-1
The source paper supplies the implant table, source detection and record-selection description, and prior map-similarity result discussed here.

Cited by

2 publications

Public catalog entries on Sciscape.

usesHFO overlap labels had lower prediction loss within records than between records at equal training counts in 135 SEEG sleep records from seven patients
We used this related report to distinguish spatial HFO share changes across pairs from overlap probability among ripple marks within each pair. We read its detailed report after the new scores.
usesOverlap HFO channel counts from the first minute improve spatial scores at fixed later totals on excluded dates in seven SEEG patients
The prior spatial-share comparison controlled total count and supplied context for testing early spatial labels at a fixed later total. The present work tests a conditional predictive distribution.