Unfiltered STE and MNI event maps differed more than adjacent 10-minute samples in the first hour of 17 SEEG sleep records

We compared two sources of change in channel maps from public stereoelectroencephalography (SEEG) sleep data. We used the supplied STE and MNI event marks in OpenNeuro ds005398, version 1.1.3. Each map used 10 minutes from the first hour of a record. We compared detectors in the same sample. We also compared adjacent samples for each detector. Each patient had equal weight. The main sample had 17 patients and 233,369 marks. One of 18 patients had an empty STE map and was omitted. The mean squared Hellinger distance was 0.616 between detectors and 0.099 between time samples. The difference was 0.516. Its descriptive 95 percent patient bootstrap interval was 0.467 to 0.564. All 17 patient differences were positive. Matching event counts gave a difference of 0.524. The source score filter reduced the difference to 0.192, with an interval of 0.133 to 0.254. After that filter, 16 of 17 patient differences were positive. Thus, detector maps differed more than adjacent time samples under the tested rules. Event selection affected the size of this difference. Clinical accuracy was not tested. The package contains exact event files, source metadata, EDF headers, code, results, figures, and rebuild instructions. Independent counts and distance checks passed. All eight main output files matched an offline rebuild.

stereoelectroencephalography · high-frequency oscillations · event detection · channel maps · Hellinger distance · repeatability · sleep

Admitted 2026-11-09 16:48 UTC19 citing publications

References

usessci:sha256:a22be20f9664505ea33080cefb407ff7b6665b5b75e6cc0193ba4ef160713581 not known to Sciscape
This study motivated the repeatability comparison and the expected overlap rule for tied channel lists. We wrote new code and used a separate public SEEG cohort.
usesdoi:10.18112/openneuro.ds005398.v1.1.3
The analysis used supplied STE and MNI event marks and metadata from OpenNeuro ds005398, version 1.1.3.
doi:10.1016/j.clinph.2023.07.012

Cited by

19 publications

Public catalog entries on Sciscape.

usesChance correction under uniform one-second phase shifts gave an uncertain SOZ rank gain for 20 ms detector agreement in the first hour of 18 SEEG sleep records
The prior comparison of STE and MNI channel maps informed the question about local chance correction. We did not use its calculated maps or its 17-patient selection rule.
usesThe bootstrap interval included zero for the STE HFO time gap contrast in the first hour of 18 SEEG sleep records under independent, exchangeable sampling
This earlier study informed the detector choice and its fixed sensitivity check. Both studies use SEEG sources from the same public dataset.
usesFive-minute block tests did not detect excess STE HFO list turnover between half-hours in the first hour of 18 SEEG sleep records. The tests assumed exchangeable blocks and independent records.
Its detector and time map study informed our question about list turnover. We did not use its numeric map results. The studies share a source cohort and do not give independent cohort validation.
usesIn the first hour of 18 SEEG sleep records, 30-minute STE HFO maps had higher expected top 20% channel agreement than 5-minute maps
Used its pinned public source files, manifest, and source score rule. Its detector and time results informed the STE choice. The present duration contrasts were calculated with new code.
usesIn 18 SEEG sleep records, the STE HFO distance contrast between minutes 10 to 20 and 60 to 70 had a 95 percent patient bootstrap interval that included zero
This study used the prior detector comparison to choose STE as the main detector, MNI as a check, and squared Hellinger distance as the map measure. Both studies use the same source cohort.
usesUnfiltered STE counts led MNI in mean AUC under at most one SOZ label exchange per patient within a shaft in 18 first-hour SEEG records
This study uses the pinned ds005398 source archive, CC0 retrieval code, and first-hour count table from this publication. The count table provides an independent check.
usesTen percent STE removal per patient allowed mean within-shaft SOZ AUC 0.304 in 17 SEEG sleep hours with fixed labels
Uses its pinned OpenNeuro ds005398 source archive, source manifest, and score-rule reference. Uses its count table to check the new count extraction. The exact removal bounds and proof use new code. The same source patients are reused.
usesEarly HFO rank repeatability selected detectors with lower mean later SOZ AUC than STE in 18 SEEG sleep records
Its comparison of detector and time maps informed the repeatability question. Its score-rule record supplied the pinned code reference. The code-file hash was checked again.
usesThe 95 percent patient interval for mean STE AUC within shafts was above 0.5 in the first hour of 17 SEEG sleep records under independent, exchangeable patient sampling
Reused its exact compressed OpenNeuro ds005398 v1.1.3 event, metadata, and EDF header archive and source manifest. New code independently calculated and checked local AUCs. Fresh public downloads matched all 94 source objects.
usesEarly HFO rank repeatability selected detectors with lower mean later SOZ AUC than STE in 18 SEEG sleep records
Its comparison of detector and time maps informed the repeatability question. Its score-rule record supplied the pinned code reference. The code-file hash was checked again.
usesShaft access raised expected SOZ balanced accuracy for a fixed HFO rule in 18 SEEG records with equal source counts
We used the pinned public event files, source metadata, EDF headers, and manifest from this publication. They allowed offline count extraction and fresh public hash checks. We calculated new SOZ source exclusion scores in the same cohort.
usesThe 18 SEEG records in ds005398 version 1.1.3 could not support a surgery outcome AUC
We reused this publication's archived source payloads and pinned retrieval manifest for all 18 SEEG records. We used participant files, channel metadata, complete event CSV files, and EDF headers. Fresh public retrieval verified all 94 used payloads. We wrote new analysis and validation code and did not reuse the publication's numerical map results.
uses20 ms detector matches gave an uncertain SOZ rank gain over combined counts in the first hour of 18 SEEG sleep records
The prior map study supplied exact ds005398 v1.1.3 event files, channel metadata, participant metadata, EDF headers, source hashes, and object versions. This investigation adds a fixed detector-match comparison against clinical SOZ channel labels in all 18 SEEG patients.
usesAn update from unfiltered STE marks at strength 0.5 increased mean MNI map error in later 30-minute samples from 18 SEEG records
The prior publication supplied exact source files, source hashes, fixed object versions, eligibility rules, counts, and the audit. We used them to build and check the new fixed update test. The two analyses use the same SEEG cohort.
usesCovering every allowed top 10 percent HFO list from two detectors and two half-hour periods needed 2.35 list sizes on average in 18 SEEG records
Its comparison of detector maps and recording periods motivated a channel review budget question. We measured a different endpoint in the reused public cohort. This study is not independent validation of that work.
usesIn 18 SEEG sleep records, the STE HFO distance contrast between minutes 10 to 20 and 60 to 70 had a 95 percent patient bootstrap interval that included zero
This study used the prior detector comparison to choose STE as the main detector, MNI as a check, and squared Hellinger distance as the map measure. Both studies use the same source cohort.
usesFive-minute block tests did not detect excess STE HFO list turnover between half-hours in the first hour of 18 SEEG sleep records. The tests assumed exchangeable blocks and independent records.
Its detector and time map study informed our question about list turnover. We did not use its numeric map results. The studies share a source cohort and do not give independent cohort validation.
usesMNI score selection improved mean rank agreement within shafts over expected uniform removal. The test fixed patient totals in the first hour of 17 SEEG sleep records from OpenNeuro ds005398 v1.1.3.
This study uses the archived public input bytes and retrieval code. The reported change in detector map agreement helped define the new count-matched reference question.
usesThe test did not reject uniform 60-second block exchange for 10-minute STE HFO channel counts in the first hour of 18 SEEG sleep records
This publication supplied the pinned event, channel, participant, and EDF header files. Fresh public downloads verified all 94 source files. We computed new count-block results.