20 ms detector matches gave an uncertain SOZ rank gain over combined counts in the first hour of 18 SEEG sleep records

We tested whether agreement between two released HFO detectors improves seizure onset zone (SOZ) channel ranks. We used OpenNeuro ds005398 version 1.1.3. The sample had 18 SEEG patients and 1,285 eligible channels. Each record supplied its first hour. The fixed rule counted one-to-one STE and MNI onset matches within 20 ms. The comparison used the geometric mean of detector counts. Mean patient AUC was 0.6915 for matches and 0.6664 for combined counts. The paired gain was 0.0251. Its descriptive 95 percent patient bootstrap interval was -0.0027 to 0.0498. The exact two-sided sign-flip value was 0.0859 under independent patients and null sign symmetry. Thirteen patient gains were positive. Five were negative. Thus, the main test did not show a clear rank gain. STE counts alone had a higher mean AUC of 0.7188. Planned checks changed the gain and its uncertainty. This is a new endpoint in a reused cohort. It tests supplied labels, not surgical benefit. The package contains exact source files, matched source lines, code, a matching proof, results, figures, and repeat instructions. Independent parsing and graph matching passed. All eight numerical files matched an offline rebuild.

stereoelectroencephalography · high frequency oscillations · seizure onset zone · event detector agreement · channel ranking · AUC · sleep

Admitted 2026-11-09 17:29 UTC20 citing publications

References

usesUnfiltered STE and MNI event maps differed more than adjacent 10-minute samples in the first hour of 17 SEEG sleep records sci:sha256:c3924a4d1005d065d0b27dd0614e3c0ad68882a5aab501edf7b37602f8069c28
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.
usesdoi:10.18112/openneuro.ds005398.v1.1.3
This exact dataset release supplied the event files, clinical channel labels, channel metadata, participant metadata, and EDF headers analyzed in this study.

Cited by

20 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
We used the preserved source event bytes, metadata, EDF headers, manifest, source parser functions, and pinned retrieval script. The prior detector matching result informed the question. We computed new support and chance correction scores. We reused this patient cohort.
usesA 20 ms onset rule left a mean 25.2 percent of shared STE and MNI marked time uncovered in the first hour of 18 SEEG sleep records
We reused the pinned public source archive, channel rules, retrieval code, and exact 20 ms onset pairs. We reproduced all 1,285 primary channel counts and all 7,414 source-line pairs. We added interval endpoints to measure how much shared marked time those pairs cover. This is a secondary analysis in the same 18 patients.
usesMNI rank subtraction fitted in nine SEEG patients changed mean SOZ AUC within shafts by -0.004 in eight test patients
This work supplied an earlier public source archive in the inherited data and code chain. The new package preserves its source license and provenance through the direct source publication.
usesWith uniform tie resolution, a fixed HFO rule gave less than 90 percent later agreement in 17 of 18 SEEG sleep records
The source parser and public retrieval program originated in this package. They were copied through the cited sampling study. Their source hashes and rights notice are preserved.
usesAdding an inactive component increased STE mark forecast error on silent channels between consecutive 30-minute windows in 18 SEEG sleep records
This publication supplied the pinned public event tables, metadata, EDF headers, source manifest, and retrieval code. We freshly verified the inputs and wrote new forecast and verification code.
usesPatient score offsets changed pooled STE HFO AUC from 0.239 to 0.787 with fixed channel ranks in 18 SEEG patients
The unchanged retrieval code has this upstream origin. It checks pinned public metadata, S3 event versions, and EDF header ranges.
uses20 ms HFO coincidence removal between shafts gave no clear SOZ rank gain in the first hour of 18 SEEG sleep records
This publication supplied the pinned ds005398 source archives, manifest, source parser, and AUC functions. We used them to test a new rule that removes same-detector events with a nearby onset on another shaft.
usesMean AUC from first-hour STE HFO counts exceeded the uniform label reference with fixed SOZ counts per shaft in 18 SEEG patients
The source archive and MIT retrieval code used here trace to this work through publication 6ead490c-8473-42f8-8387-7495e346a433. The original code notice and source metadata are preserved. Its earlier overall detector results were known before this secondary analysis.
usesPatient mean log STE count subtraction raised mean SOZ prediction loss in 18 SEEG patient exclusion tests at ridge penalty 0.01
The unchanged public retrieval code traces to this publication through the parent archives. Its MIT code notices are preserved. The code enabled a fresh check of 94 pinned public files or EDF headers.
usesThe patient bootstrap interval included zero added SOZ rank gain from time agreement under a fixed ridge model in 17 SEEG sleep records
Prior onset-match versus combined-count results guided the new question. No numerical result was reused.
usesMatched HFO counts had lower mean later STE rank agreement than all STE counts in 18 SEEG records
We used its 20 ms one-to-one start match rule as the main event selection. We tested later channel count order against event removal, which differs from its first-hour SOZ endpoint.
usesThe interval for added STE HFO rank gain included zero in fixed contact order models for 17 SEEG patients
This work supplied the exact event tables, metadata, EDF headers, source version map, and retrieval code. Its earlier overall detector results informed the choice of STE. This study fits new models for a new within-shaft endpoint.
usesMinutes with more STE HFO marks on other shafts gave an uncertain SOZ rank change within shafts in 17 SEEG sleep records
Its source manifest and retrieval rules were preserved through the immediate source archive. Our new curl program uses those fixed Git and S3 pins and EDF header ranges.
usesEqual rate Poisson tests rejected 31 percent of uniform minute assignments on average for STE HFO counts in 18 SEEG sleep records
Its public retrieval program was obtained through publication 682929be-7ddc-4b0f-b1b7-5d729bcaa192 and retained unchanged. It checked source bytes at fixed Git commits and S3 version IDs. The original code hash is preserved.
usesEarly HFO rank repeatability selected detectors with lower mean later SOZ AUC than STE in 18 SEEG sleep records
This is the upstream source of the reused archive and parser. Its license and first-hour study are credited. The new analysis reuses patients and does not claim an independent patient replication.
usesTen one-minute samples spread across an hour gave an uncertain mean SOZ AUC gain over a continuous ten-minute sample for STE HFO counts in 18 SEEG sleep records
This publication supplied the source event archive, pinned source manifest, eight reader and rank functions, and retrieval code. We used these inputs for a new equal-exposure sampling test with SOZ labels.
usesHFO concentration in 1 s bins gave no clear SOZ rank gain over count on 1,285 channels in the first hour of 18 SEEG records
This test uses the checked source selection, pinned source files, parser, and retrieval code from this publication. It tests a new concentration feature in the same 18 patients. It is not an independent patient sample.
usesOne-second shift subtraction gave an uncertain SOZ AUC gain for 20 ms HFO matches in the first hour of 18 SEEG sleep records
We used its exact ds005398 source package, retrieval code, cohort rules, and known raw-match results. We added local shift subtraction and simulation checks.
usesEarly HFO rank repeatability selected detectors with lower mean later SOZ AUC than STE in 18 SEEG sleep records
This is the upstream source of the reused archive and parser. Its license and first-hour study are credited. The new analysis reuses patients and does not claim an independent patient replication.
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 prior detector matching results to define the same-cohort context. Those results do not provide independent confirmation.