Under independent patient resampling, the 95 percent interval for mean HFO count prediction gain from a 60 second update included zero in seven SEEG patients
We tested whether the first 60 seconds improve HFO count predictions for the next 240 seconds in public SEEG sleep data. The test used 135 records from seven patients in OpenNeuro ds003498 version 1.1.1. Each target record was excluded from the source rate fit. A total count update kept spatial proportions fixed. Its equal-patient mean loss gain was 0.344 natural log units per pair per record. The 95 percent patient bootstrap interval was -0.015 to 0.844 under independent patient resampling. It included zero. Six patients had lower loss and one had higher loss. A pair update used the same early counts and the same predicted total. Its additional mean gain was 0.198, with an interval of -0.012 to 0.424. An exact score decomposition separated total count gain from spatial gain. The main mean gain remains uncertain. The package preserves source annotations, hashes, code, all fixed settings, fits, predictions, bootstrap draws, and successful reproduction checks. This is an exploratory test of reused detector marks, not a chronological forecast or a clinical outcome test.
high frequency oscillations · stereoelectroencephalography · sleep · count prediction · Poisson mean loss · patient bootstrap
References
6- usessci:sha256:74e22fdac1f333c2b0ddb2a290f381a26fb548555b4502c3a28eb34f5d53248c not known to Sciscape
Used the checksum-verified source annotations, metadata, headers, source documents, and retrieval helpers. New code computed the early-to-later mean predictions and their scores.
- usesAn excess variance rule lowered prediction loss for HFO counts in 135 excluded SEEG sleep records from seven patients under record exchangeability sci:sha256:9febe1f024cef70f3934a38192744e15a5127f925e8083529c6df7cbd708b9f1
Used the prior whole-record distribution prediction test to define a different question: whether early counts improve later mean predictions within an excluded record. The new test separates total count loss from spatial allocation loss.
- usessci:sha256:a22be20f9664505ea33080cefb407ff7b6665b5b75e6cc0193ba4ef160713581 not known to Sciscape
Used the documented rules for structural mesial bipolar pairs and short-label aliases. These rules kept pair selection independent of event counts.
- usesdoi:10.18112/openneuro.ds003498.v1.1.1
Source annotations and metadata were taken from the pinned OpenNeuro ds003498 version 1.1.1 release.
- usesdoi:10.6080/K06Q1VD5
This CRCNS iEEG-1 dataset is the original source of the recordings and HFO annotations.
- supportsdoi:10.1038/s41598-017-13064-1
This paper documents the source recordings, detector, and implant context used to establish the SEEG scope.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- usesOverlap HFO channel counts from the first minute improve spatial scores at fixed later totals on excluded dates in seven SEEG patients
The prior first-minute mean update supplied the test context and the fixed half-weight benchmark formula. The present study excludes the full source date and fits a Dirichlet-multinomial concentration from other dates. It uses the same cohort and is an extension.
- usesCorrected conformal calibration by scan date returned all selected SEEG pairs at 90 percent in seven ds003498 patients
The checked source archive, resolved headers, count parser, and contact aliases supplied the data and count rules. We did not reuse its prediction scores.