Pooling three background samples per site gave finite rank thresholds at a 5% target in six of eight SEEG patients
We tested background calibration for a brain stimulation response monitor. We used public SEEG data from eight patients and 86 selected sites in OpenNeuro ds004696 version 1.0.1. Each site used five fit pulses, three calibration pulses, and later test pulses. The tested rank rule needs at least 19 calibration scores for a finite threshold at a 5% target. With three scores per site, at least seven sites are needed. Pooling within a patient gave finite thresholds in six of eight patients. Separate-site thresholds were infinite at all 86 sites. Local noise scaling gave equal-patient means of 3.01% before-pulse calls and 56.39% after-pulse calls. The descriptive 95% intervals were 1.13% to 5.06% and 31.07% to 78.98%. The uncertainty did not establish that noise scaling was better than raw scores. Only 33 of 353 test pulses came after all pooled calibration data were available. This is an offline test on reused data. It does not establish a 5% background-call guarantee or a live device benefit. The package includes exact voltage windows, code, all scores and calls, fitted models, three figures, and source recipes. Fresh downloads of all eight source derivatives matched every stored voltage sample and event mapping. Independent arithmetic and a full analysis replay passed.
calibration · rank threshold · electrical stimulation · SEEG · response monitoring · background calls · signal normalization · reproducibility
References
5- usesSeparate thresholds at five checks raised background calls from 29.7% to 53.8% at 86 selected SEEG sites in eight patients with three calibration pulses sci:sha256:4612087085ba175f3259a38ef9b8b4e097a8c0460809d224818d598007eee499
This publication supplied the exact pulse voltage windows, event maps, source pins, and five-fit/three-calibration split. We added patient pooling, score normalization checks, and a session timing audit.
- usesAfter five calibration pulses, a 100 ms endpoint gave mean later AUC of 0.951 for fixed SEEG pairs in eight sessions sci:sha256:bfd74ee6170a5883952f42e0898b29d958be62240dad7866aafd0c75d34aa016
Its earlier endpoint comparison informed the fixed 100 ms primary endpoint. This study reuses the same cohort and supplies no independent patient replication.
- usesFive calibration pulses gave mean later signed template AUC of 0.939 compared with 0.859 for one pulse in eight SEEG sessions sci:sha256:1d2aa103275ad7400befec4c899740ede8f2eb59a09897bdc3f4d4e190e54418
This publication supplied the fixed first-five-pulse bipolar readout choices and candidate selection code. We kept these choices for all 86 sites.
- usesdoi:10.18112/openneuro.ds004696.v1.0.1
The analysis uses processed trial voltage from OpenNeuro ds004696 version 1.0.1; its saved dataset metadata records this version-specific DOI and CC0 license.
- doi:10.1523/JNEUROSCI.2201-22.2023
This is the source experiment paper for the public dataset and provides background on the recorded single-pulse responses.
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesTwo voltage checks accepted fewer correct responses than one check in five offline SEEG sessions at a 5% background calibration target
We adopted its finite rank threshold rule, explicit background sample budget, and global time audit. This prior work uses a different SEEG dataset. Our study does not claim a new rank method or an independent replication of its patient outcomes.