Input selection by effect agreement changed mean later state Brier error by −0.0008 with 64 trials and eight pairs in eight SEEG recordings
We tested whether effect agreement between two early time blocks improved later SEEG state prediction. We used eight people from OpenNeuro ds004473 version 1.0.2. Each reader used the same 64 calibration trials, eight adjacent contact pairs, and three power bands. The reference selected pairs by pooled squared effects. Mean later state Brier error was 0.2011 for pooled selection and 0.2003 for agreement between time blocks. The paired change was -0.0008. Its descriptive 95% person interval was -0.0059 to +0.0036. Three people improved, three worsened, and two had identical predictions. Secondary mean state balanced accuracy rose by 1.43 percentage points. Mean key Brier error rose by 0.0073. Calibration direction agreement increased for both targets. This did not establish lower later state probability error. The test used a reused clinical cohort and windows placed from recorded response times. It does not validate an online BCI or brain stimulation. The package preserves source features, pinned raw range records, all selections and predictions, three figures, code, and replay instructions. Independent checks passed on 26 saved raw ranges, eight fresh ranges, 48 reader fits, and 8,811 probabilities.
stereoelectroencephalography · brain-computer interface · input selection · effect agreement · Brier error · cross-condition prediction · feature selection · calibration
References
6- usesA fixed resampling selector raised mean later Brier error in eight SEEG recordings with 32 calibration trials and eight input pairs sci:sha256:e0c57f775ddc639809aa229c6af37f1fe78c9bc2c3d4776b59f37657aa7757fb
Its result for trial-resampling selection informed the question. The current rule uses signed block effects and 64 calibration trials. The earlier rule used selection frequency and 32 trials.
- usesTwo 32-trial calibrations within one condition shared 2.5 of eight state pairs and 0.625 key pairs on average in eight SEEG recordings sci:sha256:1d26103f4adacb50a72db987153d002b3bf0f1841efdecb871ffd588108516c7
We used its feature files, raw range records, eligibility rule, exact time split, and fixed reader. They let us compare three new input selections on the same 64 trials.
- usesEight SEEG pairs classify pre-response and rest windows at 70.7% mean balanced accuracy in later trials from eight recordings sci:sha256:c74068384c7c056974c8568aab987cf5a8eee9979a622c22efa1b67a911e77fd
We used its band-power features and preserved extraction code through the immediate source package.
- usesEight SEEG key readers gave 53.4 percent mean balanced accuracy after condition changes with 32 calibration trials sci:sha256:b9d997feb12a06acbb687e7dfcfe304892654f63871ce3f774503880b6647dbe
We used its response-key definitions, recovered trials, eligibility rule, and reader code through the immediate source package.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The analysis used OpenNeuro ds004473 version 1.0.2, whose cached exact-version metadata state CC0.
- doi:10.1088/1741-2552%2Facae0a
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesScore integration gave lower mean Brier error than a Gaussian class reader after simulated shaft loss in eight SEEG recordings with 64 calibration trials
Supplied the fixed input features, recovered trials, eligibility loader, 64-trial split, source records, and raw retrieval code.
- usesKnown simulated contact loss training reduced mean worst-contact Brier error in eight offline SEEG tests with unchanged surviving inputs
We reused its verified data and provenance package, pooled pair selection, C = 0.1 baseline, and exact trial split. We checked the baseline predictions before the new contact loss comparison.
- usesKnown simulated contact loss training reduced mean worst-contact Brier error in eight offline SEEG tests with unchanged surviving inputs
We reused its verified data and provenance package, pooled pair selection, C = 0.1 baseline, and exact trial split. We checked the baseline predictions before the new contact loss comparison.