Two 32-trial calibrations within one condition shared 2.5 of eight state pairs and 0.625 key pairs on average in eight SEEG recordings
We tested whether two separate 32-trial calibrations chose the same SEEG input pairs. We used eight people from public OpenNeuro ds004473 version 1.0.2. Both blocks were in the first task condition. A fixed three-band score chose eight pairs for each target and block. The maps shared a mean of 2.5 pairs for pre-response versus rest and 0.625 pairs for the actual pressed key. The mean shared fractions were 31.25 and 7.81 percent. Their paired difference was 23.44 percentage points. Its descriptive 95 percent person interval was 7.81 to 40.63 points. The exact two-sided sign-flip proportion was 0.0625. Key label controls shared about 0.62 pairs. Both maps used the same 64 fitting trials. We then tested 979 later trials. Mean decision agreement was 75.21 percent for state and 59.33 percent for key. These results measure map repeatability in this reused cohort. They do not prove a general difference between targets. The package has three figures, all pair ranks, model states, predictions, code, source records, and replay instructions. Raw checks passed on 26 saved ranges and eight fresh downloads. The study does not test an online interface or stimulation.
stereoelectroencephalography · SEEG · brain-computer interface · calibration · feature selection · map repeatability
References
5- 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 selection-frequency study gave the prior method for testing map stability under trial resampling. We now test separate observed time blocks. We keep the reader fitting trials the same for both maps.
- 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 use its three-band features, 500-sample windows, adjacent bipolar pairs, extraction code, and pinned source hashes. We keep these choices fixed for the map comparison.
- usesEight SEEG key readers gave 53.4 percent mean balanced accuracy after condition changes with 32 calibration trials sci:sha256:b9d997feb12a06acbb687e7dfcfe304892654f63871ce3f774503880b6647dbe
We use its eight SEEG feature arrays, trial records, recovered features, and source range records. We also adapt its loader and fixed reader. These inputs let us compare maps from separate calibration blocks.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The analysis uses OpenNeuro ds004473 version 1.0.2; the saved dataset metadata for this exact version identify this DOI and CC0 terms.
- doi:10.1088/1741-2552%2Facae0a
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesInput selection by effect agreement changed mean later state Brier error by −0.0008 with 64 trials and eight pairs in eight SEEG recordings
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.
- usesKnown simulated contact loss training reduced mean worst-contact Brier error in eight offline SEEG tests with unchanged surviving inputs
We used its two early 32-trial blocks, one skipped trial, and later condition test. We preserved the trial identities and checked the split against the immediate source.
- SEEG state maps shared 2.50 of eight pairs between time blocks and 2.24 between matched random groups on average in eight recordings
[Legacy relation: reproduces] Independent scores reproduced all eight state overlap counts and all eight key overlap counts at the eight-pair limit.
- usesSEEG state maps shared 2.50 of eight pairs between time blocks and 2.24 between matched random groups on average in eight recordings
We used the exact time blocks, features, pair lists, eligibility records, and raw source records. They fixed the matched random comparison.
- usesKnown simulated contact loss training reduced mean worst-contact Brier error in eight offline SEEG tests with unchanged surviving inputs
We used its two early 32-trial blocks, one skipped trial, and later condition test. We preserved the trial identities and checked the split against the immediate source.
- usesA direction rule changed mean key Brier error by -0.00036 for eight SEEG pairs after 64 calibration trials in eight people
We use its exact public SEEG feature arrays, trial rules, metadata, native voltage examples, and source retrieval records. Its calibration map result motivates the new joint pair selection test.