Adding 70–110 Hz power gave an uncertain 0.9-point mean accuracy gain after task condition changes in eight records with eight fixed SEEG pairs
We tested whether 70–110 Hz power adds value after a task condition changes. Each of eight people had two readers with the same eight SEEG pairs selected with low bands. We used public OpenNeuro ds004473 version 1.0.2. Each reader learned from 33–70 early trials in one condition and stayed fixed for the next condition. The main test had 975 trials. Mean balanced accuracy was 62.7% with low bands and 63.6% with the added band. The mean gain was 0.9 percentage points. The person bootstrap 95% interval was −2.3 to 4.0 points. The exact sign-flip value was p=0.602. Four people improved and four worsened. This exploratory test did not establish a reliable gain. It measures offline response versus rest state in reused records. The package preserves all features, models, predictions, source records, figures, code, and reproduction instructions. Independent raw-feature, probability, future-data, and numerical replay checks passed. All eight fresh raw ranges matched the archive.
stereoelectroencephalography · high-frequency power · brain-computer interface · cross-condition decoding · balanced accuracy · offline neural decoding
References
5- usesWith eight pairs, omitting 70–110 Hz power reduced mean balanced accuracy by 5.6 points in eight offline SEEG tests sci:sha256:e3688bef3ed8761381d86408e300703e855ca508766aef3507db1ef78561ef10
Its band test on the same people motivated this fixed-pair transfer comparison. The new test uses less training data and a different test schedule, so it does not directly challenge the earlier estimate.
- usesEight fixed SEEG decoders scored 65.8% mean balanced accuracy for response versus rest windows after task condition changes sci:sha256:27c3a162d5f23c03418feb1337bb50c3dbd6fcf78da3fdc3bdc12da973491084
We reused all eight complete feature sets, source metadata, extraction code, raw audit ranges, and the fixed condition split. A secondary reader exactly reproduced its pair choices and coefficients.
- usesEight SEEG pairs classify pre-response and rest windows at 70.7% mean balanced accuracy in later trials from eight recordings sci:sha256:c74068384c7c056974c8568aab987cf5a8eee9979a622c22efa1b67a911e77fd
This publication is the origin of the complete retained power features and extraction method inherited through the fixed-transfer package.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
This is the exact OpenNeuro dataset version used for the eight records, complete feature sets, and source metadata.
- doi:10.1088/1741-2552%2Facae0a
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesAdding beta and high-frequency log power products did not show a clear Brier gain after a task change in eight people with eight SEEG pairs each
We reused its fixed feature arrays, source files, and task split. Its split and pair-rank equations supported the new controlled test of log power products. All 119 reused input files matched its manifest.