64 matched calibration trials increased mean later SEEG balanced accuracy over 16 with a fixed penalty in eight offline recordings
We tested 16, 32, and 64 calibration trials in eight public SEEG recordings from OpenNeuro ds004473 version 1.0.2. Each model used eight bipolar pairs. We used nested samples with the same task and trial-group proportions. We fixed the penalty on mean model loss. Mean later balanced accuracy was 60.8%, 63.9%, and 67.8%. The 64-minus-16 gain was 7.0 percentage points. The descriptive 95% patient interval was 5.2 to 8.9 points. All eight patient means increased. The exact two-sided sign-flip p value was 0.0078125 under its stated assumptions. At 64 trials, 30.2% of rest windows were still called response. Calibration samples covered a long early interval. The result applies to offline event-aligned task states and inherited quality rules. It does not establish live intention detection. The package preserves code, features, source records, model states, scores, four figures, and reproduction commands. A second run reproduced all 304 model arrays. Raw signal and fresh public-range checks also passed.
stereoelectroencephalography · calibration size · balanced accuracy · brain-computer interface · offline decoding · OpenNeuro
References
4- usesRecent calibration did not show a clear accuracy gain in eight SEEG recordings with 32 matched calibration trials sci:sha256:6744fecf8794e7441df0d58b4b4ef770b691923a6569df596f9bbfea141dc374
This study supplied design context on recent calibration and matched task proportions. We used that context to define a separate calibration-size question.
- 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 study supplied the verified SEEG features, event windows, later test, retrieval code, and source records. We reused these files.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
This is the exact OpenNeuro dataset version used for the eight recordings.
- doi:10.1088/1741-2552%2Facae0a
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesPair selection and fitting each improved mean offline SEEG accuracy with eight pairs and 64 instead of 16 trials in eight recordings
This work supplied the public trial features, matched calibration samples, fixed penalty, and later test. Its preserved model states let us check both diagonal settings exactly.
- usesAdded fitting trials reduced Brier error more than added selection trials in eight offline SEEG tests with eight pairs and a fixed penalty
This work supplied the unchanged power features, nested samples, fixed penalty, and later test. We reproduced its diagonal model scores.
- usesCovariance terms changed mean offline SEEG balanced accuracy by −0.09 points with 16 calibration trials in eight recordings
We reused its public SEEG features, source metadata, exclusion rules, early pools, later test blocks, sample rule, and fixed logistic reader. All 2,400 task sample choices and logistic score arrays matched the parent. The new test changes the direct use of covariance terms at fixed calibration budgets.