Separate class correlation estimates increased later Brier error in eight SEEG recordings with fixed Gaussian readers
We tested whether separate class correlation estimates lower later probability error in public SEEG data. Eight people supplied 2014 retained trials from one session each. Two fixed Gaussian readers used identical means, variances, and 24 power inputs. At 50 percent retained correlation, separate estimates increased mean later Brier error from 0.1948 to 0.2104. The paired increase was 0.0156, with a 95 percent descriptive person-bootstrap interval of 0.0124 to 0.0184. Error increased in all eight people. The gamma-only interval included zero. The package preserves code, features, models, predictions, three figures, source records, and reproduction commands. Fresh source data reproduced 32 trials exactly. This is an offline task-state test from a small clinical cohort.
stereoelectroencephalography · brain-computer interfaces · Gaussian classification · correlation estimation · Brier error · probability estimation
References
3- 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 all 2014 retained trial features, trial records, exclusions, and source provenance from this publication. We kept its window and split rules. We tested separate and shared class correlation estimates while holding class means and variances fixed.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The analysis used all eight recordings from OpenNeuro ds004473 version 1.0.2. The source records preserve pinned ranges and hashes; four trials per person were freshly rebuilt from pinned source ranges.
- doi:10.1088/1741-2552%2Facae0a
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.