sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

Added fitting trials reduced Brier error more than added selection trials in eight offline SEEG tests with eight pairs and a fixed penalty

We tested two uses of added calibration trials for a SEEG reader. We used public OpenNeuro ds004473 version 1.0.2. The test used eight people, eight selected pairs, and fixed later trials. Four models crossed 16 or 64 trials for selection with 16 or 64 trials for fitting. Both steps used the same nested samples. The two mixed models each had a 64-trial union. In a planned secondary test, more fitting trials gave lower Brier probability error than more selection trials. The mean advantage was 0.0532, with a descriptive 95% person interval from 0.0326 to 0.0731. All eight people had that advantage. The advantage remained at both fixed penalty checks. Selection and fitting contributed 3.51 and 3.50 percentage points to the mean accuracy gain. The primary mixed-model difference was 0.015 points, with a descriptive 95% interval from -0.829 to 0.956 points. This result reproduces a related sci accuracy report. It does not establish equal effects. A rest control also showed a probability benefit from fitting. Thus, this benefit does not show a task-specific neural effect. The package preserves features, source range records, code, model states, scores, three figures, and reproduction instructions. These event-aligned results do not validate an online reader or brain stimulation.

stereoelectroencephalography · calibration · Brier score · balanced accuracy · logistic regression · brain-computer interfaces

Admitted 0 citing publications

References

6
  1. uses64 matched calibration trials increased mean later SEEG balanced accuracy over 16 with a fixed penalty in eight offline recordings sci:sha256:1afe3af4a53f584020860f5028d9cc451d3b18ab565a7a1b0ae0c23074d9dd0c

    This work supplied the unchanged power features, nested samples, fixed penalty, and later test. We reproduced its diagonal model scores.

  2. usesAll available SEEG pairs had an uncertain mean task-state accuracy gain over eight pairs after 32 calibration trials with L2 C=1 in eight records sci:sha256:2dfb88b8b49b59c7ed25acb61dd560d18def2a21039909a1de4bbe2388d33b0c

    This report informed the scope of training-only pair selection. We separate the eight selected pairs from the full candidate set needed for calibration.

  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

    This work supplied the feature extraction and pinned raw retrieval method through the parent package. We checked scalar raw powers, fresh raw ranges, and four fresh source trials.

  4. Pair selection and fitting each improved mean offline SEEG accuracy with eight pairs and 64 instead of 16 trials in eight recordings sci:sha256:6dcbaf5f03b104b33fcff76a2e5783b8684a52d1fad2c86e5e410f74cda45beb

    [Legacy relation: reproduces] This report already tested the same four main models. We reproduced its samples, pairs, fitted states, and 3,200 main score sets. The same people and trials were used. We added planned Brier and log-loss tests and two fixed penalty checks.

  5. usesdoi:10.18112/openneuro.ds004473.v1.0.2

    This versioned OpenNeuro dataset supplied the eight SEEG recordings and task events used in the analysis.

  6. doi:10.1088/1741-2552%2Facae0a

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.