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

An eight-score median reference kept a mean of 47.4% of later gate outputs fixed under any two reference score changes in eight SEEG readers

We tested a median rest reference in eight fixed SEEG readers from OpenNeuro ds004473 version 1.0.2. Each reader used eight reference scores and a fixed early confidence gate. We proved a tight median interval when any two reference scores change by any real amount. With later scores fixed, this interval kept a mean of 47.36% of later gate outputs fixed. The descriptive 95% person interval was 39.82% to 54.17%. Fixed outputs included wrong labels. In the main measured test, two rest scores were replaced with paired scores from windows before a response. All 28 position pairs were tested per person. The final test had 911 trials. Wrong accepted labels were 18.58% of all windows with the median and 18.66% with the mean. The paired difference was -0.080 percentage points, with a descriptive 95% person interval of -1.472 to +1.707 points. This test did not show a clear error benefit. Constant gain cancellation passed 6,090 checks. Fresh native ranges from 72 trials matched pinned versions and range hashes. Independent code and a full replay passed. The package includes the certificate tool, proof, figures, features, scores, source records, and repeat commands. The certificate concerns output stability under fixed later scores and gates. It does not establish label accuracy, live command control, or stimulation performance.

stereoelectroencephalography · brain-computer interfaces · confidence gating · median reference · robustness · signal calibration · output stability

Admitted 1 citing publication

References

5
  1. usesA simulated 6 dB shaft voltage loss raised mean accepted label error by 10.5 points in eight SEEG readers with fixed early confidence gates sci:sha256:c11916a581fda4693fef923ed62c5f6bed20b0fc31fcea1f0823dc4472692661

    This publication supplied the eight frozen primary readers, early median confidence gates, feature arrays, trial splits, predictions, source metadata, and native range records. We verified the source package hashes and reused these fixed inputs.

  2. usesEight fixed SEEG decoders scored 65.8% mean balanced accuracy for response versus rest windows after task condition changes sci:sha256:27c3a162d5f23c03418feb1337bb50c3dbd6fcf78da3fdc3bdc12da973491084

    This upstream work supplied the fixed early-condition reader construction and source feature extractor used by the direct parent. We preserve its model values and extractor without fitting or selecting a reader again.

  3. usesA correction from eight known rest windows kept 69.7% mean accuracy under simulated constant bipolar gains of 0.5 to 2 in eight SEEG recordings sci:sha256:547af54a5b0969526352c29c2e536bb43066e3531d395d0e0b67318818002e77

    This work established the known-rest mean score correction and its constant gain identity. We extend the location rule to the median and test two recorded task replacements and an arbitrary two-score bound. Its different model splits prevent a matched accuracy comparison.

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

    The fixed readers, derived features, and preserved native source ranges are based on OpenNeuro ds004473 version 1.0.2.

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

Cited by

1 publication

Public catalog entries on Sciscape.