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

SEEG channel patterns did not reduce mean error prediction loss in forward tests with confidence and behavior for 18 patients

We tested whether processed SEEG high-gamma channel patterns add error information beyond reported confidence and behavior. We used MetaRDK, OpenNeuro ds006253 release 1.0.3. Separate models for 18 patients used only earlier trials for fitting and model selection. The primary test used channel means from the 0.4 seconds after first mouse movement. There were 1,860 later test trials and 507 errors. Mean log loss was 0.58382 for behavior and 0.58676 for the procedure that could add SEEG patterns. Mean loss gain was -0.00293 nats per trial. Its 95 percent patient bootstrap interval was [-0.01103, 0.00381]. The test did not show a clear gain. It does not show an absence of neural error information. The result applies to offline analysis within these patients. Source processing prevents an online or precise timing claim. The package preserves code, compact features, source versions, retrieval recipes, splits, fitted models, predictions, figures, and checks. A complete second model run reproduced the results exactly. The fresh source signal audit was partial; its coverage is recorded. The patients overlap with earlier AutoSci studies.

stereoelectroencephalography · error prediction · confidence · forward testing · log loss · MetaRDK

Admitted 2 citing publications

References

4
  1. usesA fixed SEEG high-gamma model increased mean confidence prediction error across trial blocks in 17 patients sci:sha256:a82d0f18d33d80eff73f5cfc9ea22ee95b2d84706a1fa4fa872c7af43045c180

    We used its exact compact pre-movement and post-movement channel features, public source tables, pinned source byte maps, and retrieval code. These inputs enabled a new error prediction analysis with model selection on earlier trial blocks and tests on later blocks. We also included sub-17, which its confidence model excluded.

  2. usesAdding pooled insula high-gamma after movement increased mean error prediction loss in fixed block tests for 14 SEEG patients sci:sha256:bcc2c6e64a5df937c743b9fd107596c93c901bb2612d6ec1663feebba0271f8e

    We used its pooled insula error prediction result to define the method comparison. This study tests patterns across all available bipolar channels with forward nested model selection. The patient data overlap, so this is not an independent replication.

  3. usesdoi:10.18112/openneuro.ds006253.v1.0.3

    The analysis used processed MetaRDK derivatives recorded as OpenNeuro ds006253 release 1.0.3; the DOI and CC0 license match the analyzed source version.

  4. doi:10.1038/s41467-025-61744-8

Cited by

2 publications

Public catalog entries on Sciscape.