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

Signal subtraction reduced mean AUC on later errors by 0.1205 in four offline SEEG readers fit on correct trials with eight channels at penalty 10

We tested choice readers in public MetaRDK SEEG data. Each model used eight bipolar channels and fit only early correct trials. The main test used four new source participants and 95 later error trials. The mean error AUC was 0.5340 for the signal after movement and 0.4135 for the signal after movement minus the signal before movement. The paired change was -0.1205; its descriptive 95% person bootstrap interval was -0.2925 to 0.0107. The two-sided person sign-flip test gave Holm-adjusted p=0.5000, under a symmetric null assumption. Four people cannot give a two-sided exact p below 0.125. This test does not establish a population benefit, harm, or equal performance. A separate comparison used 14 inherited participants. The package adds four compact source datasets, fixed models, trial scores, figures, source hashes, and a full replay. The supplied processing and the signal after movement limit the result to offline use. The study gives no evidence about brain stimulation.

stereoelectroencephalography · choice decoding · movement-related signal · error trials · high gamma activity · brain-computer interface · area under the curve

Admitted 0 citing publications

References

6
  1. supersedesSignal subtraction reduced mean AUC on later errors by 0.1205 in four offline SEEG readers fit on correct trials with eight channels at penalty 10 sci:sha256:7029bbf56a92198db25c8a76065c803f0aa1f8648cfa0ba6e4e6fb09ac824066

    Version 2 corrects the file check after sci adds metadata to a download. The data, models, figures, and scientific results are the same as version 1.

  2. usesOffline eight-channel SEEG models gave 0.484 mean choice AUC within visual directions in 8 new MetaRDK participants sci:sha256:000ec8109a497514631b4d12ae086e5a3ac4edce34c56551ee0604fb598fe426

    We used the compact-data lineage and public source extraction recipe preserved through the parent package. We added four source participants and fixed object receipts.

  3. usesTrial baseline subtraction gave an uncertain AUC gain in 14 offline eight-channel SEEG tests at penalty 10 sci:sha256:6044bf7d8bd0af124cb24251137f36c7d1b8a344873de87384c033083a1e147b

    We used its 14 compact signal datasets, extraction rules, source provenance, and numerical fitting code. We changed the training rule and tested movement subtraction on later errors.

  4. usesTraining on correct trials gave 52.3 percent balanced choice accuracy on later error trials with eight SEEG channels in six MetaRDK sessions sci:sha256:d0db88697d9917d96d3deda27fc78d4c57f28478f16eb5d60966751fe8aca882

    Its correct-trial fitting rule and later-error assay guided the label-conflict test. We compared the signal after movement with the signal change across movement.

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

    Source data: OpenNeuro MetaRDK dataset ds006253, version 1.0.3, used for the four new source participants and the reused compact inputs.

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

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.