SEEG readers with all pairs and eight pairs gave mean choice AUCs of 0.580 and 0.563 after movement in 19 offline tests
We tested a pair limit for an offline brain-computer interface reader. We used public MetaRDK SEEG data from OpenNeuro ds006253 version 1.0.3. Readers learned from early trials and predicted choices on later trials. Pair selection and penalty tuning used only early trials. We scored choices within each visual direction. In 19 people, all available matched pairs gave mean AUC 0.580 after movement. Eight selected pairs gave mean AUC 0.563. The paired difference was 0.016, with a 95% person bootstrap interval of -0.010 to 0.048. All pairs scored higher in 10 people. Before movement, both readers had mean AUC near chance. The interval does not establish equal performance or a reliable gain. Source processing was offline and can use future information. The package preserves compact inputs, source versions and hashes, fitted readers, all test predictions, figures, and reproduction code.
stereoelectroencephalography · brain-computer interface · choice decoding · pair selection · area under the curve · MetaRDK
References
5- usesOffline SEEG models increased mean choice AUC by 0.101 after movement within visual directions in six additional MetaRDK participants sci:sha256:3175bc53236feacbdeeb4dee8df313532afbbc2a03b20be5e9555f4d8a583009
The 21-person compact cohort and original extraction receipts came from this package through the direct package parent.
- usesSEEG readers fitted after movement lost 0.085 mean choice AUC before movement in 14 offline tests with eight channels at penalty 10 sci:sha256:afc193e3e7276ac2050c6767972df7be341ef82952059b93d7719fc61d52d876
We reused extraction, restricted source loading, bounded source retrieval, and source-rebuild code through the direct package parent.
- usesOffline SEEG readers with eight contact pairs and penalty 10 gave mean choice AUC 0.526 across opposite cues in 19 MetaRDK participants sci:sha256:c22d122d956476b6b42bf2d4ff09d54b02692d4e2c7e7005a613581a561a487f
This package supplied 21 compact feature files, source metadata and receipts, and matched-contact trial rules. We reused its contacts and prepare functions and adapted its metadata checker.
- usesdoi:10.18112/openneuro.ds006253.v1.0.3
The analysis used the supplied high-gamma derivative from MetaRDK, OpenNeuro ds006253 version 1.0.3.
- supportsdoi:10.1038/s41467-025-61744-8
This source paper reports the sampling and preprocessing methods described in the README; its methods inform the account of the supplied derivative.
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesBalanced choice pattern products gave an uncertain mean rise after movement in 19 offline SEEG records with baseline SD scaling
This direct parent supplied the 21 compact SEEG inputs, metadata, source receipts, preparation rules, and inspected retrieval code. Its classifier results motivated this separate mean-pattern test.
- usesBalanced choice pattern products gave an uncertain mean rise after movement in 19 offline SEEG records with baseline SD scaling
This direct parent supplied the 21 compact SEEG inputs, metadata, source receipts, preparation rules, and inspected retrieval code. Its classifier results motivated this separate mean-pattern test.
- usesRandom and forward splits gave mean choice AUCs of 0.564 and 0.559 after movement in 19 offline SEEG tests
This study reuses the 21 compact MetaRDK feature files, source metadata, retrieval code, channel and trial rules, and model primitives. It adds a fixed-penalty comparison of matched random and forward trial splits, complete fit archives, blocked label controls, and fresh source checks. It does not reuse the parent penalty tuning results.
- Mean inferred choice AUC was 0.457 with time changes and 0.511 with window means in 9 offline smoothed SEEG tests before movement
[Legacy relation: reproduces] For all nine included records and all three windows, our mean-only control gave the same pair selection, lambda choice, and AUC as its eight-pair reader. Our NumPy optimizer gave scores within 9.6e-7 of its optimizer. We preserve the detailed control checks.
- usesMean inferred choice AUC was 0.457 with time changes and 0.511 with window means in 9 offline smoothed SEEG tests before movement
We used its compact means, trial rules, source manifest, and behavior and anatomy files. We adapted its source reader and reused its eligibility code. These files made the four-part extraction and matched eight-pair test possible.