Eight beta directions fitted for two seconds showed an uncertain change in scaled variance loss after stimulation in 20 patients from ds006519 v1.0.0
We tested whether eight beta directions (13 to 30 Hz) fitted for two seconds lose more scaled variance after stimulation. The source is OpenNeuro ds006519 version 1.0.0. One record failed the fixed window-length rule. The test used 34 records from 20 patients and all 1,691 stored channel occurrences. Patients had equal weight. Loss is the best test-window retained variance minus variance retained by fixed directions. Mean loss was 29.45 percentage points before stimulation and 30.89 after. The mean change was 1.44 points, with a 95% patient bootstrap interval of -1.79 to 5.98. The exact two-sided patient sign-flip value was p = 0.6433. Seven patients had an increase and thirteen had a decrease. With each window's own channel scales, the descriptive mean change was 0.03 points. There is no matched control or command score. The result does not show a causal stimulation effect or BCI accuracy. The package has covariance inputs, saved models, an encoder, three figures, source hashes, retrieval code, and replay checks.
stereo-EEG · brain-computer interfaces · spatial filters · variance retention · electrical stimulation · signal processing
References
4- usesA fixed detector found periodic voltage runs in 31 of 35 SEEG records from ds006519 version 1.0.0 sci:sha256:ea4058aa1dce526e987cca14e771bbeab34ccf9025624ac6983dbb7c42876d8e
Its tagged source audit established the version and Git-annex hash identity of the retrieval files inherited through sci 55e1205e-8b3a-4e7f-a3c4-06eee47e7d24. Its event interpretation informed the timing limit. We verified all downloaded source bytes in this run. We did not use detector labels to select the sample.
- usesA 100 ms wait gave 1.9% mean channel agreement within 1% of baseline power after causal high-gamma resets in 35 SEEG records sci:sha256:656315e237f16d31ca1b9bfbfdc600ae0f5d2e8151d32b4b687b9cb307887d42
We copied and inspected its code/retrieve.py and provenance/source_manifest.json. These files enabled a verified download of all 355 source files for the new spatial projection analysis. No earlier numerical power outcome was an input.
- usesdoi:10.18112/openneuro.ds006519.v1.0.0
The analysis used the released voltage and metadata from OpenNeuro ds006519 version 1.0.0.
- doi:10.1016/j.neuroimage.2026.121822
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesA beta SEEG model with half the channel correlation gave an uncertain mean score change after stimulation in 20 patients
This study uses the exact covariance and mean arrays, native windows, source metadata, channel rule, exclusion rule, retrieval script, and extraction script. It tests a new Gaussian voltage score. All copied inputs passed hash checks. A full raw rebuild gave exact array matches.