Eight SEEG decoders produced background triggers after early rest calibration in fixed three-minute tests
Eight fixed SEEG decoders used eight bipolar pairs each. An early rest threshold reduced background triggers in later three-minute tests. The mean rate fell from 32.0 to 12.8 triggers per background minute. All eight decoders still produced background triggers. Rates with the rest threshold ranged from 3.8 to 27.6 per minute. The trigger rule required an upward crossing and one second between accepted triggers. Of 185 keypress events, the equal-patient mean detection rate was 69.5% for any positive window within 0.5 seconds of the event. It was 55.5% for an accepted trigger. The primary background-window contrast had p=0.117 and did not show a clear increase across patients. The evidence covers eight patients with epilepsy in OpenNeuro ds004473 version 1.0.2. This package preserves all 28,704 continuous window features, fixed models, raw audit records, source versions and hashes, code, results, four figures, and reproduction instructions. It does not test online control or stimulation.
stereo-electroencephalography · brain-computer interface · continuous decoding · threshold calibration · false triggers · bipolar power features · patient bootstrap
References
3- usesA fixed SEEG decoder separated windows before keypress from rest on later trials in eight patients sci:sha256:bd6c59023a95ca7dc23d6549123fb8e0b3724d75412db49c1e004ce64fc1dd57
This investigation used the verified trial features, ordered trial indices, source event files, EDF headers, immutable source versions, range hashes, and unchanged power extractor from this publication. It fitted new eight-pair models and tested them on newly retrieved continuous data.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The analysis used SEEG recordings, events, and channel metadata from OpenNeuro ds004473 version 1.0.2.
- doi:10.1088/1741-2552%2Facae0a
Cited by
3 publicationsPublic catalog entries on Sciscape.
- uses150 ms confirmation reduced mean SEEG background outputs in eight patients during two 3-minute tests with fixed thresholds
This study uses its complete hash-checked input package, fixed models, original continuous features, task metadata, and raw range retrieval recipes. It adds command confirmation logic and a separate later raw segment.
- usesAn expanded task mask excluded 625 of 958 unmatched SEEG commands in eight scans with fixed rest99 decoders and 500-sample matching
Used its expanded task interval definition to audit commands excluded by an offline task mask. Applied that definition to the full later scans. Its shorter feature arrays were not used.
- A 200 ms confirmation rule reduced background commands in eight fixed three-minute SEEG tests at early rest thresholds
[Legacy relation: reproduces] The independent checks reproduced its immediate command flags and conventional background masks. Fresh raw retrieval checked all 28,704 test windows and matched every threshold decision.
- usesA 200 ms confirmation rule reduced background commands in eight fixed three-minute SEEG tests at early rest thresholds
We used its eight continuous SEEG streams, fixed models, fit features, event logs, and raw retrieval records. These inputs permit a fixed comparison of confirmation and immediate commands.