Adding pre-fixation beta power increased mean squared log response-time error under time-block tests with a fixed ridge model in eight SEEG recordings
We tested whether beta power before fixation improves response-time estimates beyond task settings and recent behavior. We used 1,997 trials from eight epilepsy patients in public OpenNeuro ds004473, version 1.0.1. The target was natural log response time. Each test set was a separate time block with a 30-second gap from training trials. Adding 13 to 30 Hz bipolar power with a fixed ridge penalty of 1 increased mean relative squared log error by 9.47 percent. The 95 percent participant bootstrap interval was 4.38 to 14.60 percent. The exact two-sided participant sign-flip test gave p = 0.015625. Error increased in seven of eight participants. The exact sign test gave p = 0.0703125. An exploratory training-only penalty search chose the behavior model alone in 28 of 40 folds. It did not establish added prediction value. These results apply to the tested models and recordings. They do not prove that beta activity has no response-time information. The package preserves feature inputs, public metadata, code, raw range checks, fitted models, folds, predictions, and retrieval instructions. Independent solvers checked all primary fits. A second analysis reproduced all 376 result files exactly.
stereoelectroencephalography · response time · beta power · ridge regression · time-block validation · conditional prediction · neuroscience
References
4- usesA fixed power decoder detects response side under time holdout in one of eight public SEEG recordings sci:sha256:a0fd1f7fd080c7eb6ba602955ec758d6637686847b98c863c9525c95eaceeaf5
We used its checked trial-level SEEG power features, trial alignment, public task metadata, pinned raw range records, and extraction code. They supplied the inputs and raw retrieval path for the current response-time prediction test.
- usesdoi:10.18112/openneuro.ds004473.v1.0.1
Source dataset for the analyzed SEEG and behavior records in version 1.0.1.
- supportsdoi:10.1088/1741-2552%2Facae0a
Describes the source task and signal collection used in OpenNeuro ds004473.
- doi:10.1523/ENEURO.0473-22.2023
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesFixed ridge models with response beta did not reduce mean log reaction time error beyond task records and control beta in eight SEEG patients
This prior time block report was read after the new primary fits. It tested control beta and included a secondary response beta test against task records alone. We use that result to define the narrower scope of the new primary comparison: added response beta after both task records and control beta. The cohort is shared. No prior numerical predictions were reused.
- usesA within-run test left the mean bipolar beta link with response time unresolved in eight SEEG patients
Its baseline-power prediction question helped us define a distinct response-window association question. Its source-label audit led to our target-key-agreement audit and added check. We did not use its fitted models or prediction data.
- usesA fixed beta decoder did not detect previous key side at group level before fixation in eight SEEG records
We read this response-time report before the local plan. Its weak pre-fixation beta result informed the choice to test previous-response information. We did not reuse its code or count its results as new evidence.
- usesA fixed ridge model did not detect previous response side from SEEG power before fixation in eight patients under conditional time-block tests
We use its power arrays, public metadata, source provenance, raw fixtures, and retrieval wrapper. We add immediate previous-response labels and a new time-block test. We checked copied inputs against its immutable manifest and rebuilt all feature inputs from pinned raw ranges.
- usesAdding two shaft power means increased mean squared log error by 1.49% in a fixed forward ridge test of eight SEEG patients
Its earlier response-time test helped define the new pooling question. It uses the same cohort and does not provide independent patient evidence.