A forward ridge test did not detect a mean error reduction from pre-fixation SEEG power in eight patients
We tested whether SEEG power before fixation improves reaction-time prediction. We used 2,024 eligible trials from eight patients in OpenNeuro ds004473 version 1.0.1. Models used earlier trials to predict later trials from the same patient and instruction. Both models used task information and previous behavior. One model also used bipolar power at 13 to 30 Hz and 70 to 110 Hz. A nested forward test selected the power penalty. Across 1,220 test predictions, the mean patient error reduction was -7.74%. The 95% patient bootstrap interval was -15.81% to +2.03%. Seven of eight patients had more error when power was added. Large fixed penalties gave mean changes near zero. The result limits this ridge procedure; it does not show that SEEG lacks information about response speed. The package preserves input features, code, source versions, predictions, results, raw audit samples, and offline reproduction steps.
stereo-electroencephalography · reaction-time prediction · ridge regression · forward validation · neural power · beta power · high gamma · clinical neuroscience
References
3- usesA fixed power decoder detects response side under time holdout in one of eight public SEEG recordings sci:sha256:a0fd1f7fd080c7eb6ba602955ec758d6637686847b98c863c9525c95eaceeaf5
This source supplied the archived pre-fixation power arrays, trial alignment, BIDS metadata, EDF headers, range hashes, and pinned retrieval code for eight ds004473 patients. We used these inputs for new forward reaction-time models and independently checked one raw epoch per patient.
- usesdoi:10.18112/openneuro.ds004473.v1.0.1
Source data for the eight-patient SEEG reaction-time test; the exact dataset version is recorded in the publication and pinned source metadata.
- doi:10.1088/1741-2552%2Facae0a
Cited by
2 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
Its prior reaction time test informed the task history controls and the baseline comparison. This study adds response beta with a fixed penalty and time block validation. It uses the same patients and does not reuse the prior numerical predictions.
- Adding two shaft power means increased mean squared log error by 1.49% in a fixed forward ridge test of eight SEEG patients
[Legacy relation: reproduces] The full-contact inner-selection and fixed-100 comparisons reproduce 2,440 prior prediction rows exactly. This is a numerical reproduction on the same patients.
- usesAdding two shaft power means increased mean squared log error by 1.49% in a fixed forward ridge test of eight SEEG patients
We use its nine behavior controls, forward folds, 30 second gap, training transformations, and ridge regression code. We retrieved the source feature subset through its package.