Current changes on other shafts reduced last-minute SEEG slope error by 4.3% when first-minute slopes were known in 24 participants at 20–100 Hz with 55–65 Hz removed

We tested whether current slope changes on other SEEG shafts improve a withheld last-minute slope estimate. The target first-minute slope was known. We used public OpenNeuro ds005624 version 1.0.0. The test included 24 participants, 34 records, 408 shaft records, and 2,173 pair records. We fit 20–100 Hz and removed 55–65 Hz. The estimate added the median change from other shafts to the earlier target slope. The entire target shaft was excluded from its donor set. Mean squared error fell from 0.116243 to 0.111221, a 4.32% reduction. The 95% participant bootstrap interval was 0.90% to 8.50%. Eighteen participants improved and six had higher error. All eight fixed checks are reported. An independent solver confirmed the scores. Two pinned raw record rebuilds matched all input arrays. The complete offline repeat gave 48 identical result files. The package preserves code, compact inputs, source versions, hashes, scores, uncertainty draws, and repeat instructions. This is a selected secondary analysis of two sparse excerpts. Later donor data are required. The result does not establish a forecast, actual failure recovery, or a neural cause.

stereoelectroencephalography · spectral slope · missing data · signal estimation · participant bootstrap · OpenNeuro