The adjusted test did not detect a mean late hippocampal theta slope for response time in 12 first SEEG sessions

We tested late hippocampal theta power against response time in public SEEG data from OpenNeuro ds004752 version 1.0.1. The analysis used 485 correct trials from 12 first sessions. The model adjusted for memory load, probe match, trial order, and the preceding error. The mean fitted response time change was +1.35% per one within-patient standard deviation of theta power. The 95% interval was -1.50% to +4.28%, and the exact patient sign permutation test gave p = 0.327. The primary test did not detect a mean slope under its stated assumptions. A stricter contact name rule gave +2.38% in 11 patients, which limits the primary conclusion. This contact check was descriptive. Adding theta increased internal prediction error in 11 of 12 patients under the fixed linear model. All 12 complete source EDF files matched their snapshot hashes. All 6,432 recomputed pair-window rows matched the archived feature table within the stated tolerance. The package preserves code, derived data, raw examples, source records, checks, figures, and a corrected retrieval recipe. Two new analysis runs produced equal hashes for all 20 outputs.

hippocampus · theta power · response time · stereoelectroencephalography · verbal working memory · patient-level inference