Mean hippocampal fixation theta contrast was lower after errors with prior response time control in 14 SEEG patients at load four under patient independence and null sign symmetry

We tested whether previous response time accounted for a reported hippocampal error association. We used public SEEG data from OpenNeuro ds004752 v1.0.1. The primary model used 714 trials in 14 epilepsy patients at current load four. It controlled session, previous load, trial position, previous signals, and log previous response time. The mean error coefficient was -1.051 dB for a fixed spectral contrast. The conditional 95% patient bootstrap interval was [-1.615, -0.523] dB. The exact two-sided sign test gave p = 32/16384 = 0.001953125. This test assumes independent patient coefficients and sign symmetry under the null. Adding response time changed the mean coefficient by -0.030 dB. The change interval was [-0.187, 0.152] dB; it does not prove equivalence. Fixed model checks retained negative means. This is an exploratory analysis of the same cohort as the earlier work. It does not establish a causal error effect. The package preserves source features, pinned raw retrieval, code, fitted models, and validation results. Fresh raw range checks passed for selected windows in all 59 analyzed sessions. A fresh full raw rebuild reproduced every feature array for one session. The stored epochs do not reveal the true time between successive trials.

intracranial EEG · SEEG · hippocampus · theta contrast · response time · verbal working memory · error · spectral analysis