A sign flip test did not detect a hippocampal 8–20/4–8 Hz power ratio change from fixation to late maintenance in 12 SEEG patients

We compared hippocampal power changes from fixation to late memory maintenance in public SEEG. The higher band was 8 to less than 20 Hz. The lower band was 4 to less than 8 Hz. A local plan fixed the contrast before calculation. The study used 1,000 successful trials with a zero source artifact flag. It used 44 matched bipolar pairs from two sessions in each of 12 patients with epilepsy. Loads, sessions, and patients had equal weights at their respective steps. The mean change in the log band ratio was -0.375 dB. The 95% t interval was -0.823 to 0.074 dB. An exact two-sided patient sign flip test gave p = 0.0947265625. The test did not detect a ratio change. This does not establish equal band changes. The exact test assumes independent patients and sign symmetry under the null. All planned descriptive checks are included. Fresh fixed-version public records verified both primary windows in all 1,167 source trials. Separate Fourier and ratio code verified the inputs and primary result. The package preserves spectra, raw excerpts, source metadata, retrieval identities and hashes, source code, results, and instructions. This is a task-phase association in a selected clinical sample. It does not isolate a memory mechanism or separate periodic from aperiodic power.

stereoelectroencephalography · hippocampus · working memory · power spectra · band power · sign flip test