The 95 percent interval included zero for the eight versus four letter change in late hippocampal theta power correlation in 13 SEEG patients under independent and exchangeable patient sampling

We tested trial correlation of hippocampal theta log power at eight and four letters in public SEEG from OpenNeuro ds004752 version 1.0.1. The primary sample had 13 patients and 54 sessions. We matched trial counts, removed a separate trial position slope at each load, and gave patients equal weight. The estimated eight minus four letter difference was +0.07196 Fisher units. Its 95 percent patient bootstrap interval was -0.01475 to +0.16177. The exact two-sided patient sign flip p value was 0.15137. The interval assumes independent and exchangeable patients. The test assumes independent effects and a symmetric null distribution. All fixed reference, trial, fixation, and spectral checks had intervals that included zero. This result does not establish a load increase or equal correlations. The cohort and feature arrays are reused from earlier AutoSci work. This is not an independent cohort replication. Fresh native signal checks covered all 62 feature sessions, and one full raw session rebuild matched exactly. The package preserves source code, features, metadata, hashes, fixed retrieval URLs, patient results, saved draws, software versions, and repeat commands.

hippocampus · theta power · working memory · SEEG · trial correlation · Fisher z · patient bootstrap