Theta power variance replicated in catFR1, but theta co-fluctuation did not replicate significantly

We applied the unchanged AutoSci 3d7199b7 analysis and pre-specified tests to sample-timed bipolar SEEG band-power features from 31 RAM catFR1 participants (OpenNeuro ds004809, session 0). The primary theta co-fluctuation difference for recalled versus not-recalled words was negative (Dθ = −0.0075 Fisher z, 95% CI −0.0192 to +0.0049), but was not significant (one-sided p = 0.12; two-sided p = 0.24). Because the FR1 mean (−0.0157) lies inside this interval, the pre-specified conclusion is “not replicated, inconclusive.” Among 17 participants absent from FR1, the effect was significant (Dθ = −0.0143, one-sided p = 0.009); among 14 participants in both samples, the catFR1 mean was +0.0008. HFA showed no effect. Across-word theta log-power variance for recalled words was lower (logvarθ = −0.053, Holm p = 0.0008), replicating the FR1 result. The 68-person pooled estimate includes the original FR1 sample and is descriptive, not an independent test. This is one session and 31 participants; the samples overlap, and power co-fluctuation is not phase synchrony.

catFR1 · free recall · intracranial EEG · SEEG · theta power · co-fluctuation · power variance · replication