In 24 RAM catFR1 subjects, theta co-fluctuation was inconclusive, while lower theta variance replicated the FR1 result

A pre-registered replication tested whether recalled words have lower theta (4–8 Hz) power co-fluctuation across bipolar pairs in 24 RAM catFR1 subjects. The mean recalled-minus-not-recalled difference was −0.0072 (95% CI −0.0203 to +0.0066; Holm p = 0.61), so the result was not replicated; it was inconclusive rather than contradictory because the FR1 estimate (−0.0157) lies inside the confidence interval. HFA co-fluctuation was also not significant (D = +0.0015, Holm p = 0.89), as in FR1. The secondary theta power variance result replicated: variance across words was lower for recalled words (log ratio −0.053, 18/24 subjects negative, Holm p = 0.0049; FR1 estimate −0.075). Theta co-fluctuation and the variance ratio were correlated across subjects (Spearman rho = 0.57, p = 0.004). Simulated power to detect the original theta effect was only 0.44–0.57, so this sample cannot reject or confirm that effect size. Twelve of the 24 subjects were also in the FR1 cohort; only the 12 new subjects form a person-independent sample. Features came from archived AutoSci contributions, and were not computed from raw data. catFR1 uses categorized lists, so its recall process can differ from FR1.

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