For unadjusted SEEG FR1 effects, tilt reached 80% power at 20 patients, while HFA alone needed more than 40

Empirical power curves were estimated from archived word-locked band-power features for 52 patients performing free recall (RAM FR1). Without serial-position control, the mean subsequent-memory effect was Hedges g = +0.178 for the difference between z-scored high-frequency activity (70–150 Hz) and theta power (3–8 Hz), −0.150 for theta, and +0.095 for high-frequency activity (HFA). In bootstrap simulations, 80% power required about 20 patients for the combined measure, 25 for theta, and more than 40 for HFA alone. MTL-cortex theta reached 80% power at 12 patients; MTL-cortex HFA required more than 40. With only the first four lists, tilt power did not decrease because early-list effects were larger; four lists could not resolve between-patient differences in tilt. Under the assumption that PAL1 effects equal FR1 effects, reported PAL1 HFA nulls with 12–19 patients had 0.22–0.36 power, and hippocampal HFA nulls with 20–21 patients had about 0.24–0.26 power using MTL-cortex HFA as a proxy. These nulls therefore do not show a smaller PAL1 HFA effect than in FR1. The bootstrap treats this 52-patient sample as the population, and the SME includes a serial-position confound; a position-controlled effect is expected to be smaller. Transfer to other tasks is an assumption, not a tested result. MTL cortex is not hippocampus.

SEEG · subsequent memory effect · free recall · statistical power · high-frequency activity · theta · hippocampus · medial temporal lobe · bootstrap