A weak blocked speed instruction did not change pre-cue SEEG band power in eight subjects

In eight subjects from OpenNeuro ds004473 v1.0.2, we compared log10 SEEG band power in the 0.5 s before the Go cue between blocked fast and slow deadline instructions. The fast instruction changed median reaction time little. In the pre-registered tests across seven subjects, the standardized fast-minus-slow effect was +0.014 for 70–110 Hz high-gamma power in high-gamma-responsive bipolar pairs (95% descriptive t CI −0.186 to +0.213; exact subject sign-flip p=0.92) and −0.155 for 14–30 Hz beta power in beta-responsive pairs (95% CI −0.400 to +0.090; p=0.19; Holm-adjusted p=0.38). Tests across all pairs, without covariates, and in the drift-balanced fast-slow-fast subset were also null. The result applies to an instruction that changed reaction time little; it does not test a strong speed-accuracy trade-off or establish absence of an effect. Interpretation is limited by the block-order confound, strong condition–trial-order correlations in two-block subjects, and the sample of seven to eight subjects, which supports detection only of large and consistent effects.

SEEG · speed-accuracy trade-off · band power · pre-cue activity · blocked instruction · sign-flip test · OpenNeuro ds004473