High-gamma timing shifted partly with reaction time in 3 of 8 patients with responsive SEEG pairs

In a cued left/right key-press task, we tested whether single-trial high-gamma (70–150 Hz) timing in stereoelectroencephalography (SEEG) bipolar pairs followed the go cue or the key press. We analyzed eight patients from OpenNeuro ds004473, splitting trials into independent even-position (set A) and odd-position (set B) halves. A pooled alignment slope k measured timing shifts, where 0 indicates stimulus locking and 1 response locking. The planned single-trial Woody slope failed a synthetic check (mean 0.31 for a response-locked bump with true slope 1), so we selected k on synthetic data before calculating real slopes. Nineteen pairs in three patients had significant high-gamma increases in set A; five patients had none. Their group mean k was 0.54 in set A and 0.32 in set B, above the reaction-time-shuffle null means (0.04 and 0.03; one-sided p = 0.005 and 0.010). Pair-level timing agreement between halves was supported by a post-hoc pooled Spearman correlation of 0.49 (within-subject permutation p = 0.048). The shift was partial, and one patient did not replicate. There was no support for an association between k and template peak time. Inference is limited to three patients with responsive pairs: a test across patients cannot yield p < 0.125, and the planned sign-flip p was 0.125. The result applies only to contacts with clear high-gamma increases; five patients had no responsive pair under the selection rule. Bootstrap and permutation counts were 200. Clinical contact placement and the 0.15–1.1 s reaction-time range also limit interpretation.

stereoelectroencephalography · high-gamma activity · reaction time · stimulus locking · response locking · alignment slope · OpenNeuro ds004473 · trial timing