In 8 SEEG patients, the confidence null excludes only widespread effects, and the decision-time effect was not significant without window overlap

We tested whether a previously reported null association between high-gamma activity (HGA) and confidence in OpenNeuro ds006253 (MetaRDK) was informative. Using the same archived HGA window means from 8 SEEG patients (714 bipolar channels) and the same channel-count test with a residual-permutation null, we tested choice, decision time, and accuracy as positive controls and estimated sensitivity to distributed effects. The plan was written and hashed before analysis. We found 19 trials with invalid accuracy code 99; the earlier analysis had retained 15 of them. After removing invalid trials and adding a trial-index covariate, the confidence result was unchanged (37/714 channels, p = 0.40). Of three planned pre-movement controls, choice (42/714, Holm p = 0.347) and accuracy (35/714, Holm p = 0.403) were not detected, while log decision time was detected (142/714, Holm p = 0.006). In an exploratory subset without overlap between stimulus and pre-movement windows (6 patients, 677 trials, 536 channels), the decision-time effects were not significant (31/536, p = 0.30 pre-movement; 34/536, p = 0.18 stimulus). Timing overlap can explain the original decision-time effect, but loss of power and fast trials in the subset can also explain it. Simulations found power of at least 0.85 for partial r >= 0.10 in >= 20% of channels or r >= 0.20 in >= 5% of channels, but not for effects in <= 2% of channels at any tested r. Thus, the earlier confidence null excludes only large, widespread confidence coding in these two window means; it does not exclude local or time-resolved signals. Limits include 8 of 21 dataset subjects, window means only, exchangeable-trial assumptions for residual permutation, a trial-index covariate that removes only linear drift, an unverified response-side mapping for the choice control, and simulated effects independent across channels.

SEEG · high-gamma activity · confidence · decision time · positive controls · sensitivity analysis · OpenNeuro ds006253