Recorded response yield had excess adjacent clustering under uniform cell shuffles with five time bins in 126 selected ds004883 1.0.1 participants

We tested adjacent recorded outcomes in three flanker tasks from OpenNeuro ds004883 version 1.0.1. A fixed-count correction gives exactly zero expected graph score under independent uniform allocations within fixed cells. In 126 selected people, the mean corrected score was 0.00988. The participant bootstrap 95% interval was 0.00277 to 0.01724. The Monte Carlo p value was 0.00010 from 9,999 primary shuffles. The raw mean was -0.00191; its finite-count correction was 0.01179. One, ten, and twenty time bins retained positive participant intervals. The single-response check gave -0.00356, with an interval from -0.00954 to 0.00250. That interval included zero. The primary outcome includes intervals without a response as failures. It measures correct response yield. The result does not establish excess clustering of ordinary response errors or a causal error effect. The package preserves a reusable tool, proof, 144 exact checks, source bytes and hashes, trial and pair tables, random draws, figures, and reproduction commands. All 1,041 source files passed a fresh public retrieval. A full replay matched all 31 numerical result files byte for byte. This is a new test in a previously studied cohort. Source labels and the conditional reference are stated assumptions.

flanker task · response yield · adjacent outcomes · fixed-count shuffle · finite-population covariance · error monitoring