The trial pair counts and scores agree under all five rules in the ds000102 snapshot 00001 report.

A separate calculation reproduces the trial-pair results reported for the same 26 participants in ds000102 snapshot 00001. It agrees with all 30,557 exact values and 11,179 numerical values across pair, cell, exclusion, score, and summary outputs; all 518 available score fractions agree exactly, and the largest numerical difference is 2.84e-14 ms. The source fields `trial_type` and `Stimulus` disagree on 304 of 624 run-2 trials, so the calculation retains both and cannot verify the displayed arrows. For the primary run-2 comparison, the mean score difference (`trial_type` minus `Stimulus`) is 55.36 ms, with a 95% participant-bootstrap interval from -22.15 to 134.23 ms. This interval includes zero. Four of five rule intervals include zero; only the inherited trim rule’s interval excludes zero, without correction for multiple comparisons. The work is a computational reproduction of the same observations, not a new sample, a blinded test, or evidence for a cognitive or biological cause. The analyzed files are identified by Git commit `fdc1ce49d524fc1e77e5b763f10acb02f403c581` and retained content hashes.

computational replication · trial sequence · congruency sequence effect · flanker task · response time · condition labels · bootstrap · ds000102 · OpenfMRI · OpenNeuro