The two computer programs find the same 304 run-2 label differences in ds000102.

A separately written program using rational arithmetic reproduces the same congruency label differences as an inspected prior program in the specified ds000102 files. The analysis covers all 26 participants, both runs, and 24 trials per run (1,248 trials). It finds no differences in run 1 and 304 differences among 624 run-2 trials (48.717949%); all 26 participants have 6 to 18 run-2 differences. Ten other field comparisons agree in all trials, and five output tables and twelve numerical summary groups match the prior results. A strict check records the label disagreement as one failed check among 379. These counts describe the selected archival files; they do not establish which labels match the displayed arrows, when or why the conflict arose, or a population effect. The two programs use the same data and condition meanings, and the researcher inspected the prior code and expected result before writing the second program. The result is a computational replication, not an independent cohort or blinded verification. The legacy response-time unit is not stated; conversion to BIDS seconds assumes a factor of 1,000. No labels were corrected, and no brain result was recalculated.

congruency · data audit · computational replication · OpenNeuro · ds000102 · flanker task · behavioral data · BIDS · provenance · reproducibility