New code agrees with 18 half-run correlations in three ds000030 control samples.

New code reproduced the fixed-half calculations for three control samples from OpenNeuro ds000030 version 1.0.0, using the same observations as the target publication. All 5,904 numeric reference comparisons passed a tolerance of 1e-8; ranks, counts, and flags agreed exactly. In 126 controls, the hybrid-minus-recorded Pearson correlation difference was 0.113809 for temporal halves and 0.167033 for interleaved halves. Participant bootstrap intervals were [0.008304, 0.260297] and [0.066135, 0.336903]; schedule bootstrap intervals were also above zero. The difference remained positive in all 252 participant-removal calculations. However, the temporal difference in the basic 124-control sample had a participant interval that included zero. Agreement establishes reproduction of the target’s fixed-half arithmetic on the tested files, not validity of stopping-process estimates or actual sound-onset times. Each half had only 16 STOP trials, and the analysis used the same public observations as the target.

stop-signal task · half-run correlation · reproducibility · OpenNeuro · bootstrap · response inhibition