Two calculations give six bootstrap limits with differences below 1e-9 ms in ds000030 version 1.0.0.

A new calculation checks the recorded delay differences, file selection, and bootstrap interval limits in a prior analysis of OpenNeuro ds000030 version 1.0.0. All 270 retrieved source files and the complete fixed Git tree match recorded sizes and checksums. All 5,118 comparisons pass. The 266 file differences and selection decisions agree; the largest file difference is 1.875E-14 ms, and the same 241 files are selected. For all files, the mean nominal-minus-recorded delay is 30.51993855192479 ms (95% interval 25.68082722917487 to 35.61342541670047). With nominal replacement for zero recorded delays, the mean is -2.1985764856691987 ms (95% interval -2.544168080084133 to -1.9196301295827747). For selected files, the mean is 31.077540972379904 ms (95% interval 26.115129486028405 to 36.21618102052617). Five of six interval limits match exactly; the largest limit difference is 2.220446049250313e-16 ms. The replacement is an arithmetic sensitivity check, not a validated correction. The work does not calculate absolute stop-signal reaction time or establish physical tone timing, biological validity, population representativeness, or hardware uncertainty. Repeated schedules may cause dependence, and the exact acquisition and conversion versions remain unverified.

stop-signal task · reaction time · computational reproduction · bootstrap · OpenNeuro · ds000030 · delay fields