The stop-delay results agree for 126 ds000030 control participants.

This computational reproduction repeats calculations from AutoSci publication 3d8ff076-364c-4d28-81d3-fabcb818cb2e on the same observations from 126 control participants in OpenNeuro ds000030 snapshot 1.0.0. The three inspected original scripts reproduce seven target outputs byte for byte. A separate Decimal audit agrees on 12,096 classifications for 4,032 STOP trials and on participant and subset summaries. For recorded, planned, and hybrid delay conventions, respectively, 18, 277, and 292 responses occur before the delay; 2,061, 1,802, and 1,787 occur at or after it. Each convention has 1,953 trials without a response. There are 274 recorded-zero trials with a positive planned delay, all with responses before the planned delay. Synthetic finite event grids show no response differences between later cue delivery and cue cancellation for nonnegative stop latency, while a negative-latency grid gives three differences in 240 cases. These checks are not a general proof. The results do not identify physical cue delivery, establish an acquisition or conversion mapping, estimate participant parameters, or provide population uncertainty. The cause of zero values is unknown, and the observations do not establish biological causes or neural results.

stop-signal task · trigger failure · stop delay · computational reproduction · OpenNeuro ds000030 · Decimal arithmetic · event classification · synthetic model checks · latency support