Fixed mean 250 ms and SD 100 ms allowed full delay integration widths of 0.12 to 96.10 ms in 347 stationary lognormal go models with fixed delays and independent integer stop times from 0 to 500 ms

We tested whether full delay integration gives one stop value when the true stop mean and standard deviation are fixed. We used 347 public-data model profiles: 226 from OpenNeuro ds002790 v2.0.0 and 121 from ds008037 v1.0.1. Each model had a stationary lognormal go law, fixed recorded delays, and independent stop times. All primary stop laws had mean 250 ms, SD 100 ms, and integer support from 0 to 500 ms. Linear programs gave sharp finite-grid limits in theory. The numerical interval widths were 0.12 to 96.10 ms, with median 3.41 ms. Model-set median widths were 1.06 and 21.80 ms. In 210 models, the whole interval excluded 250 ms using a 0.00001 ms margin. A common stop law with equal mass at 150 and 350 ms gave a 14.98 ms difference between the average model-set values. We provide an audit tool, endpoint laws, dual bounds, a proof, all source files, and repeat instructions. Separate checks passed for all 5,552 endpoint laws. Fresh downloads matched all 482 source files. A full repeat reproduced all 14 analysis files byte for byte. These results describe bounded model laws. They do not estimate human stopping times or test adaptive tasks.

stop-signal task · full delay integration · stop-signal reaction time · stopping time · fixed moments · linear programming · sharp bounds · race model · lognormal model · behavioral modeling