95% HC3 intervals include zero for six age slopes in two ds000030 control models.

A new Python program reproduced three age models for stop-signal reaction time (SSRT) scores from 241 participants in OpenNeuro ds000030 version 1.0.0. All 72 estimates and interval limits agreed with the target within 1e-8, and all 120,000 saved bootstrap values agreed within that tolerance. We compared percentile, bias-corrected and accelerated (BCa), and HC3 intervals for six control age slopes. All six HC3 intervals included zero; five of six BCa intervals and four of six percentile intervals had lower limits above zero. Each method's three age intervals in the all-participant model included zero. Paired slope-difference intervals included zero for all methods and models. A 20-seed check found the percentile lower-limit classification stable across those specified draws, but it supplies no new participant observations or interval-coverage evidence. These cross-sectional estimates do not establish a causal age effect. Each participant completed 32 STOP trials, the recorded fields do not measure sound onset, the intervals do not model trial or timing uncertainty, and no multiple-comparison adjustment was used. The findings retain the target's score rules, selected sample, and model assumptions.

stop-signal reaction time · age · bootstrap · HC3 · confidence intervals · OpenNeuro · computational reproduction