Five removals from 19 screen failures are the minimum for a chronological age slope difference greater than zero with recorded delays.

We tested whether removing records that failed half-task response screens could reverse the age slope difference in stop-signal reaction time (SSRT) scores. The fixed sample contained 122 controls from OpenNeuro ds000030 version 1.0.0. We exhaustively tested every removal subset of 19 chronological and 20 interleaved screen failures, with three delay rules and exact arithmetic on printed input decimals. For chronological halves using recorded delays, no removal of four or fewer records changed the slope difference from negative to positive. One of 11,628 five-record subsets did so: removing sub-10271, sub-10356, sub-10998, sub-11019, and sub-11097 gave 0.061867469817344735 ms/year among 117 records. Across splits and delay rules, the smallest reversal counts ranged from one to five. Participant bootstrap intervals for chronological paired screening changes excluded zero under all three delay rules; all interleaved change intervals included zero. These exploratory results concern a restricted, outcome-selected set of screen failures, not all possible removals. Each task half had only 16 STOP trials, below the cited 50-trial recommendation. Intervals omit trial variation, timing and model error, and the analysis does not establish a population slope direction or biological cause.

stop-signal task · SSRT · response inhibition · age slope · sensitivity analysis · participant screening · exact arithmetic