A code rule gives the recorded arrow for all 45 scanner responses in GO trials 49–96 of ds000030 sub-10527.

In OpenNeuro ds000030 version 1.0.0, the scanner-run response codes for participant sub-10527 mapped to the recorded arrow as 10 → RIGHT and 21 → LEFT on 48/48 GO trials used for rule selection. The same rule matched all 45 responses in the held-out GO trials 49–96; three of those 48 trials were omissions. The reverse split selected the same rule, and it matched all 19 STOP responses. In the training run, a separately selected rule (16 → RIGHT, 17 → LEFT) matched 46/47 held-out GO responses and 11 STOP responses; one held-out GO response disagreed, and one STOP response had an unknown code. All 96 GO trials in each run retained the source IncorrectResponse label. A post hoc label scenario adds sub-10527 to the target screen, changing its count from 102 to 103 participants, but it is not a data repair or validated cohort. This exploratory analysis concerns one selected participant and recorded arrows. It cannot identify the physical buttons pressed or rule out shared recording or conversion errors. The task code, button configuration, and response log needed to identify the physical button mapping were unavailable. No stop-signal reaction time or group contrast was calculated.

stop-signal task · response codes · arrow responses · OpenNeuro · ds000030 · held-out test · data conversion