Under fixed match rules, mean local response time increased more around errors in three ds004883 1.0.1 flanker tasks for 112 people

We compared response time changes around errors and matched correct sequences in OpenNeuro ds004883 version 1.0.1. The fixed rules retained 112 people and 9,878 errors across three flanker tasks. Controls had the same three stimulus conditions and recording segment. Their centers were within 50 trials. Prior response times differed by at most 25 ms. The mean local error change was 36.06 ms. The matched correct change was 24.37 ms. Their mean difference was 11.69 ms, with a 95% interval of 8.57 to 14.82 ms. The interval used whole-person samples within two sites. Eight planned rule checks gave positive estimates. The additional center-time match left no complete sample. A later check used the same people and errors. Adding the prior time match reduced its score by 23.27 ms compared with trial-distance ranking alone. That check was exploratory. Error center responses remained about 95 ms faster than control responses. These are conditional associations, not causal effects. No EEG voltage was analyzed. The package preserves source files, hashes, code, matches, all scores, bootstrap draws, independent checks, and plots. Both a fresh-source repeat and an archived-source repeat reproduced all 23 numerical files by hash.

post-error slowing · response time · flanker task · matched controls · behavioral measurement · OpenNeuro ds004883