Under stated trial rules, nominal 95% ranges covered later FFA conflict scores in 100 of 133 selected ds004883 1.0.1 participants

We tested whether earlier response times give useful ranges for later flanker scores. The source was OpenNeuro ds004883 version 1.0.1. The sample had 133 selected participants. The primary score was FFA conflict cost, the incongruent mean minus the congruent mean. Each range used earlier retained times, observed later retained counts, and two independent samples from the earlier times. Nominal 95% ranges covered 100 of 133 later scores: 75.2%, with a 95% person bootstrap interval of 67.7% to 82.0%. There were 29 misses below the range and four above it. Mean later minus earlier conflict cost was -9.16 ms. Blocks of ten covered 79.7%. An artificial random split covered 94.0%. These checks support sensitivity to order but do not identify a cause. The range assumes the same condition distributions in both halves and independent trials. The audit conditions on later counts and full-task selection. It does not test online prediction or EEG voltage. The package preserves fixed sources, all trial records, a reusable range function, all primary random draws, results, and reproduction checks. A separate parser agreed on all 278,853 stimulus intervals. A full replay matched all 14 analysis outputs byte for byte.

flanker task · conflict cost · response time · prediction ranges · coverage · behavioral measurement · cognitive neuroscience · bootstrap · temporal change · OpenNeuro