Uniform picture resampling increased bootstrap standard deviation by 63% for the ds003823 HRV group false alarm valence contrast

This study tested shared-picture uncertainty in 165 public recognition records from OpenNeuro ds003823, release 1.3.5. The score compared positive minus negative false alarm rates between groups assigned to increase or decrease heart rate variability. Age and picture-set weights were fixed. Adding shared uniform picture draws to person draws increased exact bootstrap standard deviation from 1.84 to 2.99 percentage points. This was a 63% increase. The estimated contrast was 2.62 points. The crossed empirical 95% interval was [-3.18, 8.52] points. Both person-only and crossed intervals contained zero. The package supplies an exact variance tool for complete blocks, a proof, source tables, saved draws, and independent checks. A fresh-source run reproduced all 37 result files byte for byte. An added audit found that 18 published false alarm means matched half the rate over new pictures within rounding precision. The findings apply to the stated finite-data draw law. Population interval coverage was not proved. The selected pictures were not a probability sample.

crossed bootstrap · shared stimuli · false alarms · emotional memory · heart rate variability · variance decomposition · OpenNeuro ds003823