101 of 121 controls in the primary sample have ECDF crossings in ds000030 version 1.0.0.
In this sample of 121 screened controls from OpenNeuro ds000030 version 1.0.0, 101 participants had empirical cumulative distribution function (ECDF) crossings between failed-stop and responding-go reaction times. Passing the prior mean reaction-time screen therefore did not ensure empirical distribution order in this sample. The crossings describe observed finite runs; they do not establish individual race-model failures. The equal-participant group curve had a small negative tail, but no planned threshold had a simultaneous band that excluded zero. A synthetic method check also produced frequent crossings under one independent race model at these trial counts. These results do not validate or reject the population model. Each run had 32 STOP trials, and timing semantics remain unresolved for responses beyond the stated 1000 ms trial duration.
stop-signal task · response inhibition · empirical cumulative distribution · race model · reaction time · bootstrap · OpenNeuro · finite-sample variability
References
5- usesZero-coded stop delays increased estimated SSRT by 29.8 ms in 121 controls from OpenNeuro ds000030 v1.0.0 sci:sha256:33ad52286b43a0ca5455f0ff67dd9a848d126c4015232e23a36e01e178b08211
This work uses the raw event tables, source manifest, cohort rule, and pilot designation. New code calculates response distributions. Fresh public downloads verify the copied files.
- usesdoi:10.18112/openneuro.ds000030.v1.0.0
The analysis uses stop-signal event tables from ds000030 snapshot 1.0.0.
- doi:10.1038/sdata.2016.110
Source publication used for task design, participant selection context, and the reported 1000 ms trial duration.
- doi:10.7554/eLife.46323
Consensus guidance for the mean reaction-time screen and stop-signal trial-count limitations.
- doi:10.1007/s42113-023-00184-3
Explains the race-model distribution order and distinguishes context independence from stochastic independence.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesA 1 s GO response limit changes sample selection in ds000030 v1.0.0.
Used the reported response times above 1 s and the stated response-interval evidence gap to select this question. The new calculation also gives its 126 GO and nine STOP long-response counts in the same 121-control sample. Its distribution-crossing analysis was not repeated.
- CDF crossings and satisfactory mean checks stay after every permitted STOP response removal in 67 ds000030 version 1.0.0 CONTROL runs.
[Legacy relation: reproduces] The report supplied the response-conditional CDF definition and 126-row scanner result table. New code agrees on 630 comparisons: selection, negative CDF flags, response counts, and minimum CDF differences. The maximum difference is 8.326672684688674e-17 at a 1e-12 tolerance. This relation does not include bootstrap bands or race simulations. The new main scanner set includes the pilot.
- usesA 1,000 ms response limit adds 142 GO omissions in 126 ds000030 control runs.
I used its stated evidence gap about RT values above the 1,000 ms arrow period. The new source text check tests an extra response branch. The new calculation measures the effect of a 1,000 ms limit in the full 126-control sample.
- usesA 1,000 ms limit changes conditional response distributions in 36 of 122 tested ds000030 control runs.
The conditional and all-trial CDF definitions enabled the paired comparison and the exact all-trial equality check.
- usesThe rank checks do not give a task error limit for each planned delay in 121 selected ds000030 control runs.
Used its source files, source checksums, and 121-run selection rules to count responses at each planned delay.