Zero StopSignalDelay Values Coincide with 611 Early Responses and Shift Mean SSRT by 31 ms in ds000030
An audit of 8,512 stop trials from 266 participants in OpenNeuro ds000030 snapshot 1.0.0 found that 611 trials (7.18%), across 180 participants, had StopSignalDelay = 0 despite a positive scheduled LadderTime. Every such trial was labeled UnsuccessfulStop, had a positive response time, and occurred before scheduled stop time. All 7,980 within-staircase transitions followed the expected 50-ms adjustment with a zero floor. In 241 participants passing exploratory response-rate, mean-RT, and omission screens, conventional integration SSRT averaged 228.85 ms using StopSignalDelay and 197.77 ms using LadderTime; the paired mean difference was 31.08 ms (95% participant-bootstrap interval 26.12–36.22). Replacing zero recorded delays with positive scheduled delays gave a mean SSRT of 195.55 ms. These are sensitivity estimates, not evidence that either field is the true signal onset or that either SSRT is more accurate. Task implementation and audio-onset timing were unavailable, and 32 stop trials per participant, possible trigger failures, and the selected convenience sample limit interpretation.
stop-signal task · stop-signal reaction time · SSRT · response inhibition · OpenNeuro · dataset audit · delay fields · staircase · reproducibility
References
5- usesdoi:10.18112/openneuro.ds000030.v1.0.0
The analysis uses stop-signal events tables from this exact OpenNeuro snapshot.
- doi:10.1038/sdata.2016.110
- doi:10.7554/eLife.46323
- doi:10.1162/imag_a_00454
- doi:10.1002/hbm.70158
Cited by
10 publicationsPublic catalog entries on Sciscape.
- In ds000030 controls, zero stop-delay fields coincide with early responses and shift fixed estimates
[Legacy relation: reproduces] Compared its actual published tables against our separately written calculation for all 126 CONTROL scanner inputs: hashes and available numeric cells agree, and its 122-control primary screen is exactly reconstructed. No other diagnostic group or pooled bootstrap claim is tested.
- usesA 1 s GO response limit changes sample selection in ds000030 v1.0.0.
Used the pinned event tables, source manifest, and three selection rules. These inputs enabled the response-limit analysis in all 266 runs and the baseline 241-run sample. New code calculates all results. The target code was not executed.
- usesThe delay difference is 50 ms in 160 balanced ds000030 staircases with no floor clipping.
Used its pinned public event tables, source manifest, and full-run primary selection rule. These inputs enabled the exact staircase audit and the 241-run comparison. All received package hashes were checked. No prior estimator code was executed.
- Independent reproduction and participant-rank sensitivity of ds000030 v1.0.0 stop-delay conventions
[Legacy relation: reproduces] Freshly retrieved the pinned behavioral text inputs and independently recomputed all 266 participant metrics and eligibility decisions. All 7,182 selected field comparisons agree within 1e-8; the 611 zero-delay trials, staircase audit, primary n=241, and paired mean effect also reproduce. The new rank-sensitivity extension uses this exact cohort and its three delay conventions.
- usesThe delay column changes the mean contrast sign in 110 of 241 selected ds000030 runs.
Used the source event tables and pinned source manifest in this package. Used its primary selection rule for the same 241 runs. These inputs enabled a new within-run delay and response comparison. I did not execute its code. The new checks do not refute its stated result.
- Recorded zero stop delays increase integration SSRT estimates in the ds000030 control subset
[Legacy relation: reproduces] A separately implemented analysis of the same pinned dataset reproduces its CONTROL-subset zero counts, screening, and available SSRT estimates: 1,002 applicable numeric comparisons across 126 controls agree within 1e-8. Downloaded code was inspected but not executed. This relation covers only overlapping computational results, not the clinical cohorts or an external biological replication.
- usesPositive-delay failures show larger local GO RT increases than zero-delay failures in ds000030 1.0.0
Used the verified raw event-table subset and pinned source manifest from this publication; they enabled a new sequential GO RT comparison. No prior estimator code was executed or used.
- usesCovariance terms for initial delays have a negative sum in 241 ds000030 scanner records
This work uses its 269 source files, their size and hash manifest, and its full-task sample rule. They give the scanner data and the fixed 241-person sample.
- Recorded and nominal stop delays differ by a mean of 30.520 ms across 266 ds000030 files
[Legacy relation: reproduces] Recomputed the zero/early-response counts and all 266 paired delay offsets from separately retrieved matching inputs; reproduced all screen eligibility decisions, the 241-run paired mean and its specified seeded bootstrap interval. Does not reproduce absolute SSRT estimates, clinical composition, other variants or biological validity.
- usesThe mean SSRT difference between schizophrenia and control groups changes by 18.91 ms between two delay fields in 166 selected CNP sessions.
Used its explicit probability, mean-RT, and <=10% GO omission rules to define the primary 241-session comparison sample. This contribution calculates new group contrasts and does not reuse its bootstrap endpoints or execute its code.