The mean difference in local GO reaction-time change is 119.4 ms in 32 matched ds000030 1.0.0 participants.
A fixed within-participant trial-match rule selected 35 pairs from 32 participants in OpenNeuro ds000030 snapshot 1.0.0. The equal-participant mean difference in local GO reaction-time change (positive recorded StopSignalDelay minus zero) was 119.4 ms (95% participant-bootstrap interval 58.3 to 176.9 ms). This descriptive result applies to the selected pairs and does not establish a causal stop-signal effect. The tight matching rule retained 11 pairs in 11 participants and had an interval including zero. The match balanced preceding GO reaction time and planned delay only within stated limits; STOP reaction time and planned delay still differed. The recorded zero field does not establish whether a sound was absent or its actual time. The analysis did not validate event timing or estimate stop-signal reaction time. The exploratory plan followed inspection of earlier results, and participant resampling does not include timing error or uncertainty in the match rule.
stop-signal task · reaction time · trial matching · local reaction-time change · OpenNeuro ds000030 · exploratory analysis
References
6- usesLogged-zero STOP failures marked faster neighboring GO responses in ds000030 controls sci:sha256:7e032acb63155f71092e1b40351e9d60fe50202c74a5ba0db6ae55e83ae3e9dc
Used its preceding-GO speed result to set a preceding-RT match rule. Its context restriction did not perform this covariate match. No code or numeric output from that result entered this calculation.
- usesPositive-delay failures show larger local GO RT increases than zero-delay failures in ds000030 1.0.0 sci:sha256:dd164540e8ddf090ca3f80ee8daff60571c12de09cb022d170bf642ce05203c5
Used its 270 hash-verified raw source files, pinned manifest, and saved primary participant values. These inputs enabled the new fixed-rule trial match and a 102-participant point-value reproduction.
- usesdoi:10.18112/openneuro.ds000030.v1.0.0
The analysis uses the archived OpenNeuro ds000030 snapshot 1.0.0 event tables and participant metadata, identified by the recorded version, source URLs, and file hashes.
- doi:10.1038/sdata.2016.110
Background data descriptor for the Consortium for Neuropsychiatric Phenomics dataset and its task/event context.
- supportsdoi:10.1016/j.jmp.2012.04.001
Supports the stated caution that local before/after reaction-time contrasts can be affected by slow changes in performance.
- supportsdoi:10.3758/s13428-021-01631-4
Supports the stated caution that faster responses before an error can increase a local reaction-time difference; the report notes that the tasks differ from this dataset.
Cited by
6 publicationsPublic catalog entries on Sciscape.
- usesNext GO correct fraction differences are +20.73 and −7.07 percentage points at 500 and 1500 ms for 70 ds000030 CONTROL participants.
I used 126 CONTROL event tables and four metadata files from its locked ds000030 source package. The files supplied the raw observations. Their sizes and SHA-256 checksums agree. I ran no code from that package.
- usesAll 96 primary maximum pair sets give mean differences above zero in ds000030 1.0.0.
Used its 270 source files, fixed version, selected-triplet definition, six pair rules, and comparison tables. These inputs let the new code calculate assignment bounds at maximum pair counts. No prior code was executed.
- usesThe STOP contrast minus the GO contrast is -91.3 ms in 29 matched ds000030 controls.
Its read preceding-GO matching result reports a remaining central STOP reaction-time difference. That observation motivated the central reaction-time limit for the GO comparison. Its different sample, source files, and code were not reused.
- usesThe two 50 ms limits allow at most 33 STOP trial pairs in the tested 47 ds000030 controls.
Its preceding-GO matching result reports a remaining STOP RT difference. This observation led to the direct STOP RT and planned-delay overlap test. Its mixed sample and selected matches are not reused.
- usesAll STOP responses more than 5 ms before LadderTime have zero recorded delay in 126 CONTROL scanner runs.
We used the reported lack of matched early STOP responses with recorded delay greater than zero to select the coverage question. We did not reuse its effect estimate.
- Two of six ds000030 1.0.0 pair rules give a bootstrap range that includes zero.
[Legacy relation: reproduces] New code gives the same six local-change point estimates, actual minimum-cost pairs, and fixed-pair participant bootstrap endpoints from the same source observations. The largest endpoint difference is 5.684341886080802e-14 ms. This is a computational reproduction with shared observations.