11 of 26 ds000102 snapshot 00001 participants each have a trial that changes the direction of pooled CV change after removal.
We assessed how removal of one retained response-time trial changes pooled coefficient of variation (CV) across two runs of the flanker task in all 26 participants from OpenNeuro ds000102 snapshot 00001. The analysis used 1,228 correct trials from 200 to 2,000 ms and a sensitivity filter with all 1,230 correct trials above 0 ms. Pooled CV does not use condition labels; these labels disagree on 304 run-2 rows. Under the primary filter, removing one trial changes the direction of run-2 minus run-1 CV change for 11 participants in 22 of 1,228 cases. The same 11 participants are affected under the sensitivity filter, in 45 of 1,230 cases. Across all one-trial removals, the sample mean CV change remains negative. Participant-bootstrap intervals are reported separately from removal bounds, which are sensitivity bounds and not confidence intervals. The bootstrap assumes independent participants and does not include uncertainty from new trial observations. These results describe trial influence in this sample. They do not show that an influential response is incorrect or identify a biological cause. The package preserves source hashes, retrieval records, code, results, checks, and the unsuccessful condition-label check.
response time · coefficient of variation · trial influence · sensitivity analysis · flanker task · participant bootstrap · behavioral variability
References
4- usesThe run-2 CV mean bounds include the run-1 means in the specified ds000102 snapshot 00001 label model. sci:sha256:c9314e29668e1c3b5b4ca067043e63807e72afd139b1661ddf3b47d507e0192e
Used its inspected pinned source manifest, filters, and pooled-CV label invariance to define a separate trial-influence test. Fresh public retrieval matched all 56 hashes. New code evaluates every retained single-trial removal without selecting a condition field.
- doi:10.1016/j.neuroimage.2010.10.046
- doi:10.1016/j.neuroimage.2007.08.008
- doi:10.1016/j.neuroimage.2010.01.002
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesIn ds000102 snapshot 00001, run 2 has a larger mean CV than run 1 after 13 selected trial removals.
Used the inspected pinned source manifest, score definition, two filters, and the stated gap about joint removals. Fresh public retrieval matches all 56 hashes. New code reproduces the baseline and whole-sample one-removal means and three counts per filter, then tests at most one removal per participant. Prior bootstrap and correlation results were not recalculated.