An offline rule that reversed binary choices below 50% confidence reduced mean accuracy in 22 ds006253 v1.0.3 records
We tested a fixed offline rule in all 22 released behavior records from OpenNeuro ds006253 v1.0.3. The rule reversed a binary choice when stored confidence was below 50%. It assumes that the alternative has the opposite correctness. The test kept 5,992 of 6,017 trials. Mean participant accuracy decreased from 72.95% to 71.33%. The mean change was -1.614 percentage points, with a 95% participant bootstrap interval from -2.955 to -0.402. Confidence still ranked correct choices above errors, with mean AUC 0.619. In the source study's 17 confidence participants, the change was -0.879 points, with an interval from -2.401 to 0.321. The full source trial filter removes all reversal candidates. The result therefore concerns all released records before that filter. The package preserves source files, hashes, code, row decisions, 50,000 primary bootstrap draws, figures, and instructions. Separate calculations passed. Fresh downloads and analysis matched all 19 numerical result files. This test does not establish a benefit from asking people to reverse choices or identify a brain mechanism.
confidence · decision-making · choice reversal · accuracy · metacognition · visual motion
References
2- usesdoi:10.18112/openneuro.ds006253.v1.0.3
The analysis uses the 22 released behavior CSV files from OpenNeuro ds006253 v1.0.3; the publication records the selected files and their hashes.
- supportsdoi:10.1038/s41467-025-61744-8
The source article provides the task description, clinical sample context, and source-study participant and confidence-analysis selection facts discussed here.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesThe primary 95% interval for current confidence log loss gain included zero in 17 MetaRDK patients under fixed ridge models with later test trials
This publication supplied exact source bytes, source hash manifests, the Git tree, annex pointers, and extracted author selection facts. We checked them against fresh public downloads. Its confidence AUC informed the question. Our new code tests the added value of current confidence on later trials.
- usesThe 95% patient bootstrap interval for current confidence gain included zero on later trials from 17 MetaRDK patients under fixed ridge models
We used its reported confidence association and offline choice reversal results to define a distinct temporal probability test. Its source package is also an upstream dependency of the reused archive.
- usesPrevious confidence reduced mean prediction error for later confidence in 17 MetaRDK patients under fixed ridge models
This package supplied the initial source data, retrieval code, and pinned source-author selection facts. All source files were retrieved again and checked. Its accuracy-reversal result was not used as evidence for the new confidence-history result.