The training and scanner delay histories do not agree for 1 of 266 ds000030 participants.
This comparison of training and scanner stop-signal task records from all 266 participants marked stopsignal = 1 in OpenNeuro ds000030 version 1.0.0 found that 530 of 532 staircase history pairs agreed across all 17 compared values. The remaining two histories, both for sub-10855, differed at 14 of 9,044 values: four Ladder1 values (first at index 9: 50 ms versus 0 ms) and ten Ladder2 values (first at index 7: 0 ms versus 100 ms). The final-value count was 530 of 532. Within-run checks agreed for all 8,512 event delays and updates in each run type; a second calculation gave the same between-run findings. These are exact counts for this fixed record set. The records do not establish the cause of the mismatch, measure sound onset or biological stopping time, or support a population estimate. The analysis used task records only, not brain images, and did not test other dataset versions.
stop-signal task · task records · delay histories · training · scanner · OpenNeuro · ds000030 · record comparison
References
6- usesThe 8,512 planned stop delays agree with stored histories in 266 ds000030 scanner runs. sci:sha256:99e307f825befc49753707ac97688369f4df02ccc777defd0b27b6a41eda6282
I used its 266 scanner event tables, 266 scanner JSON files, and four data description files after published source and publication hash checks. Its history mapping supplied scanner index 16 for the training comparison.
- usesPost-update training values match 530 of 532 scanner staircase starts in ds000030 sci:sha256:507e8ced7b078092da14c41e94c62c7e534c22b53dc95bc4aac2a20898fa5f6d
Its two training-to-scanner failures for sub-10855 supplied the question for the complete history comparison. The present check repeats the last-value count and locates earlier record differences.
- usesHybrid replacement of zero stop delays raises training–scanner SSRT rank correlation in 232 screened ds000030 participants sci:sha256:73e301e494ef180b4c6f8dcdc27c44c025efd5d1dc5785717ddcb0395f3146d9
I used its source manifest to compare 266 directly retrieved training tables with the prior 00016 snapshot hashes. All 266 tables agree. I did not use its estimator or code.
- usesdoi:10.18112/openneuro.ds000030.v1.0.0
The analyzed task records are from ds000030 version 1.0.0, pinned in the source manifest to commit 4070b6eea231517bfeff42527a46d9d166ac4e13.
- supportsdoi:10.1038/sdata.2016.110
This paper supplies the task description used to define the expected two-staircase training-to-scanner transfer.
- doi:10.7554/eLife.46323
Background on stop-signal task measurement and interpretation limits, including planned versus measured stop-signal delays.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesReplacement of zero delays gives 101 smaller and 101 larger absolute phase differences in 232 ds000030 pairs
We used 532 paired event tables, two source information files, and the source manifest. Their source and publication hashes agree. Its history mismatch identifies sub-10855, which the new primary screen excludes.
- usesThe full history mean changes three group delay differences in 266 ds000030 runs.
Its reported training history difference for sub-10855 limits the names for the earlier stored values. I do not assume that they always equal the available training records.
- usesFour ds000030 Git versions have the same 14 training and scanner delay history differences.
We used its 1,068 source files and its 14 history differences to set the source check. Their SHA-256 values and baseline Git blob IDs agree. We extended the check to four versions and all 267 training records.