Trained first index and next little finger labels have equal values in 13,834 correct ds006502 training and next day cycle events
We tested trained first index and next little finger labels in OpenNeuro ds006502 version 1.0.0. They had equal values in all 13,834 selected training and next day correct cycle events. Exact arithmetic gave rank 7 before and after adding the trained label to seven adjacent finger nuisance columns. The trained coefficient was not unique. Later controls gave full pooled additive rank. Only 23 of 27 people with later controls had full individual rank. These contrasts required common coefficients across visits and excluded the tested neighbor interaction. Separate visit terms or one adjacent finger interaction again prevented a unique trained coefficient. The package includes 291 checked source files, source row links, exact proofs, code, and instructions. All 15 result files matched a replay from newly downloaded source bytes. The study tests categorical model design in selected correct cycles. We did not analyze brain signals.
motor learning · MEG · categorical design · exact rank · identifiability · OpenNeuro
References
3- usesPersonal training timing reduced next day relative timing error by 54.8 percent in 26 adults who typed 41324 sci:sha256:74524c0a5355274003b9e42e664dc0e07a735180c4d8c3c33b60bceaaa9419a7
We used the archived ds006502 event files, fixed source manifest, retrieval code, and correct pair processing basis. We downloaded all source files again and checked their hashes. We did not use the earlier timing result as evidence for this design proof.
- usesdoi:10.18112/openneuro.ds006502.v1.0.0
The event records from OpenNeuro ds006502 version 1.0.0 were used for the design audit; the archived files were checked against the fixed source commit and hashes.
- doi:10.7554/eLife.102475
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.