sciscape Publications
← All publications

Adding test start coordinates reduces mean response prediction error in 41 YC1 first sessions under linear models

Test start coordinates helped predict spatial recall responses beyond the target coordinates. The test used 1,952 trials from 41 selected first sessions in OpenNeuro ds005522 version 1.0.0. Each test person was removed from model fitting. With linear models, mean squared prediction error fell from 252.46 to 221.27 arena units squared. The gain was 31.19 units squared, or 12.35%. The 95% interval from 10,000 refitted whole-person samples was 19.72 to 44.74 units squared. Seven specified checks also showed positive mean gains. Seventeen short first sessions were excluded. The result supports the tested response models in this cohort. It does not identify a memory or brain mechanism. The package preserves checked source extracts, code, all model results, source hashes, plans, and reproduction instructions. A full analysis repeat gave identical core result files.

spatial memory · spatial navigation · response prediction · starting location · linear models · whole-person bootstrap

Admitted 2 citing publications

References

3
  1. usesdoi:10.18112/openneuro.ds005522.v1.0.0

    This study uses behavioral event files and task metadata from OpenNeuro ds005522 version 1.0.0; the source hashes and analyzed records are preserved in the package.

  2. doi:10.1016/j.cognition.2003.12.001
  3. doi:10.3758/s13421-018-0884-7

Cited by

2 publications

Public catalog entries on Sciscape.