A complement rule with an edge limit reconstructs 7,185 stored visible probabilities in ds004917 version 1.0.1 within 0.0001

We audited all 62 event tables in OpenNeuro ds004917 version 1.0.1. A complement rule with an edge limit reconstructed all 7,185 stored visible probabilities. The maximum absolute error was 0.00002. The rule uses a visible area limit of 0.0125 for ambiguity rows. A complement substitution alone failed on 791 ambiguity decisions. The stated metadata equation failed on all 3,595 ambiguity decisions, as reported in prior work. Its inverse produced 2,585 values outside [0, 1] and 1,000 cue direction reversals. The edge constant came from inspected failures in the same dataset. No test set was reserved. The result describes stored fields. It does not establish the physical display or a brain effect. Exact rational checks passed. Offline and fresh source rebuilds matched all numerical output bytes. The package preserves original files, retrieval hashes, source code, row results, figures, plans, and reproduction instructions.

decision making · ambiguity · probability cues · OpenNeuro · fMRI · EEG · data audit · reconstruction