[BSPC V20/Draft] Independently protocol-locked nine-person BNCI2014_001 EEG and third-cohort external manuscript

[lindicaphxag-tech]

NEW data and publication milestone, not just another PDF version. BNCI2014_001 (BCI Competition IV 2a) nine subjects, 2592 trials, 22 EEG channels independently executed in Actions run 38196958135 under a repository-committed BEFORE first run protocol. All source-only lambda choices SHA-locked before any held-target AUC evaluation. Original source macro 0.789051783, proper Gaussian source-mean predictive 0.783414781, historical held-validation-person class-centred covariance 0.772387260; LW 0.752722, OAS 0.752829; source-macro minus proper +0.005637 AUC (subject bootstrap descriptive CI [-.00619,.01793]). No formal superiority/equivalence and not independently registered. Includes standalone independent nine-person archive/hash/montage/label/AUC auditor, original source-only homotopy theory, new V20 complete article and CI/full PDF build. Unseen data status cannot be guaranteed across all external research; no claim of blind registered trial, no acceptance/submission.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}