Phase-controlled ordinal regularity has mean onset AUC 0.480 under the supplied reference in two 20 s SEEG segments from each of 36 HUP patients

We tested whether ordinal regularity beyond a Fourier phase control ranks clinical onset contacts. We used OpenNeuro ds004100 version 1.1.3, 36 pure SEEG patients, and two 20-second segments per patient. The cohort has 4,260 matched contacts. The fixed score is mean entropy from 19 phase controls minus observed permutation entropy. The primary mean patient AUC under the supplied reference was 0.480; its 95 percent patient interval was 0.433 to 0.529. The patient interval includes 0.5. This test does not establish an average ranking advantage for the fixed score. It does not establish equivalence. The paired AUC difference from raw regularity was -0.021 (-0.091 to +0.050). Descriptive median file rank agreement was 0.795 for raw regularity and 0.180 for the phase-controlled score. The package preserves code, source versions and hashes, metadata, ordinal counts, every control entropy value, patient results, figures, and reproduction checks. The phase control can change amplitude distributions and transients. The source reference and biological time gap are unknown. This is a fixed contact-label test in one reused clinical cohort. It does not establish a nonlinear brain mechanism or treatment benefit.

permutation entropy · phase controls · stereoelectroencephalography · seizure onset · ordinal patterns · contact ranking