Human SEEG and BOLD lag scores did not support a shared N3 decrease in 56 patients and 16 records

We tested whether N3 sleep reduces lag asymmetry in separate human SEEG and fMRI studies. Fixed rules retained 56 epilepsy patients and 16 BOLD records. We used matched within-person sites and windows, separate physical lag scales, and person-level inference. The score was the mean squared difference between positive and negative lag correlations. We also derived its exact independent-phase reference. Mean N3-minus-wake raw changes were +0.000620 in SEEG and +0.018356 in BOLD. Their 95% person intervals were -0.001066 to +0.002236 and +0.001938 to +0.034720. Holm p values across four primary tests were 1.0 and 0.1915. Phase-excess changes were also uncertain. The fixed shared-decrease rule failed. After these results, an exploratory Gaussian-information transform changed the BOLD mean contrast to -0.003315, with interval -0.008402 to +0.000555. An exact factor audit accounts for this numerical change. It does not identify a biological mechanism. Mathematical controls show why neither phase excess nor a zero transformed score measures full-process reversibility. Cohort differences, sparse SEEG coverage, short clips, and missing BOLD motion and demographic data limit the claim. The BOLD source overlaps prior sleep literature; this is not new independent BOLD confirmation. The package preserves code, source versions and hashes, every inclusion decision, negative results, figures, and repeat instructions. A fresh primary replay matched nine numerical files exactly.

human neuroscience · SEEG · fMRI · sleep · time reversal · lag correlation · spectral reference · reproducibility