Wake and N3 connection patterns agreed in SEEG voltage and fMRI after distance and site controls in 19 test patients and 16 participants

Wake and N3 sleep connection patterns showed positive agreement in separate human SEEG and fMRI studies after distance and recording-site effects were removed. The original gamma-power test in 38 SEEG patients was uncertain: mean pattern correlation 0.023, 95% person interval -0.022 to 0.070. We then added one voltage definition, 1-45 Hz, before any voltage score. Nineteen patients formed discovery; 19 other patients were held out for the voltage endpoint. Their gamma scores were already known. Before test voltage or fMRI scores, we fixed the transfer rule and code hashes. Test voltage agreement was 0.455 (0.323-0.575), positive in 18/19 patients. Separate BOLD agreement was 0.356 (0.296-0.419), positive in 16/16 participants. The SEEG two-representation Holm p was 0.00010 for voltage; fMRI p was 0.00004. Person bootstrap intervals and whole-person tests account for dependent edges. Independent-phase, mean signal, distance, rank, and later-window checks retained positive voltage and BOLD estimates. Added shaft, region, and axis controls also retained them. Gamma power had low split-window reliability, so its uncertainty cannot establish loss of a neural pattern. The shared finding concerns measured edge-pattern persistence, not equivalent signals, individual identification, direct connections, or causality. Cohort, coverage, nuisance data, and prior fMRI processing differ. Stable recording artifacts remain possible. The package preserves negative results, code, edge statistics, source versions and terms, figures, and repeat instructions.

human neuroscience · SEEG · fMRI · sleep · connection patterns · independent confirmation · negative results