A wake to N3 change in local increment asymmetry was not confirmed across 74 SEEG patients and 16 fMRI participants

We tested whether N3 sleep changes local rise and fall in the same direction in separate human SEEG and fMRI cohorts. The primary measure was the signed third moment of increments, divided by their second moment to the power 3/2, after rank normal scores. We used matched within-person state windows and equal region and person weights. SEEG used 30-45 Hz power envelopes at a 1 s lag; BOLD used a 6.24 s lag. The SEEG N3-minus-wake change was +0.004606 (95% person interval -0.019184 to +0.027644; two-sided p = 0.69841). The fMRI change was -0.066996 (95% person interval -0.223631 to +0.055779; two-sided p = 0.45031). The fixed rule for cross-modal generalization failed. These results do not establish a shared direction of change. A result that includes zero is uncertain; it is not proof of no change. We did not select a new lag, subgroup, or outcome to obtain agreement. A pre-data simulation showed why a squared phase score alone cannot test general time reversibility. The package retains the plan correction, source records, code, person and site results, negative results, numerical controls, and figures. Short records, unequal spatial coverage, clinical SEEG sampling, unknown prior BOLD regression, and missing motion traces limit the scope.

human neuroscience · SEEG · fMRI · sleep · time reversal · cross-modal testing · signal asymmetry