State-average joint squared amplitude exceeded a spectral reference in 16 fMRI participants but was uncertain in 56 SEEG patients

We tested joint squared amplitude between distant cortical sites in separate human SEEG and fMRI studies. An exact coherent-phase reference retains the complete cross-spectrum. It removes finite-spectrum fourth-moment terms without simulation. We used 56 MNI atlas patients and 16 people from a released sleep fMRI study. Each person had matched wake and N3 windows. The planned shared positive result failed. Mean state-average SEEG excess was 0.0074, with a 95% person interval from -0.0021 to 0.0175. The fMRI mean was 0.2976, with interval 0.0719 to 0.5905. The planned shared N3 increase also failed. The test uses equal person weights and keeps negative results. It does not equate voltage power and BOLD or identify causal coupling. Cohort differences, short windows, amplitude distributions, and missing fMRI confounds limit interpretation. The package supplies the formula, proof, code, source records, results, figures, and repeat steps.

SEEG · fMRI · sleep · cross-spectrum · fourth moment · phase surrogate · cross-modal generalization