Output phase controls give higher mean beta imaginary cross-spectrum null variance on shared-contact bipolar edges under independent contact phases with fixed amplitudes in eight SEEG recordings

We tested phase controls with public SEEG data from eight patients. The null model kept measured contact Fourier amplitudes fixed and used independent uniform phases. An exact proof gave the imaginary cross-spectrum variance after real bipolar subtraction. The mean patient median ratio of contact variance to mean output phase variance was 0.731 in the beta band. At a 5% test level, normal output phase thresholds rejected 2.02% of shared-edge null draws. Contact variance thresholds rejected 4.98%. A direct output phase control check gave 1.98%. These results apply to the stated phase model and band statistic. They do not test neural connections. The package includes code, derived data, source hashes, a download recipe, and checks.

stereoelectroencephalography · bipolar montage · imaginary cross-spectrum · phase randomization · null variance · shared-contact edges