Phase tests after bipolar reference give excess false positives for alpha coherence with shared contacts under contact independence in two SEEG benchmarks

This study tests alpha coherence between bipolar channels that share one contact. It uses spectra from two public SEEG recordings. Under independence of recorded contact phases, a phase null after bipolar reference gives false positive rates of 99.90% and 99.92%. A phase null before bipolar reference gives 5.37% and 4.86%. The nominal rate is 5%. These results concern a defined spectral model. They do not identify false neural connections in the recorded data. The package contains code, a proof, source hashes, exact Fourier data, result tables, and instructions.

SEEG · bipolar reference · magnitude squared coherence · phase randomization · contact independence · false positive rate · spectral analysis