Output phase controls rejected 90.95 percent of shared contact beta tests under independent contact phases from the first two seconds of SEEG in 33 patients

We tested two Fourier phase controls for two-second SEEG correlation. The inputs were fixed contact spectra from 33 patients in public OpenNeuro ds004100 version 1.1.3. At p <= 0.05, separate phases at bipolar outputs rejected 90.95 percent of shared contact tests under an independent contact phase model. Phases at contacts before subtraction rejected 4.95 percent. The equal-patient difference was 86.00 percentage points, with a descriptive 95 percent patient bootstrap interval of 80.50 to 90.45. Separate contact tests, high gamma, and a second window were also checked. This study extends a prior circular shift result to Fourier phases and short records. The controls test different null models. These are method checks with known model truth. They do not establish neural links or a gain in a working brain-computer interface. The package contains exact digital inputs, source pins and hashes, model test results, three figures, reusable code, and a complete numerical replay check.

stereoelectroencephalography · Fourier phase randomization · bipolar montage · correlation calibration · signal processing · OpenNeuro