Fixed theta phase models gave -0.004 percent mean gamma amplitude gain between SEEG files at 72 contacts from 36 patients

We tested transfer of a theta phase model to another interictal SEEG file. We used OpenNeuro ds004100 version 1.1.3 and two fixed contacts from each of 36 patients. The model used 4 to 8 Hz phase to predict centered log amplitude at 70 to 110 Hz. Mean gain between files was -0.00403 percent. The 95 percent patient bootstrap interval was -0.04133 to 0.05098 percent. Mean gain within a file was 0.02618 percent. The difference was 0.03020 percentage points, with Holm p=0.07175 for two planned tests. Only 8 patients had positive mean gain between files. The observed means were above all 39 shifted timing controls. These controls are descriptive and do not establish an exact null. All five fixed checks gave negative mean gain between files. All 72 source files passed fresh record checks. Direct predictions reproduced all 576 transfer scores, and a full repeat gave identical numerical outputs. The 35 MB package preserves native traces, source pins, code, results, figures, and repeat commands. The result concerns this fixed model and these contacts. It does not show that phase-amplitude coupling is absent or identify its neural cause. Patient inference requires independence and stated sampling assumptions.

phase-amplitude coupling · theta phase · gamma amplitude · stereoelectroencephalography · intracranial EEG · epilepsy · cross-file transfer · signal stability