A 500 ms future buffer reduced theta phase label change compared with 50 ms in seven patients' rest SEEG with fixed reflection

We tested how future data change offline theta phase labels. We used 17 public rest SEEG files from seven patients. The labels used 4 to 10 Hz filtering, a Hilbert transform, and fixed raw reflection. Against a matched four second future reference, mean label change was 35.93 degrees with 50 ms of future data. It was 3.39 degrees with 500 ms. The mean decrease was 32.54 degrees, with equal weight for each patient. A descriptive patient bootstrap interval was 30.67 to 34.39 degrees. The decrease occurred in 7 of seven patient means. The exact sign flip probability was 0.015625, under an independent patient and symmetric null assumption. Fixed changes to reflection, past length, and extension length gave the same direction of group change. At 50 ms, 24.60% of success labels for a frozen forecast changed against the four second reference. At 500 ms, 2.35% changed. The reference defines processing stability. It does not establish biological phase truth. The package preserves native inputs, pinned sources, all query labels, code, tables, a figure, and complete numerical checks.

SEEG · theta phase · future buffer · offline phase labels · signal processing · phase sensitivity