After 20 ms blanking, future reflection changed symmetric 7 Hz phase more before pulses than controls in eight STReEF patients

We tested whether future voltage changed a phase estimate before an electrical pulse. We used public SEEG data from eight STReEF patients, 96 conditions, and 960 pulses. The sample had 6,803 readout-condition rows. Each row was a remote bipolar pair in one condition. The filter was a symmetric complex Gaussian projection at 7 Hz. The estimate was 9.77 ms before each cut. A linear bridge replaced the first 20 ms after each cut. We then replaced future voltage with reflected earlier voltage. All pre-cut samples stayed fixed. The mean absolute phase change was 59.73 degrees at pulses and 56.04 degrees at control times. The paired difference was 3.69 degrees. Its 95 percent patient bootstrap interval was 2.85 to 4.54 degrees. All eight patient differences were positive. The exact two-sided patient sign-flip test gave p = 0.0078125. Control estimates also had large future dependence. The result measures sensitivity to one defined voltage replacement. It does not give error against true brain phase. It does not test an online device. The package has three figures, all phase products, 960 audit traces, checked source identities, code, and retrieval instructions.

phase estimation · temporal smearing · SEEG · electrical stimulation · future dependence · signal processing · brain-computer interface · patient bootstrap