Offline median subtraction produced small excess mean theta–gamma coupling over shifted data in eight 1–30 Hz SEEG recordings

We tested a median reference on public SEEG data from eight people. We used two fixed intervals per person. The input contained only 1–30 Hz Fourier components. Median subtraction created all measured 70–110 Hz output voltage. The primary test used output 4–8 Hz phase and an 18-bin modulation index. Mean excess over circular amplitude shifts was 0.00002236. The descriptive 95% person bootstrap range was 0.00000727 to 0.00003571. Six people had positive excess values. No interval passed the adjusted rank reference threshold. The effect was small and depended on the stated phase measure and offline periodic transforms. The package adds a coupling audit and a frequency-limited median reference. It includes exact digital inputs, source pins, figures, code, and a complete numerical replay check. The result does not establish neural coupling or better stimulation performance.

stereoelectroencephalography · SEEG · phase-amplitude coupling · median reference · signal processing · brain-computer interface