Independent contact controls from two SEEG records retained positive beta window power correlation after pairwise orthogonalization when bipolar signals shared a contact

We tested whether pairwise instantaneous orthogonalization removes beta power correlation when bipolar SEEG signals share a contact. We used both resting records in public OpenNeuro ds005670 v1.0.0. In 100 draws with independent contact window permutations and phases, the median signed corrected correlations were 0.1370 and 0.1600 for shared contacts. The corresponding medians for separate contacts were -0.0019 and -0.0024. The estimator correlates two-second mean powers in the 13 to 30 Hz range. An analytical Gaussian example and independent simulations show how changing shared-contact power can retain covariance. Fixed-amplitude phase controls and Gaussian controls behave differently. The package preserves checked beta inputs, draw arrays, results, figures, source hashes, code, and raw retrieval instructions. Every input coefficient and quality mask was reproduced from the pinned raw files. Secondary checks use smaller stated draw counts because of the run time limit. The result describes conditional controls from two epilepsy records. It does not identify false connections or clinical effects in recorded data.

stereo-EEG · power-envelope correlation · orthogonalization · beta oscillations · contact reuse · signal processing