Shaft reversal changed top 10% positive correlation hubs in 35 of 36 SEEG epilepsy patients for 1–80 Hz signals

We tested whether shaft subtraction order changes SEEG hubs based on positive Pearson correlation strength. The public dataset was OpenNeuro ds004100 version 1.1.3. We used two interictal records for each of 36 epilepsy patients. Hubs were the top 10 percent of nodes. Whole-shaft reversals changed 1–80 Hz hub sets in 35 patients. Across 4096 uniform sign choices per patient, mean replacement was 33.95 percent. The 95 percent patient bootstrap interval was 30.68–37.12 percent. Absolute correlation strength stayed the same. We derived exact attainable strength bounds for each node and saved the signs that attain them. Fresh public EDF range checks passed for four records. A separate numerical rerun matched all six result files and 4968 saved arrays. The package preserves 72 covariance inputs, code, source versions, hashes, results, figures, and instructions. Uniform sign choices define an audit. They do not model laboratory practice. The bootstrap assumes independent patients. The study did not test seizure location or clinical outcome.

stereoelectroencephalography · SEEG · bipolar montage · shaft polarity · positive correlation · network hubs · epilepsy