Beta coupling gave a mean AUC of 0.481 for source seizure onset marks within name groups in 19 public SEEG records

We tested beta coupling against source seizure onset marks in 19 selected public SEEG records. The data came from OpenNeuro ds003876 version 1.0.2. We used adjacent contact differences and two minutes per patient. We compared marked and unmarked sites within contact name groups. A positive site had soz=yes at either endpoint. Other sites were unmarked. Each group had equal weight within a patient. Each patient had equal weight in the main result. The mean AUC was 0.481. Its descriptive 95 percent patient bootstrap interval was 0.372 to 0.590. The exact two-sided sign reversal probability was 0.746719 under independent patients and symmetric null contrasts. The planned power and lower-contact-number checks gave mean AUCs of 0.680 and 0.795. These results concern source marks. Source n/a does not prove a negative clinical label. The primary result did not give evidence for higher mark ranks from higher beta coupling. It does not prove absence of coupling or validate seizure location. The package contains code, input features, source metadata, pinned retrieval recipes, numerical results, figures, and checks. Fresh public EDF ranges rebuilt all 4,222,800 input coefficients exactly at complex64 precision. This is a new analysis of an existing cohort, with no new patients.

stereoelectroencephalography · SEEG · epilepsy · seizure onset marks · beta coupling · imaginary coherency · within-group AUC · OpenNeuro ds003876