In eight SEEG patients, spontaneous coupling did not improve held-out prediction of CCEP N1 beyond distance and shaft

We tested whether spontaneous pre-pulse SEEG coupling predicts the N1 response to single-pulse stimulation beyond distance and same-shaft status. In eight patients from OpenNeuro ds005448 v1.0.0, we analyzed 22,402 stimulated-pair/recording-contact observations, including 3,302 expert-checked N1 responses. For each patient, 40 one-second windows from the first SPESclin run supplied zero-lag 1–40 Hz signal coupling and 65–95 Hz Hilbert-envelope coupling. In five-fold cross-validation grouped by stimulated pair, adding both coupling features to distance and same-shaft status changed mean AUC by −0.0001 (95% bootstrap CI −0.0024 to +0.0016; 4/8 patients positive; exact sign-flip p = 0.95). The result was also near zero for other-shaft contacts more than 20 mm away (+0.0009; p = 0.84). The in-sample 1–40 Hz coefficient was positive in all eight patients, but coupling did not improve held-out ranking. A local-bipolar sensitivity analysis finished for only four patients and was descriptive. The result is limited to eight patients from one center, 40 seconds of between-pulse data per patient, zero-lag linear measures, and an incomplete bipolar check.

SEEG · CCEP · N1 response · spontaneous coupling · epilepsy · electrophysiology · cross-validation · target selection