Pre-pulse 10 Hz magnitude did not predict CCEP N1 in 13 patients with 1 Hz SEEG trains

A state-triggered stimulator requires a pre-pulse signal that predicts response size. We tested whether pre-pulse 10 Hz magnitude predicted single-trial CCEP N1 in a third center, OpenNeuro ds006254 (UAB; 13 SEEG patients; 5 mA pulses in 1 Hz trains), using a plan hashed before calculating study statistics. We analyzed one train per patient and 545 responsive bipolar pairs. At 10 Hz, mean patient r was -0.002 (7/13 positive; one-sided exact sign-flip p=0.55); pooled unit r was -0.003 (95% bootstrap CI -0.024 to +0.019), excluding the ds005448 estimate +0.055 and overlapping the ds004696 interval. The 6 Hz, 15 Hz, and real-minus-sham comparisons were not significant after Holm correction; we do not claim a 15 Hz effect. Sham windows 0.25 s earlier gave positive associations of similar or larger size (10 Hz +0.031; 4 Hz +0.049), showing that a small positive r alone does not establish a stimulation-linked state effect. Across three centers, a positive 10 Hz association was reported in one dataset and not in two. In this dataset, 10 Hz magnitude did not provide a usable trigger for larger N1 responses. Interpretation is limited by the 1 Hz train design, possible overlap of the preceding response tail with pre-pulse windows, few pulses per unit, one train and stimulated pair per patient, correlated units, and different stimulation currents across centers. The analyzed source files are identified by their recorded URL and window hashes; the exact OpenNeuro release is uncertain.

CCEP · SEEG · brain stimulation · N1 response · pre-pulse magnitude · sham windows · brain-computer interface