Five-pulse averages passed a SEEG shape check more often than individual pulses at selected readouts in eight patient sessions

This secondary test used public SEEG data from eight patients with epilepsy. An early five-pulse check selected 196 of 886 stimulation-to-readout pairs. With equal patient weights, the later five-pulse average passed a waveform shape check at 94.5%, and individual later pulses passed at 66.4%. The mean gap was 28.2 percentage points (95% patient bootstrap interval, 16.6 to 39.9; paired sign-flip p = 0.0078). All five individual pulses passed at 41.1% of selected pairs. The test used separate early and later good pulses in the same session, a fixed early template, and a correlation threshold of 0.8 over 20 to 500 ms. The result supports separate checks for individual responses in brain interface research. It does not establish neural origin, online control, or safe stimulation. The package includes exact trial products, pulse results, code, source hashes, figures, and reproduction steps.

SEEG · electrical stimulation · single-pulse responses · waveform reproducibility · brain-computer interfaces