In eight sessions, 33 of 183 later SEEG means that passed a 0.8 shape check failed a voltage size check at a factor of 1.25

We tested whether an early shape check also selected SEEG responses with stable voltage size. We used public OpenNeuro ds004696 version 1.0.1 data from eight patients with epilepsy. The early check selected 196 of 886 bipolar stimulation-to-readout pairs. It used the first five good pulses and a correlation cutoff of 0.8. The template was their mean. Of 183 final five-pulse means that passed the shape check, 33 had an RMS voltage ratio outside 0.8 to 1.25. With equal patient weight, shape agreement was 94.5% and joint shape and size agreement was 80.4%. The shape-only rate among selected pairs was 14.1%, with a 95% patient bootstrap interval from 8.9% to 21.7%. Individual later pulse joint agreement was 51.5%. Size failures fell to 13 at a factor of 1.5 and to 2 at a factor of 2. Every patient had a size failure at the main tolerance. One patient supplied 19 of the 33 failures. Exact waveform products gave separate shape, size, projected gain, and prediction error measures. All archived input hashes passed. A complete source derivative check covered one patient. The package preserves code, data products, source identities, result tables, figures, and reproduction instructions. The result applies to selected limbic readouts in the same clinical sessions. It does not establish perception, benefit, safe dose, or performance in a new session.

stereo-EEG · electrical stimulation · response shape · voltage size · brain-computer interface · intracranial electrophysiology · waveform agreement