Alternating templates fitted to 16 pulses reduced SEEG variation at −1 to 2 ms but did not establish a mean reduction at 3 to 15 ms in 20 recordings

We tested whether an alternating pulse class template transfers from early SEEG pulses to later individual pulses. We used one recording from each of 20 people in OpenNeuro ds005169 version 1.0.0. We fitted the first 16 pulses and tested the next 32. Classes were consecutive pulse positions, with no verified hardware label for each pulse. Alternating-only subtraction reduced mean centered voltage variation by 74.44 percent at −1 to 2 ms. The 95 percent person bootstrap interval was 60.81 to 85.56 percent. At 3 to 15 ms, the mean reduction was 2.10 percent, with an interval of −1.09 to 5.55 percent. Eight recordings had a reduction. A fixed timing grid fitted from early edges gave a similar later result. Two fit pulses increased later variation by 28.53 percent on average. We also proved which persistent added signals each fixed template preserves under an additive model. All 20 raw file hashes and reconstructed inputs matched. The package contains a correction tool, source code, input epochs, results, figures, and reproduction instructions. Lower variation does not establish artifact-free neural voltage or better interface performance.

stereo electroencephalography · SEEG · pulse template · alternating stimulation · signal processing · voltage variation · brain-computer interface · intracranial stimulation