Adding pulse interval increased mean SEEG size forecast error by 0.13% at 20 to 500 ms in eight tests that excluded one person from shared slope fitting

We tested whether pulse interval improves a forecast of SEEG voltage size after stimulation. Public OpenNeuro ds004696 version 1.0.1 supplied data from eight patients. Each test used four calibration rows from a five-pulse budget. Shared slopes used only the other seven patients. The base model used before-pulse size and pulse order. Adding log interval increased mean squared log-size error from 0.140769 to 0.140955 at 20 to 500 ms. This is a 0.13% increase. Three patients had lower error. A descriptive patient bootstrap with model refitting gave a change range from -0.20% to +0.80%. This test gives no clear forecast gain. It does not prove that interval has no neural effect. The package has three figures, all forecasts and splits, model slopes, source code, and a raw retrieval recipe. Fresh checks matched 192 source files and all 4,732 voltage matrices. A complete raw rebuild gave the same feature table. This is an offline test of a reused clinical cohort.

stereoelectroencephalography · electrical stimulation · pulse interval · forecasting · forecast error · brain-computer interface · intracranial EEG