Limits from four early pulses covered 28 percent of complete SEEG mean responses at 20 to 100 ms, averaged across five patients

We tested limits for the mean of four later SEEG stimulation pulses. The fit used four early pulses. Public OpenNeuro ds004457 1.0.2 supplied 134 sites in five patients. The response window was 20 to 100 ms. At the nominal 95 percent level, Student t limits covered 94.27 percent of individual voltage coordinates. They covered every measured coordinate at 27.67 percent of sites. Each patient had equal weight. Bonferroni limits covered every measured coordinate at 96.64 percent of sites. Their mean width was 3.61 times the mean width of the limits for each coordinate. A before pulse check also showed a large gap between coordinate and complete response coverage. The package provides a future mean interval tool, compact pulse data, coordinate forecasts, four figures, and source hashes. A fresh raw rebuild matched all saved inputs. Independent calculations and an exact replay checked the results. The statistical claim needs independent normal pulses with a common mean and covariance at each site. The cohort is reused and has only five patients. The result does not establish coverage for a new patient or an adaptive interface.

SEEG · intracranial EEG · electrical stimulation · future mean · prediction intervals · simultaneous coverage · Bonferroni correction · uncertainty