A five-pulse 1 mA SEEG template with one shared scale had more 2 mA forecast error than a three-pulse 2 mA template with a noise scale in three people at 20–500 ms

We tested whether a stored 1 mA SEEG template could reduce 2 mA response calibration. The source template used five pulses. One shared scale was fitted from three target pulses for each ordered condition. The comparator used those same target pulses and a noise scale for each readout. Two later target pulses were tests. We used STReEF release 1.0.0 from OpenNeuro. The fixed selection had three people, 31 sites, and 54 ordered conditions. At 20 to 500 ms, source transfer had 34.1%, 291.1%, and 23.7% more mean squared voltage error. The mean person increase was 116.3%. Source-energy scores gave a 32.5% mean increase. Transfer also had more error than a target template with one shared scale. Current and time changed together. One person also had separate runs. This result does not isolate a causal current effect. It tests one offline forecast rule on a reused cohort. The package preserves code, exact waveform products, native audit waveforms, four figures, provenance, and reproduction commands. All 540 retrieved source ranges matched the input hashes. All product arrays, edge tables, and eight result tables passed reconstruction checks.

SEEG · electrical stimulation · brain–computer interfaces · forecasting · calibration transfer · intracranial EEG · STReEF