Observed normalized 20 ms forecast error was lower after marked pulses in 59 matched hippocampal SEEG records from seven people

We tested a voltage reader around marked electrical stimulation in public depth SEEG data. We used OpenNeuro ds007095 version 1.0.0. We fitted each channel on the first 10 seconds of each record. The model used only past samples. Its normalized error was unchanged by constant voltage gain and offset. The fixed primary rule retained 59 records from seven people. The actual-to-control ratio of error changes was 0.898. The descriptive 95 percent person bootstrap interval was 0.815 to 0.980. The exact two-sided person sign-flip p value was 0.0625. Error fell in both periods. A stricter matching rule gave a ratio of 0.973. A 50 ms forecast gave 0.984. The records were selected by an implant detector. They cannot show that stimulation caused the change or improved a brain-computer interface. The package contains all 222 source records, source checks, code, model weights, forecasts, tables, four figures, and repeat checks.

stimulation · stereoelectroencephalography · SEEG · voltage prediction · normalized forecast error · brain-computer interface · OpenNeuro · matched control · ridge regression · responsive neurostimulation