Eight SEEG pairs on eight shafts gave lower mean relative 70–110 Hz test power in 36 epilepsy patients

We tested a limit of one bipolar SEEG pair per electrode shaft. We used public interictal records from 36 epilepsy patients in OpenNeuro ds004100 v1.1.3. Each rule selected eight pairs by power in a ten-second A sample. A separate ten-second B sample supplied test power. A and B were ordered by file name. Their time gap was unknown. The limit changed the mean clinical onset-label fraction from 21.9% to 20.5%. The paired change was -1.4 [-7.3, +4.2] percentage points, with a 95% patient bootstrap interval. This interval includes an increase and a decrease. Relative test power fell from 86.0% to 67.1%. Its change was -18.9 [-23.7, -14.2] points. A control gave both rules 16 distinct contacts. The power change remained -8.4 [-12.9, -4.5] points. The label change was +1.7 [-2.1, +5.9] points. The shaft limit gave no clear evidence of lower onset-label exposure. This does not prove equal exposure. All 72 source records were rebuilt from frozen public objects. Their features matched the inherited inputs. The package contains code, source receipts, data, patient results, and figures. Power is a surrogate. This study did not test BCI task accuracy or stimulation effects.

stereoelectroencephalography · SEEG · brain-computer interface · channel selection · power · seizure onset labels · electrode shafts