Outer and inner sEEG shaft halves gave similar offline key press detection in eight subjects

In eight subjects from OpenNeuro ds004473, we compared movement versus rest detection from the inner and outer halves of stereoelectroencephalography (sEEG) shafts. Shrinkage linear discriminant analysis used high gamma (70–150 Hz) and beta (13–30 Hz) bipolar log band power, trained on trials 1–100 and tested on later trials. Mean test AUC was 0.791 for inner pairs, 0.786 for outer pairs, and 0.828 for all pairs. The preregistered difference, outer minus inner AUC, averaged −0.006 (median −0.004; four of eight subjects positive; exact two-sided sign-flip p = 0.75). Secondary results were also inconclusive: HG-only difference −0.035 (p = 0.30), depth versus single-pair HG separation mean Spearman rho +0.005 (p = 0.91), and 0.42 of the eight best training pairs were outer (p = 0.25 versus 0.5). The data therefore showed no detectable difference between halves, but this null result does not establish equality. Each half was about 0.04 AUC below all pairs, suggesting that pair count mattered more than shaft depth in this offline, cued task. With eight subjects, only large effects were detectable. The analysis used clinical implants in epilepsy patients and movement-versus-rest decoding; results may differ for choice decoding.

stereoelectroencephalography · sEEG · brain-computer interface · movement decoding · high gamma · beta power · linear discriminant analysis · OpenNeuro