White matter contact pairs did not show lower post-train deviations in 10 SEEG patients

This offline analysis asked whether bipolar stereo-electroencephalography (SEEG) recording pairs with a white-matter (WM) contact deviated less from baseline than gray-matter (GM) pairs at equal distance after a 5 s, 43–50 Hz stimulation train. It used archived per-pair post-train z-scores from sci 10d6f526, based on OpenNeuro ds006519 v1.0.0. In 10 subjects and 21 recordings with at least two WM-touching and two GM pairs, distance-adjusted WM coefficients for log mean |z| during 0–2 s after the last pulse had medians from −0.04 to +0.06 across six features; Holm-adjusted p = 1 for all. Results were similar at 2–8 s and without distance adjustment. For 70–150 Hz power, the median coefficient was +0.05 (bootstrap 95% CI −0.08 to +0.13; ratio 0.92–1.14); for line length it was +0.06 (CI −0.07 to +0.19). The distance positive control was weak: its coefficient was negative in 6–8 of 10 subjects at 0–2 s, and only alpha had uncorrected p = 0.0098 (Holm p = 0.059). The test therefore has low sensitivity and does not exclude WM effects smaller than about 15–20% in mean |z|. WM-touching pairs were pooled because WM-only pairs were rare. This is an offline, one-train-per-recording analysis of 10 of 18 NMR-evoking patients, using archived z-scores normalized by each channel’s baseline standard deviation.

stereo-electroencephalography · SEEG · white matter · gray matter · electrical stimulation · neural recovery · brain-computer interfaces · z-scores