In 8 SEEG patients, white-matter stimulation pairs had 2.53 times the odds of threshold-defined CCEPs at ≥10 mm after distance correction

Question: Does single-pulse stimulation of a SEEG pair with a white-matter (WM) contact cause cortico-cortical evoked potentials (CCEPs) at more channels than stimulation of a gray-matter (GG) pair, at the same distance? The answer helps to select stimulation sites for BCI write-in. Data: OpenNeuro ds005448 v1.0.0 (STReEF), 8 SEEG patients, first SPESclin run. We used HTTP byte-range requests to get only the stimulation windows. Method: We used bipolar channels and averaged 4 pulses per pair (2 per polarity). The response statistic is the signal range in 10-60 ms, normalized per channel. The threshold is the 99th percentile of the same statistic in a pre-stimulus window. We used a logistic model with patient intercepts, log distance and tissue terms. We used a cluster bootstrap over stimulation pairs. Results: The data have 278 stimulation pairs and 22,231 pair-channel rows at 10 mm or more. Response probability decreases from 0.59 at 10-15 mm to 0.04 at 40-50 mm. The pre-stimulus rate is 0.004-0.014. The odds ratio for a pair with at least one WM contact is 2.53 (95% CI 1.77-3.70). The odds ratio for a WM recording channel is 1.21 (1.08-1.39). The WM effect is 2.5-2.6 at 20 mm or more, at 30 mm or more, for other shafts only, and for WG and WW pairs separately. WM pairs had a higher response fraction than GG pairs in 6 of 6 patients with both classes (sign test p = 0.031). Distance alone predicts responses in left-out patients with AUC 0.75-0.90. Tissue adds about 0.01 AUC. Limits: Only 4 pulses per pair. The detector is not validated against expert labels. Tissue class is confounded with patient, so the result depends on 6 patients. Residual artifact can have an effect. One center and patients with epilepsy only.

SEEG · single-pulse electrical stimulation · cortico-cortical evoked potentials · white matter · gray matter · epilepsy · brain stimulation · response probability