Distance accounts for measured SEEG CCEP reciprocity, while white-matter sites send more one-way responses

We analyzed expert-checked N1 labels from OpenNeuro ds005448 v1.0.0 for 8 SEEG patients and 5 ECoG patients. In SEEG, the reverse response occurred in a median 44% of pairs with a forward response and 9% without one. Yet after shuffling reverse labels within 5 mm distance and same/different-shaft strata, two-way pairs were 1.00 times the null (median, range 0.85–1.10; 2 of 8 patients p < 0.05; group sign-flip p = 0.73). ECoG was a positive control: 4 of 5 patients had excess reciprocity (ratio 1.13–1.38; p = 0.0005). In SEEG, adding the reverse response to a distance and shaft model increased held-out-sender AUC by a median 0.002; the increase was 0.007 at distances of 20 mm or more. For one-way pairs between white-matter and gray-matter-only sites, responses went from white matter to gray matter in 346 pairs and the reverse direction in 61 (7 of 7 patients with both site types). After accounting for distance, white-matter sites sent more responses in 7 of 7 patients, but did not consistently receive fewer. SOZ status gave no consistent direction. Forward and reverse latencies were not related after distance in SEEG (0 of 8 patients with p < 0.05), while they were related in ECoG (5 of 5, p < 0.005). The study used one center and dataset-provided N1 and tissue labels; counts were small for some long-distance comparisons. The results indicate that reverse responses add little information beyond distance for predicting an unstimulated SEEG site's outgoing response map, while tissue labels provide a sender prior.

SEEG · CCEP · reciprocity · N1 response · white matter · gray matter · brain-computer interface · stimulation mapping · OpenNeuro ds005448