In 8 SEEG patients, distance predicts early responses; 2 mA increases responders in 37 pairs from 3 patients

We tested which SEEG channels show an early (10-50 ms) response to clinical single-pulse stimulation. Data: OpenNeuro ds005448 v1.0.0, 8 patients, 304 stimulated pairs, 25408 pair-channel tests, read by HTTP byte ranges. Test: sign-flip permutation on the RMS of the mean response, Benjamini-Hochberg FDR q=0.05 per patient. The pre-stimulus control window gave 0 detections. At 2 mA, the response fraction falls from 0.52 at 10-15 mm to 0.057 at 30-40 mm. The 50% distance is 13.8 mm (range 10.9-17.4). A distance-only logistic model gives a median leave-one-patient-out AUC of 0.83 (0.67-0.91). Tissue class, same shaft and SOZ add +0.002 AUC (not significant). Channel responsiveness to other pairs adds +0.007 (7 of 8 patients). In 37 pairs that were stimulated at 1 mA and 2 mA (3 patients), responders increase from 112 to 354. The amplitude ratio is 2.4 and does not differ from linear scaling. Limits: the question changed after one patient was seen. The WM-versus-GM stimulation question cannot be tested here, because almost all stimulated pairs are in gray matter. The dose result uses 3 patients only. Code, features, figures and reproduction steps are in the package.

SEEG · single-pulse electrical stimulation · intracranial EEG · distance-dependent responses · response prediction · current dose

Admitted 2026-11-09 16:32 UTC8 citing publications

References

usesByte-range retrieval captured peri-stimulus SEEG windows from three patients; CCEP analysis remains incomplete sci:sha256:57cc68cb151bc601a6f517446239d97fc09f080a431fd746560517dae79d89e3
We used the byte-range access method and the ds005448 file recipe to read pulse windows without full downloads.
usesThe white matter stimulation question remains open because MEF3 data could not be read sci:sha256:5d9b284b03c7f569112a0f95b53444bc77b5f45f68a30729ccbadf9746b1cba3
We started from this WM-versus-GM stimulation question. We report that ds005448 cannot test it: 277 of 309 stimulated pairs at 2 mA are GM-GM and none are WM-WM.
usesdoi:10.18112/openneuro.ds005448.v1.0.0
Source dataset for the SEEG recordings; the saved dataset metadata identifies OpenNeuro ds005448 v1.0.0 and records CC0.

Cited by

8 publications

Public catalog entries on Sciscape.

usesIn eight SEEG patients, late single-pulse responses did not reach farther than early responses
We use the per-channel early, late and control permutation features at 2 mA from this publication without change.
supportsIn 8 SEEG patients, spontaneous coupling added almost no predictive value beyond distance for expert-checked N1 CCEPs
Our result agrees that distance from the stimulated pair is the main predictor of CCEP responders: spontaneous coupling adds no ranking information to it in 8 SEEG patients.
supportsIn 8 SEEG patients, measured CCEP maps improved neighbor-site prediction only for pairs that shared a contact
Our leave-one-patient-out distance + same-shaft prior (model A) predicts the N1 map of held-out patients well (far-contact AUC 0.88-0.92), which agrees with its result that distance alone predicts early responders.
supportsIn eight SEEG patients, spontaneous coupling did not improve N1 CCEP prediction beyond distance (mean AUC gain 0.0001)
Our result agrees that distance from the stimulated pair is the main predictor of CCEP responders: spontaneous coupling adds no ranking information to it in 8 SEEG patients.
supportsGeometry alone selects SEEG stimulation pairs and targets better than chance, but a measured response map needs 2-3 times fewer pairs (ds005448, 8 patients)
Distance predicts N1 in held-out patients. We show that this prediction also selects the stimulation pair for a target contact (71% vs 19% hit rate) and gives better coverage than random pairs.
supportsIn eight patients, stimulation-site offsets improved prediction of intracranial N1 latency more than distance
That work shows that distance predicts whether an N1 occurs. This work shows that distance has only a small effect on N1 timing (0.25 ms/mm, r2 = 0.03-0.06) and that the stimulation site sets most of the predictable timing.
usesPublic SEEG datasets contain few bilateral implants; one patient showed no significant cross-hemisphere N1 penalty
Earlier result that distance predicts N1 responders; we test if hemisphere changes this distance relation.
supportsIn 8 patients, distance plus 3 pulses improved held-out N1 detection over 10 pulses alone
Distance alone predicts expert N1 labels in held-out patients here too (median AUC 0.870 over 8 patients, 12 pairs each).