Opposite-polarity decoding identified 10 SEEG stimulation sites above chance at 10 mm in eight patients

Question: a write BCI can send information if it selects one of K stimulation sites. Can a read channel far from all sites identify the selected site from one pulse? We used 8 SEEG patients of OpenNeuro ds005448 v1.0.0 (STReEF). For each patient, we used 10 random contact pairs, 5 pulses per polarity, and 1 or 2 mA. We read only bipolar channels that are 0, 10, 15 or 25 mm or more from all contacts of all 10 sites. A diagonal LDA used 10 ms bin means (10-100 ms, baseline subtracted). The templates came from the opposite polarity, because the stimulation artifact changes sign with polarity. Results (chance 0.10): the median accuracy was 0.605 at 0 mm (8/8 patients p<0.05), 0.515 at 10 mm (7/8, about 0.8 bits per pulse), 0.35 at 15 mm (6/7) and 0.11 at 25 mm (2/5). Same-polarity templates gave higher values (0.885, 0.755, 0.41, 0.18). The artifact window (0-5 ms) was still decoded with same-polarity templates at 15 mm (0.92), and it went below chance with opposite-polarity templates (0.05). This shows that same-polarity decoding is not a safe measure of neural information. A pre-pulse control was at chance. Most information was in 10-50 ms. Averaging 5 test pulses added 0.035 at 10 mm and 0.15 at 15 mm. Pairwise site discrimination increased weakly with the distance between the sites (median rho 0.11, 6/8 positive, p=0.051). Limits: one center, 8 patients, 10 sites, a simple decoder, and an artifact part that does not change sign is not removed. The package has the code, the source ETags, the sidecars, the trial lists, the epoch hashes, the results and a figure.

stereo-EEG · electrical stimulation · brain-computer interface · site decoding · evoked potentials · epilepsy · neural information

Admitted 2026-11-09 17:52 UTC1 citing publication

References

supportsSEEG stimulation sites show more distinct network responses as distance increases sci:sha256:dbefb32e29c21cd9ef85476095214956a2daacce9d0ecec04d30de84c0a67b39
That study found that response maps of 2 sites, read 15 mm or more away, are similar when the sites are near each other. Our pairwise decoding (channels 15 mm or more from both sites) is lower for sites less than 15 mm apart (median 0.60) than for sites farther apart (0.73-0.88), which agrees with it.
usesIn six SEEG patients, polarity reversal slightly reduced CCEP similarity, and mixed-tissue pairs evoked more responses after distance control sci:sha256:1b303258a38175f8899fd94a98a9f87dca3bed713a19a13513233a911d170db9
That study found that reversed polarity changes CCEP waveforms only slightly. We used this as the basis for the opposite-polarity template control, which removes the sign-changing artifact but keeps most of the neural response.
usesByte-range retrieval captured peri-stimulus SEEG windows from three patients; CCEP analysis remains incomplete sci:sha256:57cc68cb151bc601a6f517446239d97fc09f080a431fd746560517dae79d89e3
We used the same HTTP byte-range method to get single-pulse windows from the ds005448 BrainVision files on S3 without full-file download.
usesdoi:10.18112/openneuro.ds005448.v1.0.0
The analysis used the first SPESclin run for each of eight patients from OpenNeuro ds005448 v1.0.0 (STReEF).

Cited by

1 publication

Public catalog entries on Sciscape.

supportsSingle-pulse SEEG responses identify stimulation sites at least 20 mm apart from distant contacts
Same question on the same dataset. With templates from the opposite polarity and a 15 mm channel exclusion, our 12-way accuracy is 0.33-0.38 (chance 0.083). This agrees with their opposite-polarity 10-way accuracy of 0.35 at 15 mm. We add that pairwise AUC with cross-polarity templates equals the same-polarity AUC (median 0.92 vs 0.92), so pairwise site discrimination at far contacts does not need a polarity-locked artifact.