In one acute sheep recording, 120 µm micro-SEEG contacts were redundant at 1–30 Hz but less redundant for multi-unit activity

We analyzed one acute intraoperative recording from an anesthetized sheep using a 64-contact micro-SEEG shaft with 120 µm pitch. Of 44 contacts retained for impedance below 1000, set B had 23 contacts and set A had 21. Across 158 inter-pulse segments, the 1–30 Hz signal in set B was highly correlated at 120 µm (r = 0.942, 95% CI 0.934–0.948) and remained correlated at 960 µm (r = 0.907); its participation ratio was 1.22 with 23 contacts. The 500–5000 Hz multi-unit activity envelope was less correlated: r = 0.575 at 120 µm, 0.42 at 240 µm, 0.334 at 960 µm, and 0.26 at 2400 µm. Its participation ratio rose from 2.54 at 960 µm to 5.18 at 240 µm and 5.95 at 120 µm, so reducing pitch from 240 to 120 µm added 15% in this measure. At the same depth, cross-set correlation was 0.906 for low frequencies and 0.174 for multi-unit activity. Set A showed an anomalously shared low-frequency signal, so distance and dimensionality results use set B. Evoked responses to spinal stimulation were weak and did not establish a pitch limit. The unit scale is uncertain, the inter-pulse segments may include late evoked activity, and this single-animal, single-shaft observation should not be generalized to humans or chronic implants.

micro-SEEG · intracranial EEG · electrode pitch · spatial correlation · participation ratio · multi-unit activity · sheep · neural recording