Abstracts

Rationale: In rodents, EEG recordings are typically performed with skull-penetrating electrodes placed in the epidural space, a technique that necessitates craniotomy and, when done with multiple electrodes, requires technically difficult and time-consuming fabrication and surgical implantation techniques. Moreover, craniotomies confound some experiments such as those studying mild traumatic brain injury (mTBI)-associated epileptogenesis. Therefore, it is necessary to develop convenient, cost-effective methods to record multiple channels of skull surface EEG.

Methods: We designed flexible printed circuit boards (PCBs) containing contacts to attach 16 recording electrodes as well as ground and reference electrodes. The PCBs were then produced by a commercial fabricator. We formed electrodes (d=0.7 mm, h=0.4 mm) on the PCB contacts using the flexible 2D material titanium carbide (Ti3C2Tx), MXene. The MXene arrays were tested in our previously described mice that heterozygously express a human epilepsy gene mutation (Gabra1(A322D)) and exhibit spontaneous spike wave discharges (SWDs). The arrays were attached to the intact skull through a scalp incision. For comparison, we also implanted separate mice with two standard epidural tungsten electrodes (d=0.1mm). Two-hour EEG recordings were performed weekly for three weeks and the spectral characteristics of the awake background and SWDs were determined. We then used the MXene electrode arrays to determine the effects of mTBI on physiology. One week after MXene array implantation, mice underwent a 2-hour baseline EEG study and then received either a single 40 psi closed-skull overpressure mTBI or sham exposure. They then underwent subsequent 2-hour EEG recordings at 1 hour, 1 day, and 7 days and the effects of mTBI/sham on the spatiospectral distribution of awake resting state spectral power and node degree, a graph metric of connectivity, were determined.

Results: Compared with two-channel tungsten electrodes, 16 channel MXene electrodes required significantly less time for fabrication (97±2 min vs. 179±17 min, P=0.004) and surgical implantation (54±2 min vs. 72±1 min, P< 0.001). During three weeks of recording, MXene electrodes were significantly less likely than tungsten electrodes to become dislodged (P=0.002). As expected, EEG recordings with intracranial tungsten electrodes showed higher voltage than those from surface MXene electrodes, but the two recording methodologies had similar distributions of spectral power during awake background and SWDs. Mild TBI, but not sham, reduced awake background 3 to 6 Hz (δ/θ) spectral power in the frontocentral electrodes by 140 µV