Abstracts

Rationale: There is a focused effort on the clinical implementation of cortical implants for prediction and control of seizures. Tissue impedance, as a measure of glial swelling, has been reported as a potential feature for seizure prediction (Olsson 2006, Hochman 2012). It also mediates most other electrical measures used to quantify neural activity in chronic recordings. Furthermore, state of vigilance (SOV) is a confounding factor in seizure prediction and detection efforts (Mormann 2007). We therefore undertook to monitor tissue impedance in long-term recordings in epileptic rats and analyze its variations in time and with both SOV and seizure dynamics. Methods: We implanted Long-Evans rats with cortical screw and hippocampal depth recording electrodes along with a pair of 250 μm stainless steel stimulation electrodes. The stimulation electrode was implanted in the arc of CA3 and the current return electrode was placed 2 mm anterior outside the hippocampus. A single dose of tetanus toxin was injected into the ventral hippocampus to induce chronic spontaneous recurrent seizures over months. We used a current-controlled stimulator to apply polarizing low-frequency electrical field (PLEF) stimulation every 15±5 seconds with 2 second sinusoidal packets of frequencies below 20 Hz and peak current of 30 μA. Stimulation was applied for at least 24 hours in n=3 animals. Packet impedance was fit with a polynomial spline to reduce noise and provide a time-continuous impedance series. State of vigilance was scored and seizures detected using previously established algorithms (Sunderam 2007). Results: In all animals, impedance rises during non-rapid-eye movement sleep (NREM), and falls during wake. The distribution of impedance slopes during NREM is significantly different than the distribution during wake, as determined by a two-sample Kolmogorov-Smirnov test.There is a small decrease in impedance during rapid-eye-movement sleep. In all animals, impedance falls during and after generalized seizures. In one animal, in addition to changes in impedance associated with SOV, we also observed diurnal trends, with impedance lower during the subjective day (lights-off) period of the animal and higher during the subjective night (lights-on) period. Further analysis is required to determine the independence of diurnal and SOV effects. Conclusions: We observe that the whole-cell brain impedance changes significantly over time, and has distinct variations as a function of, and on the time-scale of, sleep-wake cycles. These impedance changes should be accounted for both if impedance is used for seizure prediction as well as in the design and implementation of brain stimulation. Impedance changes may also impact other brain state measures especially used for seizure prediction. Further investigation into SOV related changes in impedance may help to shed light on the observed relationship between sleep and seizure states. Olsson T. et al. Neuroscience 140:505-515, 2006. Hochman D.W. Epilepsia, 53(Suppl. 1):18-25, 2012. Mormann F et al. Brain 130(Pt 2): 314-33, 2007. Sunderam S et al. J Neurosci Meth (163):373, 2007.