Abstracts

Rationale: Elucidation of seizure initiation, propagation, and termination dynamics would contribute greatly to seizure prediction, intervention, and prevention research. Some theories implicate a seizure focus or hippocampal subfield for seizure initiation while others point to a diffuse synchronous network spanning the hippocampus. Seizure termination has been the recent focus of several papers and reviews focussing on possible mechanisms from membrane to brain nuclei levels. Methods: Spontaneous seizure is a challenging phenomenon to observe in-vivo, requiring in-vivo microelectrodes, continuous electrophysiological recording, and behavioral observation. Data was acquired from a spontaneously seizing animal model of temporal lobe epilepsy using 32 anatomically verified (using histology and MRI) microelectrodes implanted bilaterally covering the CA1 and DG subfields. Granger Causality (GC), a novel connectivity measure borrowed from the field of economics, is used to quantify the dynamics between hippocampal subfields during spontaneous seizure. GC provides the magnitude and direction of influences between subfields. Behavioral correlates to seizure are provided by real time surveillance video of the animal. Results: Moments before seizure onset, highly synchronous activity is seen in the CA1 in one hemisphere. This activity propagates by entraining the dentate gyrus (DG) of the same hemisphere and then entraining the entire contralateral hemisphere during tonic seizure. Additionally, the primary driving influence is from the CA1 to the DG. The transition from tonic to clonic seizure is accompanied by highly synchronous causal activity between all regions in the hippocampus. After this transition, the animal experiences repetitive limb jerking and rearing for 10-20 seconds until the seizure ends. The dominant causal patterns between hippocampal subfields completely reverse after this transition where the DG now drives the CA1 in both hemispheres. Interestingly, this pattern is similar to patterns seen during normal exploratory behavior. The spike-wave discharges seen during termination of seizure may also suggest that the DG is regaining feedback inhibitory control over downstream subfields. Conclusions: This analysis reveals common initiation, propagation, and termination dynamics for all seizures. Seizure activity usually initiates within the CA1, but may originate from the unobserved entorhinal cortex. This activity then entrains the DG and then the other hemisphere and the animal is then in full tonic seizure. After seizure initiation, overall synchronization decreases but is maintained with mostly diffuse local activity. The seizure termination sequence begins when highly synchronous spike and wave discharges originating from the DG begin to entrain the CA1 subfield. These spike and wave discharges may indicate that the DG is gradually regaining feedback inhibitory control perhaps indicating the role of inhibition in seizure termination.