Abstracts

Rationale: A recent computational model (Stacey, et al. J Neurophysiol. 2009) predicts that increased random synaptic input, in the presence of pathological coupling, can provoke epileptiform activity. We test this prediction in vitro, and compare the results with simulated parameters. These two effects, synaptic noise and coupling, potentially play an important role in seizure generation, and may provide new avenues for seizure control.Methods: Intact septo-hippocampal formations of P5-7 rats were placed in submerged chambers isolating the septum from the intact hippocampus, while maintaining axonal connectivity. We recorded simultaneously from field and patch clamp electrodes within the hippocampus. In the first experiment, we placed the hippocampus in zero-magnesium solution to provoke spontaneous bursting, and monitored the activity prior to each burst. We compared the levels of noise with the computational model predictions. In a second experiment, we induced zero-magnesium bursting, then raised the magnesium concentration in the hippocampal chamber until the bursting stopped completely. We then modulated the synaptic noise arriving in the hippocampus by adding either high potassium or DC current to the septum. Since efferent septal axons pass into the hippocampal chamber but the media are isolated, this method allows us to change the afferent synaptic activity in the hippocampus without altering other cellular parameters. We modulated ephaptic coupling with either mannitol or hypotonic solution in the hippocampal chamber. Several levels and combinations of noise and coupling modulation were tested.Results: In zero-magnesium recordings, synaptic noise increased prior to spontaneous bursts in the majority of cells. The levels of noise were similar to those predicted in the model. Synaptic noise levels varied much more than the local field potential, which was often unchanged. When epileptiform bursts were halted by increased magnesium, synaptic noise and/or hippocampal coupling were able to generate bursting activity. The bursting could be stopped and started very quickly and reversibly by changing these parameters. This finding was robust, occurring in the majority of tested cells. Bursting also occurred when subthreshold levels of synaptic noise and coupling were combined, a major prediction of the computer model. Conclusions: As predicted in the computer model, these results suggest that synaptic noise and coupling play a crucial role in the generation of epileptiform bursts, and can have synergistic effects. We hypothesize that it may be possible to modulate seizure activity by controlling these two effects, which are not common targets for antiepileptic therapy.