Phase dependence of the termination of absence seizures by cerebellar input to thalamocortical networks

Abstract

Absence seizures are the most common form of epilepsy in children. They start and finish abruptly, last for 10–20 seconds and can be detected by generalised spike-and-wave discharges (GSWDs) in the electroencephalogram. These GSWDs are based on neuronal oscillations in thalamocortical networks, which can be caused by excessive inhibition in the thalamus or excessive cortical activity. Absence seizures can be triggered by switching of the normal asynchronous neuronal activity in thalamocortical networks to synchronised oscillations, and terminated by the reverse process, switching from synchronised oscillations to asynchronous activity. Experimental studies have shown that thalamic stimulation can disrupt oscillatory activity in thalamocortical networks. More recently, it was also found that optogenetic activation of neurons in the cerebellar nuclei (CN) can stop epileptic absence seizures and reset the oscillatory activity, for example in a closed loop system [1]. However, the underlying mechanism of the termination of absence seizures by CN stimulation is not yet clear. To investigate the mechanism of the termination of absence seizures by thalamic input from the CN we used computer simulations. We simulate a thalamocortical network model with adaptive exponential integrate-and-fire neurons, displaying complex intrinsic properties such as low-threshold spiking, regular spiking, fast spiking and adaptation [2]. The network activity can exhibit oscillatory or asynchronous irregular (AI) dynamics, depending on the time constants of the inhibitory synaptic conductance, which are 5 ms (AI) and 15 ms (oscillatory), respectively. An increase in the inhibitory decay time constant reflects a change from GABAA dominated inhibition to more GABAB, which can result in GSWDs, given that the “wave” components of GSWDs are related to slow GABAB -mediated K+ currents [3]. We provide CN input to all thalamocortical neurons to analyse the mechanism of reverting from abnormal oscillatory activity to the normal AI state. Our results confirm that input from the CN can control oscillatory activity in thalamocortical networks. Furthermore, they show that the effectiveness of this input exhibits phase-dependence. In our simulations, CN input terminates epileptic absence seizures most effectively when it arrives at the peak of GSWDs, while seizure termination is least efficient for input between the GSWD bursts. This finding is potentially relevant for therapeutic applications of CN stimulation to terminate seizures. However, the simulations in silico did not take several biological factors such as indirect pathways from the CN to the thalamus into account that may explain differences with animal models of epilepsy [1].