Control theory applied to neural networks illuminates synaptic basis of interictal epileptiform activity.

A brief historical account is presented of the formulation of two hypotheses that have been proposed to explain the mechanisms underlying the paroxysmal depolarizing shift (PDS) in experimental epilepsy. The two hypotheses are called the giant EPSP hypothesis and the endogenous burst hypothesis. The giant EPSP hypothesis states that the PDS (the intracellular correlate of the interictal discharge) is comprised of a larger-than-normal-strength excitatory synaptic input, whereas the endogenous burst hypothesis states that the PDS is an endogenous burst triggered by an excitatory postsynaptic potential of normal strength. Two sets of four experimentally testable predictions, which were derived from these two hypotheses for the PDS, are presented. These predictions describe the expected behavior of the PDS in response to changes in membrane potential and under conditions of voltage clamping. With the advent of single-electrode current- and voltage-clamp techniques and improved intracellular recording conditions, the testing of these predictions has become possible. Experiments are described in which each of the predictions from the two hypotheses were tested. The results strongly support the giant EPSP hypothesis and are not easily reconciled with the endogenous burst hypothesis. Because the PDS is a network-driven event, it is important to understand the properties of the neuronal network responsible for the genesis of the PDS. Others have proposed that there are three necessary conditions for epileptiform activity in any neuronal network: endogenous bursting, disinhibition, and recurrent excitatory synapses. Using control theory as a frame of reference, we argue that it is premature to raise these three phenomena to the level of general theoretical requirements for interictal activity. Insufficient quantitative information exists about the properties of neurons, synapses, and connectivity patterns in any cortical neuronal network to conclude that the three preposed requirements are necessary and sufficient general conditions for epileptiform activity. Because all of the key predictions of the giant EPSP hypothesis have now been experimentally verified, we conclude that the PDS is a large, network-driven EPSP. The current challenge to neurophysiologists is to describe is detail the properties of neurons and synapses in a cortical neuronal network and then to evaluate the relative contributions of network and individual neuronal properties to the expression of interictal epileptiform activity.