Status epilepticus as a system disturbance: Is status epilepticus due to synchronization or desynchronization?


Address correspondence to Sydney S. Cash, Department of Neurology, Massachusetts General Hospital and Harvard Medical School, WACC 730, 55 Fruit Street, Boston, MA 02114, U.S.A. E-mail:


The traditional view of seizure activity is one in which there is extreme hypersynchronization. Although what is meant by hypersynchronization is rarely explicitly and fully defined, it can be understood to imply large numbers of neurons firing together essentially simultaneously. In this discussion we explore the possibility that seizures—both self-terminating and sustained in status—are not purely synchronous in time or in space. We investigate the alternative possibility that much seizure activity represents spatiotemporal desynchronization. Furthermore, we discuss the possibility that, in contrast to canonical views of epileptic activity, a high degree of synchronization is a prerequisite for termination of the seizure rather than a marker of early and ongoing seizure activity. These ideas will be discussed with reference to results from our collaborative group based on microelectrode recordings in patients with epilepsy as well as to the many studies done by others in both patients and animal models. Finally, we will explore implications for these hypotheses in the treatment of patients with epilepsy and in status epilepticus.

Despite many decades of research in both clinical neurophysiology and basic neuroscience, our understanding of the mechanisms of epileptiform activity remains crude. Two general principles of neuronal activity have come to dominate our thinking about what is a seizure or how a seizure occurs. The first is that seizure activity arises from an imbalance of inhibitory and excitatory neuronal action. The second is that ictal activity represents a form of hypersynchrony within neural behavior. These descriptions of neuronal action have been enormously powerful in helping formulate ideas about seizure progression and in the explanation and even design of therapeutic interventions—primarily pharmacologic ones. Particularly in the past decade, however, these simplifications have been increasingly recognized as just that—simplifications that, although useful in many regards, do not fully describe the spatiotemporal dynamics of neuronal activity during seizures. Instead, an emerging body of work suggests that both self-limited seizures and status epilepticus evolve through different phases during which there are different degrees of spatiotemporal synchrony (reviewed recently in Jiruska et al., 2013).

To begin, we need to be clear about what is synchrony and desynchrony. They can be variously defined. At one extreme, at the level of individual neurons and with the most conservative definition, we can establish synchrony only when neurons are firing action potentials truly simultaneously. Desynchrony occurs when the normal degree to which neurons are firing simultaneously is disrupted. At the other extreme we could imagine synchrony to be present when either action potentials or any other neuronal activity (e.g., postsynaptic potentials) are maintaining a consistent relationship over short periods of time—closer than is the case under “normal” conditions. This could be true over the time span of 10 s of milliseconds or even much more. What determines synchrony depends in part on the kinetics of the underlying phenomenon. Action potentials lasting only a few milliseconds require a different definition of synchrony than do postsynapatic potentials lasting 10–100 s of milliseconds.

A major confound in this scheme is the overall degree of neuronal activity. As action potential firing rate increases, for example, there is necessarily a greater degree of synchrony for firing between any two neurons in any given bin of time. Is this truly increased synchrony or merely an epiphenomenon of the increased rate of activity? The answer will depend essentially on how synchrony is defined, and it is currently both possible and reasonable to use either end of the spectrum to make that definition. What is perhaps most important is to be clear about what degree of synchrony is being measured and, ultimately, what are the mechanistic causes/effects of that synchrony. With this major, caveat present we can then proceed to look for evidence that seizures, and status in particular, demonstrate either synchronization or desynchronization.

Evidence for Desynchrony at the Microscale

Seizures, by definition, go through an evolution in time. Different phases of the seizure demonstrate different activities at whatever level of analysis one chooses to examine. This is particularly true for focal seizures but also true for generalized events. Nonetheless, the degree to which this change represents changes in levels of synchrony has been unclear. To explore this issue specifically at the level of individual neurons, our group examined the behavior of neuronal populations in small cortical regions inside and outside of the seizure focus in patients with intractable focal epilepsy. Surprisingly, the early and middle phases of the seizure appear to be characterized by heterogeneity in firing activity (Truccolo et al., 2011). Similar findings are present using other methods of recording unit activity (Bower et al., 2012). In addition, these findings parallel earlier work done in a slice model of epilepsy. Netoff and Schiff (2002) used intracellular recordings to demonstrate that synaptic inputs to pairs of neurons were less correlated during initiation and maintenance of the seizure than during interictal periods. In theoretical work it was predicted that when the spiking rate is at its maximum, neurons would desynchronize and this is likely to occur at the height of the seizure (Ermentrout & Kopell, 1998; Gutkin et al., 2005). Finally, even brief epileptiform discharges seem to display a similar degree of heterogeneous activity, which is not consistent with the notion of pure hypersynchrony (Keller et al., 2010). This heterogeneity in neuronal behavior argues against homogeneous and hypersynchronous runaway excitation or widespread paroxysmal depolarization as the primary mechanism underlying seizure initiation and spread. Instead, the seizure appears to result from a complex interplay among groups of neurons with different spiking behaviors that evolve at multiple temporal and spatial scales.

Evidence for Desynchrony at the Macroscale

Seizures, by definition, go through an evolution in space as well. Measures of activity across that spatial domain also demonstrate that there is greater heterogeneity in interareal interactions between different cortical areas than might be expected from a model of pure hypersynchrony. For example, specific measures may show a decrease in synchrony in the minutes to hours preceding a seizure (Le Van Quyen et al., 2001), a spatial decorrelation during the low voltage fast activity that is seen at the beginning of certain seizures (Wendling et al., 2003), or changes in the degree of synchronization as recorded with magnetoencephalography (MEG) for many different types of seizures (Garcia Dominguez et al., 2005). To be sure, there are many examples of analysis that show higher macroscale synchronization during seizures or at rest in the seizure focus. But, the aforementioned studies (and many others) point out that the notion of hypersynchrony is certainly incomplete in its most pure form.

Furthermore, these and other studies also clearly demonstrate that the degree of synchrony or desynchrony is dynamic during the seizure. Using functional network measures based on correlation, for example, our group has observed a sequence of mildly increased synchronization at the start of the seizures, followed by desynchronization in the middle of the seizure, only to see truly heightened synchronization at the end of the event (Kramer et al., 2010). Similar results have been observed using other methods of measuring coupled activity (Schindler et al., 2007, 2008).

Greatest Synchrony in Self-Terminating Seizures May Be Present at the End of the Event

In fact, the different methods for measuring or describing synchrony seem to demonstrate that it is at the end of a typical focal seizure that generalizes that synchrony at both the scale of individual action potentials and local field potentials becomes most synchronous. As mentioned, in examining seizure networks with either eigenvalue decomposition (Schindler et al., 2008) or correlation measures (Kramer et al., 2010), it is just before the seizure terminates that synchrony appears to be greatest. The same is true at the level of the single neuron (Truccolo et al., 2011). This has led to the hypothesis that seizure termination that occurs normally is the result of this truly hypersynchronous activity. In cases of status (albeit few that they are), the dynamic changes presumably essential for seizure termination do not appear to be present and, perhaps as a result, the seizure is unable to cross a critical transition to postictal behavior (Kramer et al., 2012).

Implications for Understanding and Treating Status Epilepticus

Most of these studies exploring the true role of synchrony/desynchrony in the initiation, propagation, and termination of seizures have been done with respect to focal, self-terminating seizures. Little of the work has directly explored the neurophysiology of status epilepticus per se. Yet, it is likely that the dynamics present at both the microphysiologic and macrophysiologic levels are at least similar. Perhaps not all of the now classical phases of status (Treiman et al., 1990) show the same degree of synchrony or desynchrony that is seen in a single seizure, but the possibility that failure to reach sufficient synchrony may prevent the seizure from terminating is certainly intriguing and may provide an entirely novel approach to prevent status from evolving or in shifting a patient from the status state to a, presumably, more benign postictal state.


The author wishes to thank Drs. Cole, Truccolo, Ahmed, Kramer, and others who have been instrumental in gathering data for related works and in many discussions about these topics that have helped in the formulation of the theoretical framework explored here. Work on these topics has been supported by the NIH/NINDS, CIMIT, and the Grass Foundation.


I have no conflicts of interest to declare. I confirm that I have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.