Interaction between synaptic inhibition and glial-potassium dynamics leads to diverse seizure transition modes in biophysical models of human focal seizures

Abstract

How focal seizures initiate and evolve in human neocortex remains a fundamental problem in neuroscience. Here, we use biophysical neuronal network models of neocortical patches to study how the interaction between inhibition and extracellular potassium ([K+]o) dynamics may contribute to different types of focal seizures. Three main types of propagated focal seizures observed in recent intracortical microelectrode recordings in humans were modelled: seizures characterized by sustained (∼30−60 Hz) gamma local field potential (LFP) oscillations; seizures where the onset in the propagated site consisted of LFP spikes that later evolved into rhythmic (∼2−3 Hz) spike-wave complexes (SWCs); and seizures where a brief stage of low-amplitude fast-oscillation (∼10−20 Hz) LFPs preceded the SWC activity. Our findings are fourfold: (1) The interaction between elevated [K+]o (due to abnormal potassium buffering by glial cells) and the strength of synaptic inhibition plays a predominant role in shaping these three types of seizures. (2) Strengthening of inhibition leads to the onset of sustained narrowband gamma seizures. (3) Transition into SWC seizures is obtained either by the weakening of inhibitory synapses, or by a transient strengthening followed by an inhibitory breakdown (e.g. GABA depletion). This reduction or breakdown of inhibition among fast-spiking (FS) inhibitory interneurons increases their spiking activity and leads them eventually into depolarization block. Ictal spike-wave discharges in the model are then sustained solely by pyramidal neurons. (4) FS cell dynamics are also critical for seizures where the evolution into SWC activity is preceded by low-amplitude fast oscillations. Different levels of elevated [K+]o were important for transitions into and maintenance of sustained gamma oscillations and SWC discharges. Overall, our modelling study predicts that the interaction between inhibitory interneurons and [K+]o glial buffering under abnormal conditions may explain different types of ictal transitions and dynamics during propagated seizures in human focal epilepsy.

Predicting the spatiotemporal diversity of seizure propagation and termination in human focal epilepsy

“…This mechanism remains valid for other nonlinear functions such as sigmoidal firing-rate functions. Although LVFA oscillations have been associated with inhibitory activity at seizure onset 28 , 34 and are consistent with the inhibitory veto hypothesis 17 , 35 , our model points to a more general dynamical mechanism for how LVFA oscillations can effectively hinder seizure propagation. The dynamical mechanism is more general in the sense that other sources of LVFA oscillations, not necessarily requiring the intervention of surround inhibition (inhibitory veto), can lead to the same hampering effect.…”

“…The slow variable v ( t ) is termed the “permittivity” variable, as it represents the distance of the Epileptor state to the seizure threshold, i.e., the ability of the model to resist to seizure triggering events. The slow permittivity variable captures slowly evolving physiological processes in the brain, such as changes in the extracellular concentration of different ions 24 – 28 , metabolism 29 , 30 , and tissue oxygenation 31 , 32 . The Epileptor’s ability to transition into seizures autonomously is mediated by an excitability parameter ( u 0 ).…”

“…Here, long-range interactions through heterogeneous coupling were modeled via fast population activity ( u 1 ) at different sites of the Epileptor field model. The two approaches are not incompatible, but rather complementary, as they account for different biophysical mechanisms: Fast coupling accounts for synaptic interactions, while slow permittivity coupling accounts for slowly evolving physiological processes such as extracellular ionic diffusion involving, e.g., potassium 26 , 43 , its spatial buffering by glial cells 28 , 44 , as well as increase, via long-range projections in remote regions, of its extracellular concentration via increased firing rate 28 , 34 . In terms of dynamics, fast coupling explains seizure dynamics such as slow seizure propagation, fast SWDs propagation, and synchronous seizure termination.…”

Human seizures couple across spatial scales through travelling wave dynamics

“…(ii) We included a model of depolarization block, so that the firing rate of each population approaches zero as the voltage exceeds −20 mV; that is, we apply a Gaussian activation function, rather than the standard sigmoid function78. Some recent observations from human seizing cortex suggest an important role for depolarization block7879, while others suggest this role may depend on the type of seizure2380. We do not find a critical role for depolarization block in the large amplitude, low-frequency dynamics simulated here.…”

Inhibitory or excitatory? Optogenetic interrogation of the functional roles of GABAergic interneurons in epileptogenesis

“…Recently, we found that high frequency firing in GABAergic interneurons could cause the exhaustion of the presynaptic neurotransmitter GABA in a low Mg 2+ /high K + seizure model, therefore leading to the transition of network activity to seizure [ 57 ]. Computer simulation predicted that certain focal seizures could be triggered by GABA depletion [ 59 ]. It remains to be seen if depletion of presynaptic GABA is presented in in vivo animal models of seizure.…”