Beyond the neuron: Unveiling the role of reactive astrocytes in epileptic seizure dynamics through self-organized...

During epileptic seizures, brain activity and connectivity undergo dramatic changes. Brain networks transition from a balanced resting state to a hyperactive and hypersynchronous state. However, the mechanisms driving these state transitions remain unclear. While astrocyte-neuron interactions are increasingly recognized in seizure pathophysiology, fundamental questions about the causal role of astrocytes in seizure initiation and termination remain unanswered. To understand the dynamic network mechanisms of epileptic seizure initiation and propagation in interacting neural and astrocytic networks, we developed a biologically relevant spiking neural network model that incorporates reactive astrocytes. This model implements the transition from normal to epileptiform activity, including its spontaneous termination, without requiring manual parameter tuning or external stimulation. Using extensive numerical simulations, we demonstrate that the interplay between fast neuronal dynamics and the slower astrocyte-induced dynamic rearrangement of the neural network's functional architecture results in critical system behavior known as self-organized bistability. This self-organized dynamic regime is characterized by the spontaneous triggering of short-lived extreme synchronization events-a hallmark of epileptic seizure activity. The model generates trains of seizures, reproducing the distribution of intervals between seizures observed in neurophysiological experiments. We also show that gap junction connectivity within the astrocyte network plays a protective role against epileptic discharges. Collectively, our findings establish a unified framework where astrocyte-mediated self-organized bistability governs epileptiform dynamics, revealing glial networks as critical modulators of both seizure generation and suppression in pathological neural circuits.