An Optimal Data-Driven Method for Controlling Epileptic Seizures

The regions of the brain may be viewed as nodes in a complex network where information is dynamically transferred through synchronization. Synchronization plays an important role in learning, emotions, and motion. However, neurological disorders such as epilepsy are known to result from abnormal brain synchronization. Coupled Kuramoto model with a little integration of the neurological factors can be a suitable model of the brain network. In this paper, we present an open-loop data-driven control strategy to effectively desynchronize the activity of brain regions during a simulated seizure episode without making any assumptions about the dynamics of the brain. In order to quantify the significance of network nodes, we used an energy-based optimization problem. Then, we evaluated our control methods using a genuine connectome with 80 regions and demonstrated that our approach remarkably decreased synchrony between phases of the oscillations of the brain during the epileptic seizure. Finally, we conclude that brain epilepsy synchronization can be controlled by applying external inputs to the chosen optimal set of driver nodes.