| Computationally investigating the structural and functional properties of brain network may contribute substantially to the understanding of brain function and dysfunction. This strategy is more highlighted when the ever changing structure of brain network, in which both synaptic and intrinsic properties of neuronal network modify continuously, is a restriction for evaluation of functional consequences of purely structural changes in invivo/invitro models. In this field, network modeling is a promising technique for brain researches in all structural, functional and dynamical aspects which has been investigated many times in healthy network as well as brain diseases such as epilepsy. In the case of epilepsy, a few works have reported that changes in network structure play an important role in seizure generation and its propagation. Based on these findings, we postulated that in seizure evolution, the network's topology and its behavior change in such a way that a transition from complex firing to rhythmic bursting occurs. However, in this process, the emerged synchronized behavior and the transition mechanism route to this state is questionable. Hence, this paper deals with a goal-directed network model of seizure evolution in epileptic brain with a prescribed dynamical phase transition route. The goal of this model is a synchrony state and the transition path is intermittency. The synchronous state is satisfied by optimizing a synchrony measure and the intermittent route is satisfied by a constraint of having a scaling relationship between Lyapunov exponent and control parameter. For this, we implemented a small world network a neuronal subpopulation and evolve it with genetic algorithm to search for an optimal synchronizable network topology. Out results show that the genetic algorithm is capable of emerging to a prescribed state with a transition path which falls in an intermittent regime after a number of generation. |
|