Parallelization of the Electrodiffusion Mechanism of the Computational Model of Spreading Depression

The simulation of cerebral phenomena, such spreading depression wave (SD), involves the modeling of large-scale neuronal populations. Consequently, the processing time and the computational resource costs become elevate. The goal of this work is to apply the parallel programming techniques on the simulation of SD in a three-dimensional representation of the neuronal tissue. The model consists in the description of synaptic terminals and glial cells coupled by means of the extracellular space, where the ionic movements are described by electrodiffusion. The parallelization strategy is achieved by means of the Message Passing Interface. The code performance in a parallel environment is compared with serial processing. The results point out the model is highly parallelized and justify the use of parallel programming in this kind of modeling. Furthermore, the reproduction of the main characteristics of SD, such the propagation velocity and the extracellular ionic dynamic profiles, close to what is observed experimentally, validates the model.