Introducing heterogeneities in biological neuronal network models with distance-dependent connectivity

Sammanfattning: Computational studies of biological neuronal network dynamics are often conducted on isotropic and homogeneous network models where all neurons are assumed to be identical. Here, we look at three different spatially extended networks and introduce heterogeneities in the connectivity. The first resembles the classical isotropic approach and connections are equally probable to be made in every direction. In the other two networks we impose preferred directions in recurrent excitatory connections. All excitatory neurons in the second network has a higher probability to connect in one direction. In the third network nearby neurons prefer similar directions, but distant neurons remain uncorrelated. We further analyze the irregularity of spike-timings of individual neurons, the synchrony of the network, network oscillations, as well as the flow of activity. Finally, we show that meaningful behavior can be generated if nearby neurons prefer to connect in similar directions.