Topological classification of population activity in spiking neural network

摘要

The role of propagating waves of neuronal activity in different brain areas, particularly in cerebral cortex, is still debated in modern neuroscience. We consider spiking neural network (SNN) model with spike timing dependent plasticity (STDP). It is shown that in the SNN the topological type of neural activity depends on the average length of connections between neurons - the connectivity radius. At the small radius (mostly local connections), the activity spreads in the form of traveling waves with well shaped leading and trailing edge. Such simple wave activity can be treated as a network spike. With increase in the connectivity radius, blurring of the leading and trailing edges is observed up to the "instant" activation of the network. In this case, the sparse neurons first generate spikes, and then all neurons in the network generate multiple spikes. This type of activity can be considered as a network burst. We also analyze the velocity of neural activity propagation depending on the connectivity radius. In the case of the network burst mode, the velocity is predominantly limited by axonal delays. In turn, in the case of the network spike mode it is defined by synaptic delays. In addition, we found that the STDP facilitates network synchronization in SNNs with a predominance of local connections.