Structural plasticity and the generation of bidirectional connectivity

Miner, D., Hoffmann, F. Z., Kleberg, F., and Triesch, J. (2017). In: Van Ooyen, A., and Butz-Ostendorf, M., eds. The Rewiring Brain: A Computational Approach to Structural Plasticity in the Adult Brain. San Diego: Academic Press, pp. 247-260.

Abstract

The recurrent microcircuitry of the cortex has been shown to be highly nonrandom. One notable aspect of this is an overrepresentation of bidirectional connections between excitatory neurons as compared to a random graph. Using an established model of a selforganizing neural network with structural plasticity in its excitatory pool of neurons, we examine possible origins for this overrepresentation. We show mathematically that given the assumption of symmetric connection probabilities, any inhomogeneity in connection probability will necessarily lead to such an overrepresentation. We examine a Markov model of the competition between this and other factors favoring this overrepresentation, and factors working against it, such as the action of classical excitatory spike-timing-dependent plasticity (STDP) in most regimes. We also examine the influence of introducing inhibitory STDP to our model system. Finally, we speculate on additional factors that may contribute to the overrepresentation of bidirectional connections.