The effect of noise in models of spiny dendrites

Abstract

The dendritic tree provides the surface area for synaptic connections between the 100 billion neurons in the brain. 90% of excitatory synapses are made onto dendritic spines which are constantly changing shape and strength. This adaptation is believed to be an important factor in learning, memory and computations within the dendritic tree. The environment in which the neuron sits is inherently noisy due to the activity in nearby neurons and the stochastic nature of synaptic gating. Therefore the effects of noise is a very important aspect in any realistic model. This work provides a comprehensive study of two spiny dendrite models driven by different forms of noise in the spine dynamics or in the membrane voltage. We investigate the effect of the noise on signal propagation along the dendrite and how any correlation in the noise may affect this behaviour. We discover a difference in the results of the two models which suggests that the form of spine connectivity is important. We also show that both models have the capacity to act as a robust filter and that a branched structure can perform logic computations.