Stable memory and computation in randomly rewiring neural networks
Our brains must maintain a representation of the world over a period of time much longer than the typical lifetime of the biological components producing that representation. For example, recent research suggests that dendritic spines in the adult mouse hippocampus are transient with an average lifetime of approximately 10 days. If this is true, and if turnover is equally likely for all spines, approximately 95-percent of excitatory synapses onto a particular neuron will turn over within 30 days; however, a neurons receptive field can be relatively stable over this period. Here, we use computational modeling to ask how memories can persist in neural circuits such as the hippocampus and visual cortex in the face of synapse turnover. We demonstrate that Hebbian learning during replay of pre-synaptic activity patterns can integrate newly formed synapses into pre-existing memories. Further, we find that Hebbian learning during replay is sufficient to stabilize the receptive fields of hippocampal place cells in a model of the grid-cell-to-place-cell transformation in CA1 and of orientation-selective cells in a model of the center-surround-to-simple-cell transformation in V1. We also ask how synapse turnover affects memory in Hopfield networks with CA3-like, auto-associative properties. We find that attractors of Hopfield networks are remarkably stable if learning occurs during network reactivations. Together, these data suggest that a simple learning rule, correlative Hebbian plasticity of synaptic strengths, is sufficient to preserve neural representations in the face of synapse turnover, even in the absence of Hebbian structural plasticity.