Information transmission by stochastic synapses with short-term depression: neural coding and optimization

The ability of dynamic synapses with short-term depression to transmit the information present in the presynaptic spike train to the postsynaptic neuron is discussed. Both by minimizing the estimation error and by maximizing the information transmitted to the postsynaptic neuron it is found that for Poisson inputs dynamic synapses are not able to estimate the rate better than static ones. However, short-term depression becomes relevant when more realistic temporally correlated spike trains are used as an input. For the simple model of vesicle depletion considered here the optimal vesicle recovery time is rather low, about a hundred milliseconds for realistic values of the input parameters. All these questions are addressed by computing analytically the distribution of intervals between consecutive synaptic responses for arbitrary renewal processes. c 2002 Published by Elsevier Science B.V.

• Type: Journal
• Year: 2002
• Where: IJON
• Authors: Jaime de la Rocha, Angel Nevado, Néstor Parga
• Keywords: Dynamic Synapses, IJON 2002, Postsynaptic Neuron, Short-term Depression