Nonlinear queuing model for dynamic synapse with multiple transmitter pools

This paper proposes a novel queue model for a nonlinear dynamic synapse. It is composed of four serially connected neurotransmitter pools at the pre-synapse. Static/time-varying transfer rates are used to describe the stationary/nonstationary neurotransmitter transmitting events among four neurotransmitter pools. Neurotransmitting is released from pool, and then recycled into the used into the used transmitter pool, forming a closed queue model. The neurotransmitter release event is affected by faciliation and depression in a third-order fashion between the pre-synapse and the post-synapse. Simulation results illustrate the comparisons among dynamic synapse (DS) model, facilitation-depression (FD) model, and our proposed high-order dynamic synapse (HODS) model, and show that the proposed model fits well to the experimental data recorded at Schaffer collateral - CA1 synapses. Furthermore, an additive white Gaussian noise and noisy spikes are added in the output of the experimental data to identify the robustness of the HODS model.