2018.04.18 Modelling of Synaptic Plasticity and Inherent stochasticity
题 目：Modelling of Synaptic Plasticity and Inherent stochasticity
报告人：Professor Don Kulasiri
Head, Centre for Advanced Computational Solutions (C-fACS), Lincoln University, New Zealand
主持人: 汤超 教授
Synaptic plasticity reflects the brain’s ability to be flexible in transmitting a large repertoire of signals across synapses, the connections between neurons. In this seminar, we will briefly discuss the mathematical models we have developed related to synaptic plasticity based on the associated biochemical pathways. First, we briefly discuss a mathematical model related to bidirectional modulations of the postsynaptic response following a synaptic transmission in a postsynaptic cell. The long term forms of synaptic plasticity, named long term potentiation (LTP) and long term depression (LTD), are critical for the antithetic functions of the memory system, memory formation and removal, respectively. A common Ca2+ signalling upstream triggers both LTP and LTD. We use our simplified integrated model based on the sub-models of the indispensable synaptic proteins in the emergence of synaptic plasticity to validate and understand their potential roles in the expression of synaptic plasticity. Within this system of pathways, a synaptic protein, Ca2+/Calmodulin dependent protein kinase II (CaMKII), plays a major role in the memory formation and impairment, and it is often called “the memory molecule.” This protein has complex state transitions and facilitates the emergence of long term potentiation (LTP), which is highly correlated to memory formation. We will discuss a model of the formation of CaMKII-NMDAR complex with the full state transitions of CaMKII, including the autophosphorylation, based on ordinary differential equations. We investigate the epistemic uncertainties of parameters of this model using global sensitivity analysis (GSA). We explore the effects of parameters on the key outputs of the model to discover the most sensitive ones using GSA and partial ranking correlation coefficient (PRCC) and to understand why they are sensitive and others are not, based on the biology of the problem. We also extend the model to add presynaptic neurotransmitter vesicles release to have action potentials as inputs of different frequencies. We perform GSA on this extended model to show that the parameter sensitivities are different for the extended model as shown by PRCC landscapes. Based on the results of GSA and PRCC, we reduce the original model to a less complex model taking the most important biological processes into account. We validate the reduced model against the outputs of the original model. We show that the parameter sensitivities are dependent on the inputs and GSA would make us understand the sensitivities and the importance of the parameters.