NETWORK SIMULATION OF A LOCOMOTION NEURAL CIRCUIT

点击翻译按钮获取中文摘要

The goal of this research is to implement a natural network model for locomotion in a primitive vertebrate, the lamprey, using measured membrane properties of neurons from the intact spinal cord to calculate post- synaptic potentials. We view the network model as a modifiable framework continually suggesting and being updated by new experiments for a better understanding of CNS behavior. The membrane properties will include the effects of the voltage dependent conductances and thus provide a means to evaluate their role in the integration of synaptic input of neurons that are part of the spinal cord neural circuitry involved in locomotion. The proposed transfer function approach is a unique way to determine the activation functions for neuronal units that incorporate the non-linear membrane properties of both soma and dendritic membranes. The voltage clamp data will be used to obtain Hodgkin-Huxley type voltage- dependent kinetic parameters that will be incorporated in a full scale non- linear differential equation for the computation of potential responses of each neuronal unit. Simulations of neurophysiological activity will be done with a neural circuit consisting of the principal interneuron types on each side of the spinal cord. The stimulations will be done using both the non-linear activation functions and synaptic weights on the dendritic compartments of the unit neurons. The following permutations of the basic network will be made: 1) a varying number of each of the cell types, 2) random and specific synaptic connections, 3) variation in specific membrane properties especially with regard to site densities. The stimulations will provide an evaluation of the effect of an unique type of synaptic receptor, namely the voltage dependent NMDA (N-methyl-D- aspartate) type, on network properties. The consequences of activation on network behavior of NMDA receptors on synaptic transfer are likely to be profound. The preliminary results show that NMDA induces a sharp tuning behavior which can greatly accentuate and filter the response of impinging synaptic inputs. The NMDA effects are strongly dependent on the membrane potential and will greatly affect the nature of neuronal interactions during network activity. The proposed network simulations provide an opportunity to assess the role of NMDA receptors in neural processing using a realistic neural rather than extrapolating from isolated or tissue cultured neurons.