Noise Effect In Neuronal Encoding

In this thesis, we mainly study the dynamic features and physical mechanisms of neuron, especially, the effect of noise in neuronal encoding.A major task in the physical research for the neuronal systems is to search for general physical principles behind various nature processes by using simple models, which can catch the main factors in the concerned aspects and help us to understand the complicated behaviors happened in the brain. We research some models that describe the main characteristic of real neurons.We first study the dynamic features of neuron by effect of noise and the bifurcation. Noise is important for neuron system. In the presence of noise, the neuron produces firing with small signal. The basic features of nonlinear system are bifurcation and chaos. As a nonlinear system, there are bifurcation and chaos in neuron.Second, we study the effect of noise in the information code of HH and FHN neuronal systems. Creating spikes using FHN and HH means imposing a type of highly nonlinear filter on the attractor signal, a filter which edits out amplitude information and converts the information entirely to event timings. The purpose is to gain insight into the data processing methods used in systems which communicate through spike timings, as is conjectured for certain neural systems. We have shown by example that noise may be useful for this communication, in that it can amplify transmission of deterministic, nonlinear state information as measured by nonlinear prediction error (NPE).