Research Of Neuron And Neuronal Population’s Firing Model In Le Rat’s Primary Visual Cortex

More than80%of the vast amounts information that brain has received needs tobe processed by the visual system, and the primary visual cortex is an indispensablearea when visual system process the external stimuli. The incredible capabilities forinformation processing of animals’ visual system is the most complete, bestmechanism of all currently known information processing systems. Studies on themechanism of visual information processing systems is a major issue which isconcerned by control scientific, intelligent science, computer science, neuroscienceand other subjects. In order to study the information handing of primary visual cortex,there is urgent need to establish a model between the external stimulus and theprimary visual cortex neurons response, this study also has important significance forfurther understanding of the V1neurons information processing mechanism andestablish a high quality,efficient processing of artificial vision systems.We proposed a neuron coding model of primary visual cortex of rats whichbased on Linear-Nonlinear-Poisson(LNP) model. Model integrates multiple factorswhich affect neuronal firing activity. The data used to fitting the model parametersand to verify the performance of the model is recorded by multi-electrode array inrat’s V1area when stimulated with different orientation gratings. A series ofsimulations are designed to study the performance of the model. The main contentsare as follows:(1) Design neurophysiological experiments. Describes the structures of theexperiment platform, how to manage the surgery and to determine the stimuliprogram. The data record in the neurophysiological experiment is used to not only fitmodel parameters but also provide model’s correctness. This part of job provides asolid foundation for subsequent work.(2) Construction of the primary visual cortex neurons coding model. Firstly, wegive the theoretical framework of the primary visual cortex neurons coding model,and introduce the physiological significance of each module. Then, we describe theconstruction method for each module in details. Finally, we elaborate how to fit model’s parameters.(3) Results and analysis of model experiments. First, in order to study theperformance of the model, we design two kind of experiments, respectively, using themodel to predict the response of neurons under repeated grating stimuli and a singlegrating stimuli. When repeated grating stimuli is in the place, we also set up threegroups of control experiments to conduct an in-depth study on the impact of model’soptimizing structure. Finally, we analysis the prediction performance of our model.