Neural Encoding Model For The ON-OFF Neuron Response To SUN Stimulation In The Pigeon Optic Tectum

Birds have excellent rapid visual perception,and Optic Tecturn(OT)is an important neural nucleus for the processing of visual information in the visual pathway of birds.It is of great significance to study and construct the coding model of fast time-varying information based on OT region neurons for understanding its working mechanism of rapid perception.In this thesis,the neurons in the middle layer of the OT region of carrier pigeon are studied.Using 32-channel microelectrode array,the spike response signals of neurons stimulated by spatial uniform white noise spatial Uniform(SUN)were collected and recorded synchronously.Two models were constructed by analyzing the relationship between stimulus and response.Through the model evaluation of neuron response prediction,the results verify the hypothesis of time-varying information processing of ON path and OFF path of RGC-OT proposed by us.The main work and results of this thesis are as follows:(1)Research on SUN stimulus coding based on Generalized Linear ModelGeneralized Linear Model(GLM)is a classical model for encoding SUN stimuli.The responses of neurons are determined by the history of input stimuli,single linear filters,and action potential spikes.The maximum likelihood estimation method is used to solve the model parameters.Based on the model,the likelihood function is constructed,the conditions for convex optimization of the likelihood function are analyzed,and the model parameters are finally determined.The responses of neurons stimulated by 30 Hz and 60 Hz SUN are predicted and good prediction results are obtained.However,the model is a functional model and does not have biological credibility.(2)Research on SUN stimulation coding based on approximate Generalized Quadratic ModelBased on the ON-OFF information integration hypothesis of RGC-OT neuronal loop,a two-filter half-wave rectifier model of SUN stimulus response of ON-OFF neurons in the shallow and middle layers of OT is proposed,which can be used to predict the SUN stimulus response.In order to solve the convex optimization problem,the quadratic curve is used to approximate the half-wave rectifier,and the two-filter model is transformed into the Generalized Quadratic Model GQM,and then optimized.The prediction results show that the quadratic model is compared with the GLM model.Better prediction results are obtained,and to some extent,the hypothesis of time-varying information processing of on and OFF paths of RGC-OT is verified.