System Identification of Spiking Sensory Neurons Using Realistically Constrained Nonlinear Time Series Models

Gaussian local rate coding (GLR) transforms spike train data into time series data, making it possible to use time series models for neural system identification. A simple computational model is used to represent the dynamics of peripheral electrosensory system of an elasmobranch, and a maximum entropy criterion is used to simultaneously optimize the coding bandwidth and the structure and parameters of the computational model. The computational model may be suitable for other neural systems. The coding model is general and provides a natural definition of neural firing rate.