Modeling nonlinear neural dynamics with Volterra-Poisson kernels

A nonparametric quantitative model is introduced that captures the nonlinear dynamic properties of neural systems using input/output data. It is based on the Volterra modeling approach adapted for point-process inputs and outputs. Using input/output data, a model is presented for the CAl region of the hippocampus. The model represents reliably the nonlinear dynamic mapping performed by CAI with high accuracy. Compared to traditional descriptors of nonlinear neural dynamics, the presented model provides a generalized, comprehensive view.