To provide some background for a chunk of this workshop, I will start by saying something about the nature of spike train recordings, and will give my own summary of the traditional questions asked by statisticians and modelers. After defining the point process regression framework (based primarily on generalized linear models — GLMs) I will very briefly define integrate-and-fire neurons and biophysical, Hodgkin-Huxley neurons. With this, I will mention some points of contact between these two very different approaches, and will list several outstanding issues