Advances in recording technologies have allowed scientists to simultaneously record the individual activity of multiple neurons during sensory and behavioral experiments.  Analyzing such spike train data presents numerous challenges but could help advance our understanding of the low-level circuitry and information flow underpinning sensory processing and behavior. In this talk, I will first discuss some efforts towards analyzing neuronal function connectivity using non-linear variants of Granger causality. I will then discuss efforts on modeling neuronal spiking using Gaussian process based latent factor models.