Principal component analysis can be used to recover a dynamical portrait or trajectory showing how whole-brain neural activity evolves through a low-dimensional state space. The analysis uses regresion and PCA to estimate a three-dimensional space on trial-averaged responses, and then projects individual trial data into that space. The result is an evolution of the dynamics. In the visualization, each trace shows a single trial evolving through a recovered three-dimensional space; the two movies are from the same data set, but in one case regression is used to look at variability related to stimulus direction, and in the other it is used to examine variability related to time.