One central goal of cognitive science is to understand how the brain supports cognition. Toward this end, a great deal of effort is devoted toward computational modeling and brain imaging. The former effort is well represented at the Annual Conference, whereas the latter effort is neglected. One common criticism of brain imaging research from the cognitive science community is that it is overly focused on the "where" of cognition, as opposed to the "how" (i.e., process-level questions linking brain and behavior). Model-based analysis of fMRI data links models to the interpretation of imaging data, allowing process-level questions to be asked. The basic approach involves fitting models to behavioral data and then using internal quantities from the models as regressors in the imaging analysis. In this symposium, a broad assortment of leading researchers demonstrate the value of this approach in several domains.