Pavlovian conditioning paradigms have been a stalwart of animal research on fear learning for over a century. Recent advances in cognitive neuroscience research have led to new insights into the neural mechanisms of how humans learn to associate cues with threats, how these representations become bound to contextual features of the environment, and how they generalize to stimuli that are perceptually or conceptually related. By integrating information gleaned from patients with brain lesions, scalp electrophysiology, neuroimaging, and intracranial recordings, researchers are assembling a dynamic view of the distributed brain activity that generates conditioned fear responses. Innovative virtual reality technology, computational modeling, and multivariate analysis tools have further refined a scientific understanding of the component processes involved, which can inform future clinical interventions for treating fear- and anxiety-related disorders.