Facial expressions of emotion powerfully influence social behavior. The distributed network of brain regions thought to decode these social signals has been empirically defined using static, usually photographic, displays of such expressions. Facial emotional expressions are however highly dynamic signals that encode the emotion message in facial action patterns. This study sought to determine whether the encoding of facial expressions of emotion by static or dynamic displays is associated with different neural correlates for their decoding. We used positron emission tomography to compare patterns of brain activity in healthy men and women during the explicit judgment of emotion intensity in static and dynamic facial expressions of anger and happiness. Compared to judgments of spatial orientation for moving neutral facial expressions, the judgment of anger in dynamic expressions was associated with increased right-lateralized activity in the medial, superior, middle, and inferior frontal cortex and cerebellum, while judgments of happiness were associated with relative activation of the cuneus, temporal cortex, and the middle, medial, and superior frontal cortex. In contrast, the perception of anger or happiness in static facial expressions activated a motor, prefrontal, and parietal cortical network previously shown to be involved in motor imagery. The direct contrast of dynamic and static expressions indicated differential activation of visual area V5, superior temporal sulcus, periamygdaloid cortex, and cerebellum for dynamic angry expressions and differential activation of area V5, extrastriate cortex, brain stem, and middle temporal cortical activations for dynamic happy expressions. Thus, a distribution of neural activations is related to the analysis of emotion messages in the nearly constant biological motion of the face and differ for angry and happy expressions. Static displays of facial emotional expression may represent noncanonical stimuli that are processed for emotion content by mental strategies and neural events distinct from their more ecologically relevant dynamic counterparts.