In visual perception of emotional stimuli, low- and high-level appraisal processes have been found to engage different neural structures. Beyond emotional facial expression, emotional prosody is an important auditory cue for social interaction. Neuroimaging studies have proposed a network for emotional prosody processing that involves a right temporal input region and explicit evaluation in bilateral prefrontal areas. However, the comparison of different appraisal levels has so far relied upon using linguistic instructions during low-level processing, which might confound effects of processing level and linguistic task. In order to circumvent this problem, we examined processing of emotional prosody in meaningless speech during gender labelling (implicit, low-level appraisal) and emotion labelling (explicit, high-level appraisal). While bilateral amygdala, left superior temporal sulcus and right parietal areas showed stronger blood oxygen level-dependent (BOLD) responses during implicit processing, areas with stronger BOLD responses during explicit processing included the left inferior frontal gyrus, bilateral parietal, anterior cingulate and supplemental motor cortex. Emotional versus neutral prosody evoked BOLD responses in right superior temporal gyrus, bilateral anterior cingulate, left inferior frontal gyrus, insula and bilateral putamen. Basal ganglia and right anterior cingulate responses to emotional versus neutral prosody were particularly pronounced during explicit processing. These results are in line with an amygdala-prefrontal-cingulate network controlling different appraisal levels, and suggest a specific role of the left inferior frontal gyrus in explicit evaluation of emotional prosody. In addition to brain areas commonly related to prosody processing, our results suggest specific functions of anterior cingulate and basal ganglia in detecting emotional prosody, particularly when explicit identification is necessary.