The ability to convey emotion through variations in tone of voice (affective prosody) is crucial for normal social interaction. Our goals were to: (1) test the hypothesis that there are double dissociations in impairment of expression and recognition of affective prosody after acute right hemisphere ischemic stroke; and (2) to identify areas of acute ischemia associated with impaired expression and/or recognition of affective prosody after acute ischemic right hemisphere stroke. DESIGN/METHODS: We tested 23 patients with acute right hemisphere ischemic stroke on the expression (repetition) and recognition subtests of the Aprosodia Battery (Ross & Monnot, 2008) 2008), for which there are normative data for middle-aged and older healthy adults. For repetition, patients repeated content-neutral sentences, monosyllables, and asyllabic utterances, imitating the emotional tones of the examiner. The coefficient of variation of the fundamental frequency (F0 – CV%) was measured. Impairment in expressive prosody was indicated by reduced F0 – CV% relative to normal controls. In the recognition subtest, patients pointed to labels of the emotion (happy, sad, angry, surprised, disinterested) for each stimulus. Lesion maps were manually drawn on the mean diffusivity image generated from the DWI sequence. For lesion symptom mapping we computed permutation thresholded t-tests, using mean Z-score of the production tasks as our index of impairment. RESULTS: 9/23 patients had expressive prosody impairments; 5/9 of these patients also had impaired recognition of affective prosody; 2/9 had selective deficits in expressive prosody; in 2/9 recognition of prosody was not tested. Another 6/23 patients had selective impairment in recognition of affective prosody. In lesion symptom mapping of expressive prosody impairment, only right temporal pole survived a p < 0.05 (threshold z < -2.45) with an observed Z-score of -2.46. However, all patients with right temporal pole lesions had deficits in both expression and recognition of affective prosody. The two patients with selective deficits in expressive prosody had right frontal lesions. Patients with selective deficits in receptive prosody had posterior temporal (n=3), caudate (n=2), or thalamic (n=1) lesions. DISCUSSION: A double dissociation between impaired modulation of expressive prosody (with normal recognition of affective prosody) in some cases and impaired recognition of affective prosody (with intact expressive prosody) in other cases after acute stroke provides evidence that there are at least some distinct mechanisms underlying expression and recognition of prosody that can be selectively impaired. (see also Ross & Monnot, 2008; Bowers, Bauer, & Heilman, 1993). Patients with impaired expression and recognition of affective prosody had lesions mostly involving right anterior temporal pole, and this area was statistically associated with severity of their expressive deficit. This region may be critical for a component common to expression and recognition of prosody, such as the semantic representation of the emotion to be conveyed. The anterior temporal poles (bilaterally) have been proposed as a “semantic hub” in object meaning (Pobric, Jefferies, & Ralph, 2010), and have been proposed as a critical for understanding social concepts (Zahn et al., 2009). It would not be surprising if they have a role in representing the meaning of emotions.