Philosophers and theorists as far back as Plato have speculated that part of music's expressive power lies in acoustic cues related to the sounds of emotive voices (Kivy 2002). From the standpoint of modern cognitive neuroscience, the idea is intriguing because cues to vocal affect include musical aspects of speech, such as pitch, tempo, loudness, and timbre (voice quality; Ladd et al. Reference Ladd, Silverman, Tolkmitt, Bergmann and Scherer1985; Johnstone & Scherer Reference Johnstone, Scherer, Ohala, Hasegawa, Ohala, Granville and Bailey1999; Reference Johnstone, Scherer, Lewis and Haviland-Jones2000), and because the patterning of these cues in emotionally expressive speech and music shows striking commonalities (Ilie & Thompson Reference Ilie and Thompson2006; Juslin & Laukka Reference Juslin and Laukka2003; Patel Reference Patel2008).

In the target article, Juslin & Västfjäll (J&V) argue that one mechanism underlying emotional responses to music is emotional contagion based on speech-like affective cues in music. That is, they postulate that listeners implicitly recognize a basic emotion in music (such as sadness) from speech-like cues to affect (e.g., low tempo, pitch level, intensity, and pitch variability), and then the listeners themselves come to feel that same emotion. Of course, the link between the perception of emotion in music (emotions expressed by music) and the subjective experience of emotion (emotions felt by the self) is not obligatory. One can, for example, perceive a piece of music as expressing sadness without feeling sad. Indeed, one can respond with an emotion different from that expressed by the music (Gabrielsson Reference Gabrielsson2002). Yet empirical research suggests that for the majority of individuals there is an alignment between emotions expressed by the music and emotions felt by the listener (Evans & Schubert Reference Evans and Schubert2008). This enhances the plausibility of the emotional contagion hypothesis.

Another factor favoring the emotional contagion hypothesis for music is neurobiological evidence for emotional contagion in a different expressive domain, namely, face perception. Using functional magnetic resonance imaging (fMRI), Keysers and colleagues have found that the perception of affect in faces triggers activity in brain regions involved in experiencing similar emotions (the anterior insula and adjacent frontal operculum), particularly in empathic individuals (Jabbi et al. Reference Jabbi, Swart and Keysers2007; cf. Keysers & Gazzola Reference Keysers and Gazzola2006; van der Gaag et al. Reference van der Gaag, Minderaa and Keysers2007). This internal simulation of emotions expressed by another is thought to be part of the brain's mechanisms for empathy, giving an observer access to the subjective state of another by recruiting the observer's own neural representations of that state (Preston & de Waal Reference Preston and de Waal2002). De Waal (Reference de Waal and Bråten2007) has argued that such mechanisms have deep evolutionary roots in primates and other highly social species, whereby they help foster social bonds which promote survival in a group setting. Given the importance of both the face and the voice in conveying emotion in humans, it seems plausible that people may also experience emotional contagion from the sounds of expressive voices, though this idea awaits confirmation from neurobiological research.

Hence, what is needed at this point in order to test J&V ‘s ideas about affect in music and speech are two kinds of neurobiological studies. First, it is necessary to test the idea that listeners recognize musical affect using neural circuitry also involved in vocal affect perception (cf. Schirmer & Kotz, Reference Schirmer and Kotz2006). Second, it is necessary to determine whether listening to emotionally expressive music activates brain regions involved in experiencing similar emotions. This comment focuses on the first kind of study, and seeks to draw attention to a clinical disorder that could be a productive tool in this regard.

This disorder is “affective aprosodia” (Ross Reference Ross and Mesulaum2000). Deficits in the ability to either recognize or produce affect in spoken utterances were first described more than 30 years ago in patients who had suffered right-hemisphere lesions (Heilman et al. Reference Heilman, Scholes and Watson1975; Tucker et al. Reference Tucker, Watson and Heilman1977). Patients with receptive affective aprosodia (RAA) provide an opportunity to test the idea that musical and vocal affect perception have shared neural substrates. Specifically, such patients could be tested for their ability to recognize certain basic emotions expressed by instrumental music, such as happiness, sadness, or fear. Indeed, in testing such patients, one could use stimuli from previous studies of musical emotion recognition in normal individuals (e.g., Krumhansl Reference Krumhansl1997). If affect recognition in music and speech relies on similar brain circuits, then individuals with RAA should be impaired in recognizing the emotions expressed by music.

To my knowledge, musical affect perception in RAA is totally unexplored; yet the topic is attractive from both theoretical and practical perspectives. In terms of the latter, the salient point is that the disorder is not uncommon clinically, even though it is relatively rarely studied (K. Heilman, personal communication). In conducting research on musical affect perception in RAA, a number of conceptual and methodological issues require careful attention. The remainder of this commentary discusses a few of these issues.

First, patients with RAA should be tested for their ability to perceive affect in other modalities (e.g., faces, gestures), in order to determine whether they have a general deficit in emotion recognition or a more specific deficit in recognizing vocal affect. From the current standpoint, it is the latter deficit that is of interest. (Existing research suggests that RAA often co-occurs with problems in recognizing affect in the face, although there are cases where vocal affect perception is disproportionately affected [e.g., Adolphs et al. Reference Adolphs, Damasio and Tranel2002; Charbonneau et al. Reference Charbonneau, Scherzer, Aspirot and Cohen2002].) Second, patients with RAA will need to be tested for auditory discrimination skills, in order to determine whether they have deficits in processing of basic auditory attributes shared by speech and music, such as timbre or pitch contours (cf. Heilman et al. Reference Heilman, Bowers, Speedie and Branch Coslett1984). Thus, for example, if a patient cannot judge emotion in speech and music because of a low-level problem in timbre processing that impacts auditory perception in general, this would not point to a specific link between emotional recognition in music and speech, but simply indicate that both processes rely on similar low-level auditory attributes of sound. Third, since RAA often (but not always) involves right-hemisphere damage, control experiments for musical memory deficits will be needed (cf. Zatorre et al. Reference Zattore, Evans and Meyer1994). If these methodological challenges can be met, then RAA could form the basis of a powerful test of the vocal affect hypothesis for musical expressiveness.