Authors:

STOYANOVA, R., EWBANK, M. & CALDER, A.J.

Reference:

Year of publication:

2009

CBU number:

7089

Abstract:

When asked to categorize emotional expressions presented visually, auditorily or bimodally, participants tend to be faster and more accurate under bimodal conditions (e.g., Collignon et al., 2008). By contrast, incongruent stimuli presented in one modality shift the perception of those presented in another. However, it remains unclear how emotion perception in one channel (e.g., the face) is affected by the congruence of emotion in both of the channels that naturally occur with it (the body and the voice). We examined this question,in a series of experiments,in which we presented emotional facial expressions in concert with bodies and voices that could be congruent or incongruent with each other, as well as with the face. Our data show a graded cost of emotional bodies on facial expression categorization as a function of the similarity between the expression normally associated with the body and the facial expression that participants are asked to categorize, replicating previous findings(Aviezer et al., 2008). We extend this work by showing a similar graded effect for emotional voices. Furthermore, our data suggest that when the body provides emotional information, there is no added cost or benefit from presenting the same information in the voice. However, a voice congruent with the facial expression can significantly reduce the \\\'cost\\\' of a body that is incongruent with this expression. These results are discussed with reference to popular theories of multisensory integration and emotion perception.