Abstract
Cognitive models of face processing suggest the existence of parallel streams that are specialized for identity versus expression perception; furthermore, neuroimaging data suggests that there may be distinct anatomic correlates for each of these processes. However, there is evidence that the neural representations of expression and identity are not completely independent. We have shown that aftereffects in the perception of expression are modulated by the identity of the adapting face, suggesting both identity-dependent and identity-independent representations of facial expression. In the present experiment, we asked whether aftereffects in the perception of identity showed a similar modulation by facial expression, which would suggest the existence of expression-dependent and expression-independent representations of facial identity.
We measured the magnitude of aftereffects from three different adapting stimuli on the perception of identity in ambiguous morphed faces. ‘Same-picture/expression-congruent’ adapting stimuli were the images used to create the morphs. ‘Different-picture/expression-congruent’ adapting stimuli were different pictures of the same individuals displaying the same expression as was present in the morphs. ‘Different-picture/expression-incongruent’ adapting stimuli were pictures of the same individuals displaying different expressions than those present in the morphs. Images were cropped to eliminate the possibility of adaptation to non-face features of the image (i.e.-hair color). Images were novel faces selected from the Karolinska Database of Emotional Faces.
We found that the aftereffects on identity perception did not differ significantly between the three different types of adapting stimuli. This suggests that the neural representations activated by novel identities are largely independent of expression, in contrast to the significant identity-dependent component seen in representations of expression that we previously reported. However, unlike learned categorical expressions, novel identities would not activate learned categorical representations of identity. Thus it seems important to determine whether this asymmetric relationship between identity and expression holds true for categorical representations of familiar identities.
This work was supported by NIH grant 1R01 MH069898 and CIHR grant MOP 77615. CJF was supported by a Michael Smith Foundation for Health Research Junior Graduate Studentship and a Canadian Institutes of Health Research Canada Graduate Scholarship Doctoral Award. JJSB was supported by a Canada Research Chair and a Michael Smith Foundation for Health Research Senior Scholarship.