| Summary: | The human face conveys a wealth of information to the perceiver. Not only can we make judgments on the identity of the person, but we can also make finer interpretations about the emotional state of the individual, and what they are currently attending to. The ability to process and act upon this information effectively, facilitates successful social interactions. The key information that indicates to us how a person is feeling and what they are attending to, is their facial expression and facial viewpoint. Because of their dynamic properties, facial expression and viewpoint are described as changeable aspects of faces. The human brain contains a core and extended network of face-responsive regions. One region in the posterior superior temporal sulcus (pSTS) is thought to have a central role in the processing of these changeable aspects of faces. An outstanding question about the neural representation of faces is whether changeable aspects such as expression and viewpoint, have distinct representations within these regions, or whether there are overlapping representations. This thesis aimed to further the understanding of the neural representation of facial expression and facial viewpoint, focussing on the neural representation in the pSTS. First, this thesis investigated how expression is represented in the pSTS. A variety of behavioural evidence has shown that face images (in contrast to other objects) are processed holistically. In contrast to these findings, this chapter demonstrated that the pSTS represents facial expressions in a feature-based way, showing changes in response to any change in facial expression. However this chapter was also able to demonstrate that a region considered to be part of the extended face processing system, the inferior frontal gyrus (IFG), has a holistic representation of facial expression that reflects behavioural holistic processing. The second experimental chapter asked whether there are distinct neural representations for processing changes in expression and changes in viewpoint, across the core and extended face-responsive regions. Distinct patterns of response were found for changes in expression and viewpoint in the core regions. These representations were largely invariant to changes in identity, supporting the idea of distinct processing pathways for invariant and changeable aspects of faces. In contrast to the core regions, regions of the extended system (IFG and amygdala) were predominantly selective to changes in facial expression rather than changes in viewpoint. The third experimental chapter asked whether there was a more fine-grained representation of facial viewpoint. Distinct patterns of response to different viewpoints were found. Interestingly, there was also a strong similarity for symmetrical viewpoints in the fusiform face area (FFA) and pSTS. This similarity in the patterns of response to symmetric viewpoint directions suggests that these regions represent an intermediate step towards full viewpoint invariance. The final experimental chapter aimed to determine the relative dominance of expression and viewpoint in the neural representation of the core regions. The patterns of response were largely dominated by viewpoint, perhaps reflecting a neural coding that is linked to the visual properties of the face. Overall, these findings support the existence of distinct patterns of response to expression and viewpoint, whereby these changeable aspects of faces are represented by an overall pattern across the core face-responsive regions, rather than as discreet modules.
|
|---|
