Representation of parts and canonical face configuration in the amygdala, superior temporal sulcus (STS) and the fusiform “face area” (FFA)

Abstract

A net work of brain areas are involved in processing faces, including the amygdala, a region in the STS and a region in the fusiform gyrus also known as the FFA. It is unknown if face selectivity in these regions is based on the structural characteristics of faces. Here we asked if the canonical configuration of internal face parts (i.e. two eyes above a nose etc. typical of all faces) is critical to face selectivity in the amygdala, STS or FFA. Sixteen adults were scanned in a 3T scanner, while performing a 1-back task as they viewed images in pseudo-random blocks. Stimulus categories included natural and schematic faces that were either in their canonical configuration or with internal parts randomly rearranged within the face outline. Controls were novel objects and textures. Data were analyzed in SPM99. In each subject, face selective ROIs (natural faces > objects) were defined in the amydala, STS and fusiform gyrus. We found significantly greater responses in the bilateral amygdala and the right STS to the canonical versus rearranged natural faces. Surprisingly, the FFA responded similarly to natural faces that were either canonically configured or rearranged. However rearrangement of schematic faces significantly reduced FFA responses, suggesting a capacity to represent canonical face configuration in the absence of natural features. These results provide evidence for representation of canonical face configuration in the human amygdala and the extrastriate cortex. Furthermore, our data suggest that either the presence of natural face parts or the canonical face configuration is sufficient to evoke strong FFA responses.

Footnotes

 Supported by NIH grant DA15893.