Abstract
Previous psychophysical evidence suggests that face processing mechanisms can be adapted in an identity-specific manner. Moreover, the effect of adaptation critically depends on the identity strength of the test face, with faces closer to the mean being more affected than faces further from the mean (Anderson & Wilson, 2005). Here, we have studied the underlying neural processes of this identity-specific adaptation effect using an event-related fMRI adaptation paradigm. BOLD responses in the fusiform face area (FFA) were measured for synthetic faces with the same identity but at different distances from the mean using a 3T MRI system. Signals were assessed either without adaptation or after adapting to a strong (12%) anti-face (on the opposite side of the mean). As a control, responses were also measured for an irrelevant task (i.e. contrast judgements for a Gabor patch) using the same adaptation protocol as for the face-processing task. This allowed us to tease apart identity-specific adaptation effects on the BOLD signal from non-specific effects of the adaptation protocol. With no adaptation, the BOLD signal from the FFA showed no marked difference between the faces with different identity strengths. After adaptation, the BOLD responses to faces with identities closer to the mean face were lower than responses to faces with stronger identities. This pattern of responses is qualitatively similar to the pattern of results observed with different identity strengths using psychophysical methods, and may reflect the operation of a gain control mechanism subserving face perception.
Supported in part by the CIHR Training Grant in Vision Health Research and NIH grant #EY002158 to HRW