Abstract
Familiarity increases our ability to recognize faces in challenging viewing conditions. We investigated how brain response patterns change as unfamiliar faces become visually familiar. A pattern-based classifier was applied to discriminate the spatial and temporal characteristics of brain responses in areas implicated in familiarity and face processing (FFA, OFA, fusiform gyrus, lateral-occipital areas, and parietal areas including the precuneus). First, in a behavioral study, we developed a learning paradigm for achieving high, medium, and low levels of familiarity by varying the number of learning exposures to multiple views of a face. Recognition of the learned faces from whole-body images increased as a function of familiarization level. Next, in an fMRI experiment, a second set of participants learned faces in the high, medium, and low familiarization conditions before the scan. Neural data were recorded while participants viewed blocks of high, medium, and low familiarized faces, as well as unfamiliar faces. The pattern classifier was used to discriminate the neural activity elicited in response to viewing faces from all pairs of familiarity conditions (e.g., high vs. medium). The most accurate discrimination was found using the combination of the fusiform gyrus, lateral-occipital areas, and parietal areas, suggesting coordination of these areas in the neural coding of face familiarity. Moreover, in the FFA-OFA, discrimination scores increased as a function of increasing levels of familiarity. A pattern classification analysis applied across the temporal sequence of the block showed that the pattern of FFA-OFA response differed with familiarity. Highly familiar faces were classified from unfamiliar faces earlier in block than medium and low familiarity faces. Finally, we confirmed the lack of neural magnitude differences in FFA-OFA as a function of familiarity. This suggests that the acquisition of familiarity is characterized by differences in spatial and temporal patterns of neural activation across a network of brain areas.
Meeting abstract presented at VSS 2012