Spatiotemporal dynamics of the face recognition process

Abstract

The spatiotemporal characteristics of the face recognition process have long been a topic of debate. The goal of the present thesis was to address key questions concerning human face recognition strategies. We report results from five face recognition experiments. In the first two studies (Chapter 2), we tested if humans develop identity-specific scanpaths for highly familiar individuals. Participants were primed with the identity of the face stimulus before stimulus onset. We did not find support for identity-specific fixations or scanpaths. In the third and fourth study (Chapter 3), we tested if participants perform complimentary sampling from the opposite side of the face after their first fixation was required to be at a certain starting location on the face. Instead of complimentary sampling, we observed that participants consistently first sampled from the idiosyncratically preferred sampling location near the centre of the face. To recognize a famous face, it was sufficient to place 2-3 fixations at the centre of the face. Fixations to additional idiosyncratically preferred sampling location were needed to recognize unfamiliar faces. In the fifth study (Chapter 4), we tested if participants can be encouraged to employ an analytic or holistic face recognition strategy when they were required to start the face recognition strategy at a feature or at the centre of the face. We found that starting the face recognition process at a facial feature or at the centre of the face, did not have an effect on the choice of the face recognition strategy. Regardless of the starting location, participants first sampled from the centre of the face and then performed further sampling from additional, idiosyncratically preferred locations on the face. We also observed that participants showed a high variability in the transitions between regions of interest on a trial-by-trial basis. In summary, we suggest to describe human face recognition strategies in terms of idiosyncratically preferred primary and secondary preferred sampling locations on the face with highly variable transitions between these regions of interest.