Abstract
We know from previous research that regions within the ventral occipito-temporal cortex (VOTC) such as the occipital and fusiform face areas show a preference for faces and face related features and are involved in their recognition. Previous work has also shown that the recognition of inner face features (e.g. eyes and mouths) is best at their expected / usual location in the visual field. However, it is still unclear if this enhanced recognition performance is caused by a learned adaptation to input statistics or if it follows an innate, face-specific template. The ongoing global pandemic with a surge of people wearing face masks gives us an exceptional chance to investigate the processing of artificial face features. If the tuning properties of face-preferring areas are shaped by input statistics, they may also learn to respond to artificial features, given sufficient exposure. I will present preliminary fMRI results showing a significant overlap of VOTC activations evoked by faces and face masks. Our data also suggest that the neural activity evoked by face masks becomes increasingly similar to that evoked by faces over several months of high exposure. Overall, these findings point towards a high degree of plasticity in face-preferring areas and support the view that its tuning properties are shaped by input statistics. We are currently collecting behavioral data to further test this hypothesis, by probing whether the feature-location contingency observed for natural facial features can be found for artificial features as well.