Within-person variability promotes learning of internal facial features and facilitates perceptual discrimination and memory

Recent research indicates that exposure to within-person variability is essential for developing robust representations of new faces. For example, people perform better on a face matching task after exposure to highly variable photos, compared to less variable photos. However, the specific aspects of face processing that benefit from variability remain unclear. We investigated whether within-person variability improves the ability to match and recognise individual faces, and whether it promotes learning of internal facial features. In one exploratory and one confirmatory experiments, we tested matching and recognition performance of participants after they learned 4 individual faces in a high variability condition and another 4 in a low variability condition. Further, to assess if variability promotes robust learning of invariant facial features (e.g., eyes, nose), we compared performance with and without external facial features (full headshots vs cropped images showing only internal features). We found a large benefit of variability in the recognition task, and a smaller effect on the matching task, but the size of the benefit was comparable with and without the presence of external features. Therefore, within-person variability improves a variety of face recognition skills, and it encourages the encoding of internal facial features.