The role of variability in appearance, exposure and learning procedure in dynamic face learning

Research suggests that humans recognize familiar faces reliably but struggle to learn new ones. Uncontrolled variations—in facial appearance and viewing conditions—facilitate learning new faces in lab-based studies. Conversely, stability in appearance seems to support early learning stages in the real-world and changes in appearance disrupt recognition regardless of familiarity. To reconcile these findings, we proposed a cost-efficient learning mechanism: stability forms initial coarse representations, while variability yields refinement over time. We tested this through five pre-registered experiments (2024- 2025) using strictly controlled but ecological stimuli. Caucasian participants in different exposure groups learned two stable and two variable Caucasian faces (3, 6, 9, or 12 videos/face), via spread learning episodes (Experiments 1, 2 and 5) or grouped blocks (Experiments 3 and 4). The similarity between test and learning materials varied across experiments. Stable faces were recognized better than variable ones, but only when test images resembled stable learning material. Recognition improved with additional exposure only in the spread learning condition. Although the nature of our sample limits immediate cross-cultural generalization, our framework offers a basis for future studies across diverse populations. Overall, our findings help bridge real-world and lab-based face learning research, refining current theoretical accounts of face learning.