Abstract
In natural viewing environments, we sometimes have to recognize people at a distance; for example, when someone approaches. In such cases, human adults utilize information from both face and body for person recognition (e.g., O'Toole et al., 2011, Hahn et al., 2016). We achieve person recognition by a significant ability to generalize our representation of the face and body of a person across different visual environments. In the developmental literature, previous studies have considered infants' face recognition from birth. These studies reported that infants' face recognition is invariant to rigid or non-rigid facial changes at around 7 months of age (e.g., Fagan et al., 1976), suggesting that infants have some generalizability of face recognition across facial transformations. However, no studies have explored infants' generalizability of face recognition across different natural-viewing environments. Therefore, we studied this ability by examining infants' recognition for the face of an approaching person after learning a face of person talking. To this end, we used a familiarization/novelty-preference procedure with 5- to 7-month-olds. In Experiment 1, infants learned a face with a video of a person talking. Recognition was tested with videos of person approaching from a distance. We found that only 7-month-olds could recognize the face of approaching person. Additionally, we confirmed that this recognition memory in 7-month-olds was based on the face, eliminating low-level visual cues in Experiments 2 and 3. When infants learned a face from a talking video and were tested with a talking video, 5- and 6-month-olds recognized the familiarized face in addition to the 7-month-olds (Experiment 4). In sum, we show that the ability to recognize the face of approaching person develops at 7 months of age. Our results suggest that 7-month-old infants can generalize their face recognition memory across the natural-viewing environments, whereas this ability is limited in younger infants.
Meeting abstract presented at VSS 2018