Poster #6 - Speech hampers infants’ face recognition

• In Event: Poster Session 14

Abstract

Prior research has shown that language modulates infants’ face recognition abilities. For example, 9- and 12-month old infants can learn and recognize pictures of own-race faces when associated with their native language but not with a non-native language (Hillairet de Boisferon et al., 2021). Noteworthily, however, infants’ face recognition ability disappeared when using dynamic talking faces instead of still pictures (Hillairet de Boisferon et al., 2021). The authors hypothesized that the processing of the faces’ indexical cues – typically located around the eyes – is diminished when perceiving a talking face, since infants’ attention is shifted towards the talker’s mouth after 8 months of age (Lewkowicz & Hansen-Tift, 2012). However, Xiao et al., (2015) found that 9-month-olds can learn and recognize a dynamic chewing face, even though they look more at the moving mouth than the eyes. Is face processing different for dynamic talking and not talking faces? To investigate this question, we run a face recognition task (Visual Paired Comparison task, as in Hillairet de Boisferon et al., 2021) in a group of 10-month-olds (ongoing data collection; N=16) and a group of 20-month-old infants (N=29), and we analysed eye gaze to the face’s eyes and mouth during two familiarization conditions: a face talking and a face chewing. The chewing control condition allowed us to present a similar amount of mouth movement and audiovisual redundancy, but in the absence of speech. Results from the 20-month-old group showed that, while the chewing condition elicited even greater attention to the mouth than the talking face, infants succeeded in learning and recognizing the face, looking longer at the novel face when it was chewing [β = .10, t = 3.96, p < .01], but not when it was talking [β = -.03, t = -1.19, p > .1, see Figure 1]. Unsurprisingly, attention to the eyes and mouth did not correlate with face recognition scores. The 10-month-old group did not reach significance but showed a trend in the same direction (i.e., chewing: β = .08, t = 1.63, p = .11, and talking face: β = -.002, t = -.029, p = .9), suggesting similar results after group’s completion. These results suggest that infants’ face processing is likely hampered by the cognitive load of processing audiovisual speech rather than by increased attention at the mouth to the detriment of the eyes region. Potential explanations for these results will be presented and discussed together with ongoing follow-up experiments.

Authors

• Joan Birules, University of Grenoble
  Presenting Author

• Mathilde Fort, Université Claude Bernard Lyon 1

• David Méary, Université Grenoble Alpes

• Olivier Pascalis, Centre National de la Recherche Scientifique (CNRS)