The roles of social referential cues and visual saliency on infant sustained attention

• In Event: Enhance cognitive functions across development: factors, strategies or interventions

Abstract

Getting an infant’s attention to the item of conversation focus is crucial for communication and language learning. Previous studies of parent-infant interaction during object play demonstrated that parents play an important role in navigating infant attention to relevant objects by using various referential cues (e.g., Deak, Krasno, Triesch, Lewis & Sepeta, 2014; Suarez-Rivera, Smith, & Yu, 2019). For example, parents often actively handle and show the target object to the infant, and such active involvement during the play enriches the child’s views of the objects (Burling & Yoshida, 2019). Despite the important role of referential cues in navigating attention, we know little about how the referential cues contributing to infant sustained attention in a dynamic scene. One proposal is that parent’s active involvement may be linked with visual saliency, which early attention depends upon. The parent's hands and the handled objects are spatially close with the involvement of rabid motion, which makes the object more salient and increases the likelihood of infant sustained attention to objects (Franchak et al., 2016; Yu & Smith, 2018). The present study first identifies which type of referential cues (e.g., parent’s looking, object handling) are available during a naturalistic object play between parent and infant, and then examine how these cues and levels of saliency may facilitate infant sustained attention to the referent object together. We recruited 43 parent-infant dyads with children aged from 5 to 18 months in the Greater Houston areas, Texas. All the dyads completed a six-minute parent-infant object play session in the lab. In the play session, both parent and infant were worn the head-mounted eye-tracking devices. Their looking behaviors were recorded by eye-trackers, and other behaviors such as object handling and labeling were recorded by the wall-mounted camera and audio recorder, respectively. We annotated an individual’s target behaviors including infant’s gaze, parent’s gaze, verbal labeling, and object handling during the play. First, our preliminary analysis confirmed the active involvement of object play by parents. Results suggest that parents spent over 90% of the play time using at least one of the three targeted referential cues. Particularly, object handling alone accounted for most of the time (34.72%, SD = 0.15). Second, the cross-correlation analyses showed that parents were more likely to respond to infant object looking and then brought up appropriate labels or used hand actions rather than vice versa. Third, we are in the process of implementing the saliency distribution for each frame of the play scene (Borji & Itti, 2013), in order to further assess the role of referential input and saliency on infant sustained attention. We are going to run both the correlational analyses and hierarchical models to address whether infant sustained attention is more likely to occur when fixating salient regions with referential cues.

Authors

• Lichao Sun, University of Houston
  Presenting Author

• Hanako Yoshida, University of Houston
  Non-Presenting Author