Gaze is not Enough: Computational Analysis of Infant's Head Movement Measures the Developing Response to Social Interaction

Infant eye gaze is frequently studied because of its relevance as an indicator of early attention and learning. However, the coupling of eye gaze with an individual’s head motion is often overlooked. This paper analyzes how head motion develops within a social interaction context. To measure this interaction, we developed an approach that can estimate infant head motion from ego perspective recordings as they are typically provided by eye-tracking systems. Our method is able to quantify infant head motion from existing social interaction recordings even if the head was not explicitly tracked. Therefore, data from longitudinal studies that has been collected over the years can be reanalyzed in more detail. We applied our method to an existing longitudinal study of parent infant interaction and found that infants’ head motion in response to social interaction shows a developmental trend. Furthermore, our results indicate that this trend is less visible within gaze data alone. This suggests that head motion is an important element for understanding and measuring infants’ behavior during parent-child interactions.