The structure of emotion dynamics across early childhood

• In Event: The development of emotion organization from infancy to adolescence

Abstract

Emotions are not isolated events -- one’s current state depends on how preceding states unfold over time. Adult’s emotion dynamics follow predictable patterns, determined in part by valence -- positive emotions more often follow positive rather than negative emotions (Thornton & Tamir 2017;2020). Little is known about how children develop a pattern of emotional experiences that is likewise organized by valence. In three studies, we explored factors that interact with the organization of young children’s experience and perception of emotion dynamics.
In Studies 1a (N=148, age=0;0-5;0) and its preregistered replication 1b (N=255, age=0;0-7;1), parents were asked to imagine a situation where their child was feeling a given emotion and estimate how likely it is that the child would feel a different emotion next. To quantify the extent to which each child’s transitions were organized by valence, we created a regression model predicting the transition likelihood for each pair of emotions from their respective valence. For example, the difference in valence between sad and irritated is small, so we would predict a high transition likelihood between them, whereas the difference in valence between sad and happy is large, so we would predict a low transition likelihood between them. The extent to which reported transitions followed these predictions was our metric of valence organization (valence organization index; VOI), computed as the regression coefficient estimated by this model for each child. The valence organization index increased with age [𝛽1a=0.05, p<0.05; 𝛽1b=0.03, p<0.05] (Figure 1a-b). Over the course of early development children became increasingly likely to transition between similarly valenced emotions and less likely to transition between differently valenced emotions.
In Study 2 (N=239, age=0;2-7;0), we tested whether the organization of children’s emotion transitions interacts with their growing vocabulary and their ability to sustain attention to emotional situations for longer periods of time. Parents reported the likelihood of their child’s emotion transitions, the time-course of children’s emotion intensity, as well as whether their child produces the labels for those emotions. For each emotion, we used the time course of emotional intensity to estimate the child’s emotion half-life. We found that an emotion’s half-life predicted how well valence characterized emotion transitions. Both across and within participants, longer emotional experiences were characterized by higher likelihood of transitioning to similarly-valenced emotions (Figure 1d 𝛽=0.19, p<0.05). In addition, emotions that the child could label (1;0 – 7;0) were more likely to transition to other similarly-valenced emotions, controlling for age (Figure 1c; 𝛽=0.08, p~0.07).
In preliminary data from Study 3 (N=20, age=0;4-0;10), we tested the effects of early experience with parent emotion dynamics on infants’ perception. We measured infants’ pupillary responses to videos of transitions between same- or different-valence emotions and asked caregivers to report the likelihood of their own emotion transitions. 7-10 month-olds showed more consistent responses to transitions that were higher in likelihood (Figure 2; 𝛽=0.06, p<0.005). Together, these findings suggest that children’s language learning, development of prolonged mental states, and the statistics of emotions in their social environment interact with the organization of children’s emerging emotion dynamics.

Authors

• Mira Nencheva, Princeton University
  Presenting Author

• Mark A. Thornton, Dartmouth College
  Non-Presenting Author

• Casey Lew-Williams, Princeton University
  Non-Presenting Author

• Diana I. Tamir, Princeton University
  Non-Presenting Author