About this paper symposium

---

Panel information
Panel 12. Methods, History, Theory

Paper #1
New online method to assess infants' social responses
Author information	Role
Dr. Yiyi Wang, University of Chicago, United States	Presenting author
Marc Colomer, University of Chicago, United States	Non-presenting author
Hyesung Grace Hwang, University of California Santa Cruz, United States	Non-presenting author
Caroline Groves, New York University Grossman School of Medicine, United States	Non-presenting author
Courtney Filippi, New York University Grossman School of Medicine, United States	Non-presenting author
Amanda Woodward, University of Chicago, United States	Non-presenting author
Abstract
Observationally coding infants' social responses is crucial for studying early social development. Due to challenges in recruiting infants for in-person research, researchers have increasingly turned to online studies using platforms like “Children Helping Science” (Scott & Schulz, 2017). This approach lowers entry to allow for larger, more diverse samples and provides new possibilities for data sharing, procedure standardization, reproducibility, and analytic innovation. However, it is unclear whether and how online studies can capture meaningful variation in infants’ social behaviors. To fill this gap, we created an online, non-contingent paradigm to assess infants’ social responses to strangers. In the paradigm, infants viewed three types of videos, in which strangers would approach, play peekaboo, and make animal sounds (see Figure 1). The videos were produced by professional actresses to ensure consistency in timing and emotional tone across actresses. In total, 108 17- to 20-month-old infants were tested on this paradigm with each viewing two actresses (one speaking English and another speaking Spanish) alternately across video types. To assess infants’ social reactions to these online videos, we developed a holistic social response coding scale. The scale integrated facial expressions with interactive behaviors like imitation, approaching, avoidance, and social referencing. The scale had 5 levels, ranging from -2 (very negative/withdrawal) to 2 (very positive/approaching) (see Table 1). Two independent coders used this scale, demonstrating strong inter-rater reliability (ICC = 0.96 at the subject level and 0.81 at the trial level). Multiple pieces of evidence supported the reliability of this new method. First, infants' social responses were negatively related to their daily social fear behavior, as reported by their parents from the Toddler Behavior Assessment Questionnaire (TBAQ, Goldsmith, 1996). Infants with higher social fear showed more negative responses in the online task (r = -0.27, p = .005). Second, infants' social responses were positively correlated with the valence assessed by BabyFaceReader (r = 0.53, p < .001), a software tool that uses deep artificial neural networks to analyze facial expressions based on the Facial Action Coding System (FACS, Ekman et al., 2002). Third, infants' social responses varied as a function of the type of video they watched. They exhibited more negative responses to the approach video than to the other videos, with the coding showing time-locked response to the actress's actions (peekaboo, b = -0.11, p = .045; animal, b = -0.12, p = .045). Fourth, infants’ social responses differed based on the language group of the actresses. Infants exhibited more positive responses to the actress who spoke English (native language) compared to the other actress who spoke Spanish (foreign language), b = 0.07, p = .028. In conclusion, our online paradigm, combined with the holistic social response coding scale, offers a valuable tool to investigate infants' social responses in various fields, such as temperament (Fox et al., 2001), social categorization (Liberman et al., 2017), and emotion development (Izard et al., 1980).

Paper #2
Examining children’s body posture as an objective measure of children’s subjective experience of positive emotions
Author information	Role
Robert Hepach, Ph.D., University of Oxford, United Kingdom	Presenting author
Stella Gerdemann, MPI for Human Development, Germany	Non-presenting author
Marlene Försterling, University of Oxford, United Kingdom	Non-presenting author
Abstract
Emotions play a pivotal role in children’s social, moral and cognitive development (Buss et al., 2019). One important function of emotions is to regulate children’s social interactions, and the systematic study of so-called prosocial emotions has provided important insights into how emotions facilitate and maintain children’s social relationships (Vaish & Hepach, 2018). Crucial evidence comes from objective, physiological measures of children’s emotional experiences and while measures emotional arousal (skin conductance, pupillometry) have long been established, there exist fewer measures of emotional valence. This is in part due to the difficulty of capturing children’s experience of feeling ‘good’ vs ‘bad’ without relying on subjective verbal responses or time-intensive manual coding of children’s facial expressions. Here we provide a synthesis of work spanning 7 years across more that 10 independent studies completed in 5 separate labs that have measured changes in children body posture to capture the expression of positive (and negative) emotions. The majority of this work is published but for each of the sets of studies reviewed below, we also draw on ongoing studies to further validate the measure. The first set of studies measured changes in posture to study children (pro)social motivations, i.e., asking how ‘satisfied’ children were with the various outcomes of social interactions. This line of work has provided important insights into the development or prosocial motivations (Gerdemann et a., 2022). It also established that emotions such as pride, shame, joy, and sadness can be induced and measured in primary school children (Försterling et al., 2024). We highlight how the study of children’s posture embraces methodological advances. Whereas initial studies relied on specialized cameras, newer approaches can capture posture from ‘standard’ video recordings. The second set of studies investigate the signaling function of emotions expressed in children’s body posture. This is line of work revealed that being observed affects the extent to which children feel positive (i.e., elevated posture) after receiving a reward or after helping others. This observer-effect is more pronounced in children older than five years of age and thus dovetails with a period in development which children being to manage their reputation in the presence of others (Hepach et al., 2023). In ongoing work, we explore whether such changes in body posture are noticeable by observers and whether this as an effect on observers’ social judgments (Försterling et al., in prep) In summary, the work we present here complements research on measuring children’s affect via facial expressions. Interestingly, some of the existing work on facial expressions in young children included videos and images of children that also conveyed postural information (e.g., Aknin et al., 2015). We discuss how objective measures of body posture can go beyond subjective reports of a child’s experienced emotions and how they might depend on the specific emotion expressed. They provide a window into children’s expressed emotions and thereby enable insights into the underlying motivations of children’s (pro)social behaviour. Future research will need to further explore the feasibility of body posture analyses to study individual differences in children’s social development.

Paper #3
New method of incorporating caregivers to assess infants' social understanding
Author information	Role
Dr. Ashley J. Thomas, Ph.D., Harvard University, United States	Presenting author
Kris Brewer, Harvard University, United States	Non-presenting author
Abstract
A primary task of infancy is to learn about the individuals and relationships in their environments. While Attachment theorists have long studied infant-parent bonds and their impact on well-being and cognition, this work does not ask how infants think and learn about people beyond their primary caregivers, nor does it establish how infants think about their caregivers. At the same time, research on social cognitive development uses highly controlled experiments to study how and whether infants understand observed actions and interactions. However, this work typically uses stimuli involving unfamiliar people, puppets, or animated characters. Consequently, we have limited knowledge of how infants apply these skills to understand individuals and relationships in their extensive social networks. To bridge this gap, I propose a new method which uses video-editing techniques, and yoked control designs. This method allows comparisons between how infants think about actions and interactions of familiar people and unfamiliar people in controlled, laboratory-like studies. I will discuss four projects using this method to test the following hypotheses: (1) Infants learn about new people from their parents' interactions, but not from strangers' interactions. (2) Infants reason better about their caregivers' minds compared to unfamiliar individuals. (3) Infants track and remember known individuals more than unknown individuals. (4) Infants differentiate between caregivers and learn about observable relationships in their environments.

Paper #4
How EEG can provide novel insights into infants’ social group understanding
Author information	Role
Dr. Hyesung Grace Hwang, Ph.D., University of California Santa Cruz, United States	Presenting author
Yiyi Wang, University of Chicago, United States	Non-presenting author
Marc Colomer, University of Chicago, United States	Non-presenting author
Amanda Woodward, University of Chicago, United States	Non-presenting author
Abstract
An enduring challenge in infancy research is difficulty assessing what cognitive and affective processes underlie behavior. A prime example of this challenge is understanding how social bias emerges in infancy. Looking time experiments are predominately used with infants and show infants prefer looking at people based on social groups, such as race and language (see Liberman et al., 2017 for a review). However, it is difficult to pinpoint what underlies such looking patterns. These patterns could be due to a cognitive bias to attend to members of one’s own social group. Or infants may simply feel more comfortable with familiar ingroup members – an affective response. Or infants may show increased mirroring of ingroup than outgroup members' actions that is suggested to be a foundation for their recognition and imitation of others. From behavior alone, these possibilities are often difficult to tease apart. EEG allows observing the neural correlates of these processes: (1) top-down attention (frontal theta; Begus et al., 2015); (2) approach-withdrawal responses to strangers (Frontal Alpha Asymmetry (FAA); Fox & Davidson, 1987), and (3) mirroring of others’ actions (mu event-related desynchronization (mu ERD); Marshall & Meltzoff, 2011). High temporal resolution of EEG enables disentangling not only which processes might be at work during perception of an outgroup member but also how these processes might unfold in time, simultaneously or successively. Although EEG has traditionally been used to assess cognitive, emotional, or clinical issues in infancy, applying it to understand social understanding is a novel use of this useful tool. To demonstrate this, we first conducted a secondary analysis of a dataset on infants’ EEG responses to goal-directed actions. We analyzed 84 7- to 12-month old infants (48 White; 40 racially minoritized) and found neighborhood racial diversity modulated White infants’ EEG oscillations (b=6.83, p=.016): White infants from racially diverse neighborhoods showed greater frontal theta (top-down attention; b=4.02, p=.022) and mu ERD (mirroring; b=−4.9, p=.040) to racial outgroup than infants from racially homogenous neighborhoods; no FAA effects were found. This evidence allowed us to systematically apply EEG to investigate infants’ neural response to linguistic and racial groups. We conducted a second study examining EEG activities of 60 8- to 12-month-old English-hearing infants who viewed linguistic ingroup (English) and outgroup (French) speakers grasp objects. Mu ERD was modulated by language group and infants’ language exposure (b=6.83, p=.005): Monolingual infants showed greater mu ERD (mirroring) to their linguistic ingroup compared to outgroup (b=-.47, p=.015), but bi/multilingual infants did not. No effects of frontal theta or FAA were found. We will also present preliminary evidence from a third study designed to elicit FAA from 34 7- to 12-month infants as they viewed racial and linguistic ingroup and outgroup people approach and play peek-a-boo with them. We expect infants to show more positive FAA (indicating greater approach motivation) to ingroup than outgroup members. Together, these studies demonstrate that EEG can provide novel insights, complementing traditional behavioral paradigms in studying infants’ social understanding.

⇦ Back to session