Parent-Infant Neural Dynamics During Social Learning

• In Event: 3-014 - The Influence of Parent-Child Interaction on Child Development: a Multi-Modal Social Neuroscience Approach

Integrative Statement

During early life, social interactions between infants and caregivers provide a powerful stimulant for learning. Yet current neuroscience frameworks are ill-equipped to explain how social interactions potentiate learning in the infant brain. Social learning - learning from and with social partners – varies dynamically in relation to oneself and one’s partner. For example, infants vary in their propensity to learn from different social partners. Even for the same dyad, learning outcomes will vary across instances depending on the efficacy of social interaction. Explaining this highly-variable, but fundamental, form of early human learning therefore requires a paradigmatically-different “two-person” neuroscientific framework. Here, we present dyadic (adult-infant) neural data that exemplify a co-constructivist approach to understanding how early learning occurs in social contexts.
Forty-seven mother-infant dyads (mean age 10.7 months) participated in a social learning paradigm whilst their neural signals were concurrently measured using dual-EEG. On each trial, mothers presented their infants with a pair of novel objects, modelling a positive response to one object and a negative response to the other object (e.g. through smiling/frowning and saying “This is good/bad, I really like/dislike this one”). Infant were then offered access to both objects, and their choices were recorded (up to 16 trials). Infants were classified as being Congruent learners (N=24, positively-modelled>negative-modelled choices), Neutral learners (N=11, positive-modelled=negative-modelled choices) or Incongruent learners (N=12, negatively-modelled>positively-modelled choices). We then examined infants’ and parents’ internal and interpersonal neural connectivity (phase-locking value, PLV) in relation to infants’ (1) Learner typology and (2) trial-to-trial Learning pattern (see Figure 1).
We observed a clear dissociation between the neural connectivity patterns that captured variations across Learners and Learning: Learner typology (i.e.congruent/incongruent) was most strongly associated with infants’ own neural activity. However, trial-to-trial Learning (i.e.choice variation) was only significantly correlated to, and discriminated by, adult-infant interpersonal connectivity (p<.002). Moreover, Congruent and Incongruent learner types showed opposite learning response functions – increased interpersonal connectivity predicted a higher likelihood of positive responding in Congruent infants, but a higher likelihood of negative responding for Incongruent infants (F(9,823)=3.09,p<.01).
Finally, we assessed the impact of maternal social behaviour on infants’ social learning. To assess mother-infant compatibility/attunement, we computed mother-infant temperament dissimilarity, as well as maternal sensitivity toward her infant. To assess maternal speech input quantity and quality we computed her instruction duration and “infant-directedness”. Multiple linear regression analyses (see Fig 2) showed that infants’ neural Learner typology was significantly predicted by maternal speech quantity (β=-.51, p<.001), sensitivity (β=.26, p<.05), and (marginally) temperament compatibility (β=-.22, p=.09). By contrast, infants’ trial-to-trial Learning was significantly predicted only by maternal speech quality (β=.34, p<.05).
In conclusion, we demonstrate that dynamic variations in infants’ social learning are predicted solely by the interpersonal neural connection between parent and infant. However, infants’ learner stance (whether they are likely to accept/reject the lesson) is indexed primarily by internal neural connectivity differences. Maternal behaviour exerts differential effects on these two dissociable aspects of social learning. Our data offer an initial insight into the neural dynamics and mechanisms by which social information is internalised for action during early life.

Authors

• Victoria Leong, Nanyang Technological University, Singapore, and Cambridge University, UK
  Presenting Author

• Valdas Noreika, Cambridge University
  Non-Presenting Author

• Lorena Santamaria, Cambridge University
  Non-Presenting Author

• Kaili Clackson, Cambridge University
  Non-Presenting Author

• Stanimira Georgieva, Cambridge University
  Non-Presenting Author

• Sam Wass, University of Cambridge
  Non-Presenting Author