Using Hyperscanning Methods to Investigate the Interpersonal Neural Dynamics of Shared Attention in Infancy

• In Event: Innovations in the Developmental Study of Biobehavioral Synchrony

Abstract

Infant attention is higher during joint play with an adult partner, compared to when they engage with toys on their own (Wass et al., 2018). Following traditional, pedagogical approaches to social learning, we know much about the unidirectional neural correlates of early interaction; i.e. how adult behaviour affects infant neural activity (Parise & Csibra, 2013). To understand the neural dynamics subserving increased infant attentiveness in the presence of a social partner, however, we need to investigate infant attention from the perspective of not one, but both members of the dyad. Using data collected in hyperscanning paradigms, we present methods for exploring early interpersonal attention-brain synchronicity; the cross-dyad associations between attention in one partner and neural responses in the other.
Previously, we used cross-correlation analyses to examine the predictive association between fluctuations in parent brain activity and continuous fluctuations in infant attention, whilst engaged in joint play (Wass et al., under review). Infant attention was converted into 1s and 0s (indicating periods of attentiveness and inattentiveness towards an object), before computing time-lagged associations between infant attention and parent EEG activity, at lags -10 to +10 seconds. We found parent neural activity in theta and alpha bands significantly and positively predicted infant attention, with strongest associations over anterior areas at time-lags -3s and +3s in the theta range (Figure 1). Here, we explore a complimentary, event-based approach to investigating early interpersonal attention-brain dynamics.
We recorded dual 32-channel EEG concurrently from infants and their parents whilst they engaged in joint play with 3 toys across a table. Thirty-one participant dyads contributed data (mean infant age= 10.74 months, SD=1.15). Infant and adult gaze behaviour was recorded at 50fps, and synchronised with the EEG signal. Look onsets were coded into 5 categories: object looks, partner looks, and inattentiveness. We subsequently identified the beginning of each infant look towards an object, and recorded the length of the look. Infant and adult EEG power at theta and alpha frequencies in the 2000ms before and after look onset was extracted using a continuous Morlet wavelet transform, and averaged over 500ms time-windows. Linear mixed effects models were calculated for each time-window, examining the relationship between EEG power and infant look duration, controlling for effects of participant.
As an example of these methods, we compared how parent and infant brain activity predicted look duration where infant attention was infant- versus adult-led. Preliminary results at the time of writing (n=31) suggest that infant theta activity did not predict look length in any time window, for infant- or parent-led looks (all ps> 0.05). Whilst parent theta activity was unrelated to infant look length during infant-follower looks (all ps> 0.05), duration of infant-led attention was predicted by parent theta activity 500-1000ms after look onset (p=0.005). These methods suggest that parent neural responsivity to infant attention predicts duration of infant looking, only where attention is infant-led. Implications of these methods for investigating inter-brain synchrony are discussed, as well as their application to work exploring the inter-personal neural dynamics associated with ostensive cues.

Authors

• Emily A M Phillips, University of East London
  Non-Presenting Author

• Ira Marriott Haresign, University of East London
  Non-Presenting Author

• Megan Whitehorn, University of East London
  Non-Presenting Author

• Victoria Leong, Nanyang Technological University
  Non-Presenting Author

• Sam V Wass, University of East London
  Presenting Author