Emotional communication between parents and children is crucial during early life, yet little is known about its neural underpinnings. Here, we adopt a dual-brain connectivity approach to assess how emotional valence modulates the parent-infant neural network.Fifteen mothers modelled positive and negative emotions toward pairs of objects during social interaction with their infants (aged 10.3 months) whilst their neural activity was concurrently measured using dual-EEG. Intra-brain and inter-brain network connectivity in the 6-9 Hz (infant Alpha) range was computed during maternal expression of positive and negative emotions using directed (partial directed coherence) and non-directed (phaselocking value) connectivity metrics. Graph theoretical metrics were used to quantify differences in network topology as a function of emotional valence. Inter-brain network indices (Density, Strength and Divisibility) consistently revealed that the integration of parents' and childrens' neural processes was significantly stronger during maternal demonstrations of positive than negative emotions. Further, directed inter-brain metrics indicated that mother-to-infant directional influences were stronger during the expression of positive than negative emotions. These results suggest that the parent-infant inter-brain network is modulated by the emotional quality and tone of dyadic social interactions, and that inter-brain graph metrics may be successfully applied to examine these changes in interpersonal network topology.
(200/200 words)
Keywords:EEG hyperscanning, network connectivity, graph theory, emotional expression, mother-infant interaction atypical patterns of EEG asymmetry, commonly showing higher right frontal EEG activity than controls (Gotlib et al., 1998). Recent research has also started to examine intraindividual network topology during emotion processing using graph-theoretic measures. For example, a recent study with adults showed that EEG graph-theoretic features performed better than traditionally used EEG features (such as spectral power and asymmetry) on the automatic classification of affective neural states (Gupta et al. , 2016).Behavioral and neuroimaging studies into early development suggest that the neural architecture for the detection and prioritized processing of emotional expressions, such as fear, emerges sometime during the first year of life (Hoehl, 2013; Hoehl et al.