Diminished Social Network Connectivity in Infancy Predicts Social Communication in Toddlers with High Likelihood of Autism

Background:

Social difficulties form one of the core features of Autism Spectrum Disorder (ASD), and could be associated with variability during early social brain development. Prospective studies examining sensitive measures of social brain development are critical to further examine this. One promising candidate is theta-band EEG power; greater theta responses to faces compared to objects has been related to communication improvement in toddlers with ASD (Dawson et al., 2012). It remains unknown how infant EEG theta connectivity is modulated by social context. Moreover, it is unclear whether theta connectivity modulations relate to likelihood and diagnostic outcome of ASD, or social-communication skills.

Objectives:

The present study is the first to examine a) how infant theta connectivity is modulated by social context, and b) how these socially elicited responses associate with familial likelihood of ASD, diagnostic outcome, and social-communication abilities during toddlerhood.

Methods:

Infants viewed videos of women singing nursery rhymes and spinning toys during EEG recording at 14 months of age. EEG theta connectivity was calculated using the debiased weighted phase lag index for 4 – 5 Hz. Diagnostic outcome and communication abilities were assessed at 36 months of age using parental interviews, questionnaires, and direct observations. The final sample with clean EEG data included 17 infants with low (LL), and 51 infants with high familial likelihood for ASD (HL); 27 infants showed typical development (HL-TD), 8 infants met criteria for ASD (HL-ASD), and 16 infants showed atypical development at 36 months (HL-Atyp).

We used global connectivity (averaged across all channel pairs) and Network Based Statistics to compare whole brain and connection-level differences between social and non-social conditions within the different likelihood and outcome groups. Spearman’s correlations were used to examine associations between theta modulations and communication skills using the Socialization and Communication subscales of the Vineland Adaptive Behaviour Scales-II.

Results:

Whole brain connectivity displayed significant increases during social compared to non-social stimuli (p < .0001, η_p² = .233). This difference in whole-brain theta connectivity between social and non-social stimuli was greater in the LL group than the HL group (p = .011, η_p² = .093, Fig. 1A), whereas this difference was similar between the HL diagnostic outcome groups (p = .593, η_p² = .022).

Further, network analyses showed stronger socially elicited responses for connections between parietal and occipital areas in the LL group; connectivity in this LL network was significantly stronger than in the HL group (p < .0001, η_p² = .180), who showed modulation by social content across a different fronto-parietal network (Fig. 1B).

Finally, individual differences in connectivity within the LL network were associated with later social and communication skills in the whole group of infants (rho’s ≥ .26, p’s ≤ .033, Fig. 2).

Conclusions:

Theta connectivity is a promising measure of social brain development in infancy. Alterations in connectivity patterns are present in infants with an older sibling with autism, and relate to later developmental outcomes. Early altered connectivity may reflect atypical interactive specialisation of the social brain in the context of increased likelihood of developmental atypicalities.