Contingent Smiling Behavior in 2-to 5-Month-Old Infants with and without ASD during Dyadic Interactions

Background: Social smiling, smiling that is closely linked to vocalizations and affective changes in one’s communicative partner, is one of the most visible forms of early social engagement, emerging by the second month of life in typical development. By contrast, individuals with autism spectrum disorders (ASD) have impaired social interaction abilities, displaying reduced social smiling relative to their TD peers (e.g., Zwaigenbaum et al., 2005). In typical development, the emergence of social smiling marks a pivotal milestone, as early success in reciprocal social interaction likely plays a role in shaping social and communicative outcomes. In ASD, reduced social smiling in later life may reflect the cascading impact of disruptions in reciprocal social interaction from the first months of life. Consequently, examining early smiling behavior may provide an early diagnostic marker of ASD and inform how deviation from the normative course of development impacts the unfolding of social deficits characteristic of ASD. 

Objectives: Investigate whether infants at low-risk for ASD with typical outcomes (LR-TD), and infants at high-risk for ASD with and without later diagnoses of ASD (HR-ASD and HR-unaffected) differ in: (1) frequency of smiles; and (2) whether their smiles are time-locked to the behavior of their communicative partner.  

Methods: Two-to-five-month-old LR-TD (n = 25; mean age = 4.06 months), HR-unaffected (n = 5 ; mean age = 3.6 months), and HR-ASD (n=9; mean age = 4.36 months) infants were recorded monthly during 30-second face-to-face interactions with their caregivers. Infant facial expressions (including smiling) and gaze direction, and caregiver facial expressions and vocalizations were coded as in Lavelli & Fogel (2005). 

Results: One-way ANOVAs revealed no differences in frequency of smiling between LR-TD, HR-ASD, and HR-unaffected infants (p’s > .05). Infants also displayed similar levels of looking towards versus away from their caregiver, and caregivers showed no between-group differences in frequency of infant-directed behaviors (all p’s > .05). Peristimulus time histograms were created to examine the temporal relationship between infant smiling and specific caregiver behaviors. LR-TD infants showed a significant decrease in smiling rate when caregivers displayed neutral expressions, both with and without vocalization (74.46% and 67.39%, respectively; p’s < .05), but showed a significant increase in smiling rate when caregivers were smiling, both with and without vocalization (37.68% and 29.31%, respectively; p’s < .05). Smiling rate of HR-unaffected infants was only time-locked to moments when caregivers were vocalizing while smiling, increasing by 24.05%. Finally, smiling in HR-ASD infants was not modulated in relation to any caregiver behavior (p’s > .05).  

Conclusions: The temporal relationship between infant smiles and caregiver behaviors differed significantly between LR-TD, HR-unaffected, and HR-ASD groups, though infants in all groups displayed comparable frequency of smiling. While LR-TD infants showed changes in smiling rate that were time-locked to affective changes in their caregivers, smiling in HR-ASD infants was not contingent upon caregiver behavior. These results highlight a very early disruption to social smiling and sensitivity to social contingencies in ASD, skills that serve as an essential means of pre-verbal communication and social learning in typical development.