Coordinated Interpersonal Timing in 9-Month Olds At High and Low Risk for Autism Spectrum Disorder and Their Mothers

Background: A body of research amassed over several decades has detailed the development and characteristics of “coordinated interpersonal timing” (CIT), or the mutual influence of the temporal patterning of behavior (typically vocal) of two participants in an interaction (Jaffe et al., 2001).  Like adults, typically developing infants and their mothers adjust the timing of their vocal behavior to coordinate with their conversational partner.  This type of coordination begins as early as 4 months of age and is related to developmental outcomes (Beebe et al., 1988; Jaffe et al., 2001).   

Individuals with autism spectrum disorder (ASD) display difficulties with communication and social interaction, two skills deeply embedded in the coordinated experience of caregivers and infants.  Furthermore, due to strong genetic etiology, infant siblings of these individuals are at high risk (HR) for being diagnosed with ASD and related disorders (e.g., language delays; Zwaigenbaum et al., 2005).   

Objectives: To explore caregiver-infant vocal coordination in HR infant-caregiver dyads with the hope of further understanding how CIT may be altered in this population.

Methods: Participants include 55 nine-month-old infants (30 HR; 25 Low Risk: LR) videotaped at home during a 5-minute toy play interaction with their mothers. At 36 months, HR infants were administered the Autism Diagnostic Observation Schedule and categorized as having no diagnosis (HR-ND), a diagnosis of autism (HR-ASD), or a diagnosis of a language delay (HR-LD). 

All infant and caregiver vocalizations and periods of silence were coded on a moment-by-moment basis.  Variables included vocalizations, pauses, contingent responses, simultaneous speech, and speaking-turns (as defined by Jaffe et. al. 2001).  In order to explore moment-by-moment coordination, the durations of all vocalizations, intra-personal pauses, and switching pauses were averaged across 5-15s time periods, and dyadic cross-lagged regression models were run to examine actor and partner effects. 

Results: Analyses to date suggest that although there is substantial variability between diagnostic groups, there are no differences in the frequency or duration of infant or caregiver vocalizations, pauses, or speaking turns.  Examination of the coordination of these characteristics within dyads suggests an effect of future diagnosis.  Specifically, HR-ND infants and their mothers show strong coordination of intra-personal (r = .827, p<.05) and switching (r = .727, p<.05) pause durations, as well as speaking-turn durations (r = .727, p<.05), while HR infants classified as ASD or LD at 36 months do not.  The positive correlations among HR-ND dyads suggest mutual regulation of the timing of the exchange, similar to that which is found between adults.  Furthermore, dyadic cross-lagged regression models suggest that moment-by-moment coordination of pause duration is evident in these interactions, and that actor effects are moderated by risk (b = .06, t(64) = 2.118, p<.05).  Specifically, LR infants appear to be more predictable based on their own past behavior than HR infants. 

Conclusions: Group differences can be found in the coordination of vocal states as well as in the moment-by-moment predictability of behavior.  These results demonstrate how fine-grained analyses of vocal coordination increase our understanding of the complex interaction between infants and caregivers.