When High-Risk 2-Year-Olds without ASD Talk the Most

Background:   Recent research using home-based digital audio recordings at 9 months of age suggests that some children at high-familial risk for ASD are “hyper-vocalizers,” producing more vocalizations than low-risk peers (Swanson et al., in press). More vocalizations are associated with higher rates of conversational turn-taking, potentially contributing to a “social feedback loop” that is beneficial to children’s social development (Warlaumont, 2014). It is not yet known whether increased vocalization rates remain characteristic of high-risk children as they age, or whether these patterns differ in children ultimately diagnosed with ASD. In this study, we test “hyper-vocalization” in a new setting (clinical assessment) with older children (2-year-olds) and gold standard human coding to identify speech and non-speech vocalization.

Objectives: Determine whether vocalization patterns distinguish 2-year-olds: (1) at high-familial risk for ASD by virtue of having an older sibling with ASD, but not currently diagnosed with ASD themselves (HR-), (2) at low-familial risk for ASD with no ASD diagnosis themselves (TDC), and (3) children who have an older sibling with ASD and are also diagnosed with ASD (ASD).

Methods:   Thirty-three 2-year-olds (9 ASD, 12 TDC, 12 HR-) were administered the Communication and Symbolic Behavior Scales (CSBS; Shumway & Wetherby, 2009) as part of a longitudinal study (IBIS; Estes et al., 2015). Vocalizations were coded as “speech” or “non-speech” by two blind reliable transcribers (“vegetative” vocalizations were excluded from the current analyses). T-tests compared the percentage of total recording duration occupied by child vocalizations of any type, percentage of speech and non-speech vocalizations relative to total recording length, and number of speech and non-speech vocalizations produced per 10 minutes.

Results: HR- children vocalized for a significantly greater percentage of time (21.5%) than ASD children (13.5%; p<.05), with TDCs falling between (16.7%). All groups produced a significantly higher percentage of speech vocalizations than non-speech vocalizations (all ps<.05). The HR- group produced the highest percentage of speech sounds, with a significantly higher percentage (17%) than the ASD group (11.6%; p=.05), but not statistically more than the TDC group (14.5%), which was also not different from the ASD group. HR- children produced a greater number of speech vocalizations per 10 minutes (Mean=6.56, SD=2.38) than ASD children (Mean=4.07, SD=1.92), suggesting more frequent speech vocalizations. The TDC group mean (5.81, SD=2.71) did not differ from either risk group. Non-speech vocalization rates did not differ by group.

Conclusions:   Our data, using gold standard human coders and standardized assessment, are consistent with recent research using home-based recordings to capture vocalizations in 9-month-old infants at high- and low-risk for ASD. Our convergent data suggest that some children at high risk for ASD may possess characteristics that lead to increased vocalization. These same characteristics may simultaneously confer protection against developing ASD, in part by contributing to an active “social feedback loop” that supports social development by providing enriched opportunities for conversational turn-taking (Warlaumont et al., 2014). We anticipate coding 6- and 12-month-old assessments for each child prior to May 2017, to model developmental trajectories in speech and non-speech vocalization across risk groups.