Sensitivity to Audio-Visual Synchrony and Its Relation to Language Abilities in Children with ASD

Background:  Recent findings show that increased attention to the mouth is related to greater success in language learning among typically developing children and children with autism (Tenenbaum, Amso, Abar, & Sheinkopf, 2014; Tenenbaum, Sobel, Sheinkopf, Malle, & Morgan, 2014). What is not yet clear is the mechanism by which attention to the mouth may be related to language development. Children with impaired language abilities may have trouble processing speech because it is a complex multimodal signal comprised of both auditory and visual information. Hence, children might look away from the mouth in an attempt to simplify their perceptual experience. Previous findings suggest that children with autism show impaired recognition of audio-visual synchrony (Bebko, Weiss, Demark, & Gomez, 2006; Irwin & Brancazio, 2014; Irwin, Tornatore, Brancazio, & Whalen, 2011; Smith & Bennetto, 2007) and look less at the mouth than typically developing peers in audio-visual processing tasks (Irwin & Brancazio, 2014). 

Objectives:  (1) To replicate findings showing reduced sensitivity to audio-visual synchrony among children with autism; (2) To determine whether sensitivity in processing of audio-visual synchrony is related to language ability among children with autism. 

Methods:  Children with ADOS confirmed diagnoses of autism (AD; n = 23, M = 5.76 years, SD = 1.76) and typically developing children (TD; n = 19, M = 2.62, SD= 1.75) matched on language ability viewed twelve trials. Each trial was composed of two videos of a woman speaking in animated infant directed speech. One side of the screen displayed a video that was synchronized with the audio track and the other showed a video that was asynchronous at one of four delays (0ms, 330 ms, 660ms, and 1000ms). The dependent measure was proportion of looking time to the synchronous video. 

Results:  Repeated measure mixed general linear model was used with group (TD vs. AD) as a between subject factor and delay (0, 330, 660, 1000ms) as a within-subject factor. Results showed a significant main effect of delay (F(3,37) = 11.8, p<0.01). Post-hoc tests revealed a significant difference between the 0- and 1000-ms-delay conditions (p<0.05). No overall group differences or significant interactions were observed. Visual inspection of within group means suggests that TD children were sensitive to all delays whereas AD children were only sensitive to the 1000 ms condition (though differences were not statistically significant). Bivariate correlations revealed a marginally significant relationship between the dependent measure at 1000 ms delay and PLS Expressive Scores in the AD group (r = -.38, p<0.1); TD children showed the reverse pattern (r = 31, ns).  

Conclusions:  These preliminary results suggest that consistent with prior findings TD children are more sensitive to audio-visual synchrony than AD children. Contrary to predictions, AD children with higher expressive language abilities did not show a preference for synchronous videos. This surprising result may reflect differences in the mechanisms underlying language learning in ASD. Data collection is still underway and final analyses will explore this relationship further.