Now You See It, Now You Don't: Context-Dependent Dyadic Vulnerabilities in Infants with ASD in the First Year of Life.

Friday, May 12, 2017: 3:16 PM
Yerba Buena 9 (Marriott Marquis Hotel)
S. Macari, A. Milgramm, P. Heymann, F. E. Kane-Grade, E. Hilton and K. Chawarska, Child Study Center, Yale University School of Medicine, New Haven, CT

The majority of studies of high-risk (HR) infant siblings with ASD have found little evidence of social abnormalities at six months of age, including social smiling, response to name, and gaze to a social partner (Bryson et al, 2006; Landa et al., 2007; Ozonoff et al., 2010; Nadig et al., 2007; Rozga et al., 2011; Young et al., 2009). Context may be key to revealing vulnerabilities, however. For example, video segments in which a person addressed the viewer using child-directed speech elicited longer looks away from the screen by toddlers with ASD, while segments containing the person engaging in other activities attracted their attention comparable to controls (Chawarska et al., 2012). Little is known about the effects of context within live interactions on social behavior of infants at risk for ASD.


To examine differences between 6-12-month-old infants later diagnosed with ASD and controls in social attention and directed affect during a variety of dyadic interactions with an examiner.


184 infants at HR and low risk (LR) for ASD participated in a standardized social interaction with an examiner at 6, 9, and 12 months of age, and were evaluated for ASD by blinded psychologists at 24/36 months: HR-ASD; n=21, HR without ASD (HR-nASD; n=104) or LR typically-developing (LR-TD, n=59). Examiners administered a series of one-minute probes: speaking in motherese, singing, playing peek-a-boo, engaging in a tickle game, and demonstrating a toy. Sessions were videotaped and coded offline by blinded coders for infants’ visual attention (looking at examiner’s face) and affect (directed positive/negative affect), standardized over the length of each probe (%Face and %DirAffect). Hypotheses were evaluated using linear mixed models followed by planned contrasts.


Anaylsis of %Face revealed main effects of group (F(2,2360)=8.78, p=.001; HR-ASD<HR-nASD=LR-TD) and probe (F(4,2360=15025.02, p<.001; toy<tickle=motherese<song<peek-a-boo), a significant probe by group interaction (F(8,2360)=2.67, p=<01), and a significant age bygroup interaction (F(3,2360)=3.59, p<.01). Planned contrasts indicated that the HR-ASD group looked less at the examiner than both the HR-nonASD and LR-TD groups during motherese and tickle (ps<.05) but at rates similar to controls during peeka-a-boo, song, and toy probes. All three groups differed from each other at 6 months but not at other ages (ps<.05). Analysis of %DirAffect revealed main effects of group (F(2,2394)=7.12, p=.001; HR-ASD<HR-nASD<LR-TD), probe (F(4,2394)=154.26, p<.001; toy<song=motherese<tickle<peek-a-boo), and age (F(2,2394)=5.21, p<.01; 6mo<12mo), and a significant probe by group interaction (F(8,2394)=1.94, p=.05). Planned contrasts indicated that the HR-ASD group directed less affect to the examiner than the LR-TD group during peek-a-boo and tickle (ps<.05) but at rates similar to controls during the motherese, song, and toy probe.


Infants age 6-12 months later diagnosed with ASD expressed vulnerabilities in social attention and shared affect during certain interactions but not others. The tickle and motherese episodes elicited less modulated social attention, while the tickle and peek-a-boo episodes exposed diminished sharing of affect compared to controls. As in eye-tracking studies presenting complex social content, varied in-vivo context can elicit vulnerabilities or be an equalizer for infants in the first year who later develop ASD.