Attention Capture By Faces within a Naturalistic Scene in Toddlers with ASD

Background: Eye-tracking studies have documented atypical attention toward facial stimuli in children with autism spectrum disorder (ASD), including both diminished attentional bias (Chawarska, Macari & Shic, 2013) and atypical scanning patterns (Chawarska, Volkmar & Klin, 2010; Chawarska & Shic, 2009). Such results raise the question of whether or not faces capture the attention of individuals with ASD to the same extent as typically developing (TD) peers. Although adolescents with ASD have been shown to take longer than those with TD to fixate faces in naturalistic images (Freeth et al., 2010), the degree to which faces within naturalistic scenes capture the attention of toddlers with ASD has yet to be investigated.

Objectives: To investigate the attentional capture of faces within naturalistic scenes for TD toddlers and those with ASD.

Methods: 135 toddlers (ASD, n=80; TD, n=55) between 15 and 40 months of age (M_ASD=25.76 months, M_TD=23.16 months) viewed an eye-tracking task testing spontaneous attention toward eight static images. Each image appeared for five seconds and contained a different woman smiling and facing forward within a naturalistic scene (e.g., sitting at an office desk or preparing food in a kitchen). Faces appeared in locations equidistant from a central fixation point that preceded each trial. Primary outcome variables were the average time it took for a child to orient toward the face within a scene (Latency_Face), the average number of discrete gaze shifts toward a face (Shift_Face), and the average percentage of valid scene-viewing time spent attending to a face (%Face). All participants received the Mullen Scales of Early Learning (MSEL; Mullen, 1995), and all toddlers diagnosed with ASD received the Autism Diagnostic Observation Schedule (ADOS-G; Lord et al., 2000). 

Results: ANOVA analysis revealed no significant differences between diagnostic groups concerning the total time spent attending to the scenes (p=0.2). Linear mixed effects modeling was employed to examine the effect of diagnostic group on Latency_Face, Shift_Face and %Face, controlling for age and Mullen nonverbal developmental quotient (DQ). No significant differences were observed between toddlers with TD and ASD on Latency_Face, %Face, or Shift_Face. Effect sizes were small for all variables (Cohen’s ds<0.3). Within the ASD group, Pearson’s r correlations revealed no significant relationships between any of the three variables of interest and Mullen verbal or non-verbal DQs or ADOS total scores.

Conclusions: Toddlers with ASD exhibited no differences from TD peers in attention capture by faces within complex naturalistic scenes. The number of gaze shifts toward faces and percent of viewing time spent attending to faces during the eye-tracking session also did not differ between groups. These results add to a growing body of work indicating that social stimuli capture the attention of young children with ASD (e.g., Johnson, 2013; Elsabbagh et al., 2012). The ability to detect and orient toward faces, even within complex visual environments, appears unimpaired in toddlers with ASD. Mechanisms contributing to the documented attentional atypicalities in these toddlers, as well as relationships to attentional differences later in life, merit further exploration.