A Dynamic Eye-Tracking Paradigm: Eye Gaze Patterns in Typically Developing and Infants at-Risk for Autism

Background:  The ability to follow the gaze of others is an essential advancement in the development of social attention. Eye gaze sets the foundation for the initiation of joint visual attention and understanding the intentions, mental states, and goals of others (Gredeback et al., 2008; Jones & Klin, 2013). Deficits in eye contact and gaze have been hallmarks of autism spectrum disorders (ASD; Chawarska et al., 2012; Swanson et al., 2013). A recent focus on comparisons between typical (TD) and at-risk development of visual fixation patterns using eye-tracking paradigms may be a proxy for early indicators of gaze discontinuity and may classify the phenotypic behaviors of infants’ visual attention patterns for early identification and intervention purposes (Frank et al., 2014; Merin et al., 2007).

Objectives:   To examine preliminary individual differences in eye gaze patterns in TD and infants at-risk for autism. 

Methods:  Participants include 37 males and 31 females consisting of (N = 60) TD and (N = 8) infants at-risk for autism with chronological ages ranging from 3 to 12 months (M = 7.59, SD =3.35). Participants were considered to be at-risk if they had an older sibling diagnosed with ASD. A Tobii Studio eye-tracking model captured the visual fixation patterns during a dynamic clip of Peter Pan. The total on-screen area was divided into areas of interests (AOIs) including eyes of the characters, mouth and chin region, and a salient object (ball) region. Gaze fixation counts and duration were calculated by measuring the frequency and duration of each individual fixation made within an AOI. 

Results:   Table 1 displays results of a 2-way ANOVA that examined the effect of developmental status (TD or at-risk) and age (3, 6, 9, and 12 months) on total gaze fixation. There was a significant effect of developmental status on total gaze fixation counts within the AOIs, F (1, 60) = 5.57, p < 0.05, ω = 0.13. There was a non-significant effect of age (p = 0. 32) and an interaction term between age and developmental status (p = 0.67). Independent samples T-tests were conducted to compare average gaze duration within AOIs (eyes, mouth, and object) in TD and at-risk groups. There was a significant difference in gaze duration towards the mouth between TD (M = 5.46, SD = 5.38) and at-risk groups (M = 1.94, SD = 2.12); t (66) = 1.82, p < 0.05. A significant difference was also found for gaze duration on the eyes between TD (M = 6.80, SD = 6.16) and at-risk groups (M = 2.44, SD = 1.62); t (66) = 1.98, p< 0.01 (see Figure 1).

Conclusions:   This study was designed to understand the differences that exist in TD and infants at-risk for ASD for gaze patterns as they relate to social attention within the first year of life. Few studies focus on eye gaze trajectories during dynamic social paradigms in young infants. This study contributes to understanding early neurodevelopmental and endophenotypic mechanisms that could inform later atypical development.