20442
Developmental Plateaus in Gaze Following in 24-Month-Old Toddlers with ASD: An Eye-Tracking Study
Recent work has highlighted eye tracking’s potential to catalogue a range of critical social and communication skills in individuals with ASD (Guillon et al., 2014). Gaze-following skills are particularly interesting given their fundamental connection to joint attention and social monitoring. Although recent eye-tracking studies have examined gaze following in older children and adolescents (Freeth et al., 2010; Swanson & Siller, 2013), no reports have used eye tracking to analyze the developmental processes underlying gaze following in toddlers with ASD.
Objectives:
To use eye tracking to examine developmental changes in gaze following during emulated joint attention in toddlers with ASD between the ages of 18- and 24-months.
Methods:
Participants included 18- and 24-month-old toddlers with ASD (ASD-18, n=77, Age: M=19.3 (SD=1.5) months; ASD-24, n=80, Age: 24.0 (0.9) months), non-ASD developmental delays (DD-18, n=11, Age: 18.4 (1.1) months; DD-24, n=12, Age: 23.5 (1.3) months), and typical development (TD-18, n=87, Age: 18.3 (0.9) months; TD-24, n=89, Age: 23.9 (1.0) months). Participants were administered an eye-tracking task testing spontaneous social attention to discrete probes, including a Joint Attention (JA) probe, in which an actress repeatedly looked at the camera and then at one of four toys (see Chawarksa et al., 2012). Outcome variables included percentages of valid scene-viewing time spent attending to the actress’s head (%Head), any toy (%Toys), and the toy at which the actress looked (%Target). Looking patterns within a circuit of targets, including Toys-cued (TargetRatio), Toys-not-cued (DistractorRatio), and the actress’s head (HeadRatio), were also calculated.
Results:
Linear mixed effects modeling was employed to examine the fixed effects of diagnostic group, age group, and their interaction. Effects for %Head were not significant. %Toys revealed an age x diagnosis interaction (p<.05; toy-looking increased with age in the TD (p<.001) and DD (trend; p=.10) groups but not the ASD group (p=.73)). For %Target, a diagnosis effect (ASD<DD,TD: p=.06, p<.001; DD=TD, p=.94) and an age effect (p<.01; target-looking increased with age) were observed. Circuit analyses confirmed conversion of attentional focus from the actress’s head to the cued toy in the transition between 18- and 24-months in the TD (increasing TargetRatio, p<.01; decreasing HeadRatio, p<.05) and DD (increasing TargetRatio, p=.09) groups but not in the ASD group. For the ASD-24 group, %Toy correlated with Mullen verbal age equivalent scores (r=.31, p<.05), and DistractorRatio correlated with ADOS severity scores (r=.31, p<.05).
Conclusions:
Our results reveal developmental gains in gaze following from 18- to 24-months of age in non-ASD toddlers but not in toddlers with ASD. Although time spent attending to the cued toy increased for all groups over time, toddlers with ASD consistently attended less to the cued toy and did not demonstrate developmental improvement in gaze following. This, combined with relationships between gaze following and characterization variables in ASD, suggests that scanning patterns in response to JA probes may provide clinically meaningful indices of social ability. More broadly, these results indicate that eye tracking may help quantify atypical social information processing in toddlers with ASD and monitor changes due to development or intervention.