Influences of Others' Gaze Behaviors on Attentional Allocation during Activity Monitoring in Adults with and without ASD

Background:  Previous work implementing an activity monitoring eye-tracking task in 20-month-old toddlers suggested diminished attention towards people and their activities in toddlers with autism spectrum disorder (ASD) when compared to developmentally delayed and typically developing (TD) comparison groups (Shic et al., 2014). However, no significant differences in attentional allocation between ASD and comparison groups were found in response to varying gaze cues from actresses in the task. It is unknown if this lack of differential response to others' gaze behaviors in ASD and TD toddlers remains stable across development.

Objectives:  To explore the effects of others’ gaze behaviors on attentional allocation during activity monitoring in adults with and without autism.

Methods:  Participants included high-functioning adult males with autism (n=18; M_age=24.89±6.45; M_FSIQ=102.26 SD=14.31) and TD adult males (n=19; M_age=26.68±4.33; M_FSIQ=117.68 SD=9.70). One participant with ASD was lost due to missing data, and an additional 2 ASD and 2 TD participants were excluded due to high numbers of trials with invalid calibration (n=15 ASD, n=17 TD remaining). Participants were seen over two days to complete clinical and experimental assessments, including an eye-tracking battery. One eye-tracking task was an activity monitoring task, in which participants viewed 12 video clips, each 20s in duration, of 2 actresses interacting in naturalistic scenes, engaging in a shared activity. This task was parsed into two gaze conditions: (1) the actresses fixated on the shared activity (activity gaze condition), (2) the actresses fixated on each other (mutual gaze condition). Analyses were conducted to examine the effect of gaze condition on the proportion of time participants spent looking at the actresses’ heads (%Head).

Results:  Linear mixed model analyses revealed a significant effect of gaze condition on %Head (p<.001, d= 0.67), with both TD and ASD participants directing their gaze more to actresses’ heads during mutual gaze trials. However, there was a significant group x gaze condition interaction (p<.05), with differences in %Head in response to gaze condition significantly larger in the TD group. 

Conclusions:  Comparisons suggest that high-functioning individuals with ASD exhibit sensitivity to gaze behaviors of others during activity monitoring, but to a lesser degree when compared to TD participants. Unlike toddlers, adults with and without ASD appear to attend to actresses’ gaze information. The significant group x gaze condition interaction may be attributed to the fact that the mutual gaze condition deviates more notably from typical gaze patterns, which may drive increased fixation to heads in TD participants. It is possible that the high-functioning adults with ASD were not recognizing the deviations from typical gaze patterns to the same degree as TD participants. These findings differ from those previously observed in toddlers, suggesting that sensitivity to gaze information in ASD and TD individuals presents differently across development, with once similar responses in childhood diverging in adulthood. Examination of response patterns in school-aged children and adolescents, and across greater variability in social and cognitive functioning, is still necessary in order to understand the complete developmental trajectory of sensitivity to gaze information during activity monitoring in ASD.