Assessing Temporal and Contextual Factors Affecting Preferential Attention to Faces in Individuals with High and Low Levels of Autistic Traits

Thursday, May 12, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
C. I. Mitchell, C. Dickter and J. Burk, College of William & Mary, Williamsburg, VA

Past research on attentional preferential attention to faces (face-bias) in individuals with and without autism spectrum disorder (ASD) shows conflicting results, depending on how long stimuli are presented (the stimulus onset asynchrony (SOA)), and what types of distractor stimuli are used (Bar-Heim et al., 2006; Moore et al., 2012). Inverted faces and face-like objects such as cars are often used as “control” distractor images but there are important perceptual differences between these stimuli that might account for discrepancies in study results (Hadjikhani et al., 2009; Haxby et al., 1999). Finally, as ASD individuals process social stimuli more slowly than individuals without ASD (Dawson et al., 2005), examining face-bias at different SOAs could reveal temporal factors influencing attention to faces in this group.


This study examined preferential attention to faces in a sub-clinical sample of students with different levels of autistic traits using a dot-probe paradigm with four distinct SOAs and two types of distractor stimuli. 


Participants consisted of 105 undergraduate students (36 males; Mage= 19.10), who completed the Autism Quotient (AQ; Baron-Cohen et al., 2001), a self-report measure of autistic behaviors. Each participant completed a dot-probe task to measure attentional bias to probes following a target or distractor stimulus. Faces were target stimuli; cars and inverted faces were distractor stimuli. SOAs of 100ms, 200ms, 500ms, and 1000ms were used. A median split on AQ scores (M=20.26, SD=8.34, Median=21) divided participants into High AQ and Low AQ groups for 


Face-bias scores were created by subtracting the mean reaction time (RT) on distractor-cued trials from the RT on face-cued trials. A 2(Distractor type) x 4(SOA) x 2(AQ group) mixed ANOVA indicated a three-way interaction, F(3,309)=3.10, p=.03. This interaction was examined separately for the High (Figure 1a) and Low (Figure 1b) AQ groups. There was a significant Distractor type x SOA interaction only for the High AQ group, F(3,165)=4.29, p<.01. Pairwise comparisons indicated that High AQ participants had significantly lower face bias scores at 100ms in the inverted face condition than in the car condition, t(55)=2.90, p<.01, but marginally higher face bias scores in the inverted face than the car condition at 1000ms, t(55)=1.671, p=.10 (see Table 1 for means). Additionally, one-sample t-tests were run for High AQ and Low AQ participants for each Distractor/SOA to identify face-bias scores significantly different from zero. Low AQ participants exhibited marginally positive face-bias scores at 200ms for both car, t(48)=1.70, p=.10, and inverted face, t(48)=1.01, p=.11, distractor trials. Conversely, High AQ participants exhibited significantly negative face-bias scores in the car distractor condition at 1000ms, t(55)=1.97, p=.05, and in the inverted face condition at 100ms,t(55)=2.85, p<.01.


These results suggest that individuals who are high in autistic traits show a different temporal pattern of attention to faces depending on what kind of competing stimulus is present, with a trend over time of increasing face-bias for inverted face distractors and decreasing face-bias for car distractors. These results highlight the importance of examining temporal and contextual factors in face-bias studies when drawing conclusions.