Visual Attention and Cue Evaluation In a Modified Posner Paradigm: Relation to Social Skills and Symptom Severity

Background: Research on visual attention in young children with autism spectrum disorders (ASD) suggests differences in visual orienting and disengagement (Landry & Bryson, 2004). Visual saccade tasks with adults with ASD suggest right-localized differences in saccadic responses, without generalized abnormalities (D’Cruz et al., 2009). The connection between visual attention and social skills is unclear.

Objectives: We investigated the interaction of low-level attentional abnormalities and higher-order social attention in younger individuals with ASD.  Using a modified Posner paradigm, in which directional cues are presented to indicate subsequent targets, we controlled for (a) cue validity- cues which correctly/incorrectly direct attention toward the target, (b) disengagement from competition- cues remaining on screen during target presentation, (c) salience- manipulating the number of valid cues preceding invalid ones, and (d) laterality- right- versus left-sided targets.  

Methods: Participants ages 7-17 with ASD (n = 17) and typical development (TD; n = 24) completed measures of visual attention and social cognition, including joint attention and theory of mind.  Groups were diagnosed via ADOS and matched on age and FSIQ.  Repeated-measure MANCOVAs examined group differences in reaction time (RT) as a function of experimental condition; bivariate correlations evaluated associations between RT and social cognition.

Results: Error analyses indicated no accuracy-speed tradeoff; RT was included as a covariate in all analyses. There were main effects of validity, p = .002, competition, p = .04, and salience, p < .001.  Across groups, participants were faster with valid cues, with cues that disappeared before the target appeared, and with lower salience cues.  There was a group by condition interaction for competition, p = .001, and a trend for a group by validity interaction, p = .09.  Both interactions indicated greater cue cost in the TD group.

Laterality analyses indicated that both groups were slower for competitive cues to rightward targets.  The cost associated with competitive cues (i.e., main effect) was not significant for leftward targets; however, the TD group demonstrated significantly greater cue cost for competitive cues bilaterally.  Similarly, both groups were slower to respond to invalid cues to both sides, whereas the TD group was significantly more slowed for rightward targets. Findings are consistent with previous research suggesting left lateralization of visual attention and cue evaluation.

Correlational analyses indicated that performance with competitive cues was correlated with behavioral measures of joint attention, r = .38, p = .02.  For the ASD group, (a) invalid cue performance was correlated at the trend level with ASD symptomatology, r = .44, p = .10, and (b) RT to rightward targets was correlated with NEPSY Theory of Mind, r = -.64, p = .01, suggesting a left lateralized bias in both tasks     

Conclusions: Participants with ASD did not exhibit “attentional stickiness” (Landry & Bryson, 2004). Rather, results suggested that greater attention to cue validity and more thorough evaluation of cue predictiveness were associated with better joint attention, theory of mind, and ASD symptomatology.  The link between low-level attentional mechanisms and higher-order social cognition may lie in the intensity of cue evaluation.