Visual Search In Low Risk Infants and In the Infant Siblings of Children with Autism: The Role of Fixation Duration

Background:  Children and adults show enhanced visual search (orienting to an unusual item in an array; e.g., O’Riordan et al., 2001). During visual search, children with autism had both faster reaction times and shorter fixation durations (Joseph et al., 2009). Longer fixation durations in infancy predict poorer cognitive performance in infancy and childhood (Colombo, 1995).   

While typically developing infants have been found to orient to unusual elements in an array (Adler & Orprecio, 2006; Amso & Johnson, 2006), visual search has not been studied in the infant siblings of children with autism.

Objectives:  To examine whether infant siblings have enhanced visual search relative to low-risk infants, and to explore the role of fixation duration. The study may help find early signs of autism and improve our understanding of the broader autism phenotype.

Methods:  We have tested 56 infant siblings (16 6-month-olds, 13 12-month-olds, 17 18-month-olds, and 10 24-month-olds) and 59 low-risk infants (16 6-month-olds, 18 12-month-olds, 17 18-month-olds, and 8 24-month-olds).

Each stimulus consisted of a “plus” sign target among “L” distractors. In the Random condition, 7, 13, or 26 distractors were distributed randomly around the screen; in the Circle condition, 4, 7 or 13 distractors were distributed in a circle. All Random condition stimuli were presented before Circle condition stimuli.

Before each 2 sec stimulus, a central attention-getter was presented. Infants’ gaze was measured with a Tobii eye-tracker; data from trials lacking an adequate central fixation were excluded. ANOVAs, univariate tests, and correlations were conducted on average fixation duration in various conditions, accuracy (number of targets found), and time-to-target (time before the infant fixated the target). Infant siblings also received additional developmental tests.

Results:  Across risk groups, accuracy decreased with increasing array size in Random (p < .01) and Circle (p < .001) conditions. Averaged over array size, accuracy increased with age in Random (p < .001) and Circle (p < .05) conditions. Contrary to predictions, time-to-target for Random arrays was longer in infant siblings than in controls (p < .01).

In both Random and Circle conditions, infant siblings had longer average fixation durations than low-risk infants (p. < .05). In the infant siblings, average fixation duration was negatively correlated with a measure of cognitive development (Mullen Composite t-scores) for both Random (r = -.368, p < .05, N = 31) and Circle (r = -.628, p < .01, N = 26) conditions. Average fixation duration was positively correlated with ADOS algorithm scores (summed across domains) in both Random (r = .532, p < .05, N = 14) and Circle (r = .695, p < .05, N = 10) conditions.

Conclusions:  These preliminary results suggest that infant siblings are not better at visual search than low-risk infants; enhanced visual search may be specific to autism rather than to the broader autism phenotype or may not be evident in implicit visual search paradigms. The results support Colombo’s hypothesis that longer average looking times may be associated with poorer cognitive performance.