Visual Processing of Hierarchical Stimuli in Typically Developing Infants

Background:   Research suggests that individuals with autism may process visual stimuli differently than their typically developing peers.  Previous work has shown that children and adults with autism show a bias for detail-oriented processing; however, typical adults demonstrate a global bias when performing hierarchical tasks.  Similarly, research suggests that infant siblings of children with autism show abnormalities in tasks that elicit visual processing in the magnocellular pathway, which carries global aspects of information such as contrast sensitivity and coarse spatial information.  Despite evidence of processing differences between autistic and typical adults, little is known about the developmental trajectory of global versus local biases in perceptual processing in infants or young children.  Understanding how a global bias develops in typically developing children could offer insight into the processing differences seen between autistic and typical individuals.

Objectives:   This study was designed to determine the developmental trajectory of visual processing biases during the first year of life in typically developing infants in order to generate data that could inform future studies of infants at risk for developing autism.

Methods:   Participants were healthy infants (gestational age ≥ 37 weeks at birth) aged three and four months (N=7, mean age = 15.5 weeks), six and seven months (N=12, mean age = 30.8 weeks), and twelve and thirteen months (N=14, mean age = 57.0 weeks).  The infants watched a 110-second presentation on a 17” computer monitor.  During six 15-second familiarization trials, the infant viewed a large letter (“E” or “H”) made of smaller letters (“E” or “H”, whichever was not the global element).  The infant was then given a forced choice between the element that made up the local level and the element of the global level during two 10-second test trials.  Because infants aged three to thirteen months reliably show a novelty preference, infants were expected to look longer at the element they were not biased towards during the familiarization period.  T-tests and univariate ANOVA were used to compare familiarization times as well as element preference during test trial.

Results: There were no differences between the age groups for time spent looking during familiarization (p=0.271) or test trials (p=0.983).  Infants twelve to thirteen months significantly prefer looking to the local element, indicating a global bias during familiarization, during the test trials (p=0.01).  In contrast, three and four as well as six and seven month olds did not show a preference during test trials (p=0.294 and p=0.971, respectively).

Conclusions: Our findings suggest that by twelve and thirteen months of age, typically developing infants display a global processing bias; however, this bias is not present at three to seven months of age.  Understanding developmental change in typically developing perceptual processing will inform hypotheses regarding the perceptual and cognitive mechanisms that go awry in the development of autism.