Face Recognition Abilities in Children with ASD, Their Unaffected Siblings and Parents

Background: Difficulties in person identity recognition via images of the face among individuals with an autism spectrum disorder (ASD) are widely documented (e.g., Wolf et al., 2008).  However, reproducible estimates of the magnitude of the deficit from studies with large samples and appropriate measurement tools are lacking.  To date, there are also no family data on whether the face recognition deficits among probands are shared by unaffected siblings and parents. 
Objectives: (1)  To estimate the effect size of face recognition deficits in a suitably large sample of typically developing controls (TDCs)  and participants with ASD using a widely accessible clinical measure – the Benton Test of Facial Recognition (3Benton²).  (2) To test for face recognition deficits in unaffected parents and siblings of the ASD probands. 
Methods: 440 children, including 182 with ASD (160 males) and 258 TDCs (168 males) were administered the Benton. ASD diagnoses were confirmed with the ADOS and ADI-R.  In addition, the Benton was administered to 173 parents of children with ASD (83 mothers, and 82 fathers), and 87 siblings (from 68 families).  Benton scores were converted to standard scores that controlled for age (mean = 100, SD = 15) to facilitate comparisons.  For each participant, a Full Scale IQ score was estimated using an age appropriate version of the Wechsler scales.  
Results: Average Benton standard score for probands with ASD was more than a standard deviation lower than TDCs (79.6 vs. 96.5). However, groups were mismatched on FSIQ (94.4 vs. 109.5) and there were significant, but small positive correlations between FSIQ and Benton scores in both groups.  After covarying FSIQ, significant group differences remained (t=9.85, p < .0001), with an estimated effect size (Cohen's d) of 0.83.  Subsampling both groups to match on FSIQ (134 ASD, 123 TDC) yielded a nearly identical result: Cohen's d = .82.  Within the ASD group, there were significant, but modest correlations between Benton scores and indicators of severity of social impairment. Parents and siblings of probands with ASD scored significantly higher on the Benton compared to the probands. There were no difference between parents and TDCs, but siblings scored significantly lower than TDCs (p<.002, d=.3, after covarying IQ). 
Conclusions: The Benton is a sensitive measure of face recognition deficits among children with ASD, comparable to lengthier computerized test results (e.g., Wolf et al 2008). Deficits are independent of age and IQ and modestly correlated with degree of impairment.  Unaffected siblings have much smaller deficits in face recognition, while parents appear unaffected. This pattern of results might suggest common environmental effects on all offspring, or some shared de novo genetic contributions.