Social Visual Engagement—a Putative Autism Endophenotype—Exhibits Marked Variation and Striking MZ Twin Concordance in Typical Infants

Friday, May 13, 2016: 11:30 AM-1:30 PM
Hall A (Baltimore Convention Center)
J. N. Constantino1, A. Klin2, A. Glowinski3, S. Kennon-McGill3, C. Weichselbaum3 and W. Jones2, (1)Washington University School of Medicine, Saint Louis, MO, (2)Department of Pediatrics, Emory University School of Medicine, Marcus Autism Center, Children's Healthcare of Atlanta, Atlanta, GA, (3)Washington University in St. Louis, St. Louis, MO
Background:  Diminished eye contact has long been described as a classic symptom of autism spectrum disorder, and abnormalities in eye tracking measurements of social visual engagement (SVE) have been shown to predict autism spectrum disorders (ASD) among infants in ASD-affected families, by as early as 6 months of age (Jones and Klin, 2013).  Because ASD is among the most heritable of all neuropsychiatric disorders (Geschwind and State, 2015), with most of its genetic risk attributable to common (polygenic) factors, and since a majority of polygenic influences on ASD overlap with those influencing normal variation in social competency in the general population (Robinson et al., 2011), it is additionally important to examine the genetic structure of candidate developmental endophenotypes of the disorder in the general population. 

Objectives:  This is the first attempt to explore specific SVE predictors of ASD in an epidemiologically-ascertained sample of infants, here encompassing 56 monozygotic (MZ) twins and 112 dizygotic (DZ) twins. 

Methods:  We capitalized upon the availability of a unique epidemiologically-ascertained infant twin cohort, which has been previously reported (Marrus et al., 2015), and acquired precisely the same eye tracking indices which predicted ASD among infants at risk in Jones and Klin (2013), in order to explore the distribution and genetic structure of these risk indices in the general population. Full description of the eye-tracking methods can be found in Jones and Klin, 2013.

Results:   Quantitative SVE measurements exhibited striking MZ twin-twin concordance (on the order of 0.90) and only moderate DZ twin-twin concordance (on the order of 0.30,) indicating extremely high heritability, minimal measurement error, and wider-than-expected variation among typically-developing twins in the 18-24 month age range. This is shown in Figure 1 (attached).  To address the fact that not every child is attentive to every section of every video clip presented, we repeated the analysis incorporating only those segments for which both members of a twin pair were visually engaged.  The results were unchanged.  Finally, we conducted detailed analyses of moment-to-moment twin-twin correlation, which revealed striking homology in tracking of eye movement in X and Y coordinate space between the members of MZ twin pairs, again in contrast to DZ twin pairs, as shown for representative twin pairs in the attached Figure (2) for gaze location in X-coordinate space over time. 

Conclusions:   SVE measurements in these epidemiologically-ascertained twins reflected remarkable MZ twin-twin concordance and robust biological/ecological validity for the measurement method. By comparison, substantially lower DZ concordance confirmed that familial influence on SVE—which has been proposed as an autism endophenotype—is largely genetic in origin.   We commonly (i.e. in over one third of this normative sample) observed patterns of visual fixation which predicted the occurrence of autism among infants at high familial risk. How SVE interacts with other neurodevelopmental liabilities in the ontogeny of ASD warrants intensive subsequent study.


Jones, W. & Klin, A. Nature504, 427-431 (2013).

Geschwind DH, State MW.  Lancet Neurol. 2015 Nov;14(11):1109-20.

Robinson EB, et al.  Arch Gen Psychiatry. 2011 Nov;68(11):1113-21

Marrus, N. et al.  J Child Psychol Psychiatry 2015

See more of: Genetics
See more of: Genetics