A Twin Study of Gesture Execution, Imitation, and Identification

Background: Persons on the autism spectrum tend to be less proficient than persons not on the autism spectrum at imitating gestures. Theorists have attributed less-proficient imitation to a myriad of interpersonal and intrapersonal deficits, such as a lack of a “sense of self” or a lack of a “subjective awareness of others” (Gallese, 2006; Harris, 2003; Rogers et al., 2003). However, empirical research documents that persons on the autism spectrum tend to be equivalently proficient at identifying other persons’ gestures, including identifying the intentionality of other persons’ gestures (Aldridge et al., 2000; Carpenter et al., 2001; Hamilton et al., 2007; Russell & Hill, 2001; Sebanz et al., 2005). Moreover, empirical research documents that persons on the autism spectrum are more proficient at imitating gestures than they are at executing gestures, without the guide of another person’s demonstration (Alexander et al. 1992; Dewey et al., 2007; Hill, 1998; Mostofsky et al., 2006; Kees et al., 2006; Zoia et al., 2002). Thus, the source of their lower proficiency appears to reside in action execution, not social cognition.
Objectives: To investigate cotwins’ proficiency and similarity in executing gestures (to an examiner’s verbal request), imitating gestures (to an examiner’s demonstration), and identifying gestures (to a video demonstration) using the well-established Praxis Test of the Boston Diagnostic Aphasia Exam.
Methods: Participants comprised 45 pairs of twins, age 4;2 to 20;3, who were recruited to the study because one or both cotwins of each pair met criteria for the autism spectrum on both the SCQ and the ADOS.
Results: Across all participants, proficiency in identifying, executing, and imitating gestures increased with age (F(1,87)=28.27, p<0.001; F(1,75)=7.04, p<0.001; F(1,83)=12.54, p<0.001, respectively). Similarly, across all participants, identifying gestures (M = 85%) was more proficient than imitating gestures (M = 61%), which was more proficient than executing gestures (M = 52%). Furthermore, participants who did versus did not meet criteria for the autism spectrum differed more dramatically in their proficiency in executing (44% vs. 70%) and imitating (53% vs. 78%) gestures than in their proficiency in identifying gestures (81% vs. 94%; F(2,144)=5.02, p<0.01, with participants’ age as a covariate).
Finally, cotwins concordant for the autism spectrum were highly similar in their proficiency in executing gestures (intraclass R=.87, N=17), imitating gestures (R=.73, N=15), and identifying gestures (R=.76, N=12), whereas cotwins disconcordant for the autism spectrum were dissimilar (execution R=-.04, N=28; imitation R=-.02, N=26; identification R=-.13, N=26). Monozygotic cotwins concordant for the autism spectrum were highly similar in their proficiency in executing (R=.93, N=13), imitating (R=.93, N=11), and identifying gestures (R=.78, N=10), whereas monozygotic cotwins disconcordant for the autism spectrum were not all similar, despite their 100% genetic overlap (execution R=-.50, N=8; imitation R=-.42, N=7; identification R=-.33, N=8).
Conclusions: Performance on the Praxis Test distinguishes persons on the autism spectrum from those who are not. Moreover, praxis performance is highly similar for monozygotic cotwins who are concordant for the autism spectrum, suggesting that praxis performance taps systematic processes more specific than whatever processes contribute to meeting the diagnostic threshold.