Emotion, but Not Identity Matching, Is Affected by IQ in Autism

Background: Contrasting results are reported with respect to performance on tasks of face and emotion recognition in ASD (e.g. Harms, et.al. 2010; Weigelt, et.al 2012). Jointly with the lack of focus on faces, in particular on the eye region, during interpersonal interactions, deficient face and emotion recognition are thought to affect the social competences of individuals with autism. Recent research points to interactions between processes of face and emotion recognition in healthy individuals. A study of populations with cognitive deficits showed facilitation of same- identities on emotion matching task, but lack of facilitation of same-emotions on identity matching (Levy, et.al 2011). This result confirms earlier reports of difficulties in emotion recognition encountered by populations with cognitive impairments (e.g. Williams, et.al 2005). Such difficulties are not surprising, given that recognition of expressions is a conceptual as well as a perceptual task, requiring categorization and social understanding and hence affected by cognitive impairment. 

Objectives: Difficulties in social interactions characterize people with autism at all levels of cognitive functioning. Are these impairments reflected in the general pattern of performance of people with autism on face and emotion recognition tasks? Alternatively, will poor performance on expression tasks characterize low functioning people with autism, yet not people with autism of normal IQ? Will a similar difference in performance characterize identity tasks?

Methods: Subjects in the HI group were 24 males with autism, IQ > 85. Subjects in the LI group were 24 males with autism, IQ< 70. Mean age of the participants was 24.4 years. The groups were matched for vocabulary on the PPVT-3. Participants were tested on a matching paradigm, in which identities and expressions were manipulated simultaneously as the relevant or irrelevant dimensions. In the Identity task subjects were required to match faces with same or different expressions. In the Expression task subjects had to match expressions of same or different faces. Verbal responses were not required.

Results: Group performance was analyzed using a mixed-model ANOVA with repeated measures. In the Identity matching task, GROUP was the between-subjects variable (HI, LI) and the within-subject factors were CONGRUENCE (congruent, incongruent) and EXPRESSION (happy, disgusted). GROUP X CONGRUENCE interaction was found, as only in the HI accuracy in the congruent condition was higher than in the incongruent condition; that is, emotion recognition facilitated identity matching solely in the HI. The Expression matching task had four expressions – happy, disgusted,sad and angry. Two main effects (GROUP, CONGRUENCE) were found, as the HI group and the congruent condition yielded better accuracy; that is, facilitation of emotion matching by same-identity was not affected by IQ differences.

Conclusions: Results point to the effects of cognitive impairment on expression matching in autism. IQ did not have a similar effect on identity matching. The LI group is thus similar to other populations with intellectual handicaps, in whom the social-conceptual nature of emotion matching is impaired. These results support the decision in the forthcoming DSM-V to consider intellectual level as external to the diagnosis of autism and points to directions in intervention.