Atypical Social Modulation of Imitation in Autism Spectrum Conditions

Jennifer L. Cook 0 1 Geoffrey Bird 0 1 0 G. Bird (&) Department of Psychological Science, Birkbeck College , Malet St, London WC1E 7HX, UK 1 J. L. Cook G. Bird Institute of Cognitive Neuroscience , UCL, 17 Queen Square, London WC1N 3AR, UK Appropriate modulation of imitation according to social context is important for successful social interaction. In the present study we subliminally primed highfunctioning adults with ASC and age- and IQ-matched controls with either a pro- or non- social attitude. Following priming, an automatic imitation paradigm was used to acquire an index of imitation. Whereas imitation levels were higher for pro-socially primed relative to non-socially primed control participants, there was no difference between pro- and non- socially primed individuals with ASC. We conclude that high-functioning adults with ASC demonstrate atypical social modulation of imitation. Given the importance of imitation in social interaction we speculate that difficulties with the modulation of imitation may contribute to the social problems characteristic of ASC. - Imitation (also known as mimicry) is intricately linked with social interaction. Being imitated increases rapport (Chartrand and Bargh 1999), altruistic behavior (van Baaren et al. 2004) and trust (Bailenson and Yee 2005). Furthermore, individuals imitate more when in possession of a positive social attitude (Lakin and Chartrand 2003; Leighton et al. 2010). For example, subliminal pro-social, compared to non-social, priming results in significantly higher levels of imitation (Cook and Bird 2011; Leighton et al. 2010). Thus, imitation is bi-directionally associated with good social interaction and is therefore a key component in building social relationships with others. Crucially, successful social interaction relies on appropriate modulation of the degree of imitation according to the demands of the social situation (Lakin and Chartrand 2003). Autism Spectrum Conditions (ASC) are characterised by impairments in social interaction, language, and communication (American Psychiatric Association 1994). A number of studies have demonstrated reduced imitation and Mirror Neuron System (MNS) activity in individuals with ASC compared to control participants (Williams et al. 2004). The MNS is a network of brain areas active when an individual both executes and observes an action (Catmur et al. 2008; Iacoboni et al. 1999) and has been argued to comprise the neural mechanism that underpins imitation (Catmur et al. 2007; Heiser et al. 2003; Iacoboni et al. 1999). It has been hypothesised that a broken MNS and corresponding imitation impairment is a core feature of ASC (Williams et al. 2001). However, experimental evidence both supports (Avikainen et al. 2003; Dapretto et al. 2006; McIntosh et al. 2006; Oberman et al. 2005; Rogers et al. 2003) and opposes (Bird et al. 2007; Dinstein et al. 2010; Gowen et al. 2008; Hamilton et al. 2007; Leighton et al. 2008; Press et al. 2010; Spengler et al. 2010) the presence of an imitation impairment in ASC. Furthermore, clinical observations of high levels of echolalia (automatic repetition of speech patterns) and echopraxia (automatic imitation of observed actions) in individuals with ASC (Russell 1997; Rutter 1974; Williams et al. 2004) are incompatible with an imitation deficit in ASC, and instead suggest problems with inhibition of imitation. In response to the inconsistent literature it has been suggested that, rather than an imitation deficit per se, individuals with ASC may have difficulties with appropriately modulating levels of imitation (Hamilton 2008; Kana et al. 2011; Spengler et al. 2010). Although this hypothesis has not previously been tested, it is consistent with studies of individuals with ASC that report hypoactivity in parts of the brain thought to be involved in the modulation of imitation (Castelli et al. 2000; Spengler et al. 2010). Given the importance of appropriate levels of imitation for positive social interaction (Lakin and Chartrand 2003) this hypothesis may go some way towards explaining difficulties with social interaction in individuals with ASC. The present study used a behavioural measure of imitation, as opposed to a measure of MNS activity, to directly test the hypothesis that the social modulation of imitation is atypical in individuals with ASC. High-functioning adults with ASC and age and IQ-matched controls first completed a previously-validated (Bargh and Chartrand 2000; Cook and Bird 2011; Leighton et al. 2010) technique to unconsciously prime either a pro-social, or non-social attitude. Participants were asked to unscramble re-arranged sentences, a proportion of which were related to either prosocial attitudes (she is my friend) or non-social attitudes (he is often alone). Following this subliminal priming, participants completed an automatic imitation task. We predicted that, as in previous studies (Cook and Bird 2011; Leighton et al. 2010), pro-socially primed control participants would show increased levels of imitation relative to non-socially primed control participants. In line with the impaired modulation of imitation in ASC hypothesis, we predicted no significant difference in levels of imitation for pro-socially and non-socially primed ASC groups. 19 adults (mean 40.9 years) with ASC and 22 age- and IQmatched control individuals participated in this experiment (see Table 1 for further details). All participants had normal or corrected-to-normal vision and were screened for exclusion criteria (dyslexia, epilepsy, and any other neurological or psychiatric conditions) prior to taking part. Participants with ASC had a written diagnosis from an independent clinician, which they received no more than 5 years before taking part in this experiment, and all participants (save one for whom data was not available) scored above threshold for Autism Spectrum Disorder on the Autism Diagnostic Observation Schedule (ADOS; Lord et al. 2000). We were unable to distinguish between participants with Asperger Syndrome and Autism, as we did not have information about early development of language in our participants. Participants were randomly assigned to either the Pro- or Non-social Prime Group. ANOVAs demonstrated no main effect of, or interaction between, Prime Group and Diagnostic Group on age or full scale IQ (all ps [ 0.1). The two ASC groups did not differ with respect to ADOS score (non-social mean (SEM) = 10.00 (1.00), pro-social mean (SEM) = 9.88 (1.01), t(16) = 0.08, p = 0.93), age (non-social mean (SEM) = 41.30 (3.84), pro-social mean (SEM) = 40.56 (4.50), t(17) = 0.13, p = 0.90), full scale IQ (non-social mean (SEM) = 114.44 (4.99), pro-social mean (SEM) = 111.44 (6.11), t(16) = 0.38, p = 0.70), verbal IQ (non-social mean (SEM) = 116.22 (3.33), pro-social mean (SEM) = 112.00 (5.92), t(16) = 0.62, p = 0.54) or performance IQ (non-social mean (SEM) = 108.89 (6.57), pro-social mean (SEM) = 108.78 (6.48), t(16) = 0.12, p = 0.99). (...truncated)