J Abnorm Child Psychol DOI 10.1007/s10802-014-9874-9
Mechanisms of Imitation Impairment in Autism Spectrum Disorder Giacomo Vivanti & David Trembath & Cheryl Dissanayake
# Springer Science+Business Media New York 2014
Abstract Individuals with Autism Spectrum Disorders (ASD) have difficulties with imitation, though the nature of these remains unclear. In this study, involving 28 preschoolers with ASD (M age=48 months; 90 % male), 17 matched children with Global Developmental Delay (GDD group; M age=44 months; 53 % male) and 17 typically developing children (TD group, M age=52 months; 65 % male), we found that preschoolers with ASD 1) imitate less frequently than both typically developing children and children with GDD; 2) when they do imitate, their imitation is less accurate than that of TD children but similar to that of children with GDD; 3) unlike participants in both comparison groups, preschoolers with ASD use emulation more often than imitation when copying others’ actions; 4) they spend less time looking at the model’s face and more time looking at her actions; and 5) attentional, social and executive factors underlie different aspects of imitation difficulties in this population. Implications for developmental models of autism are discussed.
Keywords Autism . Imitation . Eye-tracking . Global developmental delay Electronic supplementary material The online version of this article (doi:10.1007/s10802-014-9874-9) contains supplementary material, which is available to authorized users. G. Vivanti (*) : D. Trembath : C. Dissanayake Olga Tennison Autism Research Centre, School of Psychological Science, La Trobe University, Bundoora CampusVictoria 3086 Melbourne, Australia e-mail: [email protected] G. Vivanti Victorian Autism Specific Early Learning and Care Centre, La Trobe University, Melbourne, Australia D. Trembath Griffith Health Institute, Griffith University, Brisbane, Australia
According to many scholars, imitation plays a key role in social cognitive development, and early disruptions in imitation might lead to sequelae affecting the organization of social learning, cognition and reciprocity (Hurley and Chater 2005; Rogers and Pennington 1991). Against this background, a number of studies have focused on the possible role of imitation difficulties in Autism Spectrum Disorders (ASD), a group of conditions characterised by early emerging deficits in social communication and behavioural flexibility (American Psychiatric Association [APA] 2013). While systematic reviews and meta-analyses support the notion of an imitative deficit in ASD (Vivanti and Hamilton 2014; Williams, Whiten and Singh 2004), the nature of such difficulties remain poorly understood. To date, a number of critical questions remain to be addressed. First, it is important to clarify whether poor imitative performance in ASD reflects difficulties in imitation accuracy or a diminished propensity to imitate, or both. Most studies in this field use paradigms in which participants are explicitly instructed to imitate, thus providing valuable data on the accuracy of imitation performance, but little information on the propensity to use imitation spontaneously in everyday life. These are two distinct aspects (e.g., a child may imitate frequently but inaccurately, and vice versa), and imitating others frequently might be more important in development than imitating them to a high degree of precision under explicit instructions. For example, participating in early socialemotional exchanges involving reciprocal imitation does not require high-fidelity matching of observed actions (Kugiumutzakis, Kokkinaki, Makrodimitraki and Vitalaki 2005; Trevarthen et al. 1999). Moreover, during childhood and throughout development, social learning is mostly mediated by spontaneous, non-elicited imitation (Bandura and Huston 1961; Vivanti and Rogers 2014). Secondly, it is important to determine whether individuals with ASD differ from those without ASD in terms of the
J Abnorm Child Psychol
specific strategies that they use to copy others’ actions. A number of distinct copying behaviours have been described (Want and Harris 2002), including stimulus enhancement (performing an action on a particular stimulus/location after observing someone acting on that same stimulus/location), emulation (reproducing the goals but not the means of observed actions), and imitation (or true imitation; reproducing both the means and the goals of the observed actions). Distinguishing between these different types of copying behaviour is relevant, because different ways of copying others’ actions might serve different functions and reflect distinct underlying processes (Vivanti and Hamilton 2014). For example, it has been suggested that true imitation might be driven by the motivation to affiliate with another person, whilst emulation and stimulus enhancement might reflect an interest in the product and instrumental function of others’ actions (Call, Carpenter and Tomasello 2005; Rogers, Young, Cook, Giolzetti and Ozonoff 2010; Vivanti and Hamilton 2014). However, only a small minority of studies have adopted this taxonomy in studies of ASD, and the results are mixed (Marsh, Pearson, Ropar and Hamilton 2013; Nielsen, Slaughter and Dissanayake 2013). Finally, evidence is inconclusive with regards to the causes of the imitation deficits. A variety of different explanations of imitation dysfunction in ASD exist (Rogers and Williams 2006; Vivanti and Hamilton 2014) that can be classified into three main explanatory models, based on their specific predictions (Colombi, Vivanti and Rogers 2011). Attentional explanations suggest that individuals with ASD fail to imitate because they do not pay attention to actions that are demonstrated to them. According to this notion, imitation difficulties in ASD reflect general abnormalities with encoding relevant information from the environment. Findings supporting this hypothesis comprise diminished/abnormal visual attention to people and their actions in ASD (Barbaro and Dissanayake 2013; Kasari et al. 1990) including during imitation tasks (Vivanti et al. 2008, 2011; Vivanti and Dissanayake 2014). A specific prediction of this account is that differences in imitation performance will be related to differences in attention to the demonstration. Social explanations suggest that imitation deficits in ASD are social in nature (i.e., they reflect difficulties with socialemotional understanding/processing), and are therefore associated with other social deficits, in particular difficulties in joint attention. This notion is supported by studies indicating a link between imitation and joint attention in ASD and typical development. For example, Tomasello and colleagues (Carpenter, Nagell and Tomasello 1998; Tomasello, Carpenter, Call, Behne and Moll 2005) indicate that imitation and joint attention both reflect mechanisms of social coordination/attunement, Hobson and Hobson (2007) propose that imitation and joint attention both reflect the underlying motivation to establish intersubjective engagement, and
Pennington, Williams, and Rogers (2006) suggest that difficulties in imitation and joint attention in ASD reflect an underlying impairment in the ability to register/appreciate correspondences between own and others’ actions. While different social explanations are grounded on different theoretical tenets, they share the prediction that imitation differences will be related to differences in other social domains, in particular joint attention. A third group of theories, which can be classified as motor/ sensorimotor explanations, emphasise the possible role of motor, praxis or sensorimotor difficulties affecting execution of actions in ASD. For example, Vanvuchelen, Roeyers, and De Weerdt (2007) suggest that a perceptual-motor impairment causes imitation problems in ASD, and Mostofsky and colleagues (MacNeil and Mostofsky 2012; Mostofsky and Ewen 2011) suggest that imitation difficulties in ASD reflect a more generalized deficit in the performance of gestures and tool use (i.e. developmental dyspraxia) due to abnormalities in the formation of perceptual-motor action models. The motor/ sensorimotor explanations share the prediction that differences in imitation in ASD will be related to differences with the performance of functional actions and tool use in nonimitative contexts. The evidence thus far provides some support for each account outlined above, but no evidence exists for the superiority of one explanation over the others (Colombi et al. 2011; Vivanti 2013). To date, no study has systematically contrasted the predictions derived from these different theories. In addition, research to date has failed to clarify whether differences in imitation reflect a deviance or a delay from typical development, and whether these differences are unique to children with ASD or shared with children with other developmental disorders. While some studies have documented imitation difficulties that were specific to autism in comparison to developmental-age matched participants (e.g., Rogers, Hepburn, Stackhouse and Wehner 2003), a recent carefully conducted study showed that young children with ASD had poor imitation skills compared with typically developing children, but their performance was similar to that of siblings with other cognitive delays not related to ASD (Young et al. 2011).This question is crucial to our understanding of the link between imitation difficulties and social impairments in ASD. The current study aims to address the critical issues summarized above. We designed a series of novel experiments with the aim to 1) clarify whether children with ASD differ from developmental and chronological age matched children without ASD in their tendency to imitate spontaneously (frequency of imitative responses) and their ability to imitate (accuracy of imitative responses); 2) identify whether children with ASD differ from participants without ASD in terms of the strategies that they use to copy others; and 3) compare the specific predictions of attentional, social, and executive accounts of imitative deficits in ASD.
J Abnorm Child Psychol
General Method
Procedures
Participants
The experimental protocol was approved by the La Trobe University Human Ethics Committee, and informed consent was obtained from all participants’ parents. Participants were tested in a quiet room in the centres where they were recruited. All the experiments were administered in the same session in a fixed random order. The average length of the experimental session was 45 min.
The participants were 28 preschoolers with an ASD (ASD group; 25 males and 3 females, 75 % Caucasian) and 17 chronological- and mental age matched toddlers with a Global Developmental Delay (GDD group; 8 males and 8 females, 83 % Caucasian). Moreover, as several imitation mechanisms tested in the present study have not been previously investigated in typically developing preschoolers, 17 age matched typically developing children was also included (TD group; 11 males and 6 females, 76 % Caucasian). Participants with ASD were recruited through the Victorian Autism Specific Early Learning and Care Centre (Victorian ASELCC), an autism specific program located at La Trobe University, while those in the TD group were recruited through the La Trobe University Community Children Centre. The DSM-IV (American Psychiatric Association 1994) diagnoses of ASD were previously made by communitybased health care professionals and confirmed by the study authors using the Social Communication Questionnaire (SCQ; Rutter, Bailey, Lord and Berument 2003) and Autism Diagnostic Observation Schedule (Lord et al. 2000), a standardized diagnostic observational instrument which quantifies autism symptoms in social reciprocity, communication, play and repetitive behaviours. Exclusionary criteria for the ASD group included the presence of major medical conditions. Participants in the GDD group were recruited through Kalparrin Early Intervention Centre, a community Early Childhood Intervention Program. These children had been assessed by professionals in the community and deemed eligible for Early Intervention services on the basis of presenting with Global Developmental Delay, defined as significant delay in two or more of developmental domains (Shevell et al. 2003). Exclusionary criteria for the GDD group included the presence of autistic features as assessed through the SCQ, and the presence of major medical conditions. Inclusion criteria for the TD group included the absence of a known history of medical conditions, and cognitive abilities within the normal range. In all groups, participants’ cognitive level was measured through the Mullen Scales of Early Learning (MSEL; Mullen 1995), a standardized measure of early development, yielding scores on the following subscales: Visual Reception, Fine Motor, Receptive Language, and Expressive Language, as well as a standard composite score. The ASD and GDD groups were matched for chronological age and scores on each subscale of the MSEL, as well as on total scores. The TD group was matched with each group on chronological age (See Table 1).
Experiment 1 The aim in Experiment 1 was to investigate whether preschool children with ASD differ from developmental and chronological age matched peers in the frequency of spontaneous imitation and in the accuracy of their imitative performance. If symptoms of ASD are linked to impairments in the ability and/or propensity to imitate others, then ASD-specific differences should be expected. Method Participants were seated in a comfortable chair, 60 cm from a computer monitor in front of a small table. A series of six 7-s video-demonstrations of actions on objects was presented on the monitor in two different fixed random orders. The same female actor was used in all videos, and two different sets of materials were used across six trials. During the video demonstration, the actor performed a simple action using some of the objects on the table in front of her (different objects were used in different trials; see Supplementary Materials Table 6 for details of the to-be-imitated actions). Materials necessary for the imitation of each trial were placed on the small table in front of the participant before the beginning of the trial. The objects and their arrangement were exactly the same as those displayed in the video. No instruction was given, and participants’ spontaneous behaviour with the objects in response to the demonstration was videorecorded for later scoring. The coding procedure involved two steps. First, whether children spontaneously imitated the modelled actions was coded using a simple yes/no (1/0) coding system. A frequency score was obtained for each participant by calculating the proportion of imitated actions out of the total imitation opportunities (i.e., the six demonstrated actions presented in the video). Second, if the modelled actions were imitated, the accuracy of imitative performance was coded, using a Likert scale based on Vivanti et al. (2011). For each participant, an accuracy score was obtained by dividing the sum of accuracy scores by the maximum possible score. Imitation items and scoring criteria are reported in Table 6 in the Supplementary Materials. Coding was conducted by a trained research assistant blind to group membership and study hypotheses. Interrater reliability between the first author and the research assistant for the coding of imitation was calculated on 20 % of the entire data set. Cohen’s kappa was 1.0 for frequency and
J Abnorm Child Psychol Table 1 Participants’ characteristics
ASD (N=28)
T test p-value
T test p-value
GDD (N=17)
TD (N=17)
ASD vs GDD
ASD vs TD
Age (months): M (SD)
48.14 (10.97)
44.40 (13.70)
52.12 (8.77)
0.25
0.49
Gender: M, F SCQ MSEL, Early learning composite: M (SD) MSEL, Visual reception T scores: M (SD) MSEL, Fine motor T scores: M (SD) MSEL, Receptive language T scores M (SD) MSEL, Expressive language T scores: M (SD) ADOS, Social communication: M (SD) ADOS, Repetitive behaviours: M (SD)
25, 3 18.77 (5.64) 64.07 (20.79) 31.86 (15.43) 28.11 (15.11) 29.36 (13.66) 31.43 (14.95) 12.93 (4.52) 4.48 (2.00)
9, 8 6.92 (4.68) 74.50 (18.02) 37.80 (15.96) 34.00 (14.05) 36.53 (15.36) 33.60 (8.83) -
11, 6 114.50 (13.47) 58.19 (13.39) 55.00 (10.97) 58.19 (8.62) 59.94 (7.72) -
Mechanisms of Imitation Impairment in Autism Spectrum Disorder Giacomo Vivanti & David Trembath & Cheryl Dissanayake
# Springer Science+Business Media New York 2014
Abstract Individuals with Autism Spectrum Disorders (ASD) have difficulties with imitation, though the nature of these remains unclear. In this study, involving 28 preschoolers with ASD (M age=48 months; 90 % male), 17 matched children with Global Developmental Delay (GDD group; M age=44 months; 53 % male) and 17 typically developing children (TD group, M age=52 months; 65 % male), we found that preschoolers with ASD 1) imitate less frequently than both typically developing children and children with GDD; 2) when they do imitate, their imitation is less accurate than that of TD children but similar to that of children with GDD; 3) unlike participants in both comparison groups, preschoolers with ASD use emulation more often than imitation when copying others’ actions; 4) they spend less time looking at the model’s face and more time looking at her actions; and 5) attentional, social and executive factors underlie different aspects of imitation difficulties in this population. Implications for developmental models of autism are discussed.
Keywords Autism . Imitation . Eye-tracking . Global developmental delay Electronic supplementary material The online version of this article (doi:10.1007/s10802-014-9874-9) contains supplementary material, which is available to authorized users. G. Vivanti (*) : D. Trembath : C. Dissanayake Olga Tennison Autism Research Centre, School of Psychological Science, La Trobe University, Bundoora CampusVictoria 3086 Melbourne, Australia e-mail: [email protected] G. Vivanti Victorian Autism Specific Early Learning and Care Centre, La Trobe University, Melbourne, Australia D. Trembath Griffith Health Institute, Griffith University, Brisbane, Australia
According to many scholars, imitation plays a key role in social cognitive development, and early disruptions in imitation might lead to sequelae affecting the organization of social learning, cognition and reciprocity (Hurley and Chater 2005; Rogers and Pennington 1991). Against this background, a number of studies have focused on the possible role of imitation difficulties in Autism Spectrum Disorders (ASD), a group of conditions characterised by early emerging deficits in social communication and behavioural flexibility (American Psychiatric Association [APA] 2013). While systematic reviews and meta-analyses support the notion of an imitative deficit in ASD (Vivanti and Hamilton 2014; Williams, Whiten and Singh 2004), the nature of such difficulties remain poorly understood. To date, a number of critical questions remain to be addressed. First, it is important to clarify whether poor imitative performance in ASD reflects difficulties in imitation accuracy or a diminished propensity to imitate, or both. Most studies in this field use paradigms in which participants are explicitly instructed to imitate, thus providing valuable data on the accuracy of imitation performance, but little information on the propensity to use imitation spontaneously in everyday life. These are two distinct aspects (e.g., a child may imitate frequently but inaccurately, and vice versa), and imitating others frequently might be more important in development than imitating them to a high degree of precision under explicit instructions. For example, participating in early socialemotional exchanges involving reciprocal imitation does not require high-fidelity matching of observed actions (Kugiumutzakis, Kokkinaki, Makrodimitraki and Vitalaki 2005; Trevarthen et al. 1999). Moreover, during childhood and throughout development, social learning is mostly mediated by spontaneous, non-elicited imitation (Bandura and Huston 1961; Vivanti and Rogers 2014). Secondly, it is important to determine whether individuals with ASD differ from those without ASD in terms of the
J Abnorm Child Psychol
specific strategies that they use to copy others’ actions. A number of distinct copying behaviours have been described (Want and Harris 2002), including stimulus enhancement (performing an action on a particular stimulus/location after observing someone acting on that same stimulus/location), emulation (reproducing the goals but not the means of observed actions), and imitation (or true imitation; reproducing both the means and the goals of the observed actions). Distinguishing between these different types of copying behaviour is relevant, because different ways of copying others’ actions might serve different functions and reflect distinct underlying processes (Vivanti and Hamilton 2014). For example, it has been suggested that true imitation might be driven by the motivation to affiliate with another person, whilst emulation and stimulus enhancement might reflect an interest in the product and instrumental function of others’ actions (Call, Carpenter and Tomasello 2005; Rogers, Young, Cook, Giolzetti and Ozonoff 2010; Vivanti and Hamilton 2014). However, only a small minority of studies have adopted this taxonomy in studies of ASD, and the results are mixed (Marsh, Pearson, Ropar and Hamilton 2013; Nielsen, Slaughter and Dissanayake 2013). Finally, evidence is inconclusive with regards to the causes of the imitation deficits. A variety of different explanations of imitation dysfunction in ASD exist (Rogers and Williams 2006; Vivanti and Hamilton 2014) that can be classified into three main explanatory models, based on their specific predictions (Colombi, Vivanti and Rogers 2011). Attentional explanations suggest that individuals with ASD fail to imitate because they do not pay attention to actions that are demonstrated to them. According to this notion, imitation difficulties in ASD reflect general abnormalities with encoding relevant information from the environment. Findings supporting this hypothesis comprise diminished/abnormal visual attention to people and their actions in ASD (Barbaro and Dissanayake 2013; Kasari et al. 1990) including during imitation tasks (Vivanti et al. 2008, 2011; Vivanti and Dissanayake 2014). A specific prediction of this account is that differences in imitation performance will be related to differences in attention to the demonstration. Social explanations suggest that imitation deficits in ASD are social in nature (i.e., they reflect difficulties with socialemotional understanding/processing), and are therefore associated with other social deficits, in particular difficulties in joint attention. This notion is supported by studies indicating a link between imitation and joint attention in ASD and typical development. For example, Tomasello and colleagues (Carpenter, Nagell and Tomasello 1998; Tomasello, Carpenter, Call, Behne and Moll 2005) indicate that imitation and joint attention both reflect mechanisms of social coordination/attunement, Hobson and Hobson (2007) propose that imitation and joint attention both reflect the underlying motivation to establish intersubjective engagement, and
Pennington, Williams, and Rogers (2006) suggest that difficulties in imitation and joint attention in ASD reflect an underlying impairment in the ability to register/appreciate correspondences between own and others’ actions. While different social explanations are grounded on different theoretical tenets, they share the prediction that imitation differences will be related to differences in other social domains, in particular joint attention. A third group of theories, which can be classified as motor/ sensorimotor explanations, emphasise the possible role of motor, praxis or sensorimotor difficulties affecting execution of actions in ASD. For example, Vanvuchelen, Roeyers, and De Weerdt (2007) suggest that a perceptual-motor impairment causes imitation problems in ASD, and Mostofsky and colleagues (MacNeil and Mostofsky 2012; Mostofsky and Ewen 2011) suggest that imitation difficulties in ASD reflect a more generalized deficit in the performance of gestures and tool use (i.e. developmental dyspraxia) due to abnormalities in the formation of perceptual-motor action models. The motor/ sensorimotor explanations share the prediction that differences in imitation in ASD will be related to differences with the performance of functional actions and tool use in nonimitative contexts. The evidence thus far provides some support for each account outlined above, but no evidence exists for the superiority of one explanation over the others (Colombi et al. 2011; Vivanti 2013). To date, no study has systematically contrasted the predictions derived from these different theories. In addition, research to date has failed to clarify whether differences in imitation reflect a deviance or a delay from typical development, and whether these differences are unique to children with ASD or shared with children with other developmental disorders. While some studies have documented imitation difficulties that were specific to autism in comparison to developmental-age matched participants (e.g., Rogers, Hepburn, Stackhouse and Wehner 2003), a recent carefully conducted study showed that young children with ASD had poor imitation skills compared with typically developing children, but their performance was similar to that of siblings with other cognitive delays not related to ASD (Young et al. 2011).This question is crucial to our understanding of the link between imitation difficulties and social impairments in ASD. The current study aims to address the critical issues summarized above. We designed a series of novel experiments with the aim to 1) clarify whether children with ASD differ from developmental and chronological age matched children without ASD in their tendency to imitate spontaneously (frequency of imitative responses) and their ability to imitate (accuracy of imitative responses); 2) identify whether children with ASD differ from participants without ASD in terms of the strategies that they use to copy others; and 3) compare the specific predictions of attentional, social, and executive accounts of imitative deficits in ASD.
J Abnorm Child Psychol
General Method
Procedures
Participants
The experimental protocol was approved by the La Trobe University Human Ethics Committee, and informed consent was obtained from all participants’ parents. Participants were tested in a quiet room in the centres where they were recruited. All the experiments were administered in the same session in a fixed random order. The average length of the experimental session was 45 min.
The participants were 28 preschoolers with an ASD (ASD group; 25 males and 3 females, 75 % Caucasian) and 17 chronological- and mental age matched toddlers with a Global Developmental Delay (GDD group; 8 males and 8 females, 83 % Caucasian). Moreover, as several imitation mechanisms tested in the present study have not been previously investigated in typically developing preschoolers, 17 age matched typically developing children was also included (TD group; 11 males and 6 females, 76 % Caucasian). Participants with ASD were recruited through the Victorian Autism Specific Early Learning and Care Centre (Victorian ASELCC), an autism specific program located at La Trobe University, while those in the TD group were recruited through the La Trobe University Community Children Centre. The DSM-IV (American Psychiatric Association 1994) diagnoses of ASD were previously made by communitybased health care professionals and confirmed by the study authors using the Social Communication Questionnaire (SCQ; Rutter, Bailey, Lord and Berument 2003) and Autism Diagnostic Observation Schedule (Lord et al. 2000), a standardized diagnostic observational instrument which quantifies autism symptoms in social reciprocity, communication, play and repetitive behaviours. Exclusionary criteria for the ASD group included the presence of major medical conditions. Participants in the GDD group were recruited through Kalparrin Early Intervention Centre, a community Early Childhood Intervention Program. These children had been assessed by professionals in the community and deemed eligible for Early Intervention services on the basis of presenting with Global Developmental Delay, defined as significant delay in two or more of developmental domains (Shevell et al. 2003). Exclusionary criteria for the GDD group included the presence of autistic features as assessed through the SCQ, and the presence of major medical conditions. Inclusion criteria for the TD group included the absence of a known history of medical conditions, and cognitive abilities within the normal range. In all groups, participants’ cognitive level was measured through the Mullen Scales of Early Learning (MSEL; Mullen 1995), a standardized measure of early development, yielding scores on the following subscales: Visual Reception, Fine Motor, Receptive Language, and Expressive Language, as well as a standard composite score. The ASD and GDD groups were matched for chronological age and scores on each subscale of the MSEL, as well as on total scores. The TD group was matched with each group on chronological age (See Table 1).
Experiment 1 The aim in Experiment 1 was to investigate whether preschool children with ASD differ from developmental and chronological age matched peers in the frequency of spontaneous imitation and in the accuracy of their imitative performance. If symptoms of ASD are linked to impairments in the ability and/or propensity to imitate others, then ASD-specific differences should be expected. Method Participants were seated in a comfortable chair, 60 cm from a computer monitor in front of a small table. A series of six 7-s video-demonstrations of actions on objects was presented on the monitor in two different fixed random orders. The same female actor was used in all videos, and two different sets of materials were used across six trials. During the video demonstration, the actor performed a simple action using some of the objects on the table in front of her (different objects were used in different trials; see Supplementary Materials Table 6 for details of the to-be-imitated actions). Materials necessary for the imitation of each trial were placed on the small table in front of the participant before the beginning of the trial. The objects and their arrangement were exactly the same as those displayed in the video. No instruction was given, and participants’ spontaneous behaviour with the objects in response to the demonstration was videorecorded for later scoring. The coding procedure involved two steps. First, whether children spontaneously imitated the modelled actions was coded using a simple yes/no (1/0) coding system. A frequency score was obtained for each participant by calculating the proportion of imitated actions out of the total imitation opportunities (i.e., the six demonstrated actions presented in the video). Second, if the modelled actions were imitated, the accuracy of imitative performance was coded, using a Likert scale based on Vivanti et al. (2011). For each participant, an accuracy score was obtained by dividing the sum of accuracy scores by the maximum possible score. Imitation items and scoring criteria are reported in Table 6 in the Supplementary Materials. Coding was conducted by a trained research assistant blind to group membership and study hypotheses. Interrater reliability between the first author and the research assistant for the coding of imitation was calculated on 20 % of the entire data set. Cohen’s kappa was 1.0 for frequency and
J Abnorm Child Psychol Table 1 Participants’ characteristics
ASD (N=28)
T test p-value
T test p-value
GDD (N=17)
TD (N=17)
ASD vs GDD
ASD vs TD
Age (months): M (SD)
48.14 (10.97)
44.40 (13.70)
52.12 (8.77)
0.25
0.49
Gender: M, F SCQ MSEL, Early learning composite: M (SD) MSEL, Visual reception T scores: M (SD) MSEL, Fine motor T scores: M (SD) MSEL, Receptive language T scores M (SD) MSEL, Expressive language T scores: M (SD) ADOS, Social communication: M (SD) ADOS, Repetitive behaviours: M (SD)
25, 3 18.77 (5.64) 64.07 (20.79) 31.86 (15.43) 28.11 (15.11) 29.36 (13.66) 31.43 (14.95) 12.93 (4.52) 4.48 (2.00)
9, 8 6.92 (4.68) 74.50 (18.02) 37.80 (15.96) 34.00 (14.05) 36.53 (15.36) 33.60 (8.83) -
11, 6 114.50 (13.47) 58.19 (13.39) 55.00 (10.97) 58.19 (8.62) 59.94 (7.72) -
