http://aut.sagepub.

com/

Autism

Recognizing faces based on inferred traits in autism spectrum disorders

Rajani Ramachandran, Peter Mitchell and Danielle Ropar Autism 2010 14: 605 originally published online 5 October 2010 DOI: 10.1177/1362361310372777 The online version of this article can be found at: http://aut.sagepub.com/content/14/6/605

Published by:
http://www.sagepublications.com

On behalf of:

The National Autistic Society

Additional services and information for Autism can be found at: Email Alerts: http://aut.sagepub.com/cgi/alerts Subscriptions: http://aut.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://aut.sagepub.com/content/14/6/605.refs.html

>> Version of Record - Dec 13, 2010 OnlineFirst Version of Record - Oct 5, 2010 What is This?

Downloaded from aut.sagepub.com by guest on March 28, 2013

Recognizing faces based on inferred traits in autism spectrum disorders

R A JA N I R A M AC H A N D R A N PETER MITCHELL DA N I E L L E RO PA R
University of Nottingham University of Nottingham University of Nottingham

autism 2010 SAGE Publications and The National Autistic Society Vol 14(6) 605618; 372777 1362-3613(2010)

A B S T R AC T

Recent ndings indicate that individuals with autism spectrum disorders (ASD) could, surprisingly, infer traits from behavioural descriptions. Now we need to know whether or not individuals with ASD are able to use trait information to identify people by their faces. In this study participants with and without ASD were presented with pairs of faces each accompanied by a sentence. One sentence allowed a trait to be inferred (e.g. This is Ross who smiled and said hello to everyone at the party.) and one allowed a fact to be inferred (e.g. This is Ben who has to bend down to enter most doors.). Subsequently, the same face stimuli were presented with a single descriptive trait, fact or name cue (e.g. friendly or tall and Ross or Ben respectively in the above examples). Participants had to choose which of the faces best related to the cue word. Participants with ASD performed surprisingly well in associating traits, facts, and names to the appropriate person signicantly above what would be expected by chance. Indeed, they performed as well as participants without ASD.

K E Y WO R D S

autism spectrum disorders; face recognition; traits

ADDRESS Correspondence should be addressed to: DA N I E L L E RO PA R , Department of Psychology, University of Nottingham, Nottingham NG9 1JA, UK. e-mail: dmr@ psychology.nottingham.ac.uk

One of the most commonly cited socio-cognitive decits in autism spectrum disorders (ASD) is impaired face recognition (Jemel et al., 2006). Individuals with ASD have difculty not only with delayed but also with immediate face recognition (Boucher and Lewis, 1992; Klin et al., 1999; Williams et al., 2005). This decit could interfere with effective social functioning given that faces play a central role in social cognition. For example, there is compelling evidence that people spontaneously form an impression of an individual, linking inferences drawn from his or her behaviour with an image of the person (Todorov and Uleman, 2004). Faces serve as an ideal representation in such an attributional process because a face,
Copyright The Author(s), 2010. Reprints and permissions: http://www.sagepub.co.uk/journalspermissions.nav DOI: 10.1177/1362361310372777
Downloaded from aut.sagepub.com by guest on March 28, 2013

605

14(6) unlike attire, gender or color is perceptually salient and unique. Associating inferred characteristics of an individual with his or her face will help in predicting how the persons behaviour will differ from others (of the same age and gender, for example). One could argue both for and against the possibility that people with ASD can infer traits. Some researchers believe that trait reasoning develops after belief-desire reasoning has already been established. While understanding of false beliefs is achieved between the age of 3 and 4 years (Wellman et al., 2001), grasping that traits determine thoughts, feelings and desires develops after 4 years (Gnepp and Chilamkurti, 1988) when children begin to understand the subjective nature of desires (Yuill and Pearson, 1998). Therefore, because individuals with ASD reportedly have difficulty with belief-desire reasoning, so they should have difficulty inferring traits. Notwithstanding, even preschoolers who have not yet acquired beliefdesire reasoning understand traits in terms of behavioural regularity. Dozier (1991), Gnepp and Chilamkurti (1988) and Yuill and Pearson (1998) reported that preschoolers as young as 3 years were able to utilize their knowledge of how a peer performed previously to make predictions about how the peer would behave in the future. Research suggests that children make these predictions using simple frequency rules about how often a particular behaviour occurs as well as more complex rules like Kellys covariation principle (Ferguson et al., 1984). Considering that people with ASD sometimes successfully learn and apply rules (Hermelin and OConnor, 1986; Klinger and Dawson, 2001; Minshew et al., 1992) they might draw on this ability to good effect when inferring traits. Hirschfeld et al. (2007) reported that children with ASD were able to use gender and race stereotypes to predict behaviour. Stereotyping involves attributing traits and preferences to a person solely because he or she belongs to a particular social group. Stereotypes are not explicitly taught to children and have a strong albeit implicit cultural and social learning component. The study by Hirschfeld et al. thus suggests that individuals with ASD are able to grasp subtle cultural cues regarding the assumed characteristics of individuals belonging to different social categories. They use this information to predict future behaviour. The study indirectly lends supports to the possibility that individuals with ASD can infer and use traits to predict behaviour. More compelling evidence emerged from a series of experiments conducted by Ramachandran et al. (2009), who presented behavioural descriptions (for example, He slipped on the rug and twisted his ankle) to teenagers and adults with typical development and with ASD. Participants with ASD performed similarly to participants with typical development,
AU T I S M

606

Downloaded from aut.sagepub.com by guest on March 28, 2013

A S S O C I AT I N G T R A I T S T O F AC E S I N AU T I S M

inferring the implied traits (clumsy for the above example) spontaneously, with minimal effort and in the absence of a reason for or benet from making the inference. This surprising result suggests that individuals with ASD may be adept at handling dispositional information which could be useful to them in understanding another persons behaviour. But, being able to infer traits might not be of much practical use unless we can also associate the traits with the actor. Although individuals with ASD appear competent in drawing trait inferences from descriptions of behaviour, we do not yet know whether they make associations between an inferred trait and an image of the person whose behaviour implies the trait. To address this research question, this experiment investigated whether individuals with ASD could identify people, represented by frontal photographs of their faces, based on inferred traits. This experiment used a forced choice reaction time procedure, based on the paradigm developed by Todorov and Uleman (Experiment 4, 2004). Each experimental trial consisted of a study phase and a test phase. In the study phase a pair of colour photographs of Caucasian males was presented along with a pair of sentences (i.e. one behavioural description under each photograph). In the test phase, immediately following the study phase, the same pair of photographs was presented either in the same or in the opposite spatial location as in the study phase (see Figure 1), with a single cue word. The cue was implied in one of the behavioural descriptions presented in the study phase. The participants were required to identify the person to whom the cue best related to.

Method
Participants Sixteen participants with a diagnosis of Aspergers syndrome (AS) took part in this study. There were 14 males and 2 females. They were recruited from a specialist college for individuals with AS in Somerset, UK. Individuals were only selected as participants if they had been diagnosed by an experienced clinician. Their latest ofcial diagnostic report obtained by the college at the time of their admission was carefully veried. Accordingly all participants met DSM-IV criteria (American Psychiatric Association, 1994) for AS. Sixteen participants without AS were also tested. For the purpose of identifying between group differences on current level of autistic features, each participant completed the Autism-Spectrum Quotient (AQ; BaronCohen et al., 2001). No participant was included or excluded from the sample based on their score on the AQ. An independent samples t test identied signicant between-group differences on AQ scores, t = 2.38,
607

Downloaded from aut.sagepub.com by guest on March 28, 2013

AU T I S M

14(6)
Test phase

Same spatial location Study phase

This is Victor who ignored his injuries and completed the race.

This is Gordon who always drives a little slower than the speed limit.

determined

Opposite spatial location Study phase

Test phase

This is Andrew who went to bed with a sore throat and a bad cold.

This is Mathew who can afford to buy only from sales at charity shops.

sick

Figure 1

Illustration of the study phase and test phase used in the experiment

p < .01. AQ is a questionnaire-based screening instrument and the lack of independent diagnosis using more elaborate instruments like ADOS may be viewed as a limitation of this study. All 32 participants were native English speakers. The two groups were matched in terms of chronological age (CA), verbal intelligence quotient (VIQ), performance intelligence quotient (PIQ) and gender. The VIQ was based on the verbal subsets (vocabulary and similarities) and PIQ on the performance subsets (block design and matrix completion) of Wechslers Abbreviated Scale of Intelligence (WASI; Psychological Corporation, 1999). Independent sample t tests did not identify any signicant difference between the two groups on CA, VIQ, PIQ or full-scale IQ, t < 1 for all. Table 1 displays participants details.

Apparatus and stimuli The experiment was programmed and presented using E-Prime on an Acer Aspire 1522WLMi laptop with 15.4 widescreen. The stimuli consisted of 18 pairs of sentences beginning with a common English name and describing a behaviour. The behaviour described in half of the sentence pairs implied traits (for example, This is Duncan who sat alone in a corner at
608

Downloaded from aut.sagepub.com by guest on March 28, 2013

A S S O C I AT I N G T R A I T S T O F AC E S I N AU T I S M

Table 1

Details of participants CA VIQ PIQ IQ AQ

Aspergers syndrome Mean SD Range Typical Mean SD Range

17.8 1.2 16.320.6 17.10 1.6 16.521

105.31 12.12 84125 103.94 10.01 78123

102.25 13.68 71118 103.56 10.54 86121

104.19 11.97 85125 104.19 8.61 84117

23 5.96 1233 18 6.36 527

Note. CA = chronological age; VIQ = verbal intelligence quotient; PIQ = performance intelligence quotient; IQ = intelligence quotient; AQ = Autism-Spectrum Quotient.

the party). The other half of the sentence pairs implied facts (for example, This is Fred who has to bend down to enter most doors). The behaviours described in trait-implying sentences refer to abstract mental characteristics of a person like shy in the example given above. On the other hand, facts refer to relatively more concrete characteristics of a person like ones physique (tall in the example given above) or ones occupation (for example, This is Charles who carried the lunch ordered by the customers to their table implies waiter). The trait-implying sentences were based on a database developed by Uleman and colleagues (1988), and the factimplying sentences were constructed by the authors. A pilot study carried out with a group of typical adults ensured that the sentences developed do imply the intended traits and facts. Trait pairs selected were: clever-friendly, daring-condent, determinedcareful, impulsive-ignorant, shy-forgetful, lazy-careless, funny-kind, nosy-fussy and greedy-rude. And the fact pairs were: waiter-student, vegetarian-actor, non-smoker-tall, deaf-wet, sick-poor, foreigner-father, pilot-barber, richslim and Scottish-short. For the experiment, traits and facts were paired by random assignment ensuring that both the trait words and fact words of a pair were matched on valance or desirability as far as possible (for example, in the pair clever-friendly, both are desirable traits, whereas both are undesirable in the case of lazy-careless). The degree of desirability/undesirability was not taken into account when matching because at least in the case of traits, evaluative extremity was not found to affect response latencies, suggesting that traits are evaluated categorically in the rst instance (Pratto and John, 1991). Thirty-six photographs (coloured, neutral expression, frontal view of the face of Caucasian males) were selected from the database, the Psychological Image Collection at Stirling (PICS), developed in the School of 609

Downloaded from aut.sagepub.com by guest on March 28, 2013

14(6) Psychology, University of Stirling, UK, which is available at http:/ /pics. psych.stir.ac.uk/. The photographs were grouped into pairs by random assignment ensuring that they were matched on age. The dimension of the photographs was 6.5 6.5 cm (length breadth). The photograph pairs were separated by 12 cm on screen. The sentences were presented directly beneath each photograph in two lines. There was a distance of at least 3 cm between the two sentences. The cues presented in the test phase were of three types: trait words (like shy), fact words (like tall) and names (like Duncan or Fred). The cue word in the test phase was presented in the same line as the sentences but in between the two photographs as illustrated in Figure 1. The textual descriptions and the words were presented in 12 point Arial black font on a white background.
AU T I S M

Procedure The participants were tested individually in a quiet room. The participants sat approximately 50 cm from the laptop and were instructed (both verbally and textually) with the following; On each slide you will be shown a pair of pictures with some information about each written below it. Read the information given one by one and then press the space bar. A pair of pictures will come up on the screen again but this time with a single word in between the two pictures. Your task is to identify the picture you think best matches the word, by pressing z if you think it is the picture on the left and m if you think it is the picture on the right, based on the information you read. Please keep your hands on the keyboard ready to respond throughout the task. You will have two practice trials. Taking the example given in Figure 1, the study phase of an experimental trial would rst show photos of Victor (on the left of the screen) and Gordon (on the right of the screen) with the sentences This is Victor who ignored his injury and completed the race and This is Gordon who always drives a little slower than the speed limit under their respective photographs. After studying the presentation the participants pressed the space bar. In the next slide the same photos appeared again but with a single word determined (test phase). Participants had to press z if they thought determined was best associated with Victor or m if they thought determined was best associated with Gordon. The position of Victors and Gordons photos, in this case, was the same in the study and test phases (i.e. left and right, respectively). However, in half the trials the position of the photos in the study and test phase was interchanged as illustrated in the other example (Figure 1). Some research suggests that individuals with ASD do not attend to social stimuli, including faces, in their environment (Begeer et al., 2006;
610

Downloaded from aut.sagepub.com by guest on March 28, 2013

A S S O C I AT I N G T R A I T S T O F AC E S I N AU T I S M

Dawson et al., 1998, 2004). Consequently, participants with ASD may perform poorly on this task if they do not pay attention to the faces in the rst place and thus are unaware of the swap in spatial location in some of the trials. In order to tackle this issue, the practice trials alone used non-social stimuli and provided feedback. Apart from being non-social these stimuli were perhaps more distinct from one another than faces (see Figure 2). Furthermore, in the practice trials participants were told that the response was incorrect if they made errors and were asked to attempt that particular trial again. The experimental trials were administered only after the participants successfully passed both practice trials. In the study phase of the experimental trials, each participant received a random order of 18 trials 9 trials with trait-implying sentence pairs and 9 trials with fact-implying sentence pairs. The test phase consisted of 18 individual words; 6 traits, 6 fact-related words and 6 names. Three of the names were from trait sentences and 3 were from fact sentences. Each participant was administered one of eight counterbalanced forms of the experiment such that each photograph was presented with both sentences of its pair equally often, each photograph pair in the test phase was presented in the same and opposite spatial orientation as the study phase an equal number of times, and the trait/fact implied in both the sentences of a pair was presented equally often in the test phase. The participants were assigned to the different forms using systematic allocation, so that the rst participant who was tested did form 1, the second did form
Practice Trial 1 Study phase Test phase

The apple was smelly and had worms in it.

The banana was hard and green in colour.

rotten

Practice Trial 2 Study phase Test phase

The pizza had steam coming up from it.

The curry had lots of chilies in it.

spicy

Figure 2

Illustration of the two practice trials which employed non-social stimuli

611

Downloaded from aut.sagepub.com by guest on March 28, 2013

14(6) 2, the third did form 3, and so on. The procedure was self paced and the test phase appeared on the screen 1500 milliseconds after the participants pressed the spacebar. The keys to be used, namely the spacebar and keys z and m were highlighted with labeled stickers; all the keys except for these were locked.
AU T I S M

Results
A score of 1 was given for each correct response. The maximum score a participant could obtain was 18 (with a maximum of 6 for traits, facts and names individually). Table 2 gives the accuracy of the participants with AS and participants with typical development on the three words when the spatial location of the photos in the two phases was the same and when opposite. One-sample t tests with a test value set as 0.5 (which is the expected accuracy by chance) were carried out on the proportion of correct responses. The results indicated that the accuracy rate of the group of individuals with AS and the group of individuals with typical development was signicantly above chance in all three word conditions irrespective of spatial location. The total scores for the three words were normally distributed and a repeated measures ANOVA with word (trait versus fact versus name) as the within-subject variable and group (AS versus Typical) as the betweensubjects variable revealed a main effect of word type, F(2, 60) = 12.6, p < .001, f = 0.21. Post hoc analysis (Bonferroni) revealed that the accuracy for the name condition was signicantly less than for the fact condition,
Table 2 Percentages of correct responses in the two groups on trait, fact and name conditions Aspergers syndrome Fact Opposite Same Total Trait Opposite Same Total Name Opposite Same Total Typical

77% 96% 86% 85% 96% 90% 84% 83% 84%

88% 90% 89% 88% 98% 93% 76% 76% 76%

612

Downloaded from aut.sagepub.com by guest on March 28, 2013

A S S O C I AT I N G T R A I T S T O F AC E S I N AU T I S M

p < .01, and the trait condition, p < .001. There was no signicant main effect of group, F < 1, and the interaction was not signicant, F(2, 60) = 1.29, p > .05. The reaction time was measured from the onset of the test phase to when the participants responded by pressing z or m. The reaction time for only the correct responses was included in the analysis. In order to meet conditions of normality the reaction time data were submitted to logarithmic transformation and then analyzed using a 3 (word: trait versus fact versus name) 2 (spatial location: same versus opposite) 2 (group: AS versus typical) mixed ANOVA. The analysis indicated a signicant main effect for word, F(2, 60) = 4.24, p < .05, f = 0.38. Post hoc analyses (Bonferroni) revealed that the reaction time for the fact condition was signicantly less than for the trait condition, p < .05. There was a signicant main effect of spatial location as well, F(2, 60) = 22.92, p < .001, f = 0.87, with the reaction time being signicantly less when the photographs in the test phase were in the same spatial location as in the study phase. A signicant main effect of group was also observed, F(2, 60) = 4.34, p < .05, f = 0.38, with the typical group being signicantly faster than the group with AS. None of the interactions were signicant. See Figure 3.

3500

TYP

AS

3000 Mean RT (ms)

2500

2000

1500 Trait Fact Cue Type

Figure 3 Mean reaction time (RT) of the typical group (TYP) and the group with Aspergers syndrome (AS) in trait, fact and name conditions

Name

613

Downloaded from aut.sagepub.com by guest on March 28, 2013

AU T I S M

14(6)

Discussion
The name condition differed from the trait and fact conditions in that the cues were explicitly presented in the study phase. In the trait and fact condition participants had to draw an inference and then associate this inference with the actor. One possible outcome we considered was that participants with ASD may have performed poorly in the trait and fact condition in comparison to the name condition due to the additional processing required. Furthermore, names involve making an association between two explicitly presented stimuli, which people with ASD are reputed to be good at doing (Williams et al., 2006). A general difculty associating any information to faces, on the other hand, would lead to poor performance in all three word conditions. A specic difculty with associating social information like traits would lead to poor performance in the trait condition but not the name or fact conditions. Accuracy was signicantly above chance for participants with AS and participants with typical development on the three word types, irrespective of spatial location. This suggests that participants of both groups were associating the inferred traits and facts as well as names with the actor. Ramachandran et al. (2009) found that individuals with ASD were drawing trait inferences spontaneously that is, with minimal effort, without reason for or benet from drawing such inferences on reading behavioural descriptions that implied traits. In this experiment participants could have made this association in the study phase or in the test phase following the presentation of the cue. If participants were making the link in the study phase, they might on seeing the cue orient towards the same location where they had seen the related face in the study phase. This would increase the reaction time for the opposite spatial location condition as participants would, on nding the incorrect face, need to re-orient to the opposite side. The results indicated that the reaction time was signicantly less when the spatial location was the same compared to when opposite for both groups of participants. Thus, it is possible that participants (including those with AS) were inferring the traits and facts implied in the behavioural descriptions and associating the inference with the face in the study phase. In the test phase they used this information to correctly identify the face. The performance of participants with AS was comparable to participants with typical development on all three word conditions. This result contradicts research that suggests difculty with face recognition in autism, especially immediate recognition (Williams et al., 2005). At least when the task demands that an impression be formed of a person, it seems that participants with AS do process faces and distinguish between people. Begeer et al. (2006) found that when attending to faces was made crucial 614

Downloaded from aut.sagepub.com by guest on March 28, 2013

A S S O C I AT I N G T R A I T S T O F AC E S I N AU T I S M

to the task, participants with ASD processed emotional expressions of people. Similarly, one possible reason for successful processing of faces by individuals with AS in this experiment may be that the task demanded elaborate processing of the face stimuli. An ANOVA revealed that the reaction time for the name condition did not differ signicantly from the trait or the fact condition in either group. However, the accuracy for the name condition was found to be signicantly less than for the trait condition and the fact condition in both groups. Thus, participants with ASD, like typical participants, were more accurate when identifying faces based on inferences they made rather than explicitly presented names. Relatively poor performance on the name condition in this experiment is consistent with existing literature on identifying faces based on names in typical participants (McWeeny et al., 1987; Scanlan and Johnston, 1997;Young et al., 1985). However, it is surprising that participants with AS were relatively effective in identifying people based on a social concept like traits which required inferences to be drawn. Both groups of participants showed faster reaction times when inferring facts than when inferring traits. This is not surprising given that the characteristics implied by fact-implying sentences were inherently less ambiguous than traits. A large number of behaviours could potentially imply a given trait and conversely a single behaviour can be perceived to imply different traits. For example,He carried the old ladys bag of groceries to her car can be perceived as kind or helpful. This is not the case with facts. The behaviour used to describe facts unambiguously refers to the particular fact. For example,He carried the lunch ordered by the customers to their table is a relatively unequivocal description of a waiter. Despite that, the accuracy for traits was found to be more than for facts, though not signicantly so. Thus, though quicker, both participants with AS and typical participants made slightly more errors identifying the face based on facts than on traits. Overall, the participants with AS performed similarly to the participants with typical development, identifying the face with relatively high levels of accuracy and speed based on inferences drawn from behavioural descriptions as well as names. Participants with AS were signicantly slower than participants with typical development. This could be explained in terms of a motor movement planning decit which is common in ASD and observed in reaction time paradigms (Rinehart et al., 2001). The similar pattern of reaction times for participants with AS and typical participants suggests a general delay in responding rather than difculty on any specic word type. In conclusion, the results of this study tell us that individuals with ASD associate inferred traits with an image of the face. The speed and accuracy of participants with ASD was comparable on traits and facts. Facts were 615

Downloaded from aut.sagepub.com by guest on March 28, 2013

14(6) relatively unambiguous while traits were sometimes about ambiguous mental characteristics. Associating the trait with the face would be difcult for individuals with ASD if they lacked understanding about the mind on a very broad level. Surprisingly, participants with ASD did not have difculty processing the relatively large amount of social information and they were more accurate when identifying faces based on inferences drawn from behavioural descriptions (facts as well as traits) than explicitly presented names. Thus, individuals with ASD associated the inferred traits with the face without difculty. In this experiment and in the experiments reported in Ramachandran et al. (2009) the traits were presented as they relate to behaviour. While the results suggest that trait inference in ASD may be a spared socio-cognitive function, at least at the level of relating to behaviour, further research could look into whether individuals with ASD also understand how traits are mediated by stable mental states. Research suggests that individuals with ASD may not spontaneously attend to relevant stimuli in the environment, particularly when the stimuli are social in nature (Dawson et al., 1998, 2004). In this experimental task the importance of paying attention to the pictures was made explicit by the use of nonsocial stimuli and feedback in the practice trials. This may have resulted in elaborate but not spontaneous processing of the face stimuli. Further research could investigate if individuals with ASD are equally successful when the instructions are less explicit. In other words, would they naturally and spontaneously attend to others behaviour and faces and make an association between the two? Irrespective of the answer to that question, it is intriguing that individuals with ASD, under some circumstances, do process faces and distinguish between people based on sometimes ambiguous social information, like mental characteristics.
AU T I S M

Acknowledgements Rajani Ramachandran is now afliated with the University of Calicut, Kerala, India. This work was supported by a scholarship received from the Commonwealth Scholarship Commission, UK (INCS-2004154) to the rst author. We are very grateful for the kind cooperation of the staff, students, and parents involved in the studies. References
American Psychiatric Association. (1994) Diagnostic and Statistical Manual of Mental Disorders (4th edition). Washington, DC: American Psychiatric Association. Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J. & Clubley, E. (2001) The Autism-Spectrum Quotient (AQ): Evidence from Asperger Syndrome/ High-Functioning Autism, Males and Females, Scientists and Mathematicians, Journal of Autism and Developmental Disorders 31: 517.

616

Downloaded from aut.sagepub.com by guest on March 28, 2013

A S S O C I AT I N G T R A I T S T O F AC E S I N AU T I S M

Begeer, S., Rieffe, C., Meerum Terwogt, M. & Stockmann, L. (2006) Attention to Facial Emotion Expressions in Children with Autism, Autism 10: 3751. Boucher, J. & Lewis,V. (1992) Unfamiliar Face Recognition in Relatively Able Autistic Children, Journal of Child psychology and Psychiatry 33: 843859. Dawson, G., Meltzoff, A.N., Osterling, J., Rinaldi, J. & Brown, E. (1998) Children with Autism Fail to Orient towards Naturally Occurring Social Stimuli, Journal of Autism and Developmental Disorders 28: 479485. Dawson, G., Toth, K., Abbott, R., Osterling, J., Munson, J., Estes, A. & Liaw, J. (2004) Early Social Attention Impairment in Autism: Social Orienting, Joint Attention, and Attention to Distress, Developmental Psychology, 40, 271283. Dozier, M. (1991) Functional Measurement Assessment of Childrens Ability to Predict Future Behavior, Child Development 62: 10911099. Ferguson, T.J., Olthof, T., Luiten, A. & Rule, B.G. (1984) Childrens Use of Observed Behavioral Frequency versus Behavioral Covariation in Ascribing Dispositions to Others, Child Development 55: 20942105. Gnepp, J. & Chilamkurti, C. (1988) Childrens Use of Personality Attributions to Predict Other Peoples Emotional and Behavioral Reactions, Child Development 59: 743754. Hermelin, B. & OConnor, N. (1986) Idiot Savant Calanderical Calculators: Rules and Regularities, Psychological Medicine 16: 885893. Hirschfeld, L., Bartmess, E., White, S. & Frith, U. (2007) Can Autistic Children Predict Behaviour by Social Stereotypes? Current Biology 17: 451452. Jemel, B., Mottron, L. & Dawson, M. (2006) Impaired Face Processing in Autism: Fact or Artifact? Journal of Autism and Developmental Disorders 36: 91106. Klin, A., Sparrow, S.S., de Bildt, A., Cicchetti, D.V., Cohen, D.J. & Volkmar, F.R. (1999) A Normed Study of Face Recognition in Autism and Related Disorders, Journal of Autism and Developmental Disorders 29: 499508. Klinger, L.G. & Dawson, G. (2001) Prototype Formation in Autism, Development and Psychopathology 13: 111124. McWeeny, K.H.,Young, A.W., Hay, D.C. & Ellis, A.W. (1987) Putting Names to Faces, British Journal of Psychology 78: 143149. Minshew, N.J., Goldstein, J., Muenz, L.R. & Payton, J.B. (1992) Neuropsychological Functioning in Non&ndashMentally Retarded Autistic Individuals, Journal of Clinical and Experimental Neuropsychology 14: 741761. Pratto, F. & John, O.P. (1991) Automatic Vigilance: The Attention-Grabbing Power of Negative Social Information, Journal of Personality and Social Psychology, 61: 380391. Ramachandran, R., Mitchell, P. & Ropar, D. (2009) Do Individuals with Autism Spectrum Disorders Infer Traits from Behavior? Journal of Child Psychology and Psychiatry 50: 871878. Rinehart, N.J., Bradshaw, J.L., Brereton, A.V., & Tonge, B.J. (2001) Movement Preparation in High-Functioning Autism and Aspergers Syndrome: A Serial Choice Reaction Time Task Involving Motor Reprogramming, Journal of Autism and Developmental Disorders 31: 7988. Scanlan, L.C. & Johnston, R.A. (1997) I Recognize Your Face, but I Cant Remember Your Name: A Grown Up Explanation, The Quarterly Journal of Experimental Psychology A 50: 183198. The Psychological Corporation (1999) Wechsler Abbreviated Scale of Intelligence (WASI) Manual. San Antonio, TX: The Psychological Corporation. The Psychological Image Collection at Stirling. Retrieved 21 September 2006 from http:/ /pics.psych.stir.ac.uk/

617

Downloaded from aut.sagepub.com by guest on March 28, 2013

AU T I S M

14(6)

Todorov, A. & Uleman, J.S. (2004) The Person Reference Process in Spontaneous Trait Inference, Journal of Personality and Social Psychology 87: 482493. Uleman, J.S. (1988) Trait inference norms. Unpublished raw data. New York University, New York. Wellman, H.M., Cross, D. & Watson, J. (2001) Meta-Analysis of Theory of Mind Development: The Truth about False Belief, Child Development 72: 655684. Williams, D.L., Goldstein, G. & Minshew N.J. (2005) Impaired Memory for Faces and Social Scenes in Autism: Clinical Implications of the Memory Disorder, Archives of Clinical Neuropsychology 20: 115. Williams, D.L., Goldstein, G. & Minshew N.J. (2006) The Prole of Memory Function in Children with Autism, Neuropsychology 20: 2129. Young, A.W., Hay, D.C. & Ellis, A.W. (1985) The Faces that Launched a Thousand Slips: Everyday Difculties and Errors in Recognising People, British Journal of Psychology 76: 495523. Yuill, N. & Pearson, A. (1998) The Developmental Bases for Trait Attribution: Childrens Understanding of Traits as Causal Mechanisms Based on Desire, Developmental Psychology 34: 574586.

618

Downloaded from aut.sagepub.com by guest on March 28, 2013