The Relationship between Form and Function

J Autism Dev Disord (2008) 38:1328–1340 DOI 10.1007/s10803-007-0520-z

ORIGINAL PAPER

The Relationship between Form and Function Level Receptive Prosodic Abilities in Autism Anna Ja¨rvinen-Pasley Æ Susan Peppe´ Æ Gavin King-Smith Æ Pamela Heaton

Published online: 3 January 2008 Ó Springer Science+Business Media, LLC 2008

Abstract Prosody can be conceived as having form (auditory-perceptual characteristics) and function (pragmatic/linguistic meaning). No known studies have examined the relationship between form- and functionlevel prosodic skills in relation to the effects of stimulus length and/or complexity upon such abilities in autism. Research in this area is both insubstantial and inconclusive. Children with autism and controls completed the receptive tasks of the Profiling Elements of Prosodic Systems in Children (PEPS-C) test, which examines both form- and function-level skills, and a sentence-level task assessing the understanding of intonation. While children with autism were unimpaired in both form and function tasks at the single-word level, they showed significantly poorer performance in the corresponding sentence-level tasks than controls. Implications for future research are discussed. Keywords Speech

Autism Language Perception Prosody

Language impairment is a hallmark feature of autism (American Psychiatric Association 1994). At the same

A. Ja¨rvinen-Pasley P. Heaton Goldsmiths College, University of London, London, UK Present Address: A. Ja¨rvinen-Pasley (&) Laboratory for Cognitive Neuroscience, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, CA 92037-1099, USA e-mail: [email protected]; [email protected] S. Peppe´ G. King-Smith Queen Margaret University, Edinburgh, UK

123

time, this disorder is associated with a significantly heterogeneous language phenotype, ranging from mutism and little functional communication (25–50% of diagnosed individuals) (Gillberg and Coleman 2000; Klinger et al. 2002) to nearly typical skills (Boucher 2003; Kjelgaard and Tager-Flusberg 2001; Lord and Paul 1997). As the language and communication impairments characterizing autism are intimately linked to the core abnormalities in social interaction (Boucher 2003), it is unsurprising that pragmatics has been identified as the only aspect of language that is universally and specifically impaired in autism (Lord and Paul 1997; Tager-Flusberg 1981). Within the pragmatic domain of language, prosody refers to the paralinguistic aspects of the spoken signal, which modulate and enhance its meaning serving affective, grammatical, and pragmatic functions (Crystal 1969). Acoustically, speech prosody is manifested by, but not limited to, variations in fundamental frequency, amplitude, and duration (Lehiste 1970). The most significant prosodic effects are produced by the linguistic use of pitch, or intonation (Lieberman 1960). Prosodic deficits appear not to be universal in autism (Simmons and Baltaxe 1975; Shriberg et al. 2001). However, since the earliest descriptions of autism spectrum disorders (Kanner 1943; Asperger 1944), atypical use of prosody has been identified as one of the hallmarks of speakers with this disorder (e.g., Baltaxe and Simmons 1985; Fay and Schuler 1980; Pronovost et al. 1966). When such abnormalities are present, they are noted early, and are highly persistent (Simmons and Baltaxe 1975). These abnormalities are highly variable, and speech may be monotonic, sing-song-like, robotic, parroted, machine-like, odd, over-exaggerated, and/or stilted, encompassing atypical use of pitch, rhythm, and intonation. Abnormal use of prosody is of significant clinical importance. Recently, Paul et al. (2005a) found associations

J Autism Dev Disord (2008) 38:1328–1340

between prosody use and socialization and communication ratings on the Vineland Adaptive Behavior Scales (Sparrow et al. 1984) and the Autism Diagnostic Observation Schedule (Lord et al. 2000) in high-functioning speakers with autism. To date, the majority of investigations into prosody in autism have focused upon expressive abilities, and surprisingly little is known about the ability of individuals with autism to interpret prosodic cues in the speech of others. A recent review article (McCann and Peppe´ 2003) concluded that the existing research literature painted an inconsistent and insubstantial picture of prosody in autism. Many studies have reported abnormalities in both the assignment and realization of stress patterns, as well as in the production of intonation patterns distinguishing utterance types (e.g., question/statement) (Baltaxe and Guthrie 1987; Baltaxe and Simmons 1985; Fine et al. 1991; Shriberg et al. 2001). However, some studies have produced mixed results. For example, Fosnot and Jun (1999) observed deficits in the use of intonation, while Baltaxe and Simmons (1985) failed to identify such abnormalities. McCann and Peppe´ (2003) suggested that these inconsistent findings may have been due to a number of shortcomings in the research, including poor diagnostic data, the use of small sample sizes, a lack of normative data and appropriate comparison groups, and the use of subjective ratings rather than objective assessment methods. Two recent studies with appropriate control data have examined both the perception and production of different aspects of prosody in children and young adults with autism (Paul et al. 2005b; Peppe´ et al. 2007). In the study by Paul et al. (2005b), stress, intonation, and phrasing, within and across both the pragmatic-affective and grammatical domains, were examined in sentence-level tasks. Twentyseven participants with autism aged from 14 to 21 years and chronological age-matched typically developing controls were tested. Although the interpretation of results from this study was somewhat limited by ceiling-level performance in both groups, the findings nevertheless showed that the prosodic deficits of participants with autism were limited to tasks involving stress. Specifically, the perception and production of both pragmatic/affective and grammatical/linguistic stress were poorer in these individuals relative to typical controls. However, the authors concluded that the apparent absence of prosodic betweengroup differences was due to ceiling effects in the perception tasks, and insensitivity of the assessment. Consistent with this was Paul et al.’s (2005b) suggestion that the prosodic differences in the task items may have been overly pronounced, resulting in low detection thresholds. In the study by Peppe´ and colleagues (2007), receptive and expressive prosodic abilities were tested using a

1329

prosody assessment battery entitled Profiling Elements of Prosodic Systems in Children (PEPS-C; Peppe´ and McCann 2003). Participants included 31 children with autism aged from 6 to 13 years, and control groups comprising typically developing children matched on verbal mental age and typical adults. The assessment allows prosodic abilities to be examined at both the form level, involving perception (auditory discrimination and the production of prosodic effects through imitation), and at the function level, requiring cognitive processing (pragmatic/interactional, affective, and grammatical/linguistic), across analogous receptive and expressive tasks. The findings of Peppe´ et al. (2007) showed that, relative to controls, children with autism exhibited significantly poorer prosodic performance overall. More specifically, results from the receptive perceptual/form tasks showed that the difficulty of children with autism concerned their tendency to perceive ‘‘same’’ auditory pairs as different. This impairment was more pronounced at the phrasal than at the single-word level. At the cognitive/function level, there was a trend towards participants with autism tending to judge questions as statements, and producing questioning intonation when a statement was required. In addition, participants with autism showed a diffuse pattern of errors in the function tasks assessing affective prosody, and exhibited pronounced deficits in their ability both to interpret and produce contrastive stress or focus. The mildest between-group differences emerged in the function tasks assessing phrasing (or ‘‘chunking’’). The findings highlighted a delay in the acquisition of receptive prosody in autism although some clearly atypical patterns were also in evidence. Comparisons of performance across the receptive and expressive domains in autism showed that receptive and expressive abilities were associated with each other, particularly in grammatical/linguistic prosodic functions. The finding of Peppe´ et al. (2007), indicating deficits in the perception of auditory-perceptual characteristics of prosody (form) in children with autism, raises important questions about the acquisition and organization of such abilities in autism. This is because form-level auditoryperceptual discrimination ability is considered to be a prerequisite of function-level cognitive prosodic abilities. Indeed, the stimuli for the form- and function-level tasks of the PEPS-C assessment differ only in the addition of lexical content at the function level (Peppe´ and McCann 2003). However, this appears to be inconsistent with findings showing both form-level deficits (Peppe´ et al. 2007) and unimpaired performance in function-level tasks assessing the understanding of the grammatical meaning of intonation (distinguishing question and statement utterances) (Paul et al. 2005b; Peppe´ et al. 2007) in individuals with autism. The study by Paul et al. (2005b) employed a sentence-level task while that of Peppe´ et al. (2007) tested

123

1330

this ability at the single-word level only. While these findings appear to be paradoxical, apparent typical performance in function-level tasks may, as previously suggested, reflect task insensitivity resulting from ceilinglevel performance, and participants with autism may well have exhibited impairments at the function level. Consistent with this suggestion are findings by Peppe´ et al. (2007), showing that some participants with autism in their study tended to judge questions as statements, with 12.9% of such participants judging all question-type stimuli as statements. Moreover, the finding showing impaired discrimination of auditory-perceptual features of prosody in individuals with autism (Peppe´ et al. 2007) is surprising within the context of a large body of evidence indicating intact or enhanced processing of non-linguistic pitch in autism (e.g., Bonnel et al. 2003; Heaton et al. 1998; Heaton 2003, 2005; Mottron et al. 2000). Speech prosody and music share significant acoustic qualities, such as fundamental frequency or temporal variations of similar period. Furthermore, recent studies have indicated that individuals with autism show superior perceptual processing of intact speech stimuli as compared to age- and intelligence-matched controls (Ja¨rvinen-Pasley et al. 2007, 2008). These findings warrant further investigation of the relationship between form and function level prosodic abilities in autism. The present study was designed to address the question of whether auditory-perceptual form- and cognitive function-level prosodic ability will dissociate in individuals with autism, and whether the length/complexity of speech stimuli will have an effect upon these processing levels in such individuals. Thus, the current experiments will specifically compare the perceptual discrimination of prosody (form) with receptive cognitive function-level abilities in children with autism and controls matched for age, nonverbal intelligence, and receptive vocabulary level. Although function-level abilities will be examined across grammatical/linguistic and affective-pragmatic tasks, a focused analysis comparing auditory-perceptual abilities (form) and those involving the understanding of grammatical use of intonation (function) across single-word and phrasal levels, will allow for a direct comparison with earlier studies (Paul et al. 2005b; Peppe´ et al. 2007). The rationale and aims of the study are: (1) In order to examine the relationship between perceptual/form- and cognitive/ function-level receptive prosodic abilities in individuals with autism, the receptive component of the PEPS-C assessment battery (Peppe´ and McCann 2003) will be administered to children with autism and their matched controls. Based upon the earlier findings of Peppe´ et al. (2007), it is hypothesized that children with autism will show an atypical organization of prosodic abilities across the form and function levels. Based upon the broader

123


existing literature, it is further hypothesized that children with autism will show lower levels of prosodic performance than controls. (2) Findings of Paul et al. (2005b) and Peppe´ et al. (2007) suggested unimpaired understanding of the grammatical meaning of intonation in individuals with autism, in spite of auditory-perceptual form-level prosodic deficits (Peppe´ et al. 2007). Thus, the impact of the length/complexity of speech stimuli upon the perceptual (form) and cognitive (function) processing levels in individuals with autism will be investigated. A sentencelevel task requiring the discrimination of question and statement utterances on the basis of intonation will be administered, to supplement the PEPS-C data.

Experiment 1: Assessing Prosodic Abilities in Children with Autism Using the PEPS-C Test Methods Participants For this study, 42 children were recruited from schools in Southwest England. They were a subset of children taking part in studies by Ja¨rvinen-Pasley et al. (2007, 2008; Ja¨rvinen-Pasley and Heaton 2007). Twenty-one children with a formal diagnosis of autistic disorder (AD) or Asperger disorder (AS) according to DSM-IV (APA 1994) or ICD-10 (World Health Organization 1993) criteria were recruited from a specialist educational establishment for children with autism. The diagnostic information was gathered from school files of documented medical diagnoses and clinical reports, and showed that each child had individually received a diagnosis by experienced clinicians within 2 years prior to conducting the study. No child had a diagnosis of Pervasive Developmental Disorder, Not Otherwise Specified, and 67% of the children had a diagnosis of AS, and the remaining 33% had a diagnosis of AD. One child with AS had, at one stage, been diagnosed as having AD but the diagnosis had been amended to that of AS because the autistic symptoms were not accompanied by a clinically significant delay in early language development. The children with a diagnosis of AD were classified as having autism accompanied by language delay. Children with Rett syndrome, Childhood Disintegrative Disorder, or autism-related medical conditions, such as Fragile X syndrome, tuberous sclerosis, and neurofibromatosis, were not included in this study. Children in both the clinical and control groups were also excluded from participation if they had a diagnosis of any medical disorder (e.g., epilepsy). The selected children all met the following criteria: they had a mono-lingual Englishspeaking home environment and were Caucasian, had no


sensorineural hearing impairment, and showed no evidence of neurological or medical abnormalities. Further, only children who showed fluent use of spoken language were included in the study. This information was gathered from the children’s teachers and verified by the experimenter in a pre-test conversation phase. Control children were matched on an individual basis to those with autism for chronological age (CA), receptive vocabulary measured by the British Picture Vocabulary Scale (BPVS) (Dunn et al. 1997), and non-verbal intelligence (NVIQ) measured by the Raven Standard Progressive Matrices (RSPM) (Raven et al. 1992). As autism is associated with intellectual impairment in 50– 75% of cases (Rapin 1997; Spence et al. 2004), the validity of research findings can be compromised by inadequate control group matching, for example, matching for CA results in mean IQ differences between participants with autism and controls. Similarly, matching for mental age results in a failure to control for potentially significant maturational factors. In the present study, therefore, the control children were matched on an individual basis to those with autism for CA, as well as on receptive vocabulary and non-verbal intelligence. Control children were recruited from a mainstream primary school and a secondary school for children with moderate learning difficulties (MLD). The standardized scores of 24% of children with autism on both the BPVS and RSPM were at least two standard deviations below the population mean (score B70). Thus, the same proportion of children in the control group had MLD. To avoid introducing a systematic bias of an accompanying disorder, only children with learning difficulties conditions that were of non-specific origin were included. No children were categorized as having language-specific disorders; instead, their language difficulties appeared to result from a combination of generally low cognitive abilities and environmental disadvantages. These children had no known history of neurological disorder or head injury. Children with severe comprehension difficulty were excluded. Participation was also ruled out if there was any reason to suspect difficulties in social development for any control child (e.g., a sibling with an autistic spectrum disorder). All children were fluent in expressive language, and showed no evidence of autistic-like behaviors or language use. All diagnostic information was again gathered from school personnel and school files of documented

1331

medical diagnoses and clinical reports, and the selected children all met the same criteria as specified for the children with autism. Approximately 76% of the control children were typically developing. They were characterized as showing average academic ability, and were perceived as having no special problems by their teachers. These children had no history of a language, neurological or medical disorder, and scored within the normative range on the BPVS and RSPM. Informed consent was obtained from the parents of all participating children. Ethical approval was obtained from the Research Ethics Committee of Goldsmiths College, University of London, which is in accordance with the guidelines of the British Psychological Society (BPS). Table 1 shows the demographic characteristics of the two groups of children. No significant between-group differences in CA (t (40) = .47, p = .64), the BPVS (t (40) = -.55, p = .59), or the RSPM (t (40) = .24, p = .81) standardized scores are in evidence.

Materials The PEPS-C test is a relatively new prosody assessment procedure. The U.K. version is described in Peppe´ and McCann (2003). The assessment follows a psycholinguistic framework (Stackhouse and Wells 1997), whereby both receptive and expressive abilities are examined in analogous tasks; however, only the receptive component was employed in the current study. The tasks are further divided into form involving perception (auditory discrimination) and communicative function requiring cognitive processing (pragmatic, affective, grammatical/linguistic). Thus, PEPSC assesses most of the aspects covered in the literature in autism (see McCann and Peppe´ 2003). The assessment is computerized, and records participants’ responses into sound and text files. Each subtest includes two example items, two practice items, and 16 test items. Only the test items count towards the participant’s total score. To avoid a heavy demand on auditory memory, only two response choices were offered for each item in receptive tasks. The response screen involves a split-screen display of attractive cartoon-type pictures (for item descriptions, see the Appendix). Because of the binary-choice design, only scores [11 and \5 can be considered as non-chance results. The developers of the PEPS-C suggest that children can be

Table 1 Characteristics of the two participant groups in study 1 (SD; range, in parentheses) M CA (years)

Sex

M VIQ (BPVS)

M NVIQ (RSPM)

Autism group (n = 21)

12.55 (2.50; 7.67–16.75)

18M; 3F

84 (19.33; 55–135)

89 (14.44; 61–119)

Control group (n = 21)

12.21 (2.15; 8.33–16.25)

13M; 9F

87 (21.32; 58–124)

88 (17.92; 61–121)

123

1332

Procedure Each child was tested on an individual basis in a quiet room in their own school. The child was seated in front of a laptop computer, within easy reach of the computer mouse. The children made judgments by clicking on the side of the computer screen that displayed the relevant response choice. The mouse click then prompted the next stimulus, in a fixed order. Where possible, the experimenter sat diagonal to the child, in such a way that both could see the computer screen. In cases where a child had immature motor skills (3 children with autism and 2 controls with MLD), the experimenter maneuvered the mouse pointer on the child’s behalf. In these instances, the child was simply asked to point to the side of the computer screen that s/he thought best matched the auditory stimulus, and the experimenter moved and clicked the mouse on that side of the computer screen. The child was told by the experimenter that s/he would be given a series of speech tasks on the computer. Prior to administering the subtests, two preliminary tasks were carried out: a vocabulary check ensured that the child was familiar with the items that would appear in the tasks. Here, pictures appear one by one on the screen, and the child was asked to name each item. The experimenter explained to the child that some of the items may be unfamiliar, and helped the child where necessary. The second preliminary task was a same/different concept check, which ensured that the child was familiar with such concepts. Here, the first screen depicts two identical red circles, with the word ‘‘same’’ written underneath, and the child was asked to state what he or she notices about the two circles. The second screen depicts a red circle and a green square, with the word ‘‘different’’ spelled underneath, and again the child was asked to comment on them. No child showed any difficulty with understanding these concepts. The PEPS-C subtests were administered in the following order: Short-item discrimination (Form), Turn-end (Grammatical intonation; Function), Affect (Affective intonation; Function), Longitem discrimination (Form), Chunking (Phrasing;

123

Function), and Focus (Contrastive stress; Function). No feedback was given on the task performance, except for positive reinforcement.

Results from PEPS-C Tasks The mean performance scores in the six PEPS-C receptive tasks for the children with autism and their age- and intelligence-matched control children are shown in Fig. 1. Error bars represent ±1 standard error mean. The children’s performance in each of the six PEPS-C tasks was compared against chance level performance (8) by applying one-sample t tests. This analysis showed that the performance of both groups of children was significantly above chance in all subtests (Autism: all p B .004; Control: all p \ .001). The scores for each PEPS-C subtest were normally distributed (Levene’s Test for Equality of Variance F statistic for all variables p C .11). A two by two repeated measures analysis of variance (ANOVA) was carried out on the data, with prosody level (form (Short- and Longitem discrimination)/function (Affect, Chunking, Focus, Turn-end)) as the within-participants variable, and diagnosis (autism/control) as the between-participants variable. This analysis revealed a significant main effect of prosody level (F (1, 40) = 7.86, p \ .01), suggesting that both groups of children exhibited relatively better performance in the form, as compared to the function, tasks; and diagnosis (F (1, 40) = 5.02, p = .03), suggesting that the autism group exhibited poorer performance overall as compared to their controls. There was no significant prosody level by diagnosis interaction (F (1, 40) = .70, n.s.), suggesting that both groups exhibited similar patterns of performance across the form and function level prosody tasks. However, as a task score [12 was deemed to represent competence level by the developers of the PEPS-C,

16

Autism group n=21 Control group n=21

14 12

Mean score

deemed to have reached competence level in a task if their score is at least 12 (75%), rather than 50%. It was therefore necessary to include at least 16 items per task, in order to obtain a reasonable number of non-chance scores. To keep the demands on attention as low as possible, the number of items was no higher than 16. The description of each of the receptive tasks of the PEPS-C assessment, together with examples of the picture stimuli used, is included in the Appendix. A demonstration of the assessment battery can be found on the following website: http://www.qmu.ac.uk/ ssrc/prosodyinasd/.


10 8 6 4 2 0 *Short-item *Long-item discr. discr. *Form-level tasks

†Affective intonation

†Chunking (phrasing)

†Focus (stress)

†Turn-end (intonation)

†Function-level tasks

Fig. 1 Means and standard error means for the receptive PEPS-C task scores for both the children with autism and their age- and intelligence-matched controls (maximum score per subtest = 16). * Denote form-level tasks and denote function-level tasks


1333

Fig. 1 shows that the controls had achieved the suggested competence level in all form-level tasks and in all but one function-level task (Focus or contrastive stress). By contrast, children with autism had failed to reach the suggested competence level in all function-level tasks. At the form level, participants with autism had reached the suggested competence level in the Short-item discrimination task, while this is not the case for the Long-item discrimination task. In order to explore more closely the between-diagnostic group differences, post hoc t tests with Bonferroni alpha adjustment were carried out for each of the six group comparisons. The results of this analysis are presented in Table 2. As the understanding of the grammatical use of intonation was of particular interest to the present study (cf. Paul et al. 2005b; Peppe´ et al. 2007), the children’s response patterns in the Turn-end task were examined. No significant between-group differences emerged in children’s performance in this task. As the data from this task was analyzed as a single score comprising both declarative and question judgments, it was of interest to examine whether this data analysis may have camouflaged important differences in response patterns between the two groups of children. Each individual child’s Turn-end task score was divided into two separate scores: the correct identification of declarative utterance types, and the correct identification of question utterance types. The means, standard deviations and ranges for the correct classification of declarative and question utterances in the Turn-end task for the children with autism and their matched controls are presented in Table 3. The percentages of correct responses are shown in parentheses. Although the number of stimulus items is small (8 per category), Table 3 shows that children with autism exhibited similar levels of performance in identifying both declarative and question utterances. By contrast, controls showed higher levels of performance with the question utterances than with the declaratives.

Table 2 Post hoc tests for differences between the test scores of the diagnostic groups in the PEPS-C tasks PEPS-C prosody task

t

PB

Short-item discrimination

-1.56

.127

Long-item discrimination

-2.45

.019*

Affect (Affective intonation)

-2.38

.022*

Chunking (Phrasing)

-2.04

.048*

Focus (Contrastive stress) Turn-end (Grammatical intonation)

-0.26 -1.32

.794 .193

* Significant at p \ .05

Table 3 Means, standard deviations, and ranges (% correct in parentheses) for the correct identification of declarative and question utterances in the Turn-end task by both the children with autism and their matched controls (maximum score per category = 8) Declarative utterances Mean

SD

Question utterances

Range Mean

SD

Range


5.81 3.20 0–8 (73%)

5.24 3.22 0–8 (66%)


5.38 3.23 0–8 (67%)

7.24 1.22 4–8 (89%)

Discussion The first experiment reported here utilizing the PEPS-C assessment sought to address the relationship between form-level prosodic skills (perceptual discrimination ability) and function-level communicative understanding, in children with autism and controls matched for age, intelligence, and receptive vocabulary. The main finding showed that both groups of children exhibited higher levels of performance in the auditory-perceptual form-level, as compared to the cognitive function-level prosody tasks, with the children with autism exhibiting significantly poorer ability overall. Further analyses of the data showed that children with autism performed significantly less well than controls at detecting auditory-perceptual changes of prosody at the phrasal level (Long-item discrimination). Interestingly, no between-group differences emerged in the Short-item discrimination task. This finding suggests that individuals with autism may have a specific difficulty with perceiving prosodic changes over longer speech stimuli, which would impact negatively upon their ability to understand the different functions of prosody. Consistent with the findings of Peppe´ et al. (2007), children with autism in the current study also exhibited impaired ability to understand affective nuances conveyed by intonation (Affect task). This finding is consistent with numerous studies reporting deficits in the understanding of vocally expressed affect in individuals with this disorder (e.g., Kujala et al. 2005; Rutherford et al. 2002). Such impairments have been linked to the core deficits in social interaction associated with autism. Results from the current study also showed that children with autism showed difficulties in understanding the grammatical function of phrasing (Chunking task). This is inconsistent with the findings of Paul et al. (2005b) and Peppe´ et al. (2007) reporting unimpaired ability to understand chunking in autism. However, unlike the findings of Paul et al. (2005b) and Peppe´ et al. (2007), no between-group differences emerged in the ability to perceive contrastive stress patterns (Focus). Consistent with the studies of Paul et al. (2005b) and Peppe´ et al. (2007), results from the Turn-end

123

1334


task indicated that both groups of children showed similar levels of understanding of questioning versus declarative intonation over single-word ‘‘conversational turns’’. However, differences in matching practices may have contributed to the differing patterns of impairment and ability observed across the studies, and camouflaged between-group differences in prosodic ability. The majority of participants with autism tested by Peppe´ et al. (2007) were younger (6–13 years) than those in the current study (7–16 years) and those in the study of Paul et al. (2005b) (14–21 years). As the controls of Peppe´ et al. (2007) comprised mental age-matched typically developing children, they may have achieved relatively low prosodic scores because they had not yet acquired the necessary skills. Since unlike in the current study and that of Paul et al. (2005b), the performance of children with autism was not compared to that of a chronological age-matched control group, their apparent level of prosodic ability may be somewhat misleading. One question that arises from the results, however, is whether the ability of individuals with autism to perceive auditory differences, and the related cognitive ability to understand the grammatical use of intonation, is limited to one-word utterances. The current results suggested that, relative to controls, children with autism showed unimpaired performance in both the Short-item discrimination and Turnend tasks. However, these children performed significantly less well than their controls at detecting prosodic changes over longer stimuli (Long-item discrimination). Further, as no corresponding sentence-level function task was included in the PEPS-C test battery, experiment 2 was designed to address this question.

Experiment 2: Assessing the Understanding of Grammatical Use of Intonation at the Sentence Level in Children with Autism Methods Participants For this experiment, 40 children were recruited from schools in Southwest England. They were again a subset of children taking part in studies by Ja¨rvinen-Pasley et al. (2007, 2008; Ja¨rvinen-Pasley and Heaton 2007). Eighty-

five percent of the children who completed the PEPS-C test also participated in this experiment. Twenty children with a formal diagnosis of AD or AS were recruited from a specialist educational establishment for children with autism, as described in experiment 1. Approximately 80% of the children had a diagnosis of AS, and the remaining 20% had a diagnosis of AD. Control children were matched on an individual basis to those with autism for CA, BPVS, and RSPM standardized scores, as described in experiment 1. Approximately 20% percent of the children in the autism group had the BPVS and RSPM standardized scores falling at least two standard deviations below the population mean (score B70). Consequently, the same proportion of children in the control group had MLD. The control children were recruited from a mainstream primary school and a secondary school for children with MLD. Again only children with MLD conditions of unspecified origin were included, and all diagnostic information was gathered from school files of documented medical diagnoses and clinical reports. The children met the same selection criteria as specified for experiment 1. Table 4 shows the demographic characteristics of the two groups of children. No significant between-group differences in CA (t (38) = .41, p = .69), the BPVS (t (38) = .18, p = .86), or the RSPM (t (38) = .96, p = .34) standardized scores are in evidence.

Materials The experiment employed an adapted paradigm of Patel et al. (1998). A version of this task was also used by Paul et al. (2005b) in their assessment of the understanding of the grammatical/linguistic function of intonation in autism. Here, 12 English sentences were recorded twice so that the sentence pairs were lexically identical but differed in prosody. For example, a sentence ‘‘He wants to leave now’’ was first read as a statement and then as a question. A native English speaking male uttered all the sentences. The sentences were treated acoustically as documented by Patel et al. (1998), with the sentence–question pairs modified so that they had the same syllable timing and amplitude patterns, resulting in sentence pairs in which fundamental frequency remained as the sole salient cue for discrimination. Two additional modifications were made, using the Praat speech editor (Boersma 2001): (1) Duration of the

Table 4 Characteristics of the two participant groups in study 2 (SD; range, in parentheses) M CA (years)

Sex

M VIQ (BPVS)

M NVIQ (RSPM)


12.03 (2.27; 7.33–16.35)

17M; 3F

88 (22.51; 55–135)

92 (17.51; 61–129)


11.87 (2.36; 7.50–16.08)

14M; 6F

87 (22.32; 55–124)

86 (17.16; 62–121)

123


final word was equalized; this word always carried the rise or fall in pitch. (2) As the words bearing rising intonation had higher amplitudes and so appeared louder, the amplitudes of the final words were equalized. A split-screen response slide showing a question mark and the word ‘‘Asking’’ spelled underneath on the left-hand-side, and a period/full stop and the word ‘‘Telling’’ spelled underneath on the right-hand-side, was constructed. The order of the 24 sentences was randomized, and the stimuli were presented using Cedrus SuperLab software (SuperLab, Cedrus, Inc., San Pedro, CA).

Procedure Each child was tested individually in a quiet room at their own school. The experimenter told the child that s/he was going to hear some sentences, some of which sounded as if the person speaking was asking the child a question, and some of which sounded as if the speaker was just making a statement, or saying what was happening. In order to ensure that s/he understood what questions and statements meant, the response slide was shown to the child. The utterance ‘‘Tim has a cat’’ was first played to the child as a statement. The experimenter told the child that the sentence was to be described as ‘‘telling’’. The sentence was then played as a question, and the experimenter told the child that the label ‘‘asking’’ applied to it. The experimenter then asked the child to comment on the differences between the two types of utterances, to ensure that s/he understood the difference between declarative and question utterances, and how they were indicated in the written format. The stimuli were then presented on the laptop computer, and the child was told to categorize each sentence. No feedback was given and the experimenter recorded the child’s responses.

Results The means, standard deviations, ranges, and percentages for correct categorization of declaring and questioning utterances for the children with autism and their controls is displayed in Table 5. The children’s scores for the classification of declarative and question utterances were compared against chance level performance (6) by applying one-sample t tests. This analysis showed that performance in both groups of children was significantly above chance with the declarative stimuli (Autism: t (19) = 3.54, p = .002; Control: t (19) = 2.33, p = .03). However, while the controls exhibited performance that was significantly above chance with the question utterances (t (19) = 3.58, p = .002), this was not the case for the children with autism (t (19) = -1.04, p = .31).

1335 Table 5 Means, standard deviations, and ranges for the correct identification of declarative and question utterances (% correct in parentheses) for the children with autism and their controls (maximum score per category = 12) Declarative utterances Mean

SD

Question utterances

Range Mean

SD

Range


8.60 3.28 2–12 (72%)

5.35 2.80 1–12 (45%)


7.85 3.12 2–12 (65%)

7.90 2.13 5–12 (66%)

Levene’s Test was carried out on the data to confirm that the distributions of scores did not depart from normality (declarative utterances: (F = .10, p = .76); and question utterances: (F = .02, p = .90)). A two by two repeated measures ANOVA was performed on the data, with utterance type (declarative/question) as the within-participants factor, and diagnosis (autism/control) as the between-participants factor. This analysis revealed a significant main effect of utterance type (F (1, 38) = 5.02, p \ .04), with more declarative utterances being correctly classified overall. The main effect of diagnosis (F (1, 38) = 2.61, n.s.) failed to reach significance, but a significant utterance type by diagnosis interaction emerged (F (1, 38) = 5.34, p \ .03). Follow-up post hoc Bonferroni alpha corrected t tests showed that controls were significantly better at discriminating question utterances than their counterparts with autism (t (38) = -3.25, p \ .005). No between-group differences emerged for the classification of declarative utterances (t (38) = .74, n.s.). Paired samples t tests revealed a significant condition effect within the autism group (t (19) = -2.80, p \ .02), with better performance occurring for the identification of declarative utterances. No significant condition effect emerged for the children in the control group (t (19) = .06, n.s.). The total numbers of declarative and question judgments made by the children within both groups were calculated, and converted into percentages. This analysis was not concerned with accuracy, but simply with data distribution. This showed that 64% of the judgments made by the autism group were declarative (and 36% question). By contrast, the controls performed in an unbiased fashion by making 50% declarative, and 50% question judgments. Thus, children with autism exhibited a declarative bias. The frequencies of children making certain numbers of declarative and question judgments of the stimuli were then calculated. Table 6 displays the clusters of children in percentages within each group making 0–6, 7–12, 13–18, and 19–24 judgments in total within each utterance type category. As each child made a total of 24 judgments, the

123

1336


Table 6 Clusters (in %) of children with autism making certain total numbers of declarative and question judgments of the stimuli, presented alongside control data (total number of judgments per child = 24) Declarative judgments Total number of judgments

Question judgments

0–6 7–12 13–18 Proportion of children (%)

19–24

0–6 7–12 13–18 Proportion of children (%)

19–24

Autism (n = 20)

5

25

35

35

45

30

20

5

Control (n = 20)

10

25

65

0

0

70

25

5

cluster categories that represent non-biased response patterns are shown in bold italics. An inspection of the proportion of children in the clusters representing non-biased response patterns (shown in bold italics in Table 6) confirms the earlier observations that, whereas approximately one-half of the children with autism exhibited a marked declarative bias, approximately 90% of the controls responded in an unbiased manner.

Discussion The second experiment reported here sought to examine the understanding of grammatical use of intonation at the sentence level in children with autism and in controls matched for age, intelligence, and receptive vocabulary. The findings showed that, similar to the study of Paul et al. (2005b), both groups exhibited similar levels of performance overall. However, the composite score including correct responses for both declarative and question utterances camouflaged important between-group differences in that the control children were significantly better than their counterparts with autism at categorizing question utterances. This pattern of results is similar to that from the Turn-end task of the PEPS-C reported earlier and also in Peppe´ et al. (2007). The controls responded in a balanced fashion and correctly identified a similar number of question and declarative utterances. By contrast, the children with autism showed a clear declarative bias. These findings suggest that children with autism showed specific difficulties in associating rising intonation with question utterances.

General Discussion The present studies were designed to address the question of whether auditory-perceptual form- and cognitive function-level prosodic ability would dissociate in individuals with autism, and whether the length/complexity of speech stimuli would have an effect upon these processing levels in such individuals. The main results showed that, while both groups exhibited significantly higher levels of

123

performance in the form than in the function level tasks of PEPS-C, children with autism showed significantly poorer prosodic abilities overall. This provided support to our hypothesis, and is consistent with existing literature. An inspection of children’s response patterns in the perceptual form-level tasks showed that, consistent with the results of Peppe´ et al. (2007), participants with autism showed a tendency to judge ‘‘same’’ auditory pairs as different (Short- and Long-item discrimination tasks), whereas the controls showed unbiased responding. At the cognitive function-level, no significant between-group differences emerged in the children’s ability to categorize declarative versus question utterances on the basis of intonation either at the single-word (Turn-end task) or the phrasal levels (experiment 2). However, significant differences in response patterns were in evidence. Unlike the controls, children with autism exhibited a clear bias towards judging question utterances as declarative, and this bias was stronger with longer stimuli (see also Peppe´ et al. 2007). The current findings may suggest a developmentally delayed pattern of prosodic processing in autism; however, the clear between-group differences in error patterns suggest some atypical patterns of development in autism, thereby lending support to our hypothesis. Although the current pattern of results may appear to suggest a dissociation between perceptual form- and cognitive function-level prosodic abilities in autism, i.e., unimpaired form- and function-level ability with short stimuli, and impaired form- but unimpaired function-level ability with longer stimuli, a close inspection of the response patterns of children with autism revealed processing abnormalities at both levels. As perceptual formlevel abilities are considered to be prerequisites for cognitive function-level abilities, these findings do not present the paradox that was suggested by the combined findings of Paul et al. (2005b) and Peppe´ et al. (2007). The current pattern of errors exhibited by children with autism highlights one major shortcoming in the paradigms commonly used in prosody tests: they often employ simple binarychoice designs that may well mask important prosodic difficulties in individuals with autism. Peppe´ et al. (2007) provided an analysis of children’s error patterns, while this was not the case in the study by Paul et al. (2005b).


The finding that children with autism fail to associate rising intonation with question utterances warrants some consideration. Although it may be suggested that deficits in perceiving auditory-perceptual cues in speech may underlie this difficulty, the opposite error pattern, i.e., a tendency to judge different items as the same, would simply suggest an inability to perceive auditory differences. Furthermore, investigations into the perception in the auditory modality in autism have reported stable memory for exact pitches (Bonnel et al. 2003; Heaton et al. 1998), enhanced processing of pitch stimuli (Heaton 2003; Mottron et al. 2000), and superior perceptual processing of intact speech stimuli (Ja¨rvinen-Pasley et al. 2007, 2008; Ja¨rvinen-Pasley and Heaton 2007). However, whereas such studies have utilized musical stimuli, pure tones, and intact speech, the stimuli of the form level tasks in PEPS-C comprised laryngographic sounds. Although these stimuli are not perceived as speech, neurological investigations have identified a pattern, in which the cortical processing of speech or vocal sounds, but not musical or environmental sounds, appears abnormal in individuals with autism relative to typical controls (e.g., Boddaert et al. 2004; Gervais et al. 2004; Lepisto¨ et al. 2005). The developmental work by Kuhl et al. (2005) has further shown that the neural mechanisms specialized for processing speech had failed to develop in pre-school children with autism who preferred a non-speech analogue to child-directed speech. Kujala et al. (2005) found evidence of impaired neural discrimination of prosody at the earliest stages of processing in participants with AS. Thus, although it is highly likely that abnormalities in the cortical auditory processing of speech stimuli contribute to the current pattern of results, identifying the exact nature of this abnormality is beyond the scope of the current paper. It may however be the case that a failure to learn to assign emotional and linguistic significance to prosodic cues is a down-stream effect of this lower-level speech-specific impairment. The results indicated that the children in the control group displayed a mild question bias in the Turn-end task of PEPS-C while performing in an unbiased fashion in experiment 2. It is important to note that this bias was the reverse of that shown by the children with autism (see also Peppe´ et al. 2007). Indeed, while a question bias implicates intact, albeit perhaps poorly tuned pragmatic function, the declarative bias shown by the children with autism suggests deficient pragmatic abilities. This is because the performance of the control children with the question stimuli suggests that they clearly understood that either information was being sought from them, or that there was questioning intonation, or both. However, children with autism may have attended to only one cue resulting in a failure to realize the importance of noticing the questioning intonation. One previous study has reported overly focused

1337

selective attention towards a limited number of available cues in complex auditory stimuli in children with autism (Schreibman et al. 1986). Whereas children with echolalia in this study tended to respond selectively to the intonation component of complex speech stimuli, and non-verbal children tended to selectively respond to the content component, typically developing controls showed either selective responding to the content, or no selective responding to either intonation or content components. Given that the function-level prosody stimuli in the current study included both intonation and content information, the pattern of findings from the autism group may reflect overly selective attention towards a limited number of available cues in speech, i.e., specifically towards the content information. This is supported by the notion that many able participants with autism remarked upon the absence of ‘‘Wh’’ question words in the test stimuli, despite it having been emphasized prior to the administration of the task that auditory-perceptual cues alone would discriminate between the two types of utterances. By contrast, the response pattern of the controls is likely to reflect the typical attention to multiple cues (i.e., to both intonation and content) in a speech stimulus. Furthermore, the tendency to judge question utterances as declarative may reflect deficits in the understanding of communicative intention, and thus in second-order theory of mind functions, on the part of children with autism, as is suggested by Relevance theory (Sperber and Wilson 1995; see also Happe´ 1993). The current findings also support the notion that ostensible social abilities in individuals with autism are usually acquired by mechanical learning (e.g., Mesibov 1986), rather than by intuition. This suggestion is consistent with the current results showing that the prosodic ability of children with autism deteriorates as the stimulus length increases, and thus becomes communicatively more loaded and complex. In conclusion, the results from the present experiments showing prosodic processing difficulties in individuals with autism are consistent with a large body of literature reporting a diffuse pattern of such abnormalities (e.g., McCann and Peppe´ 2003; Peppe´ et al. 2007; Rutherford et al. 2002). However, the current results extend the existing findings by identifying specific patterns of impairment. Specifically, abnormalities were evident both at the auditory-perceptual form and cognitive function levels. Although the current study cannot elucidate the direct relationship between the processing deficits at these two levels, a recent study by Kujala et al. (2005) found evidence for a cortical discrimination impairment of prosody in participants with AS. The current pattern of results showing a robust tendency in children with autism to judge question utterances as declarative may suggest atypically narrow attentional focus in speech processing in

123

1338

autism, possibly towards the semantic content. This processing tendency was apparent at both the single-word and phrasal levels. Indeed, while overly focused selective attention towards a single element in complex auditory stimuli may have limited influence upon musical and other auditory information, the effects of such a tendency upon linguistic information processing may well be profound (cf. Schreibman et al. 1986). Possible clinical interventions that can be derived from the current findings are that receptive prosodic skills could be targeted by exercises to enhance sensitivity to prosodic changes in individuals with autism, for example, by over-emphasis, to help individuals with autism to appreciate that acoustic variations in speech specifically function to modulate or enhance the meaning of what is being said. Furthermore, exercises should make it explicit that simultaneous attention towards perceptual features and linguistic content is required for successful communicative interpretation of utterances. Thus, intervention approaches focusing too narrowly upon individual prosodic cues may not be helpful. Future studies should focus upon: (1) identifying the atypical processing strategies utilized by individuals with autism when completing prosody tasks with differing information processing demands; (2) investigating the relationship between auditory-perceptual and functional prosodic abilities more formally in individuals with autism. The extent to which prosodic skills in experimental settings in autism generalize to naturalistic situations is also unknown. Acknowledgements This work was submitted in partial fulfillment of a Doctor of Philosophy degree at the University of London by the first author. Dr. Heaton’s work is supported by EU grant (12984) Stages in the Evolution and Development of Sign Use (SEDSU). We would like to express our warmest thanks to all the children who participated in these studies, and their parents and teachers, for kind co-operation.

Appendix Profiling Elements of Prosodic Systems in Children (PEPS-C) receptive task descriptions Examples of the actual stimuli can be found on http:// www.qmu.ac.uk/ssrc/prosodyinasd/ (1) Short-item discrimination (form): This task assesses the ability to perceive differences in intonation. The stimuli are pairs of same and different short (1–2 syllables long) laryngographic sounds, and the stimuli are composed of equal numbers of stimuli from the Affect and Turn-end function tasks. The participant’s task is to indicate their responses by clicking on the appropriate side of the computer screen (same versus different) with the mouse pointer. The same-symbol depicting two red circles appears

123


on the left side of the screen, and the different-symbol depicting a red circle and a green square appears on the right side of the screen. (2) Long-item discrimination (form): This task assesses the ability to perceive prosodic differences. The design of the task is identical to the Short-item discrimination task described above, but here the laryngographic utterances are longer (6–7 syllables long), and are composed of equal numbers of Chunking and Focus stimuli. (3) Affect (function): This task assesses the ability to understand affective or attitudinal meaning conveyed by intonation. The stimuli are single-word food items expressed as strong liking or disliking (reservation). Each auditory sample is followed by a screen depicting a happy face on the left, and a sad face on the right hand side of the screen. The participant’s task is to click on the matching face with the mouse pointer. (4) Chunking (function): This subtest assesses the ability to understand syntactically ambiguous sentences disambiguated by prosody. ‘‘Chunking’’ refers to phrasing or boundary-marking of a sentence into units or ‘‘chunks’’ for grammatical, semantic or pragmatic purposes. The stimuli employ minor phrase boundaries which can be used to differentiate items in a list. Some of the phrases use color combinations, e.g., ‘‘pink, and black and green socks’’ (signaling a boundary after the first item), or ‘‘pink and black, and green socks’’ (signaling a boundary after the second item). Other stimuli utterances consist of simple and compound food items, such as ‘‘fish, fingers, and fruit’’ versus ‘‘fish-fingers and fruit’’. The participant is required to click on a picture on the screen that depicts the utterance using the mouse pointer. (5) Focus (function): This test measures the ability to perceive contrastive stress or focus. Focus refers to the speaker’s use of emphasis to make a distinction between the most important word in the phrase and those of lesser importance. In this task, the participant is told that the person on the computer went to buy some socks, but forgot to buy one color, and it is the participant’s task to indicate the color (between two colors) that was forgotten. Examples of sentences are, ‘‘I wanted BLUE and black socks’’ versus ‘‘I wanted blue and BLACK socks’’. The participant responds by clicking on the appropriate color patch on the screen with the mouse pointer. (6) Turn-end (function): This task assesses the ability to understand questioning versus declarative intonation over single word ‘‘conversational turns’’. The stimuli consist of food words which are presented with opposing intonation either as offering (‘‘would you like some?’’) or reading (‘‘this is what I see in the book.’’). The participant’s task is to match auditory to visual stimuli by clicking either side of the computer screen with the mouse pointer. The response slide has a picture of a boy holding the food item in question on a


plate and a question mark on the left hand side of the screen, and a picture of a boy looking at a picture of the food item in question in a book on the right hand side.

References American Psychiatric Association (1994). Diagnostic and statistical manual for mental disorders (4th Ed.). Washington: APA. Asperger, H. (1944). Autistic psychopathy in childhood. (Translation and annotation by U. Frith of the original paper). In U. Frith (Ed.), Autism and Asperger Syndrome (1991, pp. 37–92). Cambridge, England: Cambridge University Press. Baltaxe, C. A. M., & Simmons, J. Q. (1985). Prosodic development in normal and autistic children. In E. Schopler & G. Mesibov (Eds.), Communication problems in autism (pp. 95–125). New York: Plenum Press. Baltaxe, C. A. M., & Guthrie, D. (1987). The use of primary sentence stress by normal, aphasic, and autistic children. Journal of Autism and Developmental Disorders, 17, 255–271. Boddaert, N., Chabane, N., Belin, P., Bourgeois, M., Royer, V., Barthe´le´my, C., Mouren-Simeoni, M.-C., Philippe, A., Brunelle, F., Samson, Y., & Zilbovicius, M. (2004). Perception of complex sounds in autism: Abnormal auditory cortical processing in children. American Journal of Psychiatry, 161, 2117–2120. Boersma, P. (2001). PRAAT, a systems for doing phonetics by computer. Glot International, 5 (9/19), 341–345. (http:// www.praat.org/). Bonnel, A., Mottron, L., Peretz, I., Trudel, M., Gallun, E., & Bonnel, A. M. (2003). Enhanced pitch sensitivity in individuals with autism: A signal detection analysis. Journal of Cognitive Neuroscience, 15, 226–235. Boucher, J. (2003). Language development in autism. International Journal of Pediatric Otorhinolaryngology, 67S1, 159–163. Crystal, D. (1969). Prosodic systems and intonation in English. The Hague: Mounton. Dunn, L. M., Whetton, C., & Pintilie, D. (1997). British picture vocabulary scale. Windsor: NFER-Nelson. Fay, W., & Schuler, A. L. (1980). Emerging language in autistic children. London: Edward Arnold. Fine, J., Bartolucci, G., Ginsberg, G., & Szatmari, P. (1991). The use of intonation to communicate in pervasive developmental disorders. Journal of Child Psychology and Psychiatry, 32, 771–782. Fosnot, S. M., & Jun, S. (1999). Prosodic characteristics in children with stuttering or autism during reading and imitation. Proceedings of the 14th International Congress of Phonetic Sciences (pp. 1925–1928). Gervais, H., Belin, P., Boddaert, N., Leboyer, M., Coez, A., Sfaello, I., Barthe´le´my, C., Brunelle, F., Samson, Y., & Zilbovicius, M. (2004). Abnormal cortical voice processing in autism. Nature Neuroscience, 7, 801–802. Gillberg, C., & Coleman, M. (2000). The biology of the autistic syndromes (3rd Ed.). London: Mac Keith Press. Happe´, F. G. E. (1993). Communicative competence and theory of mind in autism: A test of relevance theory. Cognition, 48, 101–119. Heaton, P. (2003). Pitch memory, labelling and disembedding in autism. Journal of Child Psychology and Psychiatry, 44, 1–9. Heaton, P. (2005). Interval and contour processing in autism. Journal of Autism and Developmental Disorders, 8, 1–7. Heaton, P., Hermelin, B., & Pring, L. (1998). Autism and pitch processing: A precursor for savant musical ability? Music Perception, 15, 291–305.

1339 Ja¨rvinen-Pasley, A., & Heaton, P. (2007). Evidence for reduced domain-specificity in auditory processing in autism. Developmental Science, 10(6), 786–793. Ja¨rvinen-Pasley, A., Pasley, J., & Heaton, P. (2007). Is the linguistic content of speech less salient than its perceptual features in autism? Journal of Autism and Developmental Disorders (OnlineEarly Articles). Doi: 10.1007/s10803-007-0386-0. Ja¨rvinen-Pasley, A., Wallace, G. L., Ramus, F., Happe´, F., & Heaton, P. (2008). Enhanced perceptual processing of speech in autism. Developmental Science, 11, 109–121. Kanner, L. (1943). Autistic disturbances of affective contact. Nervous Child, 2, 217–250. Kjelgaard, M., & Tager-Flusberg, H. (2001). An investigation of language impairment in autism: Implications for genetic subgroups. Language and Cognitive Processes, 16, 287–308. Klinger, L., Dawson, G., & Renner, P. (2002). Autistic disorder. In E. Mash & R. Barkley (Eds.), Child psychopathology (2nd Ed., pp. 409–454). New York: Guilford Press. Kuhl, P. K., Coffey-Corina, S., Padden, D., & Dawson, G. (2005). Links between social and linguistic processing of speech in preschool children with autism: Behavioral and electrophysiological measures. Developmental Science, 8, F9–F20. Kujala, T., Lepisto¨, T., Nieminen-von Wendt, T., Naä¨ta¨nen, P., & Naä¨ta¨nen, R. (2005). Neurophysiological evidence for cortical discrimination impairment of prosody in Asperger syndrome. Neuroscience Letters, 383, 260–265. Lehiste, I. (1970). Suprasegmentals. Cambridge: MIT Press. Lepisto¨, T., Kujala, T., Vanhala, R., Alku, P., Huotilainen, M., & Naä¨ta¨nen, R. (2005). The discrimination of and orienting to speech and non-speech sounds in children with autism. Brain Research, 1066, 147–157. Lieberman, P. (1960). Some acoustic correlates of word stress in American English. Journal of the Acoustical Society of America, 32, 451–454. Lord, C., & Paul, R. (1997). Language and communication in autism. In D. J. Cohen & F. R. Volkmar (Eds.), Handbook of autism and pervasive developmental disorders (2nd Ed., pp. 195–225). New York: John Wiley & Sons. Lord, C., Risi, S., Lambrecht, L., Cook, E. H. Jr., Leventhal, B. L., DiLavore, P. C., Pickles, A., & Rutter, M. (2000). The autism diagnostic observation schedule—generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30, 205–223. McCann, J., & Peppe´, S. (2003). Prosody in autism spectrum disorders: A critical review. International Journal of Language and Communication Disorders, 38, 325–350. Mesibov, G. B. (1986). A cognitive program for teaching social behaviours to verbal autistic adolescents and adults. In E. Schopler & G. B. Mesibov (Eds.), Social behaviour in autism (pp. 143–156). New York: Plenum Press. Mottron, L., Peretz, I., & Meńard, E. (2000). Local and global processing of music in high-functioning persons with autism: Beyond central coherence? Journal of Child Psychology and Psychiatry, 41, 1057–1065. Patel, A., Peretz, I., Tramo, M., & Labreque, R. (1998). Processing prosodic and musical patterns: A neuropsychological investigation. Brain and Language, 43, 4–11. Paul, R., Shriberg, L. D., McSweeny, J., Cicchetti, D., Klin, A., & Volkmar, F. (2005a). Brief report: Relations between prosodic performance and communication and socialization ratings in high functioning speakers with autism spectrum disorders. Journal of Autism and Developmental Disorders, 35, 861–869. Paul, R., Augustyn, A., Klin, A., & Volkmar, F. R. (2005b). Perception and production of prosody by speakers with autism

123

1340 spectrum disorder. Journal of Autism and Developmental Disorders, 35, 205–220. Peppe´, S., & McCann, J. (2003). Assessing intonation and prosody in children with atypical language development: The PEPS-C test and the revised version. Clinical Linguistics and Phonetics, 17, 345–354. Peppe´, S., McCann, J., Gibbon, J., O’Hare, A., & Rutherford, M. (2007). Receptive and expressive prosodic ability in children with high-functioning autism. Journal of Speech, Language, and Hearing Research, 50, 1015–1028. Pronovost, W., Wakstein, M., & Wakstein, D. (1966). A longitudinal study of speech behavior and language comprehension in fourteen children diagnosed as atypical or autistic. Exceptional Children, 33, 19–26. Rapin, I. (1997). Autism. New England Journal of Medicine, 337, 97– 104. Raven, J. C., Court, J. H., & Raven, J. (1992). Standard progressive matrices. OUP: Oxford Psychologists Press. Rutherford, M. D., Baron-Cohen, S., & Wheelwright, S. (2002). Reading the mind in the voice: A study with normal adults and adults with Asperger syndrome and high functioning autism. Journal of Autism and Developmental Disorders, 32, 189–194. Schreibman, L., Kohlenberg, B. S., & Britten, K. R. (1986). Differential responding to content and intonation components of a complex auditory stimulus by nonverbal and echolalic autistic children. Analysis and Intervention in Developmental Disabilities, 6, 109–125.

123

J Autism Dev Disord (2008) 38:1328–1340 Shriberg, L. D., Paul, R., McSweeny, J. L., Klin, A., Cohen, D. J., & Volkmar, F. R. (2001). Speech and prosody characteristics of adolescents and adults with high-functioning autism and Asperger syndrome. Journal of Speech, Language, and Hearing Research, 44, 1097–1115. Simmons, J. Q., & Baltaxe, C. (1975). Language patterns of adolescent autistics. Journal of Autism and Childhood Schizophrenia, 2, 333–350. Sparrow, S., Balla, D., & Cicchetti, D. (1984). Vineland adaptive behavior scales. Circle Pines: American Guidance Service. Spence, S. J., Sharifi, P., & Wiznitzer, M. (2004). Autism spectrum disorder: Screening, diagnosis, and medical evaluation. Seminars in Pediatric Neurology, 11, 186–195. Sperber, D., & Wilson, D. (1995). Relevance: Communication and cognition (2nd Ed.). Oxford: Blackwell. Stackhouse, J., & Wells, B. (1997). Children’s speech and literacy difficulties: a psycholinguistic framework. London: Whurr Publishers. Tager-Flusberg, H. (1981). On the nature of linguistic functioning in early infantile autism. Journal of Autism and Developmental Disorders, 11, 45–56. World Health Organization. (1993). The ICD-10 classification of mental and behavioural disorders. Geneva: World Health Organization.