Autism : Clarifying Core Characteristics and Aiding Diagnosis
These notes concern the identification of behavioural, cognitive, and adaptive features which may differentiate autism from other pervasive developmental disorders, and which may assist towards greater accuracy of diagnosis.
The first summary describes the use of observed adaptive behaviour for this purpose.
Next, reference is made to the patterns of non-verbal cognitive functioning in young children with autism.
Sensory anomalies are described in the third summary.
The final section concerns problems with (motor) imitation among boys with autism.
M.J.Connor August 2007
The report by Tomanik et al (2007) begins by quoting the DSM-IV definition of autism as a developmental disorder characterized by qualitative impairments in social interaction, communication, and restricted or repetitive behaviours.
The individuals with this condition will typically demonstrate deficits in cognitive and adaptive capacities, with a majority likely to experience significant learning difficulties.
These authors continue by noting how symptoms of autism may be recognisable during the first year, but many children are not diagnosed until much later, perhaps 2 or 3 years after the onset of symptoms, others are not diagnosed at all, and a further group are diagnosed incorrectly.
The problems for diagnosis include the symptom overlap among a number of disorders, the way in which the overt behaviours linked to some other condition
(such as ADHD) may mask the early autistic characteristics, the heterogeneity of cognitive and behavioural performance within an autistic population, and the changes that occur in the expression of symptoms across ages and developmental levels.
Reliability of diagnosis involves the experienced use of standardized tools, and it is noted that the Autism Diagnostic Interview –Revised (ADI-R) and the Autism Diagnostic Observational Schedule (ADOS) have become commonly preferred forms of assessment.
Both are based on DSM-IV and ICD-10 criteria, but they differ in their format and in the nature of the information gained.
The ADI-R involves parental observation of the child in different contexts, while the ADOS requires a single observation period and would not in itself produce a valid diagnosis. Instead, it should be used in conjunction with a different measure … such as the ADI-R.
Meanwhile, evidence is cited (Lord et al 2000) that the ADOS effectively discriminates between autism and non-autistic disorders but not so well between autism and pervasive developmental disorder not otherwise specified (PDDNOS).
Further, these two measures may not concur when it comes to differentiating autism from specific language impairment.
Converging evidence (eg Noterdaeme et al 2000) indicates that the ADOS is accurate in identifying children with language impairments but is lacking in sensitivity in identifying autism and produces a large number of false positives.
The two measures are more consistent in identifying individuals with mental retardation, especially among young children.
The conclusion from such data highlights the need to be aware of, and to assess for, other variables that might lead to differential diagnoses by the ADI-R and the ADOS.
The suggestion from Tomanik et al is that, as well as information concerning the child’s early history and ratings/observations of current symptoms, an awareness of the child’s cognitive and adaptive functioning could be significant for the diagnostic process. In particular, it is noted that children with autism frequently demonstrate deficits in their adaptive functioning of a greater severity than would have been predicted from their intellectual level. Compared to typically-developing children, children with autism usually show poorer adaptive behaviour … even compared to age and IQ matched controls. The deficits may be observed in such areas as communication and socialisation; and the children with autism are perceived as functioning less well in their day to day environments than children of similar ability but without autism.
This being so, an assessment of adaptive skills is likely to assist in the process of forming a diagnosis especially where “traditional” measures (such as the ADI-R and the ADOS) produce inconsistent outcomes.
Accordingly, the study by Tomanik et al set out, firstly, to examine the agreement between the ADI-R and the ADOS in reaching an autistic diagnosis; and, secondly, to determine whether autistic diagnosis could be improved by the addition of information about adaptive behaviours.
The participants included 129 children and adolescents ranging in age from 7 to 18 years. Mean age was 12 and mean IQ was low average (88).
The autism group included 77 individuals who met the ADI-R criteria for autism. A “non-autistic” comparison group included 52 individuals of similar age and IQ but who did not meet the ADI-R criteria.
Around half of the participants met the ADOS criteria for autism, 15% met the ADOS criteria for PDDNOS, and the remainder did not meet the ADOS criteria for either disorder.
The outcomes indicated that the matching of ADI-R and ADOS diagnoses was adequate with 75% of the individuals being correctly classified. Of the children misclassified, most received a diagnosis of autism on the ADI-R but not on the ADOS. The ADOS was confirmed, therefore, as having a lower sensitivity.
The authors accepted that children who did get a false negative classification may have been mild or atypical cases of autism. It was also considered possible that children who might earlier have met the ADI-R criteria had shown improvement in the symptoms with age or in response to intervention so that they did not meet the ADOS criteria … and this view is consistent with existing research which has shown a greater agreement between the two assessment tools in the case of younger children.
Meanwhile, the differences were thought also to reflect the disparity in the type of information gained. For example, the ADOS involves a single sample of the child’s performance and classification errors might arise form atypical behaviour during the assessment occasion … and the authors’ review of clinical records indicated that many of the false positive classifications involved children described as shy, anxious, inattentive, or oppositional during the assessment.
The children with autism displayed significant impairments in their communication, socialisation, and daily living skills compared to non-autistic children … again consistent with existing data.
As hypothesised, the inclusion of measures of adaptive functioning improved diagnostic accuracy
The implication drawn was that, when the ADI-R and the ADOS give discrepant results (eg a child may be deemed autistic on the ADI-R but not autistic on the ADOS), the use of a measure of adaptive functioning would usefully be included in order to increase the accuracy of diagnosis.
Where a discrepancy does arise, the further information from measures of adaptive functioning (such as the Vineland Scales) would serve to determine whether or not to apply a positive diagnosis of autism … and the current findings suggest that children with the higher scores for adaptive functioning are more likely to be part of a non-autistic group.
Patterns of Non-Verbal Cognitive Functioning in Young children with ASD
The research by Kuschner et al (2007) is introduced by a description of the uneven pattern of cognitive abilities that is commonly observed among individuals with ASD.
The concern is that some overall measure of intellectual functioning may mask an idiosyncratic profile of strengths and weaknesses (which may offer some cues concerning how best to provide intervention).
For example, they note that there is often a higher score for non-verbal intelligence than for verbal intelligence but, within each of these elements, there may be a wide scattering of subtest scores. In other words, a higher score for non-verbal IQ does not necessarily suggest an overall superiority in visuo-spatial skills, nor does a depressed verbal IQ necessarily suggest generally poor language-related skills.
Their review of relevant research findings from the use of the Wechsler Scales with individuals with autism shows that an uneven pattern of verbal abilities is common, with adequate or good rote-memory skills (reflected in high scores in Digit Span) but weaknesses in Comprehension. In the performance scale, strengths are frequently observed in Block Design, but greater difficulties are seen in Picture Arrangement or Coding.
(One might tentatively suggest, at this point, that the significant issue is the manipulation or organisation or production of material… ie it is one thing simply to listen to a set of numbers and attempt to repeat them, and quite another to generate the necessary responses in Comprehension or to determine the underlying story in order to select the order of the items in Picture Arrangement. Meanwhile, the very lack of a holistic approach produces some advantage in copying the patterns in Block Design.)
Kuschner et al go on to cite the general implications behind existing research data that, compared to what might have been predicted from overall mental age, individuals with autism show relatively enhanced non-verbal skills, with a particular strength in focusing upon the details of a stimulus rather than upon the whole stimulus
(local aspects rather than gestalt).
This may be interpreted (see, for example, Frith and Happé 1994) as a weak or absent quest for “central coherence” which may be reflected both at a simple perceptual level and at a higher-order level in respect of using context, resulting in deficits in extracting the overall meaning or “gist” in language and social interactions.
However, findings have been inconsistent when it comes to supporting the hypothesis that this processing style is a core characteristic of autism … possibly because of variations in methodology across the various research tasks, differing demands upon the attention of the participants, or the use of participants covering a wide range of ages and levels of ability or with additional difficulties.
The implication drawn by Kuschner et al was the need for clarification whether this cognitive/perceptual style and strength is present in young children with autism.
If this focus on detail rather than upon the whole picture is present, it may help to explain the detail-focused style in other domains, such as the restricted pattern of interests. However, there may be support for the view that non-verbal skills are not universally enhanced in autism and that difficulties arise when the non-verbal task involves a conceptual skill rather than a perceptual skill.
Their own study, therefore, set out to explore this described pattern of relative strength in detail-focused perceptual skills, and relative weakness in conceptual skills involved in abstract reasoning, with a particular interest in determining whether such a pattern would be evident among young children.
The participants included 16 children with ASD, 16 children with non-autistic developmental delay, and 16 typically-developing children. The ages of children in the 2 target groups ranged from 3-5 to 5-5, and they were matched to each other and to the control group in chronological age, mental age and socioeconomic status.
The measures included the Social Communication Questionnaire; the Modified Checklist for Autism in Toddlers; and the Leiter International Performance Scale- Revised.
The results extended previous research by showing an uneven profile of nonverbal abilities in this sample of young children with autism. This sample did have relative strengths on perceptual tests such as figure-ground discrimination, and weaknesses on conceptual tests such as sequential ordering.
This kind of perceptual-conceptual split was in contrast to the more even profiles observed among the children in the two other groups.
This was taken as confirmation that visuo-spatial “disembedding” is a relative (and specific) strength in individuals with autism, and as providing support for theories concerned with weak central coherence and an absence of global preference when dealing with sensory input.
The pattern of results also supported theories concerned with a difficulty in autism with concept formation or the initiation of new ideas or principles, and those concerned with problems among individuals with autism in making meaningful connections between different stimuli or bits of information.
Any differences in the precise pattern of scores in this current study and in the study by Tsatsanis et al (2003) … current scores in Figure Ground and Form Completion were higher than those of Tsatsanis et al …. were thought possibly to reflect age differences in the two samples, or to indicate that relatively enhanced performance in nonverbal perceptual tasks is more evident among higher functioning children with ASD.
In any event, the authors believe that the observed pattern could represent a particular element of autism … ie that visuo-spatial disembedding and detail-focused processing are specific and relative strengths among children with autism; and abstraction and concept formation are specific and relative weaknesses.
However, it is accepted that individuals differ and that this pattern is not universally strong; thus highlighting the heterogeneity of children with ASD and the possibility of cognitive subtypes within this population.
Nevertheless, the profile described indicates at least a common primacy of these visuo-spatial strengths and conceptual weaknesses, with implications for not only identifying initial categories of need but also for the development of subsequent symptoms in autism … such as a failure to cope with social interaction at a broad level, or a restricted set of behaviours and interests.
In respect of intervention and education, the need is to look beyond relatively “gross” measures of ability and functioning and to assess for idiosyncratic strengths and weaknesses in order to determine the nature of therapeutic and curricular targets.
Sensory Style and Deficits
In their study of sensory abnormalities among individuals with autism, Leekam et al (2007) refer to self-reports as commonly describing unusual sensory experience such as insensitivity to pain or atypical reactions to sensory input across the modalities.
Meanwhile, they note how psychological explanations of autistic behaviours highlight sensory and perceptual problems, such as impairments in attention or in gaining the full picture of a stimulus “event” (and focusing instead upon component details), and the difficulty in integrating the stimulus information to which there is access.
However, these authors argue, there remains a limitation in the systematic evidence concerning the patterns of atypical sensory behaviours in autism even if it has become clear that sensory disturbances are common.
Many of the existing studies have taken the form of a comparison between the performance of children with autism against that of typically-developing children; and have confirmed the frequency of sensory anomalies among children with autism and Asperger Syndrome … but there are few studies which have examined their precise nature compared to what may be found among children belonging to other clinical groups and who are functioning at a similar cognitive level.
One such study (Freeman et al 1981) showed that low IQ children with autism were not readily differentiable from non-autistic children with a similar level of intellectual functioning; but differences were marked when high IQ children with autism were compared to typically-developing controls.
A recent study by Rogers et al (2003) involved young children as participants, matched for mental age, representing those with autism, Fragile X syndrome, developmental delay, and typical development.
Their results indicated that the children with autism and Fragile X had more sensory symptoms than either of the other two groups in respect of tactile sensitivity, taste/smell sensitivity, and auditory filtering (but not of auditory/visual sensitivity).
The current authors used the Diagnostic Interview for Social and Communication Disorders (DISCO) in their own studies. This measure is designed to be used with parents or carers in gathering information about sensory abnormalities in three domains … proximal (touch, taste, smell, kinaesthetic); auditory; and visual.
The first study compared children with autism, learning disability, language impairment, and typically-developing controls; and the second involved children and adults with autism in examining the patterns of sensory abnormalities within the wider autism population (including whether differences in their nature or frequency are associated with age or IQ levels).
In study 1, parents of the four different groups of children were interviewed. The groups included 16 low and 17 high functioning children with autism, 19 children with developmental disability, 15 with language impairment, and 15 typically- developing children. Their ages ranged from 34 to 140 months.
The results indicated that children with autism were more affected by sensory abnormalities than children in the two other clinical groups, and this applied to both the frequency and the pattern of symptoms. Comparison children may have had sensory abnormalities in one domain, but the children with autism showed anomalies across all sensory domains.
This supported the earlier-cited findings of significant group differences for total sensory symptomatology, and for the specific domains of smell, taste, and touch.
Further, children with autism could be differentiated from non-autistic children in respect of the seeking after self-stimulation (such as by smelling objects or people, or by smearing, and playing with, saliva).
The lack of significant effects for auditory symptoms was an unexpected outcome in that several of the control children had unusual responses to sound (with the possibility that sound stimuli are less sensitive in differentiating between typical and atypical responses). However, it was also speculated that the magnitude of auditory sensitivity and the impact upon day to day functioning may be more significant for the children with autism than for other children.
What was marked was the frequency of sensory symptoms among the higher functioning children with autism. These children had more sensory symptoms overall and were more affected by multiple sensory symptoms across the domains.
The low functioning group did not differ from the developmentally-delayed group.
The second study was concerned with whether the pattern of sensory behaviours/symptoms change with age and ability level.
The participants totalled 200 individuals, including children from 32 months to adults, who were assessed by means of the DISCO, and also by various cognitive and language tests in order to have a true picture of the range of abilities.
The results here indicated that more than 90% of the participants with autism had sensory abnormalities involving multiple modalities. In particular, the “proximal” domains of touch, smell, and taste that distinguished the autism and non-autism groups in study 1 did not change with age or IQ in study 2.
However, not all proximal domains remained the same across age and IQ. For example, other oral symptoms were found significantly less frequently in older participants, and kinaesthetic and mixed symptoms did show a relationship with age and ability.
In sum, the second study was said to offer two main findings.
Sensory abnormalities continue to impact upon the great majority of people with autism into adulthood; but that some sensory symptoms change with age and IQ.
This involves some reduction of overall symptomatology, but an increase in sensitivity to some forms of sensory stimulation such as being touched.
The overall conclusions by Leekam et al highlight the significance of sensory abnormalities for people with autism.
Children with autism differ from comparison children in overall sensory scores and in the specific domains of smell, taste, touch, and vision. It can further be stated that these abnormalities remain constant over time and across a range of ability levels.
In other words, these sensory issues may not be limited to children and adults with autism, but are more prevalent in this population. They are also persistent across age and IQ even if some differences can be observed for certain symptoms.
The implication is that one might be able to make use of the prevalence of sensory symptoms in respect of the assessment, identification, and management of children with autism.
The authors suggest particularly the significance of recognising the possible stress and distress among individuals with autism who have sensory sensitivities but who may not be able to explain their feelings or demonstrate their source. Such awareness could then serve to organise or modify the environment in order to reduce stress and improve observable behaviours … perhaps programmes of desensitisation could be employed … and parents and carers would benefit from an increased understanding of behaviours in their children which appear to be rejecting of their presence but which may reflect, for example, an oversensitivity to touch.
The findings of multi-modal sensory abnormalities may also help to explain the problems in “higher-order” perception, and the characteristic problems in the integration of sensory/perceptual input, which will have implications for the social and communicative anomalies characteristic of autism.
Autism and Motor Imitation Problems
Imitating the actions and movements of other people is a significant means by which to demonstrate social and non-social behaviours before children have access to adequate receptive and expressive language.
Further, as noted by Vanvuchelen et al (2007), DSM-IV includes problems with imitative and social play as one of the diagnostic criteria for autistic spectrum disorder.
Following a review of relevant studies, these authors argue that the available evidence suggests that motor imitation problems are characteristic and specific in autism. Imitation problems, albeit of a subtle nature, may persist beyond childhood … as highlighted by Hobson and Lee (1999) who found that the essential elements of some action may be reproduced but not the quality or the “style” be it, for example, harsh or gentle.
It has also been shown that problems are more marked with gestural imitation
(some action not involving the use of some object) than with procedural imitation
(the copying of an action that does involve the use of some object).
The current authors’ interest is in the underlying mechanisms involved in gestural imitation among individuals with autism where research is difficult to interpret since imitation is not some unitary phenomenon, but each imitative act will involve cognitive, representational, and visual-motor aspects.
They suggest that one way of advancing knowledge in this field is by the use of studies where task variables are manipulated, notably meaningful compared to non-meaningful single or sequential actions.
Meaningful (transitive) gestures are symbolic and linked to language as illustrated by an action such as pretending to comb one’s hair (which would demonstrate a degree of symbolisation and of knowledge of the function of the object/tool in question).
Less meaningful (intransitive) gestures do not involve an imaginary tool or object, and involve some simple communication, such as waving goodbye. Such gestures are non-symbolic nor linked to previous knowledge or experience and require simply an adequate visual perception and a capacity to reproduce the motor action.
Another approach involves the investigation of the relationship between problems in imitation and problems in motor coordination … noting that clumsiness has often been regarded as a feature of Asperger Syndrome.
Their own study explored the hypotheses that … 1. If children are poorer at imitating non-meaningful than meaningful gestures, this reflects a lack of cognitive representation of the meaning and purpose of the gesture, but the opposite situation reflects a lack of visual-motor organisation; and … 2. If children perform more poorly on a general motor test than on a gestural imitation task, this would suggest a general motor inability, but the opposite situation would be suggestive of a specific dyspraxic deficit.
The participants included 2 groups of boys with autism. The first included low functioning individuals (with IQs below 80) ranging in age from 5-1 to 6-11; and the second included higher functioning individuals with IQs above 80 ranging in age from 7-5 to 10-5.
Each group was matched to a control group comprising, respectively, boys from a school for children with learning difficulties whose IQs ranged from 55 to 78; and typically developing boys drawn from mainstream schools.
The assessment measures included the Gestural Imitation Test comprising meaningful items in respect both of transitive gestures (involving an imaginary object) and intransitive gestures for conveying some communication (such as a salute), together with non-meaningful gestures involving single hand postures and sequences of hand postures. The boys’ motor performance was examined with a standardised measure appropriate to their cognitive level … either the Peabody Developmental Motor Scales or the Movement Assessment Battery for Children.
The results indicated that all the boys with autism, whatever the level of functioning, had more problems than the non-autistic controls in imitating non-meaningful gestures compared with meaningful gestures.
This finding was taken as support for the view that mainly perceptual-motor elements are relevant for the gestural imitation problems in autism. There was no support for the view that it is largely a matter of cognitive weaknesses in respect of conceptualisation that underlie the imitation problems.
The authors suggested that adding meaning to the gestures is likely to improve the imitation among individuals with autism so that, for example, being able to imitate the action of combing one’s hair with an imaginary comb indicates an understanding of the meaning of the action and the symbolisation of the object use.
Further, previous experience of the real action is also likely to improve imitation.
On the other hand, the participants would have had no experience of the non-meaningful gestures and they could depend only upon accurate visual and perceptual-motor skills. Difficulties were thought to have arisen from inadequate perception or memory.
The authors went on to argue that the finding of significance in perceptual-motor elements for imitation is supported by the established relationship between imitation performance and general motor ability in children with autism.
In the first place, the motor impairment was more apparent than the imitation problems among the participants with autism compared to non-autistic controls. In other words, imitation difficulties are part of a more general motor problem rather than a specific dyspraxic-type problem. In the second place, the motor performance of the low functioning children with autism could explain most of the variance in the imitation of meaningful gestures. However, in the high functioning group, only the imitation scores on non-meaningful sequences of hand postures correlated significantly with scores on the Movement Assessment Battery for Children.
There remains some question why the present results showed a stronger relation between imitation performance and motor ability in younger children with a lower developmental level than in older children with a higher developmental level.
The suggestion was that motor ability was assessed with standardised measures suitable for the developmental levels of the children whereas the imitation test involved non-standardised items which did not take account of variation in developmental levels and may have been very easy for the higher functioning children.
On the other hand, boys with autism performed more poorly on imitation tasks than controls of a similar level, with the implication that motor imitation problems are a feature of autism (and with support for the view that imitation in autism is more impaired among younger children and those with lower levels of development).
The authors concluded by acknowledging some limitations in their study, such as the small sample sizes (eg 8 boys in the low functioning autism group, and 17 in the high functioning autism group), or the use of different IQ measures to match the different ages, or the ceiling effect operating on the high functioning children in the imitation test which will have reduced the variation among scores.
Nevertheless, it was recommended that diagnostic practices could be improved by the inclusion of assessments of motor and imitation skills … and by the availability of a standardised test of imitation skill for use with children of various ages and developmental levels (and it is their goal to produce such an instrument).
* * * * *
M.J.Connor August 2007
Freeman B., Ritvo C., Schroth P. et al 1981 Behavioural characteristics of high and low-IQ autistic children. American Journal of Psychiatry 138(1) 25-29
Frith U. and Happé F. 1994 Autism : beyond theory of mind. Cognition 50 (1-3) 115-132
Hobson R. and Lee P. 1999 Imitation and identification in autism. Journal of Child Psychology and Psychiatry 40(4) 649-659
Kuschner E., Bennetto L., and Yost K. 2007 Patterns of nonverbal cognitive functioning in young children with autism spectrum disorders. Journal of Autism and Developmental Disorders 37 795-807
Leekam S., Nieto C., Libby S., Wing L., and Gould G. 2007 Describing the sensory abnormalities of children and adults with autism. Journal of Autism and Developmental Disorders 37 894-910
Lord C., Risi S., Lambrecht L., et al 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 24 659-685
Noterdaeme M., Sitter S., Mildenberger K., and Amorosa H. 2000 Diagnostic assessment of communicative and interactive behaviours in children with autism and receptive language disorder. European Child and Adolescent Psychiatry 9 295-300
Rogers S., Hepburn S., and Wehner E. 2003 Parent reports of sensory symptoms in toddlers with autism and those with other developmental disorders. Journal of Autism and Developmental Disorders 33(6) 631-642
Tomanik S., Pearson D., Loveland K., Lane D., and Shaw J. 2007 Improving the reliability of autism diagnoses : examining the utility of adaptive behaviour. Journal of Autism and Developmental Disorders 37 921-928
Tsatsanis K., Dartnall N., Cicchetti D. et al 2003 Concurrent validity and classification accuracy of the Leiter and Leiter-R in low functioning children with autism. Journal of Autism and Developmental Disorders 33(1) 23-30
Vanvuchelen M., Roeyers H., and Weerdt W. 2007 Nature of motor imitation problems in school-aged boys with autism. Autism 11(3) 225-240
© Mike Connor 2007.
Back to NAS Surrey Branch Welcome Page