These notes are a summary of a series of articles which appeared in a current journal (The Journal of Child Psychology and Psychiatry) given over, to a large extent, to the subject of autism.

Initially, there is a precis of a history of research findings and clinical practice, followed by reference to work dealing with "executive function" or a deficit in generativity of novel responses, and with hypotheses concerning "central coherence" and "hierarchisation".

A further paper concerns the prevalence of Gilles de la Tourette syndrome in children and young people with autism; and the final summary covers the issue whether accurate diagnosis is viable with children below 3 years of age.

M.J.Connor March 1999

Rutter begins his history of practice and policy by describing the work of Kanner who identified the key features of autism. However, it is recognised that Kanner would have been influenced by the general atmosphere and thinking operating at the time as a result of which autism came to be seen as an early form of schizophrenia, and the source was seen to be within inappropriate parenting. Rutter comments that these assumptions were not based upon firm evidence so that, for example, there was no testing of the theory that autism could somehow be caused by coldness of parenting styles; however, treatment focused upon compensating for the alleged damage resulting from poor management or non-attachment.

In the early period of working with autism, in the 1950s and 1960s, the emphasis was upon clarifying the symptoms by which to identify autism as a specific syndrome. The question did not concern how to recognise autistic characteristics but how to compare and contrast autism with other disorders.

There were found to be three areas of behaviour present in the majority of autistic children, namely, the failure to develop social relationships, language delay or disorder, and ritualistic or compulsive behaviours linked to repetitive and stereotyped play.

The range of long-term follow-up studies showed consistency in the general pattern of autistic behaviour over a long time scale, but the studies also showed how cases would vary dramatically in the degrees of social impairment so that, while most autistic individuals remained markedly handicapped, a minority were able to obtain regular employment. The most important predictor of outcome was the initial level of measured ability, and the overall level of language competence was also shown to be significant. Adequate or better social functioning in adult life was unlikely if the child had not developed useful speech by the age of 5 years.

An unexpected finding from these studies was the frequency with which children with autism were found to have developed epilepsy during the follow-up period, most often during adolescence. This prompted the view that autism did not represent some psychogenic disorder, but that it resulted from some organic brain dysfunction.

Meanwhile psychological assessments showed the distinctive nature of the autistic child's pattern of cognitive functioning. For example, particular weaknesses were observed in scores of verbal or sequencing skills, with the conclusion that the deficit was not in language per se but in particular elements of language such as abstraction, rapid comprehension, or making inferences.

The significance of these findings was for switching the emphasis in treatment from a psychotherapeutic approach to educational and behavioural approaches. For example, operant-learning principles were used in attempts to modify the behaviour of autistic children, and it is interesting to see that among the pioneers of such an approach was Lovaas (1967). However, there was some concern about the use of such methods because the changes in behaviour did not necessarily generalise from a clinical setting to the outside world, and there was some anxiety about the use of punitive techniques to modify social behaviour. There was also the concern that insufficient regard was taken of the significance of underlying organic disorder or damage; and it was further feared lest the behaviour of the children would become dependent upon extrinsic rewards.

A further change in treatment style arose from the recognition that it was necessary to plan intervention with a view to fostering development as opposed to "curing" some psychosis. Therefore, for example, efforts were directed towards how best to foster social development or how to encourage language and communication skills to enable the child to engage in meaningful interaction.

Rutter describes how during the 1970s and early 1980s there were two major trends. Firstly, it was recognised that research and clinical practice would usefully develop similar and systematic assessment approaches. Accordingly, there were developed a range of assessment schedules such as the Autism Diagnostic Interview and the Autism and Diagnostic Observation Schedule.

Secondly, there was continuing recognition of the range of types and levels of difficulty within autism, along with a need to know how to differentiate an autistic disorder from other forms of developmental disorders. In particular, reference is made to Rett syndrome given the frequency with which girls diagnosed with this syndrome have also been seen as having autism, although subsequent analysis was able to demonstrate significant differences. There were also found to be frequent symptoms of an autistic kind shown by children with Fragile X syndrome.

Meanwhile attention was directed towards cases showing marked behavioural deterioration with the possible implications that such symptoms were associated with a form of acquired brain disease, although subsequent medical investigations proved negative. Profound mental retardation and autism were likely to be confounded as a result of shared characteristics; and there was some further interest in whether or not one could differentiate a particular form of autism observed in children of normal intelligence, as first described by Asperger in the 1940s.

The issue of language disorder was also further investigated with the conclusion that autism was more than a matter of some developmental disorder of receptive language, and that it involved a wider cognitive deficit that includes a language problem but extends far beyond that.

At this time, a number of studies investigated possible medical causes of autism and a range of reports emerged to suggest that autism can be associated with some specific medical condition. Also at this time it was suggested that autistic syndromes linked to some medical condition were atypical, and differentiable from those cases of autism which did not have a link with some other medical condition. The final strand in respect of aetiology at this time was the growing evidence for a genetic factors in autism.

Psychological studies reaffirmed the importance of low measured ability as a predictor of outcome with some research devoted to investigating whether low scores might be the result of social withdrawal; and the outcome of various studies was that many autistic children have a general cognitive deficit that is not simply a result of social withdrawal, but it was equally evident that low IQ could not in itself account for autism.

The research, therefore, began to focus on ways in which some cognitive deficit might underlie problems in a social interaction and a number of theories arose such as that which linked autism to perceptual inconstancy, or sensorimotor delay, or over selectivity in attention. However, it was not clear whether these observed characteristics were a function of low IQ or of autism; and when such theories were examined within appropriately controlled studies it was found that the overall level of cognitive deficit was more significant than the diagnosis of autism.

In the 1980s there was first observed that approach which suggested that autistic children lack the ability to show or feel empathy and that the underlying key to autism is this socioemotional deficit. In a related approach, attention was focused upon the issue of mind reading .... the inability to recognise the opinions or feelings or points of view of other people with implications for a failure to develop social relationships or communication.

Treatment approaches increasingly showed an emphasis upon behavioural and educational strategies and evidence was obtained to show that autistic children progressed best in settings that provided a very structured organisation with a focus on the teaching of specific skills. However there was still concern about the lack of a generalisation of what was learned in the school setting to other settings.

It was during this period that the TEACCH programme became established (Schopler and Reichler 1971); and a further innovation involved home-based treatment programmes (Howlin et al 1973) which emphasised a clear analysis of the child's behaviour along with assisting parents to be consistent in their handling style and to teach the child particular social and communication skills.

What appeared to be less helpful was the onset of a number of claims about the particular effectiveness of treatment. For example, a number of authors argued that many cases of autism could be cured by ensuring a secure parental attachment or by re-establishing appropriate forms of parenting. As a result of such claims expectations were roused but none of the claims were supported by a evidence.

Similarly exaggerated claims were made for the effect of fenfluramine which can reduce serotonin levels in the blood. This was held to be significant in the light of the observations that a significant number of autistic children do have unusually high serotonin levels. Again no hard evidence was found to support the claims; just as no hard evidence was found to justify the assertion (Rimland 1987) that high doses of the vitamin B6 can bring about significant improvements for some autistic children.

During the later 1980s into the 1990s, the studies of aetiology were marked by evidence for significant genetic influence. For example, twin studies in Scandinavia and Britain showed a great difference in the concordance rate for monozygotic compared to Dizygotic pairs.

On the other hand, it is still the case that the great majority of the siblings of autistic children do not have autism, and this apparent paradox can be explained in terms of the observation that several genes are involved. Thus, while many family members will have some of the genes that lead to susceptibility to autism, they will not have all of them, and will not become autistic.

Findings are also consistent in respect of autistic cases that go beyond the "traditional" syndrome such that social and communication deficits are found in individuals of normal intelligence. Such individuals may represent what has come to be called "the broader phenotype of autism". It has been noted that there is an increased rate of affected disorders among the relatives of autistic individuals with the possible implication that depression or anxiety might play a part of this broader picture.

Findings of this kind tended to counter the previously held view that one either had autism or one did not. Instead it became increasingly evident that autism operates as a continuum

During this period there appeared a large number of reports which suggested some link between autism and some particular medical abnormality. The problem has been that most of the reports concerned individual cases so that very few of the associations have been replicated. Meanwhile autism has been associated with a whole host of metabolic deficits and a range of infections during the pre-natal period or early post-natal period.. Problems exist in interpreting these kinds of reports although it is possible that there may be some common ground in the aetiology of certain of these conditions and autism even if the nature of that common ground is not yet clear.

Nevertheless Rutter suggests that despite all the possibly misleading inferences one should not ignore the biological findings and there are some associations that do appear to be valid. For example, there seems consistent evidence that tuberous sclerosis is linked with autism, and that the link with fragile X may not be as great as earlier thought, but is still significant.

One of the major requirements in order to establish appropriate intervention is early diagnosis, but a number of studies have shown that, even if some autistic characteristics may be observed in the first year, reliable diagnoses cannot be made until the child is around two years of age and the diagnosis is made more difficult when mental retardation is severe.

Evidence for the nature of autism has increasingly focused upon the impaired ability to understand mental states although a number of questions have been raised. For example, if the cognitive deficit underlying autism is as specific as this, one would ask why there is such an association between autism and general mental retardation. One would also ask why autism can be detected very early in the child's life if theory of mind is the critical issue and if a theory of mind does not usually develop until the child is around four years of age. There is also the question of the relationship between theory of mind and language, and it is further noted that many children with autism can pass theory of mind tests.

Meanwhile it has been found that many autistic individuals have some weakness in central coherence in that they have problems in planning or organising; on the other hand, such weaknesses have been found in a range of disorders and it has yet to be found whether there is a particular type of coherence or planning deficit that is particularly linked to autism.

In respect of treatment the trend has continued in respect of developmentally oriented behavioural and educational strategies although this decade has been marked by some controversies.

Firstly there is the claim from Lovaas (1987) for the effectiveness of his intensive home based behavioural programme with the implication that recovery can be achieved. However, his study has been subject to much criticism, and there is a consensus that an independent controlled study needs to be undertaken before one can form a valid view of the Lovaas approach.

Secondly, a whole series or specific therapies have been put forward as likely to bring about marked benefits. These include auditory integration therapy and facilitated communication, but again there is a lack of empirical evidence supporting the claims.

Thirdly, there has been an emphasis upon early intervention, and while it is accepted that early intervention is generally logical and one would seek to avoid secondary problems, there is still some question over the real effectiveness of particularly early treatments.

During the late 1990s, it is suggested that the most exciting development relates to genetic findings It is anticipated that evidence will emerge for the location of those several genes which combine to give an increased vulnerability to autism.

Findings also exist in respect of similarities between autism and the behavioural patterns of children who have experienced profound deprivation.

Meanwhile long-term studies have demonstrated the high frequency of social deficits among boys found to have severe difficulties in receptive language during their early childhood, and this has re-raised the questions about the precise linkage between autism and semantic pragmatic disorder..

In other words, evidence accumulates for a huge variability in clinical patterns with the implication for the need to identify meaningful differences between the various types of conditions which share the overall autistic diagnosis.

No great changes are reported in therapeutic styles although Rutter refers to the increased attention to strategies that may help adults again and maintain employment and become socially independent. Further there are said to be a number of drugs that can produce some benefit although no drug has been found to produce major behavioural improvements in most individuals with autism.

Rutter's conclusion emphasises the need to ensure that research has some impact upon actual services. The important thing is to ensure that research findings are replicated and tested so that hard evidence can become available by which to help children and by which to avoid "excesses" and any tendency to go down unhelpful avenues.

A number of reasons are given for the significance of the difficulty in producing new ideas or behaviours. For example, problems of this kind would be consistent with the lack of spontaneity often observed in autism, or the limitations in speech and difficulty among autistic children to participate in pretend play.

The prediction is made that, if an autistic person cannot generate new responses, there will be poor performance on those tasks that provide limited cues to guide behaviour, and that this disadvantage would be most marked in those tasks that require the individual to go beyond existing knowledge and to produce truly new responses.

Existing studies have commonly involved fluency tasks where, for example, subjects are asked to produce as many words as they can beginning with a given letter. There are also tests involving the fluency of ideas where the subject may be asked to name as many as possible uses of some particular object. A third type of task is that of design fluency in which subjects are asked to produce as many different designs as they can in a given time. These two latter types of task do not depend upon stored knowledge but require genuinely original responses.

Turner recognises that failure on tests of this kind may not represent simply some difficulty in producing new ideas, but problems might result also from a failure to inhibit a concentration upon what is known to be the established function of the object in question, and there may be some difficulty in shifting attention away from that familiar feature of the stimulus.

In any event, there have been few studies which have looked at the fluency of responding in people with autism, although what little information does exist suggests a difficulty within this sample. For example, the study of Boucher (1988) involved asking subjects to think of as many different words as they could in a minute and it was found that children with autism performed very poorly, indicating some weakness in developing an appropriate approach or strategy to cope with a new problem. Turner's own study set out to compare the fluency performance of children and adults with autism in tests which tapped the ability to retrieve stored knowledge and to generate new and imaginative responses.

The results indicated an overall pattern supportive of the view that people with autism are limited in their ability to generate multiple and novel responses. The word fluency tasks produced clear evidence of difficulty in the autistic sample although there was no difference between autistic and controls in the proportion of inappropriate or repeated responses. Further investigation indicated that the autistic subjects tended to produce responses which were clustered in groups of words which were semantically or phonetically similar, while many of the control subjects adopted 'organised" strategies by which to improve performance.

In the tests of idea fluency, the autistic subjects showed marked deficits in the ability to generate new responses and ideas.

On the other hand in the design fluency task there was no significant difference between the numbers of responses produced by the autistic subjects and those produced by the control subjects. However the autistic children produced high rates of inappropriate or repeated responses which supported the suggestion of some primary impairment in the ability to regulate response behaviour.

The explanation offered for the observed discrepancy between these two sets of results was that the design fluency test placed greater demands upon the system responsible for monitoring and inhibiting behaviour than the test of ideational fluency.

A paradoxical finding was that the high rate of repetitive responding in the autistic sample could be accompanied by an understanding that such responding was inappropriate. The situation was described as similar to that found among patients with frontal lobe injuries who were able to recognise their performance m sorting tasks as incorrect but continue with that inappropriate response.

Turner concludes by restating the view that an impaired capacity to generate new ideas and action plans will lead to be limited spontaneous behaviour and to behaviour that is characterised by repetition Evidence is cited that individuals with autism do lack imagination and have restricted behavioural repertoires.

It is also argued that the difficulty in producing new behaviour will disrupt the completion of routine tasks where there is some unusual circumstances. The example is given of the difficulty that an autistic individual might have even in a familiar social situation in recognising that different people around him/her may have different beliefs and behave in different ways, so that (s)he will have problems in interpreting their behaviour and responding accordingly.

Some of the problems in theory of mind may be explicable in these terms in that mentalising involves not just the ability to recognise that other people have ideas or opinions that differ from one's own, but also the capacity to work out (by combining knowledge and experience in novel ways) what will be the consequence of having those ideas or opinions.

The final comment from Turner is that a broadening of the study of executive functioning to include the presence or absence of the capacity to generate spontaneous and new responses may give further insight into the particular symptom profile and behaviours characteristic of autism.

The signs might include abnormal responses to sensory stimulation, or to savant abilities, but particular reference is made to perceptual and visuo-spatial functioning. For example, evidence is cited that autistic children perform better than would be predicted on the block design test suggesting a perceptual orientation towards component parts rather than the whole of the design, and some lack of central coherence.

Central coherence theory holds that individuals with autism have deficits in integrating information into higher level of holistic information. For example, the autistic person may be more able than controls matched for mental age at finding hidden figures within a larger drawing, and this kind of superior performance is interpreted as indicating a focus upon component parts and, therefore, some absence of global interference.

Reference is also made to the hypothesis which relates to a deficit in working within some hierarchy. For example a savant autistic individual may have particular skills in producing a drawing of a building such that the overall proportions and the details are accurate even if the individual begins drawing with local details in contrast to the non-autistic artist who will almost invariably begin with the overall aspect of the building and come to the fine and local details later.

These two theories were tested by means of presenting two tasks of visual perception to a group of high functioning individuals with autism and a group of children and young people matched for age, IQ, gender, and laterality.

The first task explored the issue of hierarchy and subjects were asked to look at large figures in the shape of particular letters formed from similar or different smaller letters. For example, the stimulus of a large letter H could be made either of smaller Hs or of different letters. The speed and accuracy of the subjects were recorded in responding to instructions such as "Find me an A made of H's", or "Show me when the H's make an A".

The results observed were not in accord either with the central coherence or with the hierarchy deficit theories. The subjects with autism and the control subjects displayed similar interference effects of the incongruent information, although it was noted that the autistic subjects displayed longer reaction times.

In the second task, subjects were presented with geometrical figures by means of which to assess holistic effects upon segmentation. The subjects were invited to examine a geometric figure paired with either the same figure or a slightly different figure which was split into two or more component parts, and they were asked to indicate by pressing one of two buttons whether the two figures were similar or not similar.

Again, the results did not match expectations based upon central coherence or hierarchy deficit hypotheses. The autistic subjects did not display relative superiority on this hidden figure task.

The authors sought to explain these discrepant results on the basis of differences in the nature of the tasks they used compared to tasks used in earlier studies. For example, typical earlier studies have examined the performance of autistic subjects on tasks such as identifying visual illusions or the completion of jigsaw puzzles, and it is argued that these types of tasks involve a lengthy timescale. However the tasks used in this study involved very short reaction times of a second or less rather than several minutes.

It was, therefore, suggested that the differences in findings reflect differences in processing which are not necessarily linked to a global/local processing dichotomy. The tasks here do not reflect or tap skills of higher order perceptual information processing.

The authors conclude that the autistic individual may not be differentiable from controls in respect of early and simple forms of global processing, but the deficits by which autistic individuals may be characterised are associated with more complex global processing associated with later or higher order information.

Autism is also a neuro-developmental disorder which can be contrasted with Tourette's syndrome but the two conditions do share a number of behavioural features. For example, echolalia is common in both conditions as are certain types of obsessive or compulsive behaviours. Unusual motor behaviours may be observed in both conditions, The authors quote a number of studies which have described cases in which autism and Tourette's syndrome have occurred together. In particular, reference is made to the cases in which motor or vocal tics in adults with autism have developed following the withdrawal of certain drug treatments. It is possible that the development of tics is secondary to the ending of medication but it is also possible that the medication will have masked or delayed the appearance of tics.

A case is quoted in which Tourette's syndrome developed in a child with autism with no history of medication and this has given rise to the consideration of a possible common neuro-chemical abnormality. Speculation also focuses upon a possible common aetiology including chromosome or genetic abnormalities. A rough estimate of the prevalence of autism is three or four per ten thousand and the prevalence for Tourette's syndrome is around two per ten thousand. Therefore, if the two conditions are truly independent, the rate of their coming together by chance would be around five per one hundred million. However the reported prevalence rates in existing studies far exceed this chance expectation especially when the autistic sample is widened to include Asperger syndrome or the range of autistic spectrum disorders.

This presents study sought to gain an indication of the rate of Tourette's syndrome in a special school population of children and young people with autism, thus to avoid sampling bias which would result if one relied entirely upon data from out-patient clinics.

A total of thirty seven children were studied, with an age range from ten to eighteen.

The rate of Tourette's syndrome in this special school population was observed to be over eight per cent which far exceeds what would be expected from chance.

It was also recognised that only one specialist school was involved and that the sample of children and young people with autism was small, so that one would hesitate to make firm conclusions, Nevertheless there appeared to be justification for a large-scale study of comorbidity.

In any event, it was concluded that the link between the two conditions may not reflect any genetic factor. Instead, the relatively high rate of Tourette's syndrome in children with autism could reflect some other common factor such as a neuro-chemical abnormality; but it was concluded that the higher rate of Tourette's syndrome does not reflect a chance occurrence nor an artefact of medication.

In the light of the frequency of cases in which very young children are put forward for possible formal assessment, it is timely to see the work of Stone et al (1999) which discusses whether autism can be diagnosed accurately in children below three years of age. They also note the heightened interest in early diagnosis stemming from the reported success of interventions which can be begun at a very early age. It is also suggested that an early diagnosis is important because access to some specialised intervention for autism may be limited to those cases where there has been a specific diagnosis set down.

It is recognised that many factors can complicate diagnosis but the authors quote a number of studies which have demonstrated the stability of early diagnosis over time, despite the fact that there has been some question over the accuracy of diagnoses made in clinical settings only, and despite the fact that some of the criteria set down by the American Psychiatric Association are not appropriate for very young children because of the developmental delays, notably in speech.

This present study was designed to investigate the accuracy of diagnosis among very young children by means of examining the extent of agreement among different specialists and by monitoring the stability of the diagnosis over a twelve month period. A total of sixty five children were recruited, from three diagnostic clinics, who met the criteria concerning age, referral for social and or language development, and access to independent diagnoses by two specialists.

An examination of the means of diagnosis suggested that at least three of the DSM4 criteria were not appropriate. These were.... Failure to develop peer relationships, impaired conversational skill, and stereotyped language. The characteristics that were found to be most significant were those concerning impairments in social and communication skills or those involving restricted and repetitive activities.

The general conclusion noted that there was consistency of diagnosis when it is a matter of discriminating an autistic spectrum disorder from no disorder, but less consistency when it comes to a more specific discrimination between autism compared to some alternative but pervasive developmental disorder.

Of the children who were diagnosed as PDD at age two, ninety two per cent were still on the autistic spectrum a year later although only around half retained that particular diagnosis. Of the children diagnosed with autism at age two, ninety six per cent remained on the autistic spectrum a year later but seventy two per cent retained the autistic diagnosis. The authors offered a number of possible explanations for changes in diagnosis, including the suggestion that actual changes in symptoms may occur as result of development or of intervention. It is also recognised that the changes may simply reflect an error of judgement at either point in time, or inaccurate parental reporting, or atypical behaviour in the child.

However, it was noted that all the children who left the autistic spectrum continued to show developmental difficulties particularly in a language which implies in itself the desirability of further clarification of the diagnostic overlap between autistic spectrum disorder and language disorder.

The results gained in the study support the view that social impairments are central to autism and autistic spectrum disorder. In particular, impairments in the use of non-verbal behaviours and a lack of social and emotional reciprocity were considered to be particularly significant diagnostically. On the other hand, repetitive or compulsive behaviours showed more variation from one child to another, although the most commonly endorsed behaviour by the diagnosticians was a preoccupation with stereotyped and restricted patterns of interest.

In summary, diagnoses of autism can be achieved accurately in children below three years of age; and the most prevalent diagnostic characteristics are deficits in the use of non-verbal social and communicative behaviours, a lack of social or emotional reciprocity, and delayed acquisition of spoken language. The authors add one note of caution to the effect that early diagnosis is not an end in itself. The whole point of early diagnosis is linked to ongoing attempts to identify the most appropriate and effective early intervention strategy.

M.J.Connor March 1999

I

References

Baron-Cohen S., Mortimore C., Moriarty J., Izaguirre J., and Robertson M. 1999 The prevalence of Gilles de la Tourette syndrome in children and adolescents with autism. Journal of Child Psychology and Psychiatry 40(2) 213-218

Boucher J. 1988 Word fluency in high functioning autistic children. Journal of Autism and developmental Disorders 18 637-645

Howlin P., Marchant R., Rutter M., Berger M., Hersov L., and Yule W. 1973 A home based approach to the treatment of autistic children. Journal of Autism and Childhood Schizophrenia 3 308-336

Lovaas 0. 1967 Behaviour therapy approach to the treatment of childhood schizophrenia. In Hill J. (Ed.) Minnesota Symposia on Child Psychology Vol.1. University of Minneapolis Press

Mottron L., Burack J., Stauder J., and Robaey P. 1999 Perceptual processing among high functioning persons with autism. Journal of Child Psychology and Psychiatry 40(2) 203-211

Rimland B. 1987 Megavitamin B6 and magnesium in the treatment of autistic children and adults. In Schopler E. and Mesibov G. (Eds.) Neurobiological Issues in Autism. New York : Plenum Press

Rutter M. 1999 The Emanuel Miller Memorial Lecture: Autism - two way interplay between research and clinical work. Journal of Child Psychology and Psychiatry 40(2) 169-188

Schopler E. and Reichler R. 1971 Parents as therapists in the treatment of psychotic children. Journal of Autism and Childhood Schizophrenia 1 87-102

Stone W., Lee E., et al 1999 Can autism be diagnosed accurately in children under 3 years? Journal of Child Psychology and Psychiatry 40(2) 219-226

Turner M. 1999 Generating novel ideas fluency performance in high functioning and learning disabled individuals with autism. Journal of Child Psychology and Psychiatry. 40(2)189-201