This set of summaries of recently published material begins with evidence concerning the very early stage of neonatal life to which the source of autism may be traced and the 'mechanism' that is implicated.

A further study highlights the significance of heightened levels of particular chemical messengers in the brain as a means of very early identification of children who are likely to develop the signs and symptoms of autism.

The issue of the link between the MMR vaccine and autism is revisited, with the conclusion that there is no evidence for any such link.

A particular aspect of autistic functioning - face perception is explored, with the hypothesis that autistics and non-autistics can be differentiated by CNS anatomical or functional features, with implications for understanding the range of levels or types of symptoms.

Finally, reference is made to auditory integration training, and to the results of a recent study which failed to demonstrate any benefits for children with autism.


The Early Origins of Autism

Rodier (2000) describes her initial interest in the etiology of autism as being stimulated by a research presentation concerning the effects of thalidomide. The authors (Stromland et al 1994) examined visual skills in a sample of adults who had been exposed to this drug during their antenatal growth. One incidental and striking finding was that around 5 per cent of this group had been diagnosed with autism which is markedly higher than the rate within the population as whole.

Rodier notes that existing theories have linked the source of autism to late gestation or early infant life, but this finding suggested that autism has its origins in the very early part of pregnancy when the central nervous system is just beginning to develop.

The genetic inheritance pattern is seen to be inconsistent in that siblings of autistic individuals have something like a 5 per cent chance of being diagnosed with the same disorder which is much greater than the risk in the general population but much less than the 50 percent chance that would be expected if the disorder was caused by a single dominant mutation, or the 25 per cent chance that would be linked to a single recessive mutation. The implication is that a number of genes are involved.

Meanwhile, relatives of individuals diagnosed with autism may not meet all the diagnostic criteria but still show some of the symptoms of autism.

Twin studies confirm the influence of heredity but suggest that environmental factors are also significant, as illustrated by the fact that there is not 100 per cent chance of an identical twin being diagnosed as autistic if the other twin has already been so-diagnosed. Instead the risk to that second twin is around 60 per cent.

Reference is made to a number of environmental risk factors such as exposure to German measles or to certain toxic substances. Further, individuals with some genetic conditions such as tuberous sclerosis also have a greater chance of developing autism. On the other hand, such factors cannot underlie more than a small number of cases, nor can one explain why the effects do not impact upon both members of a pair of twins rather than just one.

The exposure to thalidomide may prove another and significant risk factor, but the study of individuals who have shown some of the physical effects is significant because it is possible to identify with great accuracy the age when a defect is induced, given the existing knowledge about when the various organs begin to develop during the pregnancy. For example, the thumbs may be affected from day 22 after conception, the ears between days 20 to 33, or the arms and legs between days 25 and 35.

In the thalidomide study, most of the individuals with physical malformations and autism were found to have defects in the external part of their ears but no problems with the arms or legs, with the implication that the problems arose very early during pregnancy, between 20 and 24 days after conception.

In other words the critical period with such cases of autism is much earlier than previously thought. At this stage of gestation, there are relatively few neurons formed and most of these are linked to the cranial nerves which operate the muscles of the eyes, ears, or face.

These neurons are located in the brain stem and, because they develop at the same time as the external ears, the prediction would have it that the thalidomide victims with autism would also suffer from dysfunctions of the cranial nerves. This prediction was found to be correct in that all the subjects with autism had abnormalities of eye movement or facial expression or both.

The question was then posed whether the cases of autism in the thalidomide victims are similar to cases of unknown etiology. Rodier describes how the few studies that have examined non-behavioural features of autism concur in concluding that there are commonly minor physical and neurological anomalies which are similar to those noted in the thalidomide-linked cases of autism. For example, there may be minor defects of the external ears as well as dysfunctions of eye movement, or lack of facial expression

The implication drawn is that the nerve dysfunctions in individuals with autism are linked to an early brain injury that not only affects the cranial nerves but also has a secondary impact upon subsequent brain development. In other words, the initial brain injury somehow inhibits the full development, or the inter-connections, of other brain regions, including those involved in speech, and the combined effects result in the observable symptoms of autism.

It is recognised that the brain stem has not figured largely in studies of autism, and has been associated with fairly basic functions such as breathing, balance, or motor co-ordination. Many of the observable signs significant to an autistic diagnosis, such as impaired language, or failure to recognise social cues, have been linked to higher level regions of the brain such as the cerebral cortex.

Nevertheless, some of the common autistic symptoms, such as the lack of emotional expression, or sleep disturbance, or over sensitivity to touch and sound, may indeed originate in brain regions which do deal with more basic functions. There is also evidence that the most consistently observed abnormality in the brains of people with autism is a reduction in the number of neurons in the cerebellum, which is that part of the brain which is critically involved in the control of muscle movement.

Rodier goes on to explain how her own research team had the opportunity to study (in a post-mortem examination) the brain stem of a person with autism of unknown etiology. They noted the virtual absence of two particular structures, the facial nucleus which controls the muscles linked to facial expression, and the superior olive which is involved in the processing of auditory information. Further observation revealed that a considerable segment of the brain stem was absent.

This striking abnormality was similar to abnormalities observed in animal studies where there were also linked symptoms such as ear malformation and absence of eye movement control. The alteration of the brains of these animals had been brought about by the elimination of the working of a particular gene (Hoxa1) so that its role in early development could be studied. The question was posed whether the human version of this gene could be one of those implicated in autism. Existing evidence has it that this gene is active in the brain stem when the first neurons are forming and is responsible for producing a protein which modulates the activity of other genes. Further, this gene appears to be active only during the early development and this fits the pattern of autism which seems to be fairly stable after childhood, which would not be the case if the gene were active throughout life since this would involve additional problems or increased problems with age.

The gene in question is located on chromosome 7, and the research team have identified two variant alleles of the gene whose presence does seem significantly to differentiate people with autism from other family members or unrelated individuals without autism.

The problem is that this gene appears to be only one of many involved in autistic spectrum disorders, and the allele studied may be expressed in a range of ways ; and its presence does not guarantee that symptoms of autism will be observed.

The implication is for a need to continue seeking further variants in this gene along with variations in other genes which may predispose the carriers to autism. Currently there is interest in regions on chromosomes 7 and 15 in terms both of variants that will increase and decrease the risk of autism.

A practical problem is that one cannot accurately predict the probability of having a child with autism just by examining the parents because of the interaction of so many genes in this disorder. However, it may be possible to develop tests for the siblings of people with autism who may be anxious that their own children may inherit the disorder. One might look for well established genetic risk factors in both the family member with autism and the unaffected sibling, and if the sibling is found not to have a number of high risk alleles, then reassurance might be offered concerning the subsequent probability of an autistic child.

Incidentally, in the same article, there is reference to work undertaken by researchers at York University and the Hospital for Sick Children in Toronto which has identified a particular autism-related behaviour which may have diagnostic significance. It has been found that autistic children respond in an unusual but characteristic way to a task involving reactions to visual stimuli.

Using children aged between four and seven years, the experimenters encouraged subjects to look at three-screen panel. When a flashing light appeared on the middle screen, all the children looked at that screen. When the light appeared on the left or right screen, all the children shifted their gaze there.

However, in a second test the lights on the middle screen continued to flash while the lights appeared on the other screen, and the children without autism shifted their gaze to the new stimulus but the autistic individuals remained focused upon the first stimulus and did not switch their gaze to the new one.

The suggestion was that this was not a characteristic found among children with other forms of brain damage, and that this kind of behaviour which appears specific to children with autism supports the view that low-level brain functions are involved in the etiology of autism since this particular operation (i.e. the shift of attention from one stimulus to another) appears very early in infants, as young as three to four months of age, and since it is recognised already that this kind of mental activity is mediated by a primitive part of the brain, probably the cerebellum or the brain stem, or both.


Early Identification

On the theme of signs and symptoms which may be detectable very early in neonatal life, one notes the work of a research team at the National Institute of Neurological Disorders, in Maryland USA, (Nelson et al 2000) who have identified a set of chemical 'markers' for children at high risk of autism.

It has been found that children who are diagnosed with autism or mental retardation are more likely, during the early days after birth, to show raised levels of neuropeptides or neurotrophins compared either with children who are diagnosed with cerebral palsy, or with normally developing control groups of children.

This is held to be a significant finding since it may provide the opportunity for giving compensatory treatment to children virtually from birth. Currently, a diagnosis of autism is unlikely to be made until the children are around four or five years of age, at which time they may already be said to have missed the chance of important pre-school special provision. In particular, language development and social skill may be enhanced by early intervention.

Similar work has been continuing at the California Birth Defects Monitoring Programme (Grether et al.... reported by Lonsdale and McKie 2000) where there is an established and routine procedure for analysing blood samples taken from neonates in order to screen for phenylketonuria. Unusually, the samples are preserved providing the opportunity to reanalyse the blood of those children who were subsequently diagnosed with autism as well as a sample of children who have shown normal development.

Of the autistic sample, 96 percent had unusually high levels of four of these proteins (neuropeptides) which act as a chemical messengers in the brain, in contrast to the control group of children where not a single case of elevated levels of the proteins was observed.

One possible implication is for routine testing for autism among infants, although it is recognised that there is some risk of false-positive findings. Similarly, the prospect of treating children found to be at risk by bio-medical intervention (such as modifying of diets in order to reduce the production of these proteins) may be realistic; although caution is required in the light of the history, in the field of autism, of a number of treatments that have been claimed initially to be the answer only to be shown subsequently to be of little significance.


Autism and the MMR Vaccination

Although there is consistent evidence to demonstrate that there is no causal link between the MMR vaccination and autism, the issue has tended to persist and there is much anecdotal reporting of parental insistence that such a link does exist.

These beliefs on the part of some parents may be maintained by reports of the finding that 24 of 25 children with autistic enterocolitis screened independently were found to have the measles genetic marker when only 1 of 15 normally developing control children were found to have the virus.

These reports appeared recently in the national press, carrying the possible implication that, for the great majority of children, the MMR vaccination is quite safe ; but that there could be a small minority of children who have some genetic susceptibility and who may be harmed by the process.

The original paper of Wakefield et al (1998) referred to a sample of cases where autistic symptoms are accompanied by gastro-intestinal problems, the children concerned have a history of allergic conditions such as asthma or eczema, and there is a family history of auto-immune disease. There is a view that when the immune system is affected, it attacks the body, and areas of the small and large intestine are inflamed. If the gut wall is damaged, harmful substances may pass via the bloodstream to the brain. However, it remains unclear what might be the actual mechanism by which the MMR vaccine could produce this kind of outcome.

In any event, there has been stimulated a further series of publication, including, in America, a summary by The National Institute of Child Health and Human Development (2000) which emphasises that there is no scientific evidence for any causal link between any vaccine and autism. This paper described the anecdotal nature of the information in the work of Wakefield et al, and the lack of any supporting or replicated evidence, as well as the point that behavioural problems appeared before the onset of symptoms of inflammatory bowel disease.... that is, the effects preceded the alleged cause!

Although there is some evidence that gastrointestinal problems may co-exist with autism (for example, the study of Horvath et al [1999] suggested that unrecognised problems of this type may contribute to behavioural disorders), this is not the same as postulating some causal association; and Horvath et al themselves are quoted as concluding that no evidence exists for any causal gastrointestinal dysfunction - CNS impact - autism linkage.

Reference is also made to the 1999 study carried out under the auspices of the UK Committee on the Safety of Medicine which evaluated a very large sample of reports of patients allegedly developing autism, Crohn's Disease, and similar disorders after the MMR or MR vaccine. Problems of a lack of control groups, bias in case selection, and variation in the quality of data were acknowledged, but the conclusion clearly stated that there was no support for the suggested causal association, nor reason for concern about the safety of the vaccines.

Taylor et al (1999) identified all known cases of autistic spectrum disorder diagnosed since 1979 in a number of London districts and correlated the cases to vaccination histories. Their findings showed a gradual increase in ASD cases since 1979, but no marked 'jump' in cases after the introduction of the MMR vaccination in 1988.

Further, where diagnoses of ASD were made, the timing of the diagnosis was about the same for the children concerned irrespective of whether the vaccination was carried out at 18 months or later; and at age 2 years, the MMR vaccination coverage among ASD cases was virtually identical to the coverage of children in the same birth cohorts across the districts. The authors also found that the first diagnosis of ASD (or the first signs of behavioural regression) were not more likely to occur within time periods following the MMR vaccination than within other time periods.

All of this offered further evidence for a lack of association between ASD and the vaccination.

A further study, Gillberg and Heijbel (1998), compared autism rates in children in two communities in Sweden. They found no difference in the prevalence of autism between children born before the introduction of the MMR vaccination and those born after it was routinely available.

It is concluded that any link between MMR (or any other vaccine) and autism would be a very unusual event given the millions of children who have been vaccinated without any negative impact upon health and development.

Further, there is no scientific data to indicate that there is any benefit from separating the MMR vaccine into individual components, but there may be some risk to children if the rubella vaccine was delayed.

Advice from the American Centres for Disease Control and Prevention (National Immunization Programme 2000) suggests that, albeit allowing for caution in some infrequent cases where family history demonstrates some clear risk, even a younger sibling, or the offspring, of someone who experienced some adverse reaction to a vaccination usually can and should be vaccinated. The adverse reactions are commonly 'local' or limited to short term symptoms such as mild fevers or rashes.

The advice goes on to indicate that, even if measles virus were found consistently to be present in the intestines of children with autism, there would still be no reason to hold that measles causes autism. To put this the other way.... if measles were implicated in the etiology of autism, it would be necessary to demonstrate that no measles virus is present in the intestines of normally developing children who have the same history of measles infection and vaccination. In any event, it is re-stated that no evidence exists for any link between viral intestinal inflammation and the neurological and behavioural components of autism, nor could one explain the 'mechanics' of such a link.


Autism and Face Perception

Moving on from the etiology of autism to a consideration of further means of delineating the behavioural or cognitive elements, one notes the work of Schultz et al (2000) who found that individuals with autistic spectrum disorder show atypical brain activity during visual tests comparing the perception of faces with perceptions of patterns or objects.

These authors note that, in respect of visual stimuli, faces form a special class. Face perception appears to be a matter of analysing the whole, compared to object perception which seems to be a matter of the analysis of separate parts.

They go on to cite evidence that a particular part of the brain (the fusiform gyrus -FG) becomes active in respect of face perception, whereas other brain regions (such as the inferior temporal gyri - ITO) seem to be more involved in object perception.

Their own research used magnetic resonance imaging to examine brain activity in samples of individuals with autism and with Asperger syndrome, and in 2 control samples, when making perceptual judgments on pairs of faces or objects.

During face discrimination, the control groups consistently showed focal areas of activation in the right FG ; the 2 autistic spectrum disorder groups did not.

In addition, the control individuals demonstrated a strong association between ITG activation and object perception ; but in the autistic spectrum disorder samples, these areas of the brain were used more than was the case among the controls when faces were being examined In other words, it appears that the perception of faces among autistic subjects is like the perceptual processing of objects among individuals who do not have autistic spectrum disorder.

It was concluded that this finding fits well with the clinical experience of autism in that individuals so-affected have very little interest in people, and it was suggested that this disinterest is matched by the manner in which visual processing is organised in the brain i.e. people are indeed treated as objects.

In a commentary on these findings, Pierce and Courchesne (2000) referred to the desirability of further studies seeking to link social processing deficits and anatomical or functional deficits.

For example, one might ask whether the decreased activation of the FG during face processing, perhaps based upon differential attention or innate preference, might lead to some maldevelopment of face processing systems.

In support of this hypothesis, reference was made to the study of Gauthier et al (1999) who found increased FG activation after normally developing individuals were trained rapidly to identify a novel set of nonsense objects, with the implication that the FG is an area involved in the processing of very familiar objects. For most individuals, faces would be part of this class of familiar objects ; but, in autism, the decreased rate of FO activation may be attributable to the limited social interaction and to the lack of eye contact.

Similarly, if the FG is related to experience and expertise and not specifically linked to face processing, then one would anticipate that some people with autism who have a particular area of interest or expertise, such as Maps, would display FO activation when shown an item in their area of expertise. In other words, the face processing system would be found to be functioning in autism, but actually used for processing a particular class of objects rather than faces.

A further hypothesis would link trauma in the developing brain to some atypical organisation of functions in the brain. The neuroplasticity may compensate for the defect suffered but it may also produce arrangements that are unusual in terms of the location of functions.

The individuality of these compensatory mechanisms might explain the different degrees of cognitive or behavioural impairment among individuals with autism.

The authors conclude, therefore, that it may eventually be demonstrated that people with autism differ in respect of the functional abnormalities in processing faces or other socially relevant information. Some individuals may show reduced activation in the brain areas normally involved, and others may show heightened activation in unusual or scattered brain areas.

These individual differences may provide important clues to explain differences in symptoms, outcomes, and responsiveness to treatment.


Auditory Integration Training

Finally, in this present set of notes, attention is drawn to the work of Mudford et al (2000) who investigated claims for the efficacy of auditory integration training among children with autism.

(See the earlier MJC notes on this topic - April 1995)

These authors review the small number of studies into this intervention and in particular quote Rimland and Edelson (1994) who examined the impact of auditory integration training on children and adolescents with autism pair-matched to control for age, gender, ear sensitivity, and history of ear infections. Although there was a practical problem in that there was no information about the general level of functioning, it was concluded that those individuals receiving the treatment subsequently displayed significantly fewer problem behaviours compared to control individuals who were not treated.

A further study by the same authors a year later claimed a similar decrease in problem behaviour among autistic children, with the greater improvements found among children at the lower level of functioning according to tests of non-verbal ability.

Mudford et al surnmarised existing evidence as providing some support for the original and anecdotal reports about the efficacy of this treatment, although other studies, such as that of Gillberg et al (1997), found no benefits from auditory integration training on the behaviour of people with autism.

Their own research involving sixteen children with autism failed to replicate the results of Rimland and Edelson. Similarly, cognitive improvements in both verbal and non-verbal areas that have been cited for auditory integration training were not replicated.

In seeking to interpret these differences, it was hypothesised that the subjects may have differed in important ways. For example, there may have been a difference in the overall levels of cognitive or adaptive functioning. There was also the possibility that the results of the 1994 study may have been linked to the costs that accrued to the parents of those children included in the research. In other words, there may have been some positive response bias established in the parents and reflected in their ratings of the behaviours of their children.

It was also recognised that the outcome measures focused upon group performance, when advocates in favour of auditory integration training could claim some benefit for individual children from the training in this present study, even if the benefit was submerged in the intragroup variation. To counter this possibility, the authors followed up any parental reports of positive changes, and two children were identified as speaking more. However, standardised testing did not show any gains ; on the contrary, these children appeared to fall further behind their chronological peers over the course of the study.

The conclusion, therefore, held that no children could be identified as benefiting from this treatment. There was even the suggestion that the treatment could have some negative impact given the potentially damaging sound levels emitted during the sessions; and that the time and costs involved might only lead to false hopes and subsequent disenchantment.

M.J.Connor June 2000


Gauthier I., Tarr M., Anderson A., Skudlarski P., and Gore J. 1999 Activation of the middle flisiform face area increases with expertise in recognising familiar objects. Nat Neurosci 2 568-573

Gillberg C. and Heijbel H. 1998 MMR and autism. Autism 2 423-424

Gillberg C., Johansson M., Steffenburg S., and Berlin O. 1997 Auditory integration training in children with autism. Autism 1 97-100

Horvath K., Papadimitriou J., Raboztyn A., Drachenberg C., and Tildon J. 1999 Gastrointestinanal abnormalities in children with autistic disorders. Journal of Pediatrics 135(5) 559-563

Lonsdale K. and McKie R. 2000 Test can pinpoint autism at birth. The Observer: 21.5.2000

Mudford O., Cross B., et al 2000 Auditory integration training for children with autism : no behavioural benefit detected. American Journal on Mental Retardation 105(2) 118-129

National Institute of Child Health and Human Development 2000 Autism and vaccine research. www.nichd. .htm

Nelson K. et al 2000 Brain chemical levels and early identification of autism. Paper presented to the American Academy of Neurology : San Diego : May 2000

Pierce K. and Courchesne E. 2000 Exploring the neurofunctional organisation of face processing in autism. Archives of General psychiatry 57(4) 344-346

Rimland B. and Edelson S. 1994 The effects of auditory integration training on autism. American Journal of Speech-Language Pathology 5 16-24

Rodier P.2000 The early origins of autism. Scientific American 282(2) 38-45

Schultz R., Gauthier I., Kim A. et al 2000 Abnormal ventral temporal cortical activity during face discrimination among individuals with autism and Asperger syndrome. Archives of General Psychiatry 57(4) 331-340

Stromland K., Nordin V., Miller M., Akerstrom B., and Gillberg C. 1994 Autism in thalidomide embryopathy. Developmental Medicine and Child Neurology 36(4) 351-356

Taylor B., Miller E., Faniington C. et al 1999 Autism and the MMR vaccination: no epidemiological evidence for a causal association. Lancet 353 2026-2029

Wakefield A., Murch S., et al 1998 Ileal lymphoid nodular hyperplasia, non-specific colitis, and regressive developmental disorder in children. Lancet 351 637-641

This article is reproduced by kind permission of the author.

© Mike Connor 2000.

Back to NAS Surrey Branch Welcome Page