that way confusion lies

Whichever way you look at it and whatever you think causes it, autism is a complex phenomenon. So it isn’t surprising that different researchers have different conceptual models of how autistic characteristics and their causes are related. Different conceptual models might confuse the public, but researchers are generally well aware of what other researchers think and why. But confusion can arise when conceptual models appear similar but are subtly different.

Here’s Mary Coleman in an interview with Adam Feinstein: “I came to realize that autism is like mental retardation – it’s a final common pathway in the brain affected by many different underlying causes, many different disease entities.” (p.147).

Feinstein says Coleman is ‘convinced that autism will eventually be found to consist of dozens – possibly hundreds – of different subtypes, each the result of a specific medical condition. She told me: “I’m not talking about co-morbidity. I’m talking about the basic disease which causes the autistic symptoms.”’ (p.194)

Now Isabelle Rapin, also interviewed by Feinstein:

Autism is a behavioural, not a medical diagnosis…I agree entirely with the concept of autism as a disorder of the developing brain. What I emphasize is that autism is defined behaviourally, in the same sense that dementia is behaviourally defined…but is, of course, medically caused. There… is no single cause, but many. I disagree with the idea that autism is a ‘disease’ – in the sense that it has one defined cause. I have always believed it reflects dysfunction of particular circuitry in the brain, whatever the biological cause of the dysfunction.” (p 194)

Thomas Kemper, who with Margaret Bauman carried out some pioneering studies of brain development in autism, says;

I still think there’s a unifying feature or group of features about children with autism: the way they use language, the way they interact socially, their play skills. There is a key group of features. I accept that it is a spectrum. But we should still be hunting for what is similar, the unifying characteristic. Even if Mary Coleman refers to 70 different kinds of autism, she’s still calling it autism. There may be different ways of getting there, but I’m not ready to give up on the hunt for some core, unifying feature of the brains of children with autism.” (p. 210)

fan-in/fan-out model

It seems to me that each of the above researchers subscribes to essentially the same ‘fan-in/fan-out’ model; that many different biological causes can produce impairments in a final common pathway that varies widely in its manifestations. Each has a different but important emphasis. Coleman emphasises the differences between the causal disease entities; Rapin the distinction between behavioural characteristics and their biological causes; and Kemper the similarities in behaviour between children.

The term ‘autism’ is a construct that refers to a triad of types of symptom. Whether that maps on to a single ‘thing’ in the real word remains to be seen, but it’s not safe to assume that it does. Jon Brock has addressed this point elegantly in his blog here.

What I find intriguing is that each of these researchers subscribes to the idea of a ‘final common pathway’, ‘dysfunction of a particular circuitry in the brain’, ‘some core, unifying feature of the brains of children with autism’. What, if anything, all people diagnosed with autism have in common is a moot point. The issue of what it is they actually have in common is dealt with explicitly by John Morton and Uta Frith in their work on causal modelling. I plan to examine this in more detail in a later post, but for the moment I’ll concentrate on the confusion in conceptual models.

That way confusion lies

It’s bad enough having to take into account subtle differences in the same conceptual model. It gets even more confusing when the same terminology has different meanings. I’ve noted some terms that are each used in different ways in the autism literature.

1. autism

can refer to either;

autistic characteristics – impairments in social interaction, communication and flexible behaviour


a medical condition assumed to cause autistic characteristics.

Using the same label for both is like diagnosing patients with a sore throat, rash and fever as having ‘sore-throat-rash-and-fever disorder’. It’s an accurate and valid label – all patients with those symptoms definitely have them – but using that term for both cause and effect wouldn’t make it clear whether you were talking about the symptoms or what causes them. It also implies that all patients have the same cause for their symptoms, when we know that there are many different causes for sore throats, rashes and fevers. In the case of autism the jury is still out on the links between cause and effect, so it’s not safe to assume, implicitly or explicitly, that there is only one underlying cause.

2. disease, illness, disorder, condition

are often used interchangeably.

As far as I can ascertain, the usual technical meanings of these terms are as follows;

disease – an infection that impairs normal function

illness – the patient feels unwell. Patients can have impaired function but not be ill, or be ill but not have impaired function.

disorder – a disruption in normal functioning. Normality is an interesting concept and one I plan to discuss in a later post.

condition – the patient’s state from a medical perspective. The patient doesn’t necessarily have a disease, illness or disorder.

The term ‘autistic spectrum condition’ (ASC) is sometimes used in place of ‘autistic spectrum disorder’ (ASD) because ‘condition’ doesn’t imply there is anything ‘abnormal’ about the person.

3. autism is a disorder with a heterogeneous aetiology

can mean either;

different things cause autistic characteristics in different people


different things cause autistic characteristics in an individual.

This phrase is often found in the introductory sections of research papers, but it’s not always clear what it means. One theory for the origins of autism is that autistic characteristics are triggered by environmental factors if a relevant genetic predisposition is present. It’s possible that several genes might be involved. So the causes of autism within an individual can be heterogeneous. Researchers using the phrase might be saying that different people probably have different causes for their autistic characteristics or might be assuming that every person diagnosed with autism has the same heterogeneous aetiology for their symptoms.

Enough of conceptual models for the moment. Bearing in mind the ambiguities in conceptual models and terminology, I next want to turn my attention to theories about the causes of autism.


Feinstein, A (2010). A History of Autism. Wiley Blackwell.


dual diagnosis: autism and other medical conditions

Many theories have been advanced for the causes of autism. Before looking at them in more detail, I want to examine a broader issue; the association between known medical disorders and autistic characteristics.

In their book The biology of the autistic syndromes, Christopher Gillberg and Mary Coleman devote an entire section (six chapters) to medical conditions associated with a formal diagnosis of autism. (My copy is of the 2nd edition published in 1992; the 1st edition came out in 1985, the 3rd in 2000 and the 4th, entitled The Autisms was published in 2012.) The authors don’t speculate on how the medical conditions they list might be related to autism, but it’s clear from this and later works that they are of the view that there are ‘autisms’, rather than a single disorder with different manifestations. From the preface of The Autisms;

Although Kanner, who first accurately described autism in 1943, likely assumed it was a single disease entity…over time it became clear that this was not so, as multiple studies… were published. Instead autism was found to be a specific set of symptoms found in a conglomeration of different diseases some of which overlapped with the intellectual disability/mental retardation syndromes.” (p.xi)

I think Gillberg and Coleman are suggesting that autistic characteristics (which by definition form a syndrome) arise as a consequence of a range of medical conditions, implying that the causes for autistic characteristics are probably different in different patients – hence the authors referring to ‘autisms’ in the plural. In 1974 Coleman had studied a group of 78 autistic children. In an interview with Adam Feinstein for A History of Autism, she says;

As a result of this [1974] study, I came to realize that autism is like mental retardation – it’s a final common pathway in the brain affected by many different underlying causes, many different disease entities.” (p.147)

I’ve summarised below the ‘disease entities’ listed by Gillberg and Coleman as associated with autism. (The list is from the 1992 edition; later research has expanded on it.)

Gillberg & Coleman's disease entities associated with autism

Gillberg and Coleman are by no means the only authors to identify an association between autistic characteristics and other medical disorders. Here are the ‘organic conditions’ reported by Wing and Gould (1979) in children who met Kanner’s criteria in their Camberwell study:

maternal rubella
infantile spasms
severe perinatal complications
multiple congenital abnormalities
unconfirmed viral illness under 1 year
gastroenteritis under 1 year
unconfirmed head injury under 1 year
operation for cleft palate under 1 year.

Michael Rutter and Eric Schopler (1988) comment:

The last issue concerns the question of etiological heterogeneity within the field of autism syndromes. We have already noted that there is undoubted heterogeneity. The very fact that the clinical picture of autism can arise from diseases as diverse as congenital rubella, tuberous sclerosis, encephalopathy, infantile spasms with hypsarrhythmia, cerebral lipoidosis and neurofibromatosis makes that clear.” [references omitted](p.28)

Patterns of association

Because autistic characteristics involve high-level processes, the relationships between autistic disorder and other medical disorders are likely to be complex and varied. However, the relationships between autistic characteristics, somatic symptoms and underlying causes will fall into one of four basic patterns, as shown below.

relationship between autistic characteristics, somatic symptoms and underlying causes

The section on other medical conditions in Gillberg and Coleman’s book was originally entitled “Disease entities that have a subgroup of patients with autistic symptoms” – in other words, not all children diagnosed with a particular disease entity were autistic, suggesting that for each disease entity:


the sub-group of children with autistic symptoms had a distinct second disorder (an autistic syndrome)


the sub-group of children with autistic symptoms had a predisposition to develop those symptoms and that they have been triggered by the disease entity.

Patterns of association and diagnosis

Depending on how strong the association is between autistic characteristics and a particular medical disorder, diagnosis is likely to proceed as follows;

A. If a medical disorder (e.g. disorder A) is frequently associated with autistic symptoms, it’s likely that the medical disorder will be assumed to account for the autistic symptoms and the child will get a single diagnosis of disorder A.

B. If a medical disorder (e.g. disorder B) is only occasionally associated with autistic symptoms, it’s likely that the medical disorder will be assumed not to account for the autistic symptoms and the child will get a dual diagnosis of disorder B and autistic disorder.

C. If the child has somatic symptoms that don’t fit the criteria for any known medical disorder, unless severe, those symptoms are likely to be overlooked and the child will get a diagnosis relating only to their behaviour – autistic disorder.

Each of these options is problematic if the cause of the child’s symptoms is to be identified, because in each group – A, B or C – the autistic characteristics might or might not be caused by whatever causes the somatic symptoms. The only ways to find out would be either to investigate all the symptoms shown by each individual, or to look at the differences between individuals who are autistic and those who aren’t, in respect of each of the other medical disorders.

What isn’t helpful is to assume that everybody’s autism has the same cause.

Coleman, M & Gillberg, C. (2012) The Autisms (4th edn). Oxford University Press.
Feinstein, A. (2010). A History of Autism. Wiley-Blackwell.
Gillberg, C. & Coleman, M. (1992) The Biology of the Autistic Syndromes (2nd edition). Mac Keith Press.
Rutter, M. & Schopler, E. (1988). Concepts and diagnostic issues in E. Schopler & G.B. Mesibov (eds.) Diagnosis and assessment in autism. Springer.
Wing, L. & Gould, J. (1979). Severe impairments of social interaction and associated abnormalities in children: Epidemiology and classification, Journal of Autism and Childhood Schizophrenia, 9, 11-29.