Viewpoint Expert opinions on trends and controversies in autism research.
Opinion / Viewpoint

The elusive essence of autism

by  /  28 January 2014

CONNECTIONS’s columnist Jon Brock explores the connections between the brain, social skills and autism.Read more columns »

Autism research is a new discipline. The condition itself was defined only 70 years ago, and it wasn’t until the late 1960s that there was a concentrated effort to address autism experimentally.

We’ve come a long way since then, but we’re still poking around at the edges of the scientific problem. I don’t think we’ve even really figured out what the questions are, let alone come up with good answers.

New sciences take time to mature. All the same, there may be lessons to be learned from older, more established fields of research.

One of the key challenges facing autism research is the heterogeneity within the condition. No two people with autism are alike, and apart from their diagnostic label, they may have little if anything in common.

This heterogeneity is widely acknowledged by researchers. And yet autism research still focuses on the label itself, considering whether on average, people with autism are different from people who don’t have that diagnostic label.

Underlying this approach is what philosophers would call an “essentialist” view of autism — somehow, beneath all the variation, people with autism share some essential property that sets them apart from those without a diagnosis. As SFARI senior scientist Alan Packer put it in a 2012 article: “Autism is a complex, heterogeneous disorder. But the core phenotype, which can be recognized to some degree in any individual on the autism spectrum, nonetheless suggests that there must be some common underpinnings.”

Attributed to the ancient Greek philosopher Plato, essentialism holds that all members of a category (whether chairs, rabbits or types of people) are mere variations on an ‘ideal’ category member. The lesson from history, however, is that essentialist thinking is a barrier to scientific progress.

The clearest example of this comes from evolutionary biology. Darwin’s breakthrough in understanding evolution required him to stop thinking of members of a species as being variations on a prototype. A species, he realized, is defined by its variation, not by its average.

Statistical limits:

The misleading nature of averages is beautifully illustrated in an essay by evolutionary biologist Stephen Jay Gould titled “The Median Isn’t the Message.” Gould discusses with remarkable humor his own diagnosis with a rare and terminal cancer — and the meaning (or otherwise) of life-expectancy statistics.

Having initially discovered that patients with his form of cancer had an average of eight months to live from the time of diagnosis, Gould took inspiration from his scientific and statistical training. He realized that the terrifying statistic obscured a huge amount of variability, that the variation was meaningful and that all indications were that he would live much longer than the average patient. He wrote:

We still carry the historical baggage of a Platonic heritage that seeks sharp essences and definite boundaries. (Thus we hope to find an unambiguous “beginning of life” or “definition of death,” although nature often comes to us as irreducible continua.) This Platonic heritage, with its emphasis on clear distinctions and separated immutable entities, leads us to view statistical measures of central tendency wrongly, indeed opposite to the appropriate interpretation in our actual world of variation, shadings, and continua. In short, we view means and medians as the hard “realities,” and the variation that permits their calculation as a set of transient and imperfect measurements of this hidden essence …

But all evolutionary biologists know that variation itself is nature’s only irreducible essence. Variation is the hard reality, not a set of imperfect measures for a central tendency. Means and medians are the abstractions.

The lesson we should take from evolutionary biology is that autism isn’t the average of people with an autism diagnosis. As researchers, we have to take heterogeneity seriously as the object of our investigation, rather than treating it as an excuse for inconsistent results or an inconvenience in our quest to identify the essence of autism.

As for Gould, he died in 2002, a full 20 years after his diagnosis and of an entirely different cancer. Among his many words of wisdom, the following quote is perhaps most appropriate as we continue trying to get to grips with this thing we call autism: “The most erroneous questions are those we think we know best — and therefore never scrutinize or question.”

Jon Brock is a research fellow at Macquarie University in Sydney, Australia. He also blogs regularly on his website, Cracking the Enigma. Read more Connections columns at »

14 responses to “The elusive essence of autism”

  1. Ramon Cererols says:

    I think the essence of autism should not be sought at the level of phenotype but at an underlying global mechanism which pervades the brain and so affects different brain functions. You can see the idea in this image ( and its development in this article (
    Ramon Cererols ( (

  2. mosaicofminds says:

    I recently read Ido Kedar’s book (Ido in Autismland) and was struck by his observation that he and Temple Grandin were almost exact opposites in abilities and experience, and yet were both had the same diagnostic label.

    Would a good solution be to investigate specific, pre-determined phenotypes of interest? For example, nonverbal people, or those with early self-taught reading, or those with specific sensory processing characteristics (reduced visual crowding? Perfect pitch? Hyperacusis? Tactile sensitivity?), or those with specific social characteristics (e.g., alexithymia, or prosopagnosia, or hyper-emotionality, or anxiety). Current research suggests some meaningful groupings, which when better defined could be compared to see if there is any meaningful similarity between different subgroups…

    How do you approach this problem?

    • Jon Brock says:

      I haven’t read Ido’s book but you’re not the first person I know who was struck by his comparison of himself to Temple Grandin. The question then is what do they really have in common other than the diagnostic label we give them – or the words we use to describe their “symptoms”.

      I think you’re absolutely right in your suggestions for looking at subgroups within autism. It’s something we try and do in our research (although we often look at continuous rather than categorical variability). I think it’s really important for theory development. At present, our theories are trying to explain findings from different studies of different groups of people with autism. But we first need to establish that there’s a link between the findings at the level of individuals.

      I probably need to emphasise that this isn’t a new idea. There are lots of other researchers thinking about and looking at variability and trying to tease apart subgroups within autism. It’s just that it’s often seen as being secondary to the core function of autism research which is to find out what differentiates autistic and non-autistic people.

      And even when people are thinking about subgroups, there seems to be an underlying assumption that these are subgroups *of autism* and that the subgroups still share this autistic “essence”.

  3. Joe says:

    A few things. First, Jones and Klin (2009) have argued that heterogeneity in the autism spectrum increases with development. I would agree with this, with an aside that co-morbid intellectual disability is a major source of variability in autism that is often present from the first two years of life. If true, when looking at causes/etiology, this would argue against paying a great deal of attention to individual differences – and would instead suggest a focus on core autism characteristics early in life. Second, the field of autism research is currently/actively addressing heterogeneity in (adolescents and adults) with autism by conducting a number of large-scale behavioral and cognitive phenotyping projects. Related to this, third, the history of research on other populations of similarly complex etiology suggests that behavioral and cognitive phenotyping may very well only complicate matters more. For example, major phenotyping (and associated genotyping) efforts for schizophrenia resulted in multiple models for sub-typing individuals on the schizophrenia spectrum, none of which was clearly better than the next and none of which were strongly or consistently related to intervention markers (e.g., genetics, neurobiology). Indeed, the current history of autism research is amazingly similar to the previous, largely failed, efforts in schizophrenia research. From behavioral phenotyping, to complex genetics, to the dorsal stream hypothesis, to neural connectivity and whole brain models of cause. What has helped individuals with schizophrenia the most are developments in behavioral and social support.

    • Jon Brock says:

      Hi Joe. Thanks for your comments!

      It’s an interesting point you make about heterogeneity increasing with age.

      The big question is whether this should be thought of as:

      (a) different children having the same thing (autism) but that thing being expressed differently as they get older.

      (b) different infants and toddlers having different conditions that we currently don’t know how to tease apart and consequently give the same label to (and in fact is questionable how well we can tease apart autism from other diagnoses in young children).

      I’m not sure which of those Klin and Jones were advocating.

      As for the lessons from schizophrenia, it does seem that many of the issues that we struggle with in autism research have been encountered already in research on other conditions, particularly schizophrenia. Certainly, we should look at other research fields and avoid their mistakes or blind alleys as much as possible.

  4. Sue Gerrard says:

    Great post. Trust Stephen Jay Gould to shed light on the topic. As soon as autism stopped being just a descriptive term and became reified, autism research lost its way.

    What I want to know is what’s causing my son’s specific developmental difficulties, what, if anything can be done about them and what support he can get.

    I don’t want to know, because it isn’t helpful, about the core phenotype of a label whose only essential property appears to be that it’s pretty arbitrary.

  5. Joe says:

    Sue, as parents that’s what we all want to know. The challenge for medical researchers is that there may be hundreds of root causes of autism. Further, doctors are only looking at the explicit symptoms to diagnose autism when the underlying cause is unknown, not understood in the least, and thus untreatable in nearly all cases. It’s a huge challenge.

    It is also clear that some very normal and some very intelligent, gifted individuals develop in such a way at young ages that they show clear symptoms autism but in the end are not really autistic. This is misdiagnosis but they likely meet all of the diagnostic criteria when tested as toddlers – the professionals are just following the guidelines. It is not their mistake. Professionals just cannot diagnose properly becauze the root cause is not understood.

    Parents like us end up lost and frustrated but I really appreciate all that is being done to attempt to understand.

    This is not identified by the medical community because there is no root cause understanding of autism.

  6. RAJensen says:

    One of my favorites is Stephan J. Gould. His theory of punctuated equilibrium states that the mutation rates in natural history are stable for millennia, punctuated by periods of rapid evolutionary change All males generate sperm mutations (deletions and duplications). The average de novo mutation rate is 1.20 × 10[-8] per nucleotide per generation with the paternal effect of an increase, doubling every 16.6 years. Environmental risk factors are associated with the production of sperm mutations and increasing levels of exposure produce increased frequency and increasing paternal age also increases the frequency of sperm mutations. Environmental pathogens that have been demonstrated to induce sperm mutations and deletions and presence in brain tissues include PCB and DDT congeners, benzene exposure, POP’s and PBDE’s Have we reached to the point where the general population’s rapidly expanding mutation rate has reached the point of no return? Thus, we may well be entering a period of punctuated equilibrium.

  7. Seth Bittker says:

    If the research mentioned is focus is on the biochemistry that characterizes autism and its evolution, I think the research is likely to be productive. If instead the focus is on brain studies, brain structures, and neuronal impulses I doubt it will lead to anything of significance. This is because it is the biochemical dysfunction in autism that leads to the neurological dysfunction and behavior that we generally call autism.

    Indeed there is a biochemical phenotype that is common among many with autism and it features a Th2 skew to the immune system, dysregulated monoamine neurotransmitters, endothelial damage, and oxidative stress. The key question that I think people should focus on is what about the modern world induces this biochemical dysfunction in so many kids.

    • mosaicofminds says:

      Seth, how do you know what you’re finding the biochemistry *of* if there is no reasonably well-defined population to study? Especially given that there are so many multiple comparisons in genetics research and the chance of finding false positives appears to be quite large, even when you know what you’re looking for? I suppose one could find the biochemistry of specific social or behavioral traits (but again, this is a bio-behavioral measure)? To start with genetics would seem to be putting the cart before the horse, unless I’m missing something about how genetics research works?

      • Seth Bittker says:

        Dear mosaicofminds,

        On the question you raised, autism is generally defined by its behavioral traits. Given this behavioral defintion if you take those that have autism and measure a number of metabolites in the blood and urine and compare those with controls, there are some general patterns that can be observed. This is documented in published research.

        One finding is that those with autism have a higher level of Th2 immune cells than controls. Another finding is that autism typically features elevations in various neurotransmitter monoamines. Another finding is that autism frequently features higher levels of neurotrophins (these are responsible for neuronal growth). We could go on.

        Some of these items mentioned above may be responsible for the neurological dysfunction that we typcially refer to as autism. What I am suggesting is that research that looks at what environmental inputs could result in this biochemical gestalt is likely to be much more productive in the long run than research that looks at macroscopic brain structures and nerve conduction as these would seem to be the downstream results of the dysnfunctional biochemistry.

  8. ASD Dad says:

    Treat the person that stands in front of you not the person you think they are and recognize them as holisitic.

  9. Franco says:

    Being myself aspie, I wholly agree with the article. And the concept ties in with the issue of competing connectivity theories (“intense world” deriving from hyperconnectivity and long-distance underconnectivity). From the aspies I have interchanged info with, it seems that both theories are valid, each describing a segment of the aspie population.

    As an aside, “intense world” seems to apply preferentially to women, while long-distance underconnectivity to men, based mainly on the ability/inability of dealing with a task and a stimulus, or two different stimuli, at the same time. E.g. I cannot read, much less write, when listening to music, and I cannot focus on any task if I do not completely disconnect from the passage of time.

    IMV, the underlying essence of ASD is an atypical brain connectivity. This atypicality can take many different forms, which accounts for the heterogeneity within the condition.

  10. Richard Caruana says:

    I have a 7 year old son with ASD. He is a very happy boy who has altered speech development and problems focussing his attention.
    I was in the car waiting for my daughter to finish dancing class and it began to pour down with rain. I began to form an analogy with the problem my son had.
    the rain (electrical brain interference) hit the windscreen (cognitive center in brain) and caused the image of traffic lights across the street (thought-form/s ) to lose focus. When the rain stopped ( the impeding condition stopped) the image refocused.
    rca08207 AT bigpond dot net dot au

Leave a Reply

Your email address will not be published. Required fields are marked *