Hacker News Reader

> She simply argues that the symptoms are caused by numerous unknown diverse causes in the brain, and the concept of a single disorder (ASD), or even a spectrum of related disorders (ASDs, the DSM-IV PDDs, and more recent ASD sub-typing proposals), isn't helpful in understanding those causes.

I think this is a dangerous misunderstanding of the reason for the grouping of autism-spectrum disorders under a common umbrella. The danger lies in the (all-too-common) tendency of people to think that the rejection of autism's categories means "autism doesn't exist" or something similar. The purpose of such categories is not to help understand the causes, but to help cement the reality that autism is a group of very serious disorders with similar traits.

As an analogy, many cancers that are grouped under a single name -- such as breast cancer -- are actually caused by very different genetic alterations. To use these differences as a justification to reject the concept of breast cancer would be ludicrous.

> I think this is a dangerous misunderstanding of the reason for the grouping of autism-spectrum disorders under a common umbrella

Who are you accusing of a "dangerous misunderstanding" here – her, or me? If you haven't read her work, I don't know how you can fairly accuse her of misunderstanding anything; likewise, I think having read at least some of her work is a prerequisite to be able to fairly accuse me of having misunderstood her

If your accusation is that she doesn't understand "the reason for the grouping" may I point out that she played a central role in the drafting of the PDDs section in the DSM-III-R [1]. As a former member of the DSM working group on autism, Waterhouse has a very good knowledge of the reasons and history behind the development of the "ASD" diagnostic label. Waterhouse may or may not be right, but her argument cannot be dismissed anywhere near as easily as you seem to think it can.

[1] https://link.springer.com/article/10.1007/BF01046326

> Who are you accusing of a "dangerous misunderstanding" here – her, or me?

I wasn't accusing anyone specifically, merely commenting on the idea you shared. However, since you asked, I would say that the misunderstanding is on both your parts -- yours, for interpreting her arguments to mean "ASD doesn't exist" (as far as I can tell, she makes no such claim); and hers, for arguing that the formal definitions and categorizations of ASD need to be biologically valid (which they can't be, as we don't understand enough about the biology for that to be possible).

It's common for many fields of research to start with "stamp-collecting", where people just identify and group diseases and conditions and birds based on shared characteristics; and later, as the field becomes more mature, to develop more meaningful categories. I think this same concept applies to ASD.

> yours, for interpreting her arguments to mean "ASD doesn't exist" (as far as I can tell, she makes no such claim);

To quote page 431 of her book Rethinking Autism (my emphasis):

> Taken together, these three claims and associated lines of evidence argue against the existence of autism as a single disorder, spectrum, or set of autism subgroups. If autism symptoms are not one disorder, and are not many disorders, what are they? The most parsimonious and least speculative view is that autism symptoms must be symptoms.

Her book is quite explicit in its claim that autism spectrum disorder does not exist. Her conclusion (p. 433) is given under the heading "AUTISM SYMPTOMS WITHOUT A DISORDER". The whole point of her book is to argue that the disorder "Autism Spectrum Disorder" does not exist, but the symptoms do. You might say she rejects the "Spectrum" and the "Disorder" in "ASD", but keeps the "Autism" as a reference to a set of symptoms – however, her "autism" is narrower than the DSM-5's "autism", since it only includes social impairment, and considers repetitive behaviours/restricted interests and sensory issues, to be distinct symptom sets coequal with attention-deficit/hyperactive-impulsive symptoms, intellectual disability, functional language impairment, seizures, see p.434-435

To quote the blurb on the back cover "Rethinking Autism... draws the potentially shocking conclusion that 'Autism' does not exist as a single disorder. The conglomeration of symptoms exists, but like fever, those symptoms aren't a disease in themselves..." (But, the book actually goes further than the blurb says it does, because it not only rejects the idea that autism exists as a single disorder, it also rejects the idea that autism exists as multiple disorders or a spectrum of disorders, as my quote above demonstrates.)

> and hers, for arguing that the formal definitions and categorizations of ASD need to be biologically valid

Biologically valid categories would be a lot more useful than biologically invalid categories, and one of her major points is that the research community's fixation on the later is getting in the way of actually discovering the former ([1], [2]). She argues that, in place of the current widespread practice of doing research with ASD-defined samples (a confirmed ASD diagnosis as a study inclusion criteria), samples should be based on all individuals showing neurodevelopmental symptoms (mixing together ASD, ADHD, epilepsy, intellectual disability, etc, into a single sample) and then trying to study the variation in the sample in order to derive biologically valid categories to replace the existing biologically invalid ones such as ASD. [3] is an example of a study along the lines of what she recommends (indeed, they've obviously read [1] since they reference it)

[1] https://link.springer.com/article/10.1007/s40489-016-0085-x

[2] https://pubmed.ncbi.nlm.nih.gov/28714261/

[3] https://www.nature.com/articles/s41398-019-0631-2

> The danger lies in the (all-too-common) tendency of people to think that the rejection of autism's categories means "autism doesn't exist" or something similar.

Serious question, do you actually have any evidence for that postulation?

Even with the analogy you have listed doesn't show evidence that people would consider non-existence or reject care from individuals suffering what we today characterize as autism.

To be fair, the article did actually suggest an etiological difference between "high-functioning" and "the most severely disabled" cases.

From TFA:

"Investigating the most severely disabled autistic individuals you almost always discover damaging genetic mutations or early life trauma, such as foetal alcohol syndrome. These are clear cases of biological dysfunction. On the other hand, the less severely disabled individuals (who would once have been called Asperger’s or “high-functioning”) show none of those biological signs of dysfunction, instead showing evidence we expect from functional adaptations: the associated genes are common and complex, brain differences are subtle, the characteristics appear early in life when they are guaranteed to affect reproduction, and the prevalence is high enough that at least one person per Dunbar-sized hunter-gatherer social group of one hundred and fifty would show the same traits – in which case, every one of our ancestors would have known an autistic person. These biological signs are those we expect to see from adaptations, not dysfunction. The question we are led to ask is what autism’s function could have been."

Indeed, this sounds a lot like cancer, which has some similar effects (even on different organs), and can be induced by a variety of methods (genetic, carcinogen, etc), but is fundamentally many different conditions often requiring different treatments.

The desire to categorize and use reductive reasoning is powerful and seductive, but not always effective until you really understand the system. With human biology and brain we're a long way from that.

Waterhouse argues that the boundary between ASD and non-ASD conditions (such as ADHD, intellectual disability, epilepsy, etc) is invalid and needs to be abandoned. I'm not aware that anyone is making a really comparable argument with respect to cancer. Cancers derived from different organ tissues are usually clearly distinguishable from each other, something which is far less true when talking about different psychiatric/psychological/neurodevelopmental diagnoses.

I've not read her book (though it's going on my list), but the second citation in your first comment suggests that it's just on the research side that they should abandon current classifications, but that the DSM definitions are still needed on the clinical side.

Does she propose a better classification system, something like what she'd replace it with in the next DSM if it were up to her? Because if we don't yet know how to split current ASD diagnoses into more specific classifications, isn't it better to have this rough grouping than to say "this person has an unknown disorder"?

> but the second citation in your first comment suggests that it's just on the research side that they should abandon current classifications, but that the DSM definitions are still needed on the clinical side.

Her book is largely focused on research, it has little to say about clinical practice. The paper she coauthored which I cited does argue, yes, that the current DSM diagnosis should be retained in clinical practice for the time being, even while they think it should be largely abandoned in research. However, I think their justifications for its continued clinical use are somewhat undermined by their arguments against its research use. That paper says "The ASD diagnosis remains necessary in the clinic to assign a child to early behavioral intervention and to explain a child’s condition". The paper argues that ASD lacks biological validity and construct validity – how can something which lacks validity actually explain anything? A diagnosis lacking validity must in itself be causally inert; and if the diagnosis is reduced to a symptom set as Waterhouse's book argues, then using it to explain symptoms becomes circular ("the symptoms are explained by the symptoms").

> Does she propose a better classification system, something like what she'd replace it with in the next DSM if it were up to her?

Her book proposes the definition of two phenotypes (not disorders) on p. 434-435. The first is "Neurodevelopmental Social Impairment Only Phenotype", which is basically equivalent to DSM-5 ASD criterion A (social impairment present from childhood), but with an exclusion for atypical sensory behaviours, atypical motor behaviours, rigidity in behaviours or interests, atypical language development (functional language impairment), ADHD symptoms, intellectual disability, developmental delay, or seizures. This phenotype is broadly similar to DSM-5 social (pragmatic) communication disorder (SCD), except it makes no claim to be a disorder, only a phenotype (symptom set), and also that it has broader exclusions – DSM-5 SCD doesn't exclude ADHD symptoms or seizures, for example, her first phenotype does. Her second phenotype is "Neurodevelopmental Social Impairment Multi-symptom Phenotype" which requires neurodevelopment social impairment combined with one or more of the exclusions from her first phenotype.

As well as criticising the DSM-5 for proposing a disorder as opposed to a mere symptom set / phenotype, she also criticises its two-domain model, which requires both social communication symptoms and RRBI/sensory symptoms for a diagnosis. She instead proposes a model with one core domain (social communication) and a set of optional domains, none of which are required, and of which RRBI symptoms and sensory symptoms are just two additional symptom domains out of several. She justifies this on the grounds of clinical experience that social communication impairment can occur without RRBIs or sensory issues – this clinical experience is reflected in the DSM-5 diagnosis of SCD.

The relationship between SCD and ASD is arguably one of the most poorly thought out aspects of the DSM-5, and in practice many people who strictly speaking should get an SCD diagnosis get given an ASD diagnosis instead, despite not actually fully meeting the ASD criteria – many clinicians are very hesitant to use the SCD diagnosis even when the letter of the DSM-5 would indicate it, due to SCD's lack of public awareness and unclear eligibility for funding. SCD itself is a strange lacuna in Waterhouse's book – despite SCD's obvious relevance to her argument, she hardly ever mentions it by name.

Don't know if you get reply notifications, but I forgot to reply the other day saying thanks for this comment. I'm looking forward to reading more on the subject!