Trial By Error: Does Functional Neurology Disorder Account for a Third of Outpatient Neurology Consults?

By David Tuller, DrPH

Functional neurological disorder, or FND, is the new-ish name for the hoary Freudian construct known as conversion disorder. For decades, psychiatrists informed patients that they were “converting” their emotional distress and anxieties into physical symptoms like tremors, seizures, sensory and cognitive deficits, a halting gait, or other physical dysfunctions. The impossibility of proving such claims did not seem to impact psychiatry’s acceptance of the concept or the certainty with which it was applied to patients with unexplained conditions.

Other terms that have been used to describe such symptoms are psychosomatic or psychogenic disorders, hysteria, and the like. These descriptions, which highlight the belief that the conditions are driven by psychological disturbances, have often been regarded as condescending and dismissive. More recently, however, the more neutral-sounding term “functional” has been increasingly adopted to make the diagnosis more palatable to patients, among other reasons. That makes some sense. After all, these syndromes feature deficits in function that resemble standard neurological complaints. And some or many patients who receive a diagnosis of FND, often after years of feeling belittled, overlooked or abandoned by the medical profession, might experience significant relief just from being given an actual  name for what they are experiencing. I get that.

I have previously written posts questioning the term FND because it has seemed to be a so-called “wastebasket” diagnosis–just like conversion disorder. Whatever could not clearly be identified as a dysfunction explained by current pathophysiological understanding seemed at risk of being called FND, as it had previously been labeled conversion disorder.

Yet the FND experts have addressed this issue. The change in name has been accompanied as well by changes in the description of the condition in the latest version of the so-called “psychiatric bible,” the Diagnostic and Statistical Manual. Although conversion disorder remains the heading of the category, the DSM-5, published in 2013, now includes “functional neurological symptom disorder” as a synonym. (Note: After posting this earlier, I found out that the DSM-5 last month swapped out the names, with functional neurological symptom disorder becoming the main heading and conversion disorder offered as a synonym.) According to the new criteria, identification of psychological distress, once considered to be a necessary component of conversion disorder, is no longer a requirement for diagnosis—although proponents say it can still be a factor in the genesis of the condition. (In a 2020 presentation on FND, for example, Mark Hallett from the US National Institutes of Health noted in one slide that the condition is “sometimes due in part to a psychological cause.”)

Moreover, FND experts have been advocating the notion that the condition is now a “rule-in” diagnosis based on “positive” signs—not just a dumping ground for all symptoms that don’t fit standard or recognizable neurological or pathophysiological patterns. This need for confirmatory evidence has also been incorporated into the DSM-5 description. The requirement has implications for calculating prevalence rates for FND.

In 2016, Alan Carson and Alexander Lehn, two leading FND experts from the UK and Australia, respectively, made this point explicit in the abstract of a chapter on the epidemiology of the condition:

“The recent changes in DSM-5 to a definition based on positive identification of physical symptoms which are incongruent and inconsistent with neurologic disease and the lack of need for any psychopathology represent a significant step forward in clarifying the disorder. On this basis, FND account for approximately 6% of neurology outpatient contacts.”

The “positive identification” of FND is conducted through the use of strategies designed, for example, to demonstrate that patients whose attention is distracted can perform physical actions they are otherwise incapable of executing voluntarily. Assuming that these signs accurately identify FND, then the prevalence of the condition in outpatient neurology consultations is, as Carson and Lehn noted, “approximately 6%.” This data point for what they called “typical FND cases” is from a major study of Scottish neurology practices, which Carson and Lehn reference in the text of their chapter. (While they rounded up to 6% in the abstract, they used a more precise figure of 5.4% in the text of the chapter.)

But that doesn’t seem to be the most widely used estimate for this disorder. FND experts have a tendency to cite a higher figure—16%–as the prevalence among outpatient neurology consultations. Here, for example, is a statement from a 2021 paper, whose authors included Alan Carson: “FND is among the most common causes of neurological disability, diagnosed in approximately 16% of outpatient neurology consultations.”

Huh? How can the prevalence of FND diagnoses in outpatient neurology consultations be both 5.4% and 16%?


Where does that 16% prevalence rate come from?

The data for the 16% estimate is also from the Scottish neurological study, whose results were reported in multiple papers. That study investigated  the diagnoses rendered to a consecutive series of 3781 outpatients attending neurology clinics. According to a 2009 paper, 1144 of the patients, or 30%, had symptoms that could not be explained at all or only “somewhat” by what the authors called “organic disease.” Of those 1144 patients, only 18% had what the paper called “conversion symptoms” and what Carson and Lehn called “typical FND cases”—presumably, those that could be ruled in through positive signs. (As far as I could tell, the papers did not explicitly mention the use of rule-in signs by the neurologists being surveyed.) That 18% of the sub-group represented less than 6% of the total sample of 3781 patients attending outpatient neurology clinics—the source of the figure cited in the chapter by Carson and Lehn.

But in a subsequent paper in 2010, the investigators presented the Scottish neurological data somewhat differently. In that paper, they highlighted a category of “functional and psychological symptoms” comprising 16% of the 3781 cases. In other words, they apparently lumped together the “approximately 6%” comprising identified FND cases with another 10% of patients assessed as exhibiting “psychological symptoms.” Yet the paper provided no clear explanation for this decision. I assume the investigators assumed that the “psychological symptoms” were linked to the functional deficits, even absent rule-in signs.

Here is a paragraph from the abstract of the 2010 paper:

“The commonest categories of diagnosis made were: headache (19%), functional and psychological symptoms (16%), epilepsy (14%), peripheral nerve disorders (11%), miscellaneous neurological disorders (10%), demyelination (7%), spinal disorders (6%), Parkinson’s disease/movement disorders (6%), and syncope (4%).”

As in the 2021 paper, experts have since repeatedly cited this Scottish research in declaring that the prevalence of FND among outpatient neurology visits is 16%. They have argued, on this basis, that FND is the “second-most” common neurology complaint in outpatient clinics, after headache. Obviously, if they used the Scottish study’s 5.4% figure for “typical” FND cases, as Carson and Lehn, the disorder would clearly not be the second-most common neurological presentation. According the the relevant data, it would rank much lower on the list, after epilepsy, peripheral nerve disorders, miscellaneous neurological disorders demyelination, spinal disorders and Parkinson’s disease/movement disorders.

The higher prevalence claim is problematic. Either FND is a rule-in diagnosis based on positive signs, in which case it is not the second-most-common neurology presentation in outpatient clinics, according to the data collected by top experts in the field. Or it is not a rule-in diagnosis based on positive signs and can instead be applied to many more people than are identified on that basis. By including in the FND category people with “psychological symptoms” combined with neurological deficits that have presumably not been diagnosed by rule-in strategies, the authors have tripled the reported prevalence.

Some articles and resources on FND refer to yet another statistic, suggesting that up to a third of outpatient neurology consultations involve the disorder. Check out this statement from Alexander Lehn–remember, he co-wrote the chapter that included the 5.4% figure for “typical FND cases”–during a 2020 discussion on an Australian podcast: “One in three presentations to neurology outpatient clinics are due to functional disorders. One in three.”

No equivocation there! One in three is more than six times the 5.4% prevalence rate.

.That statistic presumably comes from the same Scottish study, with its 30% prevalence, or not quite a third, of patients whose conditions were not explained or only “somewhat” explained by what the authors refer to as “organic disease.” (A side note: Given that FND experts themselves are clearly differentiating between “organic disease” and “psychological” causation, why do they insist that others who make the same distinction are engaging in unwarranted and specious “mind-body dualism”? Asking for a friend…)

The 30% figure takes into account all 1144 of the patients whose disorders neurologists could not explain by “organic disease,” including the 5.4% rate of “typical FND cases” and the 10% rate of cases involving “psychological symptoms” but apparently with no rule-in signs. The additional 14% of that 30% cohort presumably have neither rule-in signs nor “psychological symptoms”—just unexplained neurological deficits. That means these 30% and 16% prevalence claims take us right back to the wastebasket diagnosis. With these expansive interpretations of what constitutes FND, anything that appears “functional”—in other words, anything the neurologists can’t explain through standard methodologies—can be counted as FND, whether identified by rule-in signs or not.

From a scientific and medical perspective, this inflation of prevalence data is problematic and confusing. What is the basis for assuming that the 10% with “psychological symptoms”–but apparently without positive, rule-in signs–also have FND? And how about the additional 14% on top of that who have not been identified as experiencing “psychological symptoms” and whose disorders are also unexplained by “organic disease”?

If FND is now a rule-in diagnosis, per the DSM-5 and the leading experts, then people without positive identification should not be given the label or imputed to belong to the category. Yet it seems, from how the data in the Scottish study have been interpreted, that these patients are at risk of being diagnosed with FND even in the absence of any such positive signs–just as they might previously have been tossed into the conversion disorder trash bin.

In short, the documented prevalence of FND among outpatients attending neurology clinics, according to the revised description of the condition and the data collected by top experts, is less than 6%–not 16% and certainly not a third. In the field of oncology, people who have not been identified as having cancer through positive diagnostic tests are not routinely included in cancer statistics. Why are neurologists doing otherwise with FND?

(One final point: Professor Michael Sharpe of PACE fame was the senior author on the Scottish neurological studies cited here. Just saying.)

18 thoughts on “Trial By Error: Does Functional Neurology Disorder Account for a Third of Outpatient Neurology Consults?”

  1. Lady Shambles

    Oh my! Well I’ve been aware of this ridiculous situation where so-called ‘FND Experts’ have played fast and loose with their own prevalence figures for a while now, so I’m glad you’ve taken the trouble to chronical this situation so clearly. Having said that, and with a ME brain having to take the strain, it’s still a complicated narrative. The upshot is incredibly simple … they are trying to claim 3 different things which are mutually exclusive at the same time and depending on the audience in question. But I’m sure for many reading about this shilly-shallying for the first time, it will be quite a mental slog.

    As I read your account a few, disparate, observations came to mind. They’re not especially earth-shatteringly insightful but I will list them nevertheless (as ever, apologies for tangled syntax, it’s a reflection of declining cognitive ability) :

    i) Wrt the nonsense of ‘positive sign’ diagnosis which doesn’t necessitate psychopathology how exactly can the FND ‘experts’ explain the biological process leading to that positive diagnosis? After all they are on record (lots of it screen grabbed) as *clearly* describing FND as synonymous with ‘conversion disorder’ as you have observed too. At Medlineplus ‘conversion disorder’ is described thus: “Conversion disorder is a mental condition in which a person has blindness, paralysis, or other nervous system (neurologic) symptoms that cannot be explained by medical evaluation”. Remove the ‘mental’ dynamic and via what route does FND, as defined by ‘positive signs’, actually occur? There doesn’t seem to be any rationale for the aetiology of even the 5.4%?

    ii) Wrt you observation here: “The “positive identification” of FND is conducted through the use of strategies designed, for example, to demonstrate that patients whose attention is distracted can perform physical actions they are otherwise incapable of executing voluntarily”. Given that in the main the so-called ‘experts’ on FND are neurologists, have they not considered that gating might be the critical consideration in such examples? And have they investigated to those ends? And if they have can they provide the evidence base for that? And should there be an evidence base that rules out gating problems being an issue (or another organic and as yet untested for / can’t yet be tested for because our understanding of biology isn’t yet good enough) what is their rationale for the process causing these presentations given the ‘mental’ element has to be removed (in other words we’re right back to question i… ‘circumlocution office’ eat your heart out!).

    iii) As someone who was told that “people like you aren’t positive for SFPN using punch biopsy, so therefore we won’t test you” (go figure! ..the gaslighting of my ‘local sfpn expert’ in Wales), but who later accessed a punch biopsy (the lab for which is, ironically for some, at KCL) which was positive & despite a similar argument against me accessing a punch biopsy from another very senior UK neurologist (‘gaslighting #2), I’m guessing that there could be shed-load of people with SFPN, for example, who could expand that 11% prevalence by taking away from the gaslighted ‘FND’ prevalence population? What I’m saying rather clumsily is that if so-called expert neurologists can’t even do their day job properly and test their patients using existing testing options properly, then how can they know how many positive organic diagnoses they’ve missed?

    iv) “Asking for a friend”. Love that! I’d like to know too 😉

    v) Getting back to the chaos of the moveable feast of the prevalence rates…. Given, as I suspect, this is just another BPS faux diagnosis (be it by ‘positive’ or ‘exclusionary’ diagnostic frameworks), which is another way of saying the whole of this body of work is (in my opinion) very possibly fraudulent, isn’t this ‘chaos’ simply a reflection of that age old story of those who tell porkies finding it hard to remember their lies? Authors are contradicting *themselves* for heavens sake. It’s absurd. Back in my day had an undergraduate made so many errors they would fail…it’s as simple as that.

  2. “- One or more symptoms of altered voluntary motor or sensory function.
    – Clinical findings provide evidence of incompatibility between the symptom and recognized neurological or medical conditions.
    – The symptom or deficit is not better explained by another medical or mental disorder.
    – The symptom or deficit causes clinically significant distress or impairment in social, occupational, or other important areas of functioning or warrants medical evaluation.”

    (My apologies if the version of dsm-5 I’m quoting here is out of date and they’ve changed the criteria).

    These criteria have features often seen in the DSM.

    The distress criterion (“clinically significant distress”) occurs in a lot of DSM definitions, and is about how serious the symptoms have to be to merit a diagnosis. It is not saying anything about causes. The idea here is that less severe versions of many DSM criteria are extremely common in the population, and it’s only worth diagnosing the people who have sufficiently severe symptoms to need treating.

    Excluding known medical diagnoses is also a common feature of the DSM. If you know your patient has brain cancer, and that explains his symptoms, then diagnose him with brain cancer and don’t bother with FND. The converse doesn’t follow. From the literature on PNES, it appears that some patients diagnosed with PNES subsequently turn out to have brain cancer. Presumably, at initial diagnosis their tumour was not sufficiently advanced to be detectable, but as the disease progressed it got easier to diagnose. Particularly in the patients that have died from their cancer.

    This kind of DSM diagnosis is always a grab bag of stuff that potentially gets rediagnosed as symptoms progress (patients brain tumour is now big enough to show up on medical imaging etc.)

  3. The “evidence of incompatibility” criterion is interesting. Presumably to meet it you need not a patient who might or might not have a brain tumour that’s too small to show up, but a patient positively exhibiting symptoms that brain cancer patients definitely don’t have.

    This is still, of course, open to the possibility that the patient has some other medical condition that the person diagnosing them didn’t think of, has different symptoms, but also gets progressively worse and easier to diagnose.

    Real patients have a wider range of symptoms than you might think just from reading the DSM. “Brain cancer patients don’t usually have this symptom, so patient definitely doesn’t have a brain tumour” is a fairly shaky argument.


    Analogies between human beings and computer systems are usually a bit suspect, so I say this with some caution. If you were diagnosing a fault in some robotic system, the symptoms might be a clue as to where in the software/hardware system the problem was.

  4. P.s. A “thought experiment” on DSM distress criteria: in the U.K. ( and many other countries) you aren’t allowed to drive a car if you have epilepsy and the seizures aren’t under control (can get your license back once the seizures have stopped).

    So, patient has occasional seizures. EEG suggests this is not epilepsy (EEG sometimes misses it). Patient says the occasional seizures are only a minor nuisance, don’t trouble them, and it’s really inconvenient not being allowed to drive a car. So, as the medical practitioner diagnosing this person, do you decide the “distress criterion’ part of the diagnosis isn’t met, and write a letter saying they don’t have FND, which the patient can then show to their motor vehicle insurer and the driver vehicle licensing agency to get their license back?

    (More seriously: I think the failure to address this question is a major flaw in some of the published research on PNES. Did those patients get their driving license back? What happened?)

  5. David wrote:
    “(One final point: Professor Michael Sharpe of PACE fame was the senior author on the Scottish neurological studies cited here. Just saying.)”

    Wasn’t Sharpe also a member of the Somatic Symptoms Disorders Work Group for DSM-5 that I imagine had a hand in the changes (to positive identification) in the DSM mentioned by Lehn and Carson in that Epidemiology chapter noted above? Separately, along with Jon Stone, James Levenson and W. Curt LaFrance -, Sharpe appears to have wanted and made a case for the inclusion of positive features or signs or symptoms in the definition of ‘conversion disorder’, with a renaming of it to ‘functional neurological disorder’. It looks like the authors of that article didn’t quite get what they wanted in DSM-5 but that things might be moving in that direction, if one of Jon Stone’s tweets is anything to go by -

    Given the contents of that article, might Sharpe be displeased (in the current DSM-5 era) if he came across prevalence figures for FND that included patients who hadn’t been positively identified? Stone, Carson, Lehn and others too perhaps? Is it worth formally asking him or Levenson, who it seems was also a member of the DSM-5 Somatic Symptoms Disorders Work Group -, for their expert opinion on this? (I appreciate that they don’t have to give it, but it would be good to know where they stand.)

  6. Following on from my last comment, I think it would be good if the Science Media Centre - or some other organization with similar purported aims, took this up and asked the experts for us. It could help iron out what is going on here.

  7. Lady Shambles

    I wrote: “…so I’m glad you’ve taken the trouble to chronical this situation..”

    apologies, obviously it ought to be ‘chronicle’.


    Eliot Slater, Diagnosis of ‘Hysteria’, 1965.

    The basic objections we are outlining in this thread have a very long history. Slater, in 1965, cites even earlier sources. Slater also gives data on the number of patients who, despite being initially diagnosed with “hysteria”, had developed “clear signs of neurological disease” at follow up (4 out of 24). Cerebral tumour was one of the conditions misdiagnosed as hysteria in Slater’s sample.

  9. The paper by Gatfield and Guze cited by Eliot Slater in 1965 predates the current DSM 5 criteria, obviously. But if diagnoses according to DSM 5 criterIa have misdiagnosis rates as bad as were seen by Gatfield and Guze, then, in my opinion, it is basically impossible to diagnose anyone with FND. It’s now an explicit diagnostic criterion that the patient have symptoms incompatible with recognised neurological conditions etc,; but if you know from previous follow up studies that about 1 in 6 patients (as seen by Gatfield and Guze) with such symptoms will eventually be diagnosed with a recognised neurological condition, the symptoms clearly aren’t incompatible, and you can’t diagnose FND.

  10. Jon Stone’s Twitter sums up the situation quite nicely:

    “FND is core health service business”

    I think we should take him at his word that it is all about money. If any patients are actually helped in some way, that is a desirable side effect.

  11. Ah, but didn’t Jon Stone and his colleagues deal with Eliot Slater’s conclusion in 2005 - ? And, going by that article, FND misdiagnosis rates these days would seem to be only around 4%. How can we possibly argue with that?

  12. Some of the statistical issues seen here also occur, for example, with covid-19. With covid-19, You’ve got different tests with different false positive and false negative rates (PCR, RAT). Maybe you aren’t able to test the whole population, and are only testing the people who end up in hospital. This can leave you with considerable uncertainty about what the actual current rate of prevalence is.

    With FND, available tests are rubbish (high false positive rate of patients who in fact have other medical conditions). Making the criteria stricter might improve the false positives, but at the expense of false negatives. So you end up with an estimate of the prevalence being anywhere between 1/3 and 5%ish, depending on test criteria.

    This is pretty unsatisfactory, but the basic problem occurs elsewhere: if you don’t have a good test for the thing you’re interested in, you don’t have a clear idea of it’s prevalence.

  13. Based on historic precedent, a pragmatic assessment of the prevalence of FND should be 0%
    I say this because historically psychiatry has always made the wrong assumptions wrt unexplained illnesses, which, with the evolution of technology and medical science eventually finds to be caused (either directly or indirectly) by some kind of pathogen.
    (Do they not learn from their past mistakes or is it the case that the medical profession quietly encourages them to be the “cork in the bottle”?)
    It would therefore seem logical that medicine should adopt a strong presumption against any kind of psychiatric diagnosis, if the possibility exists of pathogen-driven disease.
    Therefore, whilst this proposal would not be welcomed by certain members of the medical profession, it is by far the best position to adopt in order to protect the interests of patients.

  14. Pingback: Trial By Error: More On Those Inflated Prevalence Rates for Functional Neurological Disorder

  15. Thanks, Paul–that’s very possible. I’m trying to point out that, even taking their words at face value, they have amped up the prevalence rates. To do that, it’s not even necessary to analyze what’s up with the “approximately 6%” they themselves have identified as “rule-in” FND.

  16. Steve Hawkins

    Good to see that you’re branching out into other territory expanding psychiatry-generated unprovable conditions. It’s truly shocking how easy it is for them to gaslight the patients, the medical professions, and the media with their made up nonsense that nobody can prove a negative for while research on neurological testing is still in its infancy. (I took years getting an appointment to see a top autonomic specialist at the top UK hospital, only to be told that they didn’t really have any tests for peripheral nervous conditions, and that their department really only did blood pressure. When I asked if they knew what might be causing my odd blood pressure behaviour, I was told apologetically that, most of the time, they can’t explain it. I guess all those of us they can’t explain would be ‘diagnosed’ with FND by Stone et al.)

    Also: if they are using distractability as ‘proof’ of FND, then they are going to be dismissing Parkinson’s patients, who have been shown to be able to pick their feet up if fitted with glasses that make it look like there is something in the way that needs stepping over.
    There is not just a separation of mind and body, but also a separation of intentional and automatic nervous control, and not being able to perform a task at will does not imply that there is a psychological problem, any more than when we cannot immediately find the right word when we need it.

    ‘Psychological medicine’ is still a largely opinion and prestige based pseudoscientific but highly profitable industry, every bit as dodgy as most of the supplements and ‘complementary medicine’ purveyors.

  17. Pingback: Trial By Error: A Letter About the Inflated Prevalence Rate of Functional Neurological Disorder

Comments are closed.

Scroll to Top