Trial By Error: A Letter Regarding Inflated Prevalence Rates for Functional Neurological Disorder

By David Tuller, DrPH

I have posted previously about how papers on functional neurological disorder (FND) have routinely mis-cited a seminal 2010 study in asserting that the diagnosis is the second-most-common presentation at neurology clinics, with a rate of 16%. In fact, the 2010 study found that only 5.5% had FND, the new name for the antiquated Freudian construct of conversion disorder. Many others apparently had symptoms that the neurologists could not explain, but that doesn’t automatically translate to a diagnosis of FND, which requires positive rule-in signs from clinical examinations.

Nonetheless, this claim—that Stone et al found FND to be the second-most-common diagnosis at neurology clinics—has gained credibility through repetition in the medical literature. But repetition doesn’t make it true.

This morning, I sent a letter to the editor of the journal NeuroImage: Clinical, asking for a correction in a 2021 paper claiming that FND is the “2nd most common” presentation at neurology clinics. Several colleagues co-signed the letter.


Dear Professor Zalesky–

For years, leading neurologists have noted that functional neurological disorder (FND), the new term for what was formerly called conversion disorder, is not a diagnosis of exclusion but a rule-in diagnosis requiring positive signs found during clinical examination. This approach was enshrined in the fifth iteration of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), which was published in 2013. Nonetheless, FND experts appear to be overlooking this definition when making claims about rates of the diagnosis at neurology clinics.

For example, a 2021 article from Perez et al in NeuroImage: Clinical“Neuroimaging in functional neurological disorder: state of the field and research agenda,” highlights the rule-in requirement and then asserts that FND “is the 2nd most common outpatient neurologic diagnosis” (1). The reference for this assertion is a seminal and widely cited paper from Stone et al called “Who is referred to neurology clinics?—the diagnoses made in 3781 new patients,” which was published in 2010 in Clinical Neurology and Neurosurgery” (2). 

Yet Stone et al, one of several papers arising from a research project called the Scottish Neurological Symptoms Study (SNSS), does not support the claim. According to the study, the second-most-common category of presentations at neurology clinics, after headache, was a grouping called “functional and psychological symptoms,” at 16%. An examination of this heterogeneous grab-bag of conditions indicates that it does not easily equate to what is called FND, per the DSM-5 criteria.

In the SNSS, 209 of the 3781 patients, or 5.5%, were diagnosed with “functional” symptoms such as sensory or motor disorders or non-epileptic seizures. These were identified in a related 2009 paper as cases of “conversion” symptoms (3); they would now be indisputably classified as FND. At the 5.5% rate, FND would be way down on the list of diagnoses mentioned in Stone et al, after headache (19%), epilepsy (14%), peripheral nerve disorders (11%), miscellaneous neurological disorders (10%), multiple sclerosis/demyelination (7%), spinal disorders (6%) and Parkinson’s disease/movement disorders (6%).

The other 10% of the 16% in Stone et al’s second-most-common grouping, who were defined collectively as having “psychological” symptoms, fell into a hodge-podge of sub-groups, including hyperventilation, anxiety and depression, atypical facial/temporomandibular joint pain, post-head injury symptoms, fibromyalgia, repetitive strain injury, and alcohol excess, among others. Also lumped in with this “psychological” cohort were cases identified as “non-organic” and “no diagnosis.” Stone et al presented no evidence that any of these patients met or could have met the rule-in criteria for a more strictly defined FND diagnosis. Arguments that members of this 10% also had FND are grounded in speculation and assumption, not fact.

In subsequent articles, two of the co-authors of Stone et al endorsed the much lower rate for the specific clinical entity known as FND. In 2016, Professor Alan Carson, the second author of Stone et al, wrote the following with a co-author in the abstract of an account of the epidemiology of FND, published as a chapter of the Handbook of Clinical Neurology (4):

“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 text of the chapter mentioned the SNSS data and gave the exact rate for “typical FND cases” as 5.4%. (It is not clear why the figure was not 5.5%.)

In 2018, several FND experts, including Professor Carson and Professor Jon Stone, the lead author of Stone et al, published a paper in JAMA Neurology called Current concepts in diagnosis and treatment of functional neurological disorders” (5). Referencing the SNSS, the paper included the following statement: In a well-designed consecutive series of 3781 outpatients of neurology clinics, 5.4% had a primary diagnosis of FND.” (Again, it is not clear why this figure was not 5.5%.) The 2018 paper noted that a larger group of patients in Stone et al had “symptoms that were described as only somewhat or not at all explained by disease.”

While these other unexplained symptoms or ailments might be called “functional” disorders in the current lexicon, they cannot reasonably be said to be equivalent to a diagnosis of FND, given the DSM-5 requirement for rule-in clinical signs. Otherwise, the 2016 and 2018 papers would have included these functional disorders as part of the FND total from Stone et al and reported that the rate was 16%, not “approximately 6%” or 5.4%.

The data have not changed since these 2016 and 2018 papers were published. Why is Stone et al now being framed, in Perez et al as well as other publications, as evidence that FND, as defined in DSM-5, is the second-most-common presentation at neurology clinics?

The popular site, maintained by Professor Stone, sheds some light on how a claim of “second commonest reason to see a neurologist,” at a rate of 16%, could be derived from the data in Stone et al. According to the site:

“In [an] older study of 3781 new appointments across Scotland, there were 209 patients who had clear FND and another 200 who had additional functional disorder diagnoses including dizziness and cognitive symptoms which could also be included now within FND. Other patients presented with diagnoses like migraine, but the neurologists thought the main issue was an associated functional disorder. So, anything from 6-16% of patients could be said to have a functional disorder depending on how that was defined. The upper limit of that estimate would make it the second commonest reason to see a neurologist.”

First, this passage confirms the relevant point. Only 209 patients out of 3781, or 5.5%, had “clear FND”–way below the level that would be required for this diagnosis to be the second-most common. Second, the 16% figure represents merely the “upper limit” of a broad possible range of estimated rates—and not rates for “clear FND” but for the fuzzier and more expansive construct of “functional disorder depending on how that was defined.”

It is not appropriate to retroactively re-interpret the data from Stone et al and effectively triple the reported rate of “clear FND” from 5.5% to 16%–thus vaulting this diagnosis into second place on the list. FND experts presumably believe the higher number is a better reflection of current diagnostic rates. notes, for example, that Stone et al’s recruitment of patients occurred two decades ago and that “recognition of FND has improved” since then. But this argument, even if valid, does not justify the decision to inflate Stone et al’s reported FND rate beyond what the data showed.  

Professor Stone made a salient observation during a 2021 podcast produced by the Encephalitis Society“Some people think that FND is a condition you diagnose when someone has neurological symptoms but you can’t find a brain disease to go along with it. And that’s absolutely not the case. Some people [i.e. clinicians] do that, but if they’re doing it like that then they’re doing it wrong.”

If clinicians who regard FND as a diagnosis of exclusion and ignore the need for rule-in signs are “doing it wrong,” per Professor Stone, then surely those who cite Stone et al to assert that FND is the second-most-common reason to see a neurologist, with a 16% rate, are also “doing it wrong.” It is confusing, not to mention epidemiologically incoherent, when FND experts report divergent rates in different papers while citing the exact same set of data.

This is especially so when the lead and second authors of Stone et al have previously made statements about the study’s rate of “typical FND” that do not support the greater claims disseminated in more recent publications, such as Perez et al. Unless the changes in the DSM-5 and the requirement for positive rule-in clinical signs are meaningless, the statement that Stone et al found FND to be the second-most-common diagnosis is categorically untrue. The citation in Perez et al—or rather, mis-citation–should be corrected.

Thank you for your attention to this matter. (The paper’s corresponding and senior authors have been cc’d on this letter, as have the co-signatories of the letter and one of the journal’s associate editors.)


David Davies-Payne
Department of Radiology
Starship Children’s Hospital
Auckland, New Zealand

Jonathan Edwards
Department of Medicine
University College London
London, England, UK

Keith Geraghty
Centre for Primary Care and Health Services Research
Faculty of Biology, Medicine and Health
University of Manchester
Manchester, England, UK

Calliope Hollingue
Center for Autism and Related Disorders/Kennedy Krieger Institute
Dept of Mental Health/Johns Hopkins Bloomberg School of Public Health 
Johns Hopkins University
Baltimore, MD, USA

Mady Hornig
Department of Epidemiology
Columbia University Mailman School of Public Health
New York, NY, USA

Brian Hughes
School of Psychology
University of Galway
Galway, Ireland

Asad Khan
North West Lung Centre
Manchester University Hospitals
Manchester, England, UK

David Putrino
Department of Rehabilitation Medicine
Icahn School of Medicine at Mt Sinai
New York, NY, USA.

John Swartzberg
Division of Infectious Diseases and Vaccinology
School of Public Health
University of California, Berkeley
Berkeley, CA, USA.

David Tuller (corresponding author)
Center for Global Public Health
School of Public Health
University of California, Berkeley
Berkeley, CA, USA


1. Perez D, Nicholson T, Asadi-Pooya A, et al. Neuroimaging in functional neurological disorder: state of the field and research agenda. NeuroImage: Clinical 2021; 30: 102623.

2. Stone J, Carson A, Duncan R, et al. Who is referred to neurology clinics?—The diagnoses made in 3781 new patients. Clinical Neurology and Neurosurgery 2010; 112: 747–751.

3. Stone J, Carson A, Duncan R, et al. Symptoms ‘unexplained by organic disease’ in 1144 new neurology out-patients: how often does the diagnosis change at follow-up? Brain 2009; 132: 2878-88.

4. Carson A, Lehn A. Epidemiology. Handbook of Clinical Neurology 2016; 139: 47–60.

5. Espay A, Aybek S, Carson A, et al. Current concepts in diagnosis and treatment of functional neurological disorder. JAMA Neurology 2018; 75(9): 1132–1141.

9 thoughts on “Trial By Error: A Letter Regarding Inflated Prevalence Rates for Functional Neurological Disorder”

  1. What trust can we have in authors who say different things in different places? We can forgive a small error or typo, of course, but giving out epidemiological information that varies to the tune of a factor of 3 or around 10%, and that happening repeatedly? How can that possibly be ok? It’ll be interesting to see what the editor says in reply, assuming he does reply, which he most definitely should, I think.

    I don’t know about having Long covid all wrong as Shure suggested in her article - , it looks like neurologists may have FND all wrong and that a lot needs ironing out now. It seems (from that article as well as the issues in David’s blog/letter) that some people are reluctant to move on from equating unexplained symptoms with psychogenic illness or FND/conversion disorder, can’t properly grasp that things changed with DSM-5 and that the old conversion-based model with diagnosis by exclusion no longer exists, even if the word ‘conversion’ hasn’t quite given up the ghost yet. So why write like it still does?

    If diagnostic criteria change then isn’t it obvious that old prevalence and misdiagnosis data will no longer apply and should therefore no longer be cited? Prevalence and misdiagnosis rates may have risen or fallen, but they’re unlikely to be the same. (For example, misdiagnosis rates could well have risen if doctors no longer feel the need to investigate and exclude everything else before diagnosing their patients with FND.) These measures relating to FND need to be re-evaluated in the light of the change to diagnosis via ‘rule-in’ signs. That’s surely just common sense, isn’t it?

  2. And…. following on from my last comment….

    If all these myths about FND - have now been now debunked then how can FND still be regarded as second most common based on Stone et al 2010? It seems that the majority of people (close to two thirds) in the second most common group in that study hadn’t been positively identified via rule-in signs (see myth 1), rather they had been diagnosed by exclusion, plus they apparently had ‘psychological symptoms’. But myth 8 tells us that FND is not exclusively a psychological problem, that people can have FND who don’t have psychological issues and that psychological issues may not be relevant in those who have them. Things have moved on and, these days, people with e.g. excess alcohol or hyperventilation shouldn’t receive a FND diagnosis without having rule-in signs.

  3. David wrote:
    “Nonetheless, this claim—that Stone et al found FND to be the second-most-common diagnosis at neurology clinics—has gained credibility through repetition in the medical literature. But repetition doesn’t make it true.”

    But still it’s repeated, it seems. Here’s a recent paper claiming 16% - to add to the growing list.

  4. Mike Fraumeni

    CT, David probably won’t recognize this last comment of yours as legitimate. David, as I’ve come to known, definitely has his own unique set of biases. Right David? Don’t post this, of course, we want your fans to know you as strictly “scientific” . 😉

  5. The 16% claim in this paper - is not scientific in my book. The group with 16% in Stone et al was seemingly made up of people with ‘functional and psychological symptoms’, not positively diagnosed FND. Around a third of them (5.5% of the total number of outpatient referrals) had ‘conversion symptoms’ which would now be called positively identified FND. If Stone or Carson or any of the other authors of that paper can tell us how the other 10.5% were positively identified using their rule-in signs, then please do. I, for one, would be all ears. But I’ve read the explanation regarding prevalence on the website - and to me, with a first-class science degree, the text provides neither a satisfactory nor scientifically cogent explanation. As David points out, – “It is not appropriate to retroactively re-interpret the data from Stone et al and effectively triple the reported rate of “clear FND” from 5.5% to 16%–thus vaulting this diagnosis into second place on the list.” And it seems that the other well-qualified people who signed the letter agree.

    In addition, it looks to me from this video - (from 8.10 mins) like Stone may appreciate that the second most common claim for FND is somewhat of a stretch but that claim, or the 16% rate, have been presented repeatedly in medical science journals in relation to FND, like here - and here - without, as far as I can see, any explanation that the second place position or 16% rate is for functional disorders and not for those which would these days be called FND. (I’m not sure how ‘functional disorders’ describes symptoms of alcohol excess or anxiety/depression, for example, but let’s leave that for another day.)

  6. Lady Shambles

    The video that CT linked to raises another issue. If FND prevalence is only around a third of what it’s been reported/assumed to be then couldn’t that potentially affect medico-legal cases to a degree? Large sums of money can presumably change hands when it comes to compensation for harms e.g. caused as a result of accidents due to negligence. If inaccurate prevalence information influenced legal judgments even to a small degree then surely that would be a very serious matter? Medics might not be all that concerned about what’s true and what’s not, but lawyers should be.

  7. I wrote above –
    “But still it’s repeated, it seems. Here’s a recent paper claiming 16% - to add to the growing list.”

    But wait – here’s a recent paper claiming 5.4% - . So who’s right? Are the experts divided now in their scientific opinion?

    It seems it’s not quite as simple as that. Here’s a recent paper that claims that FND is second most common - . Remembering that second most common equates to the 16% prevalence rate, it looks like one of the FND experts is being consistent – consistently wrong maybe, but consistent – while another expert appears to me unable to make up his mind across his papers.

    And it’s not just that one (male) expert who comes across as conflicted. For equality’s sake, here’s a recent paper by a bunch of female authors - – that claims that FND is second most common. Two of its co-authors also co-authored the paper with the 5.4% claim.

    So what is going on here? Perhaps a psychiatrist or neuropsychiatrist could help me out?

  8. Lady Shambles

    CT. The cynic in me might observe that people who lie have to be capable of remembering their lies consistently over time to be believed. Could it be that simple? Phony construct requires phony ‘data’ sets? Just pondering…

  9. Lady Shambles – who knows? I think I’d probably be more inclined towards hubris than loss of memory, but I don’t claim to be well-versed in such matters and therefore don’t think it’s really for me to judge. I’d be thrilled though if a psychiatrist/neuropsychiatrist was kind enough to offer their expert opinion.

Comments are closed.

Scroll to Top