By David Tuller, DrPH
This morning I sent the following letter to CNS Spectrums, a neuroscience journal published by Cambridge University Press. Subject line: “Correction needed for etiological statement in ‘Functional neurological symptoms occur commonly in healthy adults: implications for the pathophysiology of FND.'” Depending on the response, or lack of one, I might also formally submit the letter for publication.
**********
Dear Editor—
In the introduction of a recent paper in CNS Spectrums, “Functional neurological symptoms occur commonly in healthy adults: implications for the pathophysiology of FND,” the authors write that “functional symptoms are neurological symptoms which are generated by abnormal brain processing.” (FND refers to “functional neurological disorder.”)
However, this description represents the authors’ theory about what is causing such symptoms. It is not a proven fact, and should not be disseminated as if it were. In reality, the authors do not know what causes functional neurological symptoms or FND and are offering their best guess. But the statement conveys a level of certainty that cannot be justified based on the existing state of evidence.
It would be different if the authors provided data to support this categorical claim, but they do not. Nor does the article cited by the authors—a 2022 paper in Lancet Neurology, “Functional Neurological Disorder: New Phenotypes, Common Mechanisms.” That paper, at least, framed the issue in a more cautious and appropriate manner by referring to the “current understanding” of the factors that many investigators believe are causing these complex conditions.
It is harmful to the scientific process when authors present their theories as documented and indisputable facts. The definitive assertion about the cause of functional neurological symptoms should therefore be corrected. (I have cc’d the corresponding author and the editors-in-chief.)
Thank you for your consideration of this matter.
Best–David
David Tuller, DrPH
Senior Fellow in Public Health and Journalism
Center for Global Public Health
School of Public Health
University of California, Berkeley
Well said Mr. Tuller.
David, I hope you read this, accept you are wrong, and issue an apology to the journal and authors.
The sentence “Functional symptoms are neurological symptoms which are generated by abnormal brain processing” is NOT an unsupported causal claim. It is a basic mechanistic description. This isn’t about semantics…etiology and mechanism (pathophysiology) are very different things! You must know this.
There is a very obvious difference between saying symptoms are generated by abnormal brain processing and claiming to know the cause of FND. The first broadly describes the mechanism producing the symptoms which has plenty of evidence – “the how”. The second would explain why that mechanism developed in a particular person – “the why”. No one claims to fully understand “the why” & the authors certainly do not with that sentence.
To use an example the neurology field has somehow managed without correction letters: “seizures are produced by abnormal electrical discharges in the brain” does not require a disclaimer acknowledging that epilepsy’s etiology remains incompletely understood.
The journal does not need to issue a correction. You need to correct your misunderstanding. There is no way for you to twist this one.
But what happens if you ask AI if functional symptoms are generated by abnormal processing in the brain? The answer I got from AI was in the affirmative and it provided me with several links including this one -https://www.ninds.nih.gov/health-information/disorders/functional-neurologic-disorder which has an interesting first paragraph. If I’m not mistaken, participants in the studies into Havana Syndrome tested negative for significant/consistent changes on functional neuroimaging and yet we were told that many (41%) met the criteria for FND (see here-https://www.nih.gov/news-events/news-releases/nih-studies-find-severe-symptoms-havana-syndrome-no-evidence-mri-detectable-brain-injury-or-biological-abnormalities)? I think this is important – it seems that they met the FND criteria despite there being no good evidence that there were changes in the connections between their brain regions. This group didn’t appear to fit or add weight to the abnormal brain processing theory and yet it’s fine to present that theory as fact? Surely they either had something else – not FND – or abnormal brain processing isn’t behind FND/functional symptoms and shouldn’t be conveyed as such ? Now I appreciate that there could be different factions within neurology and neuropsychiatry who hold different views on what causes FND and functional symptoms but, if so, I would have expected to have seen some pushback from one faction or the other on this apparent inconsistency surrounding Havana Syndrome. Perhaps I missed it?
Getting back to AI, I’d say it poses an enormous threat to science if it relays theory as fact because journal editors haven’t pulled authors up when they’ve represented their theories/hunches/guesses as indisputable truth? Unless journal editors shape up, won’t we all be sunk?
I’ve just read a comment above. There may well be evidence, loads of evidence even, of symptoms being generated by abnormal processing in the brains of some FND patients, but if there is no evidence of this in others who’ve been diagnosed with FND yet have tested negative for this then I’d say that there is either something wrong with the claim or there is something wrong with those people’s diagnosis. Which is it?
Adding to my comment above: I’m not convinced that the evidence is there anyway of causation rather than association, but that isn’t the point I was making here. If the theory doesn’t hold true for all FND patients then surely it doesn’t stack up for FND, unless some of them are not functional/FND patients at all and faulty diagnostic criteria have let them down?
CT, I am sorry but there is a whole lot of faulty logic in your reply.
First, there are significant limitations to using AI. It’s very easy to get it to respond in a way that satisfies our priors and for it to leave out important information. It is only as good as the person prompting it.
Second, compared to the FND fMRI studies, did the researchers looking at the functional connectivity in people with Havana syndrome use the same techniques, same type of experimental conditions, and look for changes in networks repeatedly shown to be abnormal in FND? Were all research participants said to have Havana syndrome diagnosed with FND? I am going to assume no.
I hope you can see why you can’t use the negative results in Havana syndrome to make any useful hypothesis or conclusions about FND. Regardless, none of that would change the fact that David confused mechanism with etiology and was wrong for sending that letter to the journal.
Zachary, you’re so cute when you’re on your high horse. I don’t need instructions from you on the difference between “etiology” and “mechanism,” but thank you anyway. The home page of neurosymptoms.org has this to say: FND and functional neurological symptoms are “Caused by a PROBLEM with the FUNCTIONING of the nervous system,” and they involve “a ‘software’ issue of the brain, not the hardware.” That statement says it very clearly and directly: the symptoms are “caused” by malfunctioning software and the “hardware” is apparently irrelevant.
So if you have a problelm with the use of the word “cause” in this context (I didn’t use “etiology” in the letter), I suggest you take it up with Professor Stone. The sentence in question in the new paper makes the same definitive claim about what is “generating”/producing/creating–ie “causing”–the symptoms. You also assert that there is plenty of evidence for this statement, so you think the categorical assertion is just fine. But that doesn’t make it true. It’s still an assumption or theory.
Because we are also told in the recent Perez paper what has become obvious from all the research about structural changes in recent years–that this whole domain is “a “software” AND a “hardware” problem.” That means that the factors causing the symptoms could relate to the structural changes and not, as always asserted, the “software” problems. Moreover, the Perez paper says the research supports that FND is “in part” a brain network disorder–a much more modest claim than that it is 100% a “software”/brain activity problem, as the home page of neurosymptoms.org still asserts.
No one has claimed function isn’t involved. But obviously neither you nor Stone nor Perez nor the authors of the new paper know what exactly is producing what, despite your strong support of the theory that software/brain network activity is generating ALL of the symptoms. I’m sorry, but brain scan associations do not provide proof that one thing is generating or producing or causing another. That’s an assumption.
Have a good day!
Me, faulty logic? I’m bruised!!
The limitations of AI and the fact that it doesn’t appear to have an in depth knowledge of the material it’s scanning, and may not interrogate it well enough, are exactly why I think journal editors need to ensure that they don’t allow theories or hunches to be presented as the undisputable truth in their publications.
With regard to Havana Syndrome, I have no idea if the researchers looked at the functional connectivity in the participants by using the same techniques and conditions that other FND researchers have used, but if the HS researchers were associated with or were members of the FND Society, or founding members even, then I’d have thought that they would have used the most up to date state of the art/field FND techniques and conditions, and probably should have done, given the high profile of the investigation/s. As I noted above, I understand that 41% of the participants met the criteria for FND, so I assume that close to 41% of those with Havana Syndrome did not test positive for fMRI functional connectivity changes and yet met the FND criteria? I’d love FND specialists in the know to come on here and explain how I’ve got this wrong, if I have, but that’s how it comes across to me.
I don’t see why I shouldn’t use the negative results in Havana Syndrome studies to conclude that those participants were either wrongly labeled/categorized with FND or that the abnormal brain processing theory doesn’t cover all those diagnosed under current FND criteria so doesn’t stack up against those criteria. Confusion will likely abound if researchers morph hypothesis into truth/fact before all sufferers and subtypes of FND/functional symptoms have been found to consistently display functional connectivity/abnormal processing differences on MRI. (And even if that day did ever come, the issue of causation vs association would still need to be ironed out, I think.)