Differential Diagnostic Dogma In the news Clinical Reasoning Research Studies Unconfirmed Diagnosis

Havana Syndrome: When an Unexplained Neurological Illness Becomes FND

Author profile image FND Nope Moderator
0 comments

In 2024, researchers from the US National Institutes of Health (NIH) published two studies investigating government personnel and family members who had experienced what became known as Havana syndrome, officially described as anomalous health incidents (AHIs). These individuals reported headaches, dizziness, cognitive difficulties, tinnitus, balance problems and other neurological symptoms, sometimes following unusual auditory or sensory experiences. The investigations found no significant group-level differences in brain imaging or most clinical and laboratory measures when compared with healthy controls. Yet 28% of the clinical study participants were diagnosed with functional neurological disorder (FND), and 41% were classified as having either FND or somatic symptom disorder.

That conclusion deserves scrutiny, not because the symptoms were imaginary, but because the cause of those symptoms remained unresolved. The researchers identified clinical patterns and assigned diagnoses, but the findings did not establish what originally happened to these individuals. There is an important difference between describing a patient's current presentation and explaining how that presentation developed. In this case, that distinction risks disappearing behind a diagnostic label.

The June 2024 Neurology Today article provides a detailed account of the research, its historical context and the disagreements surrounding its interpretation.


The neuroimaging study examined 81 affected individuals and compared them with 49 controls. Researchers found no significant differences in brain volumes, structural measurements, diffusion imaging or functional connectivity. A separate clinical study examined 86 affected individuals and 30 controls, finding no significant differences across most cognitive, neurological and biomarker assessments. However, the clinical study did identify differences involving balance, as well as fatigue, post-traumatic stress symptoms and depression. Describing the entire investigation as showing nothing objectively abnormal would therefore be inaccurate.

Of the 86 participants in the clinical investigation, 24 were diagnosed with FND. Many of these diagnoses involved persistent postural-perceptual dizziness (PPPD), a condition characterised by persistent dizziness, unsteadiness and sensitivity to movement or complex visual environments. PPPD can develop following vestibular disorders and other precipitating illnesses or events. Its presence does not automatically establish the cause of the original disturbance.

This is where the interpretation becomes particularly questionable. If somebody develops persistent dizziness following an unidentified event, diagnosing PPPD may describe their current condition, but it does not necessarily explain the event that preceded it. Similarly, identifying signs classified as functional does not establish that an underlying physiological disturbance was absent. The NIH findings cannot resolve that distinction simply by attaching a diagnostic category to the resulting symptoms.

There was also a substantial disagreement between successive investigations. Earlier studies conducted by researchers at the University of Pennsylvania reported neurological abnormalities and differences in brain imaging among affected individuals. A 2018 study described findings suggestive of widespread brain network injury, while a 2019 study reported differences in white matter volume, regional brain measurements, cerebellar microstructure and functional connectivity. Those studies attracted methodological criticism, and the subsequent NIH research did not reproduce their findings.

But failure to reproduce an earlier finding is not the same as demonstrating that every participant was neurologically unaffected. The studies examined different populations at different times after symptom onset, using different methods. The NIH researchers themselves acknowledged that these differences limited direct comparisons. It would be equally unjustified to treat the earlier imaging findings as definitive proof of injury or the later negative findings as definitive proof that no injury occurred.

The accompanying JAMA editorial by Stanford researcher David Relman identified a particularly serious limitation: the affected population may not have represented one homogeneous condition. Different individuals could have experienced different mechanisms, injuries or illnesses, yet were investigated collectively under the broad category of anomalous health incidents.

That matters because group averages can conceal abnormalities affecting smaller subsets of participants. If only a proportion of the group experienced a particular neurological injury, comparing everyone collectively against healthy controls could dilute the measurable difference. A negative group-level result cannot establish that every individual within that group was free of an abnormality.

Relman also questioned whether the researchers had enough information to distinguish meaningful subgroups. He argued that PPPD could not adequately explain certain cases involving strongly location-dependent auditory and vestibular phenomena. Earlier assessments had considered whether PPPD might have developed as a consequence of an initial injury rather than being the complete explanation for the illness.

That distinction is fundamental. A syndrome occurring after an event is not necessarily an explanation for the event itself. Treating the later clinical presentation as the complete diagnosis risks replacing an unresolved causal question with a descriptive label.

The directed-energy hypothesis remains disputed. A 2020 National Academies assessment considered directed, pulsed radiofrequency energy a plausible mechanism for some reported cases, while subsequent US intelligence assessments largely argued against foreign adversary involvement. Neither position establishes the medical cause of every reported illness. Evidence about whether a foreign government conducted an attack is also not interchangeable with evidence about whether an individual experienced a physiological disturbance.

The Neurology Today article additionally reports that some participants were surprised by, or resistant to, their FND diagnoses. The discussion then turns towards how clinicians should communicate such diagnoses. But better communication cannot resolve a disputed causal explanation. A diagnosis does not become more scientifically secure because it is delivered sympathetically, explained with an animation or endorsed by specialists.

The more important question is whether the diagnosis adequately accounts for the available evidence, including findings that contradict or remain unexplained by it.

This case is particularly relevant to the common phenotype hypothesis. The hypothesis questions whether neurological signs classified as functional necessarily identify a distinct underlying disorder, or whether similar clinical patterns might arise from multiple different neurological and physiological disturbances.

The Havana syndrome investigations cannot establish which interpretation is correct. They do, however, illustrate why the distinction matters. Some participants displayed symptoms and signs that clinicians classified as functional, while the origins of their illnesses remained uncertain and researchers disagreed about the significance of earlier findings. The possibility of different underlying mechanisms was explicitly raised in the accompanying editorial.

Calling such presentations FND does not automatically distinguish between an independent functional disorder, a consequence of an unidentified neurological disturbance, or a common phenotype shared across different conditions. Those are competing explanations requiring evidence, not distinctions that can be settled through terminology.

The same issue arises when another neurological condition is subsequently identified and the original FND diagnosis is retained as a supposed comorbidity. Coexistence is possible, but so are misdiagnosis and overlapping clinical phenotypes. The mere presence of signs labelled functional does not determine which explanation applies. This broader problem is also examined in the discussion of POTS and FND, where the identification of two diagnostic labels raises questions about whether they genuinely represent two separate disorders.

There is an uncomfortable irony in the Havana syndrome investigations. Researchers were attempting to resolve an unexplained neurological illness, yet some of the resulting diagnoses introduced another explanatory problem: whether the identified clinical patterns were independent disorders or consequences of something that had not been established.

None of this proves that Havana syndrome resulted from directed-energy attacks, that the earlier imaging studies were correct, or that every FND diagnosis in the NIH cohort was mistaken. Making those claims would repeat the very error being criticised: claiming greater certainty than the evidence permits.

What the evidence does demonstrate is that substantial uncertainty remained, that different research groups reached conflicting findings, and that a leading scientific commentator warned against treating the affected population as though it represented one clearly understood condition.

Against that background, the headline that many Havana syndrome patients had functional disorders risks communicating considerably more certainty than the investigations actually established.

The bottom line is simple: finding no abnormality on a particular investigation does not establish that no abnormality exists. Identifying a functional phenotype does not establish its cause. And assigning a diagnosis does not make the unresolved questions disappear.

If the original neurological disturbance remains unexplained, then it remains unexplained. Calling the resulting symptoms FND does not magically fill that evidentiary gap.

References