FND and Autism: Does Diagnostic Overlap Establish Comorbidity?

FND Nope Moderator

28 Sep 2026 0 comments

A 2025 study published in the Journal of Neurology examined the relationship between Functional Neurological Disorder (FND) and autism using electronic health records from the TriNetX research network. The researchers identified 220,312 individuals with an FND diagnosis, 674,971 with an autism diagnosis, and 5,152 who had received both. Autism was recorded approximately 5.7 times more frequently among patients diagnosed with FND than in the wider database population. The study also found elevated rates of ADHD, psychiatric diagnoses and functional seizures among patients carrying both diagnoses.

The authors suggest that these findings may reflect shared vulnerabilities and a distinct clinical profile. However, there is a distinction between finding two diagnoses in the same medical record and establishing that two independent conditions are present. The study demonstrates the former, but its methodology cannot independently establish the latter.

What the study found

Among patients diagnosed with FND, 2.3% also had an autism diagnosis. The association was particularly pronounced in children, with autism recorded in 9.6% of children diagnosed with FND compared with 1.8% of the background population. Functional seizure diagnoses were also more common in the dual-diagnosed group. Among adults, 52% had a functional seizure code, compared with 44% of patients diagnosed with FND alone.

Psychiatric diagnoses were prominent. Among adults with both FND and autism recorded, 78% had an anxiety disorder, 77% a mood disorder, 35% PTSD and 50% ADHD. These differences warrant investigation, but the study's reliance on historical diagnostic codes raises questions about what they actually represent.

When historical diagnoses become comorbidities

One of the most important limitations appears in the methodology. The researchers explain that TriNetX retains cumulative diagnoses and acknowledge:

We cannot be certain that revised diagnoses lead to the removal of previous diagnoses.

Consider a patient initially diagnosed with FND whose symptoms are subsequently attributed to another neurological condition. Even if the original diagnosis is withdrawn clinically, its historical code may remain in the database. Such a patient could still contribute to statistics describing FND alongside another condition. The study cannot tell us how frequently this happens because it did not independently reassess the diagnoses used to construct its cohorts.

This matters beyond the present research. FND literature frequently cites low diagnostic reversal rates as evidence of diagnostic accuracy. But if revised diagnoses are not consistently removed from medical records, retrospective studies may continue counting them as comorbidities rather than diagnostic reversals. The researchers do not establish that this happened in their cohort, but their methodology cannot distinguish those possibilities. A diagnosis remaining in a medical record is not evidence that it remains clinically valid.

Coexistence does not settle the differential

The possibility that FND and autism can coexist is not in dispute. The question is whether their coexistence was independently established in the patients being studied. This is familiar territory. In our earlier article on dysautonomia as a differential diagnosis for PNES, we discussed 17 patients whose previous functional seizure diagnoses were considered misdiagnoses after autonomic abnormalities were identified. The studies concern different populations, but the distinction is relevant to both.

Finding another condition does not automatically invalidate FND. Equally, the possibility of coexistence does not establish that the original FND diagnosis remains correct. What matters is whether independent evidence for FND remains after the alternative explanation has been investigated. The autism study does not examine that question. It assumes the recorded diagnostic classifications for the purposes of its analysis.

The common phenotype hypothesis

Another possibility is that some manifestations assigned to different diagnostic categories arise from shared underlying processes rather than two independent conditions. The researchers themselves discuss potential shared mechanisms involving attention, interoception, alexithymia and emotional regulation. They consider whether these characteristics might increase vulnerability to developing FND, but there is another interpretation worth investigating: could some manifestations classified as FND represent expressions of an underlying neurodevelopmental or neurological phenotype, rather than a separate disorder?

The distinction is important. A shared mechanism might increase the likelihood of two genuinely distinct conditions occurring together. Alternatively, the same underlying process might produce manifestations that clinicians classify under different diagnostic labels. The study did not investigate these possibilities directly. Its findings are compatible with genuine comorbidity, overlapping phenotypes and some degree of diagnostic misclassification. Before explaining why two conditions occur together, we should establish that two conditions are actually required to explain the presentation.

Functional seizures and the epilepsy question

The researchers found that functional seizure diagnoses were more frequent among autistic individuals with FND. They also acknowledge that epilepsy is relatively common in autism, citing an estimated prevalence of approximately 12.1%. Yet epilepsy was not investigated in their dataset. Their explanation is that the database retains historical diagnoses, complicating interpretation.

That raises an interesting methodological question. If historical diagnostic codes make it difficult to establish whether epilepsy is present, what establishes the continued validity of the historical FND codes? The study does not report whether individual functional seizure diagnoses were independently confirmed, whether epilepsy had been excluded in each case, or whether seizure classifications had subsequently changed. This does not suggest that every functional seizure diagnosis in an autistic patient is incorrect. It means that the study cannot independently establish the diagnostic accuracy of the seizure classifications it reports. Given that functional seizures were among its principal findings, that limitation deserves attention.

Psychiatric comorbidity and causation

The researchers report elevated rates of anxiety, mood disorders, PTSD and other psychiatric diagnoses among patients carrying both FND and autism diagnoses. However, they acknowledge that they could not establish whether psychiatric diagnoses preceded or followed the onset of FND. That distinction changes how the findings can be interpreted. Anxiety might precede neurological symptoms, but it might also develop after prolonged illness, loss of independence, unsuccessful treatment or years of diagnostic uncertainty. Both situations produce psychiatric comorbidity in a medical record, but they do not support the same causal interpretation.

The authors also acknowledge that some personality disorder diagnoses in autistic individuals may reflect earlier misdiagnoses before autism was recognised. This returns us to the problem of historical coding. If a psychiatric diagnosis can remain recorded after being reconsidered, the same possibility should be investigated when interpreting FND diagnoses. The presence of psychiatric comorbidity establishes neither the cause of the neurological symptoms nor the correctness of the original diagnostic attribution.

ADHD: another diagnosis deserving scrutiny

One of the more striking findings concerns ADHD. Among adults, ADHD was recorded in 50% of patients with both FND and autism, compared with 36% of autistic patients without FND and 13% of patients diagnosed with FND alone. The authors discuss whether attentional dysregulation might interact with proposed FND mechanisms. However, before treating ADHD as an explanatory factor, it is worth examining the diagnostic category being introduced into the argument.

Like FND, ADHD presents epistemic questions that deserve more attention than they sometimes receive. ADHD is diagnosed through clinical assessment of reported and observed patterns of behaviour, developmental history and impairment. There is no definitive biological test that establishes the diagnosis in an individual patient. Its defining characteristics, including difficulties with attention, activity regulation and executive functioning, are not exclusive to ADHD. Similar difficulties may arise in other neurodevelopmental conditions, psychiatric disorders, sleep disturbances and neurological illnesses. The diagnostic task is therefore not merely to identify these characteristics, but to establish that their particular classification is justified and that alternative explanations have been adequately considered.

This becomes particularly relevant when ADHD is recorded alongside autism. Overlapping characteristics can be classified under more than one diagnostic label, but the presence of multiple labels does not independently establish multiple underlying conditions. The same question raised about FND applies here: are we identifying distinct disorders, or are we sometimes dividing a broader phenotype into several diagnostic categories? The TriNetX study cannot resolve that question because it relies on existing diagnostic codes rather than independently reassessing the clinical presentations behind them.

There is also a potential circularity in using ADHD to explain FND through attentional mechanisms. If attentional characteristics contribute to an ADHD diagnosis, and attention-dependent changes in neurological symptoms contribute to an FND diagnosis, finding the two diagnoses together does not establish that one condition predisposes a patient to the other. It may reflect genuinely separate conditions, overlapping manifestations, shared underlying processes or the way clinicians interpret similar observations in different contexts. The study did not investigate which explanation applies.

This raises a further question about positive FND signs. Some signs used to diagnose functional movement disorders involve changes in symptoms associated with attention, distraction or altered performance. How specific are these observations in autistic patients and those diagnosed with ADHD? Have the relevant signs been adequately evaluated in populations whose attentional characteristics may differ from those of the patients in whom the signs were originally studied? The authors did not examine positive signs, so their findings cannot answer these questions.

ADHD therefore deserves a raised eyebrow here, just as FND does. Not because this study demonstrates that either diagnosis is incorrect, but because it treats recorded diagnostic categories as established inputs and then explores possible relationships between them. An association between labels is not independent confirmation of the conditions those labels are intended to represent. Nor does it establish a causal mechanism connecting them. Before attentional dysregulation is invoked to explain the relationship between FND and autism, the diagnostic specificity of the categories involved deserves scrutiny.

If everything involves the brain, what makes something psychiatric?

This also connects to our earlier article, The FND Dualism Debate May Be Starting With the Wrong Question. If FND is explained through attention, expectation and altered brain networks, while ADHD, autism, anxiety and other psychiatric diagnoses are also understood through processes involving the brain, what exactly distinguishes the explanations being offered? Calling one process psychiatric and another neurological does not establish that they represent separate mechanisms. Nor does describing a psychological process in the language of brain networks necessarily provide a new explanation. If everything under discussion can be described as brain-based, what makes one diagnosis psychiatric and another neurological, and what evidence justifies that distinction?

The concern is not simply where these conditions belong in medicine. It is whether several diagnostic labels are being treated as independent explanations when they may describe overlapping manifestations or different levels of the same underlying process. Before using one diagnosis to explain another, we need to establish what each diagnosis independently contributes to our understanding of the patient's symptoms. Otherwise, we risk accumulating labels without accumulating explanations.

Are we measuring comorbidity or diagnostic intensity?

The authors acknowledge that patients receiving both diagnoses may represent more clinically complex or visible cases. That introduces another possible explanation for the elevated rates of additional diagnoses. Patients receiving extensive medical attention have more opportunities to accumulate diagnostic codes, while those with fewer healthcare encounters may have conditions that remain unrecognised or unrecorded. The study does not establish how much healthcare utilisation or diagnostic intensity contributed to the differences between groups.

The comparison groups also differed in age and sex distribution. Although the authors performed sex-stratified analyses, these differences and other potential confounders complicate interpretation. Furthermore, the researchers acknowledge that most statistically significant associations had small effect sizes. A large database can identify relatively small differences with considerable statistical confidence, but that does not establish their clinical importance or explain their cause.

When the diagnosis becomes the evidence

The broader concern is how diagnostic assumptions can pass into research without being independently examined. The researchers begin with clinician-assigned diagnoses, identify patients carrying particular combinations of those diagnoses, and investigate the resulting associations. This is a legitimate method for describing patterns in healthcare records, but it is not a method for independently validating the diagnoses used to construct the study population.

The authors acknowledge difficulties with FND coding, incomplete historical documentation and the possibility that revised diagnoses remain recorded. Those limitations are particularly relevant when the resulting associations are interpreted as evidence of distinct clinical conditions or shared underlying mechanisms. The study can tell us how frequently particular diagnostic codes appear together. It cannot establish that every recorded diagnosis remains correct, or that the manifestations attributed to each condition are independent.

This is closely related to the concern raised in our earlier discussion of diagnostic reversal: if another explanation is discovered but FND remains recorded as a comorbidity, has the original diagnosis actually survived scrutiny, or has its status simply never been reconsidered? A retrospective database cannot resolve that question unless diagnostic revisions are reliably captured.

Conclusion

Smythe and colleagues have identified an association between recorded FND and autism diagnoses in a large healthcare database. The findings concerning ADHD, psychiatric diagnoses and functional seizures warrant further investigation. However, the study cannot establish how many historical FND diagnoses remained valid, whether alternative neurological explanations had been adequately investigated, or whether some manifestations attributed to FND reflected a shared underlying phenotype. It also cannot determine whether the psychiatric diagnoses preceded or followed the neurological symptoms.

These are not minor questions when the research is being used to understand why two diagnoses appear together. The common phenotype hypothesis offers one possible explanation. Genuine comorbidity offers another. Diagnostic misclassification and differences in healthcare utilisation may also contribute. The present study cannot distinguish their relative contributions.

The question is not whether FND and autism can coexist. It is whether the evidence establishes that both diagnoses are necessary in the patients being counted. Until that question is addressed, diagnostic co-occurrence should not be mistaken for independent diagnostic confirmation.

Source

Smythe L, Asan L, Nicholson TR, Happé F, Edwards MJ. Co-occurring functional neurological disorder and autism: an exploratory study of comorbidities in a retrospective cohort study using TriNetX. Journal of Neurology. 2025;272:653. Read the full study.

DOI: 10.1007/s00415-025-13385-6