When a Psychiatric Diagnosis Is Uncertain, Look Beyond Psychiatry

Psychiatric diagnoses are often made from patterns of symptoms, behaviour, history and clinical observation. Those tools are important—but when a diagnosis is questionable, atypical, unusually severe, treatment-resistant or accompanied by cognitive or neurological changes, they should not necessarily be the end of the investigation.

A questionable psychiatric diagnosis should trigger a serious search for neurological, cognitive, metabolic and endocrine explanations.

This is not an argument that psychiatric illness is “not real,” nor that every person with depression, psychosis, bipolar disorder or another psychiatric condition needs an enormous battery of tests. It is an argument for diagnostic humility: before permanently attributing unexplained changes in behaviour, personality, cognition or perception to a psychiatric disorder, clinicians should ask whether the brain and the body’s underlying physiology have been adequately investigated.

Start with objective cognitive testing

A detailed neuropsychological or neurological cognitive assessment can provide information that a conventional psychiatric interview may miss.

Memory, attention, processing speed, language, executive functioning, visuospatial abilities and working memory can be measured systematically. Patterns of impairment may point toward neurological disease, acquired brain injury, neurodegenerative processes or other conditions that can present with apparently psychiatric symptoms.

This is particularly important when someone reports a significant change from their previous cognitive functioning. The question should not simply be, “Does this person meet the criteria for a psychiatric disorder?” It should also be, “What has happened to this person’s brain function?”

Modern neuropsychiatric assessment explicitly recognises the importance of cognitive, neurological and laboratory information in understanding difficult presentations. 

MRI should not be dismissed when the presentation is unusual

When there is a credible possibility of neurological disease or brain injury, high-quality brain MRI deserves consideration.

MRI is generally more informative than CT for many structural brain abnormalities and provides better visualisation of grey and white matter and certain deeper brain structures. Neuropsychiatric references identify circumstances such as new-onset psychosis, cognitive decline, new personality change, seizures, traumatic brain injury and treatment-refractory presentations as situations in which structural imaging may warrant consideration.

The point is not that an MRI will automatically reveal the cause. It often will not. Nor should the phrase “high-resolution MRI” become a promise that subtle psychiatric symptoms can be explained by a scan.

The point is simpler: if neurological damage or disease is genuinely in the differential diagnosis, it should be investigated rather than assumed away. NICE material on first-episode psychosis likewise recognises neurological examination and targeted laboratory and EEG investigation, with MRI or CT considered when an organic cause is suspected.

EEG deserves consideration when electrical brain dysfunction is plausible

Some neurological disorders can produce behavioural, perceptual or cognitive symptoms that can be mistaken for psychiatric illness.

Seizure disorders are an obvious example. EEG is particularly relevant where there are episodes that are sudden, stereotyped, episodic or otherwise suggestive of abnormal electrical activity. Psychiatric assessment guidance notes that continuous or ambulatory EEG can sometimes be necessary to document abnormal activity when ordinary testing does not capture an event. 

That is why, where the clinical history makes intermittent electrical dysfunction plausible, a normal short routine EEG should not automatically end the neurological investigation.

In selected cases, prolonged ambulatory or video EEG, including a 48-hour study where clinically justified may provide substantially more opportunity to capture abnormalities than a brief recording.

But this distinction matters: a 48-hour EEG should be clinically indicated, not imposed indiscriminately on every psychiatric patient. Current psychiatric guidance generally describes EEG as an investigation to use when neurological history or examination raises the possibility of a seizure disorder or other neurological process. 

Endocrine and metabolic causes should be taken seriously

The brain does not operate independently of the rest of the body.

Thyroid dysfunction, metabolic abnormalities and other systemic illnesses can produce changes in mood, cognition, energy, sleep, behaviour and mental state. Consequently, laboratory investigation belongs in the differential diagnosis of unexplained psychiatric and cognitive presentations.

There is already recognition within psychiatric practice that physical and laboratory assessment is an important component of diagnosis and management. Standard investigations can include thyroid function and metabolic testing, with additional testing guided by the individual’s presentation. The principle should therefore be:

Do not diagnose the mind while ignoring the body.

Where symptoms suggest an endocrine disorder, the appropriate endocrine investigation should be pursued rather than assuming that psychiatric symptoms are necessarily primary.

This is not “psychiatry versus neurology”

The false choice is between saying that somebody has a psychiatric illness and saying that they have a neurological illness. Sometimes the answer is both. Sometimes neither initial diagnosis is correct.

A person can have a psychiatric disorder and a neurological disorder. Brain injury can coexist with depression. Epilepsy can coexist with anxiety. Endocrine disease can produce psychiatric symptoms. Medication, substances, sleep disorders and systemic illness can complicate an otherwise genuine psychiatric condition.

Neuropsychiatry exists precisely because the boundary between neurological and psychiatric presentations is not as clean as diagnostic labels can sometimes imply. The professional literature explicitly advocates an integrated approach rather than a rigid division between “psychiatric” and “neurological” disease.

The standard should be diagnostic confidence—not diagnostic convenience

The strongest case is not for testing everyone. It is for testing appropriately when the diagnosis does not adequately explain the person in front of you.

Red flags might include:

• a sudden or unusually late onset of psychiatric symptoms;
• substantial or unexplained cognitive decline;
• new personality or behavioural change;
• episodes suggesting seizures or altered consciousness;
• focal neurological symptoms or abnormal neurological findings;
• a history of significant head injury;
• an unusually atypical clinical presentation;
• rapid deterioration;
• unexpected treatment resistance; or
• a substantial discrepancy between the psychiatric diagnosis and the person’s observed cognitive or neurological functioning.

These are precisely the sorts of circumstances in which psychiatric literature recommends considering additional medical or neurological evaluation. 

A more rigorous standard

When a psychiatric diagnosis is questionable, the appropriate response should not be to simply repeat the same psychiatric assessment until the original diagnosis appears more convincing.

It should be to expand the differential diagnosis. That may mean comprehensive cognitive testing, neurological examination, appropriate blood and endocrine investigations, high-quality brain MRI, and, when the history warrants it prolonged EEG monitoring such as a 48-hour ambulatory study.

None of these tests is infallible. None should be treated as a magical machine capable of proving or disproving a psychiatric diagnosis. And indiscriminate testing can create false positives, unnecessary anxiety and inappropriate treatment.

But the opposite error can be far more consequential: mistaking an unexplained neurological, cognitive or physiological disorder for a purely psychiatric one and never looking for anything else.

A psychiatric label should be a clinical hypothesis supported by evidence—not a reason to stop investigating.

When the diagnosis is questionable, the brain deserves to be examined, cognition deserves to be measured, electrical activity deserves to be considered, and the body’s physiology deserves to be investigated.

Good psychiatry should not fear neurological investigation. It should demand it when the clinical picture warrants it.