I once discussed intelligence among psychiatric patients with my former psychiatrist. My impression was simple: people with psychiatric disorders seemed, on average, to have lower intelligence than the general population.
He disagreed.
At the time, my argument was largely anecdotal. Since then, however, I have had considerably more contact with psychiatric populations. I spent time in a mental hospital, surrounded by people with a wide variety of psychiatric diagnoses. I encountered another psychiatric population through an organization for people with mental illness.
My subjective impression remained the same.
Of course, this is not scientific evidence. Personal observations are vulnerable to selection bias, confirmation bias and countless other distortions. A psychiatric hospital in particular is anything but a representative sample of everyone with a psychiatric disorder.
Fortunately, we do not have to settle the question with anecdotes. There is actual research.
And the general picture emerging from that research is surprisingly clear: lower intelligence and psychiatric illness are associated at the population level.
That does not mean that psychiatric patients are unintelligent. It does not mean that every psychiatric disorder is associated with reduced IQ. And it certainly does not allow us to infer an individual’s intelligence from a psychiatric diagnosis.
It means something much more specific: when we compare sufficiently large populations, lower IQ is associated with a greater probability of psychiatric illness, and psychiatric populations tend to show lower cognitive ability on average.
Almost three million people
Perhaps the strongest evidence comes from a major 2025 systematic review and meta-analysis published in Communications Psychology.
Instead of examining one psychiatric diagnosis, the researchers attempted to synthesize the broader literature connecting intelligence with later health.
They analyzed 49 studies representing more than 2.9 million people.
Importantly, intelligence was measured early in life—during childhood, adolescence or early adulthood, before age 21. Researchers then examined subsequent physical and mental illness.
The overall relationship was substantial enough to be measurable across this enormous population.
A decrease of one standard deviation in IQ—approximately 15 IQ points—was associated with a 22 percent greater risk of subsequently developing mental or physical illness.
Lower early-life intelligence predicted multiple conditions, including schizophrenia and depression.
The relationship weakened when researchers statistically controlled for education and when healthcare quality was higher. However, it did not disappear.
That is important.
The association cannot simply be summarized as “poor people receive less education, score worse on IQ tests and become mentally ill.” Education and socioeconomic circumstances clearly matter, but they do not appear to explain the entire relationship.
Lower cognitive ability itself—or biological and environmental factors associated with it—appears to be part of the picture.
Looking at psychiatric patients from the other direction
Other studies approach the question from the opposite direction.
Instead of asking whether people with lower IQ are more likely to develop psychiatric disorders, researchers can take people who eventually enter psychiatric care and examine how their earlier intelligence compares with that of people who do not.
A study published in The British Journal of Psychiatry examined 3,289 men from the Copenhagen Perinatal Cohort.
Researchers had access to intelligence-test results from early adulthood and later psychiatric hospitalization records. They identified 350 men who subsequently appeared in the Danish Psychiatric Central Register and compared them with 2,939 men who did not.
The researchers examined nine broad diagnostic categories.
Lower IQ was associated particularly clearly with schizophrenia and related disorders and other psychotic disorders. Associations were also found with adjustment disorders, personality disorders, alcohol-related disorders and substance-use disorders, although the statistical picture became more complicated after accounting for comorbidity.
Mood disorders and neurotic disorders did not show a statistically significant association with lower IQ in that particular sample.
The authors nevertheless reached an important general conclusion: most of the broad categories of mental disorder they studied were associated with relatively low premorbid IQ.
Even more directly, they listed one of the clinical implications of their research: mean IQ-test performance among psychiatric patients was relatively low.
That is remarkably close to the proposition my psychiatrist rejected.
Lower IQ can precede mental illness
There is an important detail here. One might reasonably argue that severe psychiatric illness itself reduces cognitive performance.
That certainly can happen.
Depression can interfere with attention and memory. Psychosis can profoundly disrupt thinking. Psychiatric medications can produce sedation or cognitive side effects. Long periods of illness can interrupt education, employment and intellectual activity. Substance abuse can produce additional cognitive damage.
But this cannot explain the entire association.
In many studies, intelligence was measured before the psychiatric illness became clinically apparent.
The 2025 meta-analysis deliberately examined IQ measured before age 21 and subsequent disease. The Copenhagen study similarly had intelligence measurements from early adulthood that could precede psychiatric hospitalization.
This establishes temporal ordering in those datasets: lower measured intelligence was already present before the later recorded illness.
It does not establish simple causation. But it makes the hypothesis that the entire IQ difference is merely a consequence of psychiatric treatment or years of mental illness untenable.
Borderline IQ and psychiatric disorders
Another study, published in BMC Psychology in 2020, examined the relationship from yet another perspective.
Researchers compared young adults with borderline intellectual functioning—IQ scores between 70 and 84—with a reference group scoring 85 or higher.
The difference was enormous.
Among participants with borderline IQ, 53 percent met research criteria for at least one psychiatric diagnosis. Among the reference group, the corresponding figure was only 12 percent.
The estimated odds ratio was 6.2.
ADHD and anxiety disorders were particularly common in the lower-IQ group.
This was a relatively small convenience sample, so it would be a serious mistake to take those percentages and assume they describe the general population. The study itself had important limitations.
But its direction agrees with the broader literature: lower cognitive ability is associated with greater psychiatric vulnerability.
This does not mean “mentally ill people are stupid”
This is where statistical literacy becomes essential.
Saying that two populations have different average IQs is not remotely equivalent to saying that every member of one population has a lower IQ than every member of another.
Imagine that one population has an average IQ of 100 and another an average of 95. There would still be enormous overlap between their distributions.
There would be highly intelligent people in the second population and people with very low intelligence in the first.
The same principle applies to psychiatric illness.
There are brilliant people with schizophrenia, depression, bipolar disorder, OCD, autism, personality disorders and virtually every other psychiatric diagnosis. There are also people of very low intelligence without psychiatric illness.
A psychiatric diagnosis therefore tells us very little about the IQ of a particular individual.
Population statistics answer a different question. They ask whether the distributions differ. The evidence indicates that, broadly speaking, they do.
Psychiatric illness is not a single category
There is another major qualification. “Mental illness” is an extraordinarily broad label.
Schizophrenia, major depression, generalized anxiety disorder, anorexia nervosa, bipolar disorder, personality disorders, substance-use disorders and OCD are not manifestations of one underlying disease.
We should therefore not expect identical relationships with intelligence.
Indeed, the Copenhagen study illustrates this beautifully. Psychotic disorders showed a clear relationship with lower premorbid IQ, whereas mood and neurotic disorders did not show a statistically significant association in that sample.
The 2025 meta-analysis nevertheless found that lower early-life IQ predicted illness across several different categories, including schizophrenia and depression.
The most reasonable general conclusion is therefore not that every psychiatric diagnosis lowers average IQ by some universal number of points.
There is no such number.
Rather, lower intelligence appears to be a broad but highly heterogeneous correlate and risk marker for psychiatric morbidity.
The strength of the relationship depends on the disorder, population and circumstances.
Why would IQ and mental illness be connected?
There is unlikely to be a single explanation.
One possibility is cognitive reserve. Greater cognitive ability may provide people with more resources for solving problems, adapting to stressful circumstances, navigating complicated social environments and developing effective coping strategies.
Someone with lower cognitive ability may have fewer such resources when confronted with the same environmental stress.
But causality could also run through common causes.
Brain development influences both cognition and psychiatric vulnerability. Genetic factors, prenatal development, childhood adversity, neurological insults, environmental toxins, nutrition, socioeconomic circumstances and education can potentially influence both.
Then there is reverse causation.
Some psychiatric disorders begin affecting cognition before they become severe enough to produce a formal diagnosis. A person may therefore score lower on an IQ test because an emerging neurodevelopmental or psychiatric process has already begun.
Finally, psychiatric illness itself can produce additional cognitive impairment after onset.
These mechanisms are not mutually exclusive.
The real explanation is almost certainly a combination of them.
Why my experience in a psychiatric hospital probably exaggerated the difference
There is one reason I would still not use my own hospital experience as evidence for the size of this effect.
Psychiatric hospitals are extremely selected environments.
Imagine two people with the same diagnosis.
One has an IQ of 125, responds well to medication, understands the illness, follows treatment recommendations, maintains employment, has supportive relationships and successfully organizes everyday life.
The other has severe cognitive impairment, poor executive functioning, difficulty understanding treatment, substance-use problems and limited ability to live independently.
Which one is more likely to spend substantial amounts of time inside the psychiatric system?
Obviously, many other variables matter, but the second person is likely to be considerably more visible in institutional settings.
This creates selection bias.
The psychiatric population one encounters in a hospital is not equivalent to the psychiatric population living in society.
The same problem can affect organizations providing services to people with serious mental illness. People functioning perfectly well despite a diagnosis may have relatively little reason to use intensive psychiatric or social services.
My observations may therefore have reflected a genuine population difference while simultaneously making that difference appear much larger than it really is.
IQ is also not the same thing as cognitive functioning
There is another complication.
When we interact with someone, we do not directly perceive their IQ.
We perceive speech, memory, attention, judgment, knowledge, inhibition, processing speed, social cognition and executive functioning.
Psychiatric disorders can interfere with many of these abilities without producing an equivalent reduction in general intelligence.
A severely depressed person may respond slowly because of psychomotor retardation. Someone experiencing psychosis may have difficulty maintaining a coherent conversation. A heavily medicated patient may appear cognitively dull because of sedation. Someone with executive dysfunction may struggle to organize relatively simple everyday activities despite having perfectly normal abstract reasoning ability.
Consequently, what I interpreted as lower intelligence during my psychiatric hospitalization may partly have been genuine differences in general cognitive ability and partly acute symptoms, medication effects or impairments in specific cognitive domains.
This distinction matters scientifically.
What the evidence actually allows us to say
My former psychiatrist was correct if the intended argument was that psychiatric patients should not simply be characterized as unintelligent.
That would be an indefensible stereotype.
But if the argument was that there is no general statistical relationship between lower intelligence and psychiatric illness, the scientific literature makes that position much harder to defend.
A 2025 meta-analysis encompassing more than 2.9 million participants found that a 15-point disadvantage in early-life IQ was associated with a 22 percent higher subsequent risk of mental or physical illness.
A longitudinal Danish study found relatively low premorbid IQ across most broad psychiatric diagnostic categories it examined, particularly psychotic disorders.
And a smaller study of young adults found psychiatric diagnoses in 53 percent of participants with borderline IQ compared with 12 percent of its reference group.
These are very different research designs, populations and questions.
Yet they point broadly in the same direction.
The general conclusion
So what can we reasonably conclude?
There is a real population-level association between intelligence and mental health.
Lower IQ is associated with an increased probability of psychiatric illness, and at least some psychiatric populations have lower average IQ than comparison populations. Crucially, part of this relationship exists before psychiatric hospitalization and even before clinically recognized illness, which means it cannot simply be dismissed as a consequence of medication, institutionalization or chronic psychiatric disease.
At the same time, the relationship is probabilistic, not deterministic. It differs substantially among diagnoses. Social and educational factors explain part of it. Shared neurodevelopmental causes may explain another part. Psychiatric illness itself may subsequently worsen cognitive functioning.
And the overlap between populations remains enormous.
Therefore, neither extreme is scientifically justified.
“Mentally ill people are unintelligent” is false.
“There is no relationship between IQ and psychiatric illness” is also inconsistent with the available evidence.
The much less sensational conclusion is the scientifically interesting one: psychiatric morbidity and lower general cognitive ability are associated at the population level, through a complicated mixture of vulnerability, shared causes, environmental influences and consequences of illness.
My experiences in a psychiatric hospital and an organization for people with mental illness cannot prove that conclusion.
Large-scale research can. And increasingly, it does.
Leave a Reply