Fifteen years ago, a string of women at a neurological hospital in London all showed signs of the same strange illness. Some were stiff and stuporous, while others were experiencing seizures or issues with movement. But their experiences had all begun with what seemed like textbook episodes of psychosis, complete with agitation, hallucinations and delusions. In the throes of those symptoms, some had gone to emergency departments or psychiatric hospitals.
Their early notes read to neuropsychiatrist Thomas Pollak as classic cases of mental ill health. But it turned out that these people actually had a condition called autoimmune encephalitis, inflammation of the brain caused by an assault from the immune system. The fact that an autoimmune condition could produce psychosis shattered the usual divide between psychiatric and neurological illnesses and “kind of blew my mind”, says Pollak.
In the years since, Pollak, now at King’s College London, has helped define an emerging field of study, one beginning to show that autoimmune disease plays a larger role in mental illness than conventional wisdom suggests. This link isn’t completely new. Studies have long shown that people with schizophrenia are prone to autoimmune disease and vice versa. But the work of Pollak and others has dramatically expanded the scope of potential overlap, suggesting that the autoimmune system may play a role in many conditions, from post-traumatic stress disorder and obsessive-compulsive disorder to depression and even dementia.
All this suggests an exciting possibility: that some cases of apparent mental health conditions could be treated with drugs that target the immune system, a solution that few doctors routinely consider. “The immune system is playing a role in behaviour much more than we appreciate,” says psychiatrist Andrew Miller at the Emory University School of Medicine in Georgia.
How the immune system attacks the brain
The human immune system is a double-edged sword. The same thing that neutralises foreign pathogens like bacteria and viruses can misfire, turning the weapons in its arsenal, such as antibodies, cytokines and T-cells, against the body’s own tissues. Broadly, this phenomenon is called autoimmunity. “We know that every single organ system is affected by autoimmunity,” says Christopher Bartley, who leads the Translational Immunopsychiatry Unit at the US National Institutes of Health. The brain is no exception.
The relationship between the immune system and mental health is most clear when it comes to psychotic conditions, such as schizophrenia. About 20 years ago, scientists first described a type of autoimmune encephalitis called anti-NMDAR encephalitis, where a specific type of antibody binds to receptors in the brain, causing neuropsychiatric symptoms including delusions, hallucinations, memory loss, bizarre behaviour and more.
Despite the similarity in symptoms, the two conditions have very different treatments. Encephalitis can be treated with medications that bring the immune system back in check, whereas people diagnosed with schizophrenia are typically given antipsychotic drugs that don’t work for up to a third of individuals.
A man who lost decades of memories to autoimmune encephalitis holds up a slide film of himself as a college student
David Goldman/Associated Press/Alamy
Autoimmune encephalitis is “just a wonderful diagnosis to be able to make, because you can literally get people better and transform their lives with relatively simple treatments”, says Belinda Lennox, a psychiatrist at the University of Oxford who studies psychotic illness. But it is a diagnosis that can be missed if people who seem to have mental health conditions aren’t screened for autoantibodies – malfunctioning immune cells that attack the body – or other signs of autoimmune dysfunction.
The women Pollak saw were referred to the hospital once they started to show neurological symptoms, such as seizures and catatonia, but some people with autoimmune encephalitis can seem like they have schizophrenia for months or years. As a result, people have languished in psychiatric hospitals or under inappropriate treatments. Recently, a 12-year-old girl in the UK died by suicide after doctors failed to carry out a lumbar puncture to test for the illness, which an inquest jury found “possibly contributed” to her death.
Autoimmune diseases including multiple sclerosis and lupus are known to result in neurological symptoms, and researchers have long hypothesised that autoimmunity might play a role in schizophrenia. Since the discovery of anti-NMDAR encephalitis, interest in the overlap between immunology and psychology has intensified.
Occasionally, scientists can pinpoint exactly how autoimmune reactions lead to mental ill health. In anti-NMDAR encephalitis, for example, antibodies clearly attack the brain. For other people, autoimmune reactions seem to be relevant, but scientists don’t yet know precisely how.
Lennox’s research suggests around 5 per cent of those diagnosed with schizophrenia have antibodies in their blood, even if they don’t meet the high bar of an autoimmune encephalitis diagnosis. She is currently running a clinical trial to see how many people with acute psychosis benefit from immunomodulating treatments, specifically intravenous immunoglobulin (IVIG) and the monoclonal antibody rituximab.
Pollak thinks that no more than 1 per cent of people with acute psychosis have symptoms directly and concretely caused by antibodies that lead to autoimmune encephalitis. “But if you ask what I consider to be the more curious and interesting and less-blinkered question” – how many are experiencing some kind of brain-related autoimmunity or inflammation – then “the numbers start to get a little bit bigger”, he says.
Bartley is looking at the problem from a wider angle, too. In his view, focusing on the autoantibodies known to cause conditions like encephalitis isn’t thinking big enough. He thinks there could be other, currently unknown autoantibodies contributing to psychosis and other forms of psychiatric illness. “Rather than testing for a dozen autoantibodies in schizophrenia that people have tested for 200 times,” he says, “why don’t we widen the aperture and try to test for as many possible autoantibodies as we can?”
Antibody factories
The human body may be capable of creating a quintillion types of antibodies. Many aren’t harmful. But Bartley’s hypothesis is that, within the massive pool of possibilities, an untold number of autoantibodies may contribute to symptoms of mental illness. His lab recently identified three novel autoantibodies that may do so, and is in the process of publishing a paper on that finding, he says. In a paper published last year, neurologists at the Charité-Universitätsmedizin Berlin hospital described several others.
Bartley’s lab is working to prove the connection. If animals develop symptoms when they are introduced to cloned versions of the autoantibodies, that would be an important signal, he says. If removing the autoantibodies causes symptoms to disappear, that is an even stronger sign that hints at potential treatments for humans.
The idea that as-yet-unknown autoimmune factors may contribute to some mental health conditions tracks with the experience of Anthony Zoghbi at the Baylor College of Medicine in Texas. In 2018, he was part of a research team involved in an extraordinary case. At its centre was a woman who had been institutionalised in a psychiatric hospital in New York state for about two decades with what was thought to be schizophrenia.
Her symptoms were so severe and resistant to treatment that Zogbhi’s team eventually referred her to Columbia University in New York City for a full medical workup. Eventually, a team of specialists found that she had biomarkers suggestive of the autoimmune disease lupus, though she didn’t have the condition’s hallmark symptoms.
Even though the researchers didn’t know exactly what condition the woman had, they treated her with drugs that dampen the immune system, and it worked. After years spent in a near-catatonic state, she began to recover within months of starting the immune therapies.
The case shook Zoghbi to the core. The treatment was experimental, breaking with the medical system’s preference for testing drugs at scale and working to understand the mechanisms behind medicine. But it also did for this patient what no psychiatric treatment had achieved in 20 years.
Experiences like this one hint that known forms of autoimmune-related mental health conditions may be only the tip of the iceberg. “You can only diagnose what you have diagnostic tests for,” says Zoghbi.
It could be that many different autoimmune processes are causing or contributing to symptoms of mental illness – not only in the relatively well-researched realm of psychosis, but also across psychiatric conditions ranging from obsessive-compulsive disorder to depression. In a small 2025 study, for example, researchers found autoantibodies in the blood serum of eight out of 20 veterans with both post-traumatic stress disorder and a history of a traumatic brain injury.
Experts aren’t arguing that all or even most people who have been diagnosed with mental health conditions actually have an autoimmune disease. Pollak, for one, finds himself walking a tightrope as research accumulates. As he has written, there is a risk not only in ignoring autoimmune-related mental illness, but also in overdiagnosing it.
There is a real danger, says Pollak, in blaming everything on the immune system and throwing treatments – many with high price tags and long lists of side effects – at people without careful analysis. “I have had patients sell assets to try to get treatment in another country [when] I have been absolutely convinced from the start that that treatment would not work,” he says.
Probably only a small fraction of people with symptoms of mental illness fall into this camp. “But the stakes are so high for that very small fraction that we have to get to the answer,” says Zoghbi.
A new treatment paradigm
Trying to find effective treatments for mental health conditions can feel frustrating. “What we have in psychiatry is basically chemotherapy for the brain,” says Miller. Blunt instruments work fairly well across the board, but don’t achieve precise results – and can come with brutal side effects.
In most cases, it isn’t even clear why psychiatric drugs work; doctors know only that they do. To that end, much of Pollak’s current research focuses on better understanding what antipsychotic drugs do to the body, and specifically how they affect the immune system.
Together with researchers including Katharina Schmack, who studies psychosis at the Francis Crick Institute in the UK, Pollak launched a project to find out. Schmack and her team trigger the immune system of mice to attack their brains, in an approximation of what they think happens in some people with psychosis. Working with these models allows the researchers to analyse how antipsychotic drugs affect the mice’s immune system and behaviour.
If their research suggests that antipsychotics work by altering the immune system, that would further support a link between the immune system and mental illness – and the idea of treating more people with drugs that directly target the immune system. “The good thing about immune drugs is [that] we already have a whole array of drugs available,” says Schmack. For example, the immunosuppressant methotrexate, which is already used to treat autoimmune conditions including rheumatoid arthritis and psoriasis, is under investigation as schizophrenia treatment.
Some immunomodulating drugs are already used against autoimmune encephalitis, particularly in countries like Germany, where a national research network ensures people are more commonly screened for these conditions. Therapies include plasmapheresis – a procedure through which plasma containing harmful antibodies is filtered from the rest of the blood, discarded and replaced with new fluid – and drugs such as corticosteroids, IVIG and rituximab.
Plasmapheresis, a blood-filtering treatment, can be key to addressing autoimmune conditions
Vo Trung Dung/Look at Sciences/Science Photo Library
Other countries may soon join Germany in expanding screening. In the US, researchers at Columbia University have embarked on an ambitious effort to screen every person institutionalised in New York state’s psychiatric hospitals – a system that houses around 3000 people – for biomarkers of 12 autoimmune, metabolic or genetic underpinnings of mental health conditions. If initial blood tests are positive, people will be recommended for follow-up testing, such as a lumbar puncture. They could then receive “treatments that are fundamentally different than the standard treatments that are given for people with severe mental illness”, says psychiatrist Steven Kushner at Columbia University.
It is too soon to say what proportion of people will test positive for any of those conditions, he says. But “to me, it doesn’t matter how small that percentage is”, he says. “If [the number] is non-zero, this is worthwhile to do.”
A non-zero percentage would also suggest that similar screening processes would be worth duplicating in other health systems, says Kushner. If even a few cases were caught and treated quickly, those people might “avoid many years currently lost to mental health disability”, he says. “To identify the treatable entities as early as possible in their trajectory is a major responsibility for the field.”
The solution isn’t to replace mental healthcare with immune treatments, says Pollak, but to find an approach that blends the two. Identifying and properly treating people with an autoimmune condition could revolutionise their care and lead to a sea change in our conception of what mental illness is and can be – benefits that shouldn’t be understated. But many will still benefit from psychotherapy and other cornerstones of traditional psychiatric medicine, says Pollak. There is no reason to cast those aside.
Ultimately, it is clear that autoimmune and mental health conditions blend together more than we ever realised, and if people who appear to be experiencing the latter can also access neurological screenings, that will inevitably transform treatments – to the extent that some will be cured of seemingly intractable illnesses. Although that won’t apply to the majority, the effects are too profound to be ignored. “A smart future medicine,” says Pollak, “is going to combine all of these different aspects at the same time.”
Need a listening ear? UK Samaritans: 116123 (samaritans.org); US Suicide & Crisis Lifeline: 988 (988lifeline.org). Visit bit.ly/SuicideHelplines for services in other countries.
Topics: