When a young woman earning her Ph.D. in biostatistics came to see psychiatrist Michael Gandal with symptoms of psychosis, she became the fifth person in her immediate family to be diagnosed with a neurodevelopmental or psychiatric condition—in her case, schizophrenia. One of her brothers is autistic and another has attention deficit hyperactivity disorder (ADHD) and Tourette syndrome. Their mother has anxiety and depression and their father has depression.
Gandal has seen this pattern before. “If one person, say, has a diagnosis of schizophrenia in their family, not only are other people in the extended family more likely to have diagnosis of schizophrenia but they are also more likely to have a diagnosis of bipolar disorder, autism or major depression,” Gandal says. This propensity runs in families.
The latest edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM), the mental health field’s diagnostic standard, describes nearly 300 distinct mental disorders, each with its own characteristic symptoms. Yet increasing evidence suggests the lines between them are blurry at best. Individuals with mental disorders often have symptoms of many different conditions, either simultaneously or at different times in their lives. What’s more, as the family patterns suggest, the genes linked with these conditions overlap. “Everything is genetically correlated,” says Robert Plomin, a behavioral geneticist at King’s College London. “The same genes are affecting a lot of different disorders.”
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
In fact, it’s difficult to find specific causes of any sort, be they genetic or environmental, for individual psychiatric disorders, says Duke University psychologist Avshalom Caspi. The same is not true of most neurological disorders, such as epilepsy or multiple sclerosis. Those conditions seem to have a distinct genetic and biological origin, in contrast to the “deeply interconnected nature for psychiatric disorders,” according to findings published in 2018.
Put another way, scientists say there exists a propensity to develop any of a range of psychiatric problems. They call this predisposition the general psychopathology factor, or p factor. This shared tendency is not a minor contributor to the extent to which someone develops symptoms of mental illness. In fact, it explains about 40 percent of the risk.
The concept is akin to general cognitive ability, or g, which predicts scores on tests of skills such as spatial ability or verbal fluency. And it suggests that what joins mental health conditions is at least as important as what divides them. The concept of p, Caspi says, “is almost a clarion call for focusing on what is common rather than being preoccupied with what is distinct.”
A few researchers are calling for the erasure of hard boundaries between psychiatric conditions, which could have dramatic consequences for both the diagnosis and treatment of these disorders. “I think this will be the end of diagnostic classification schemes,” Plomin says.
Although that is not likely to happen soon, given that doctors and insurers rely on the DSM’s diagnostic codes, researchers have proposed alternative schemes that are more in line with the concept of p. At the very least, some experts say, mental health research, including clinical trials of treatments, should break free of its DSM silos and encompass multiple diagnoses. “We should be looking at psychopathology without the blinkers of the DSM,” says Patrick McGorry, a psychiatrist and professor of youth mental health at the University of Melbourne in Australia.
Some scientists have already cast off the DSM’s blinkers. Their efforts to uncover the genes or brain signatures that may underlie the p factor could inch us closer to a deeper understanding of psychiatric illness. “If you can find a biology that the p factor is working through, then in theory, if you could find ways to target the biology,” the resulting treatment would likely be effective for many psychiatric disorders, Gandal says.
Others argue that the p factor does not necessarily reflect a common cause of psychiatric illness. It may instead represent qualities present in different types of mental illness, akin to a symptom such as fever that emerges in response to various viral illnesses, says Roman Kotov, a clinical psychologist and a professor of psychiatry at Stony Brook University.
And the pursuit of a p factor was sharply criticized in a recent article in the journal Nature Reviews Psychology. The paper’s authors questioned the statistical model used to justify the p factor’s validity. This model, they contend, will tend to confirm the presence of a p factor regardless of whether one exists. “It’s very clear at this point that the evidence people are using to claim that they’ve found a p factor is not sufficient to prove its existence,” says lead author Ashley Watts, a psychologist at Vanderbilt University.
Yet the data backing the DSM’s version of reality is at least as flimsy, experts say. The disorders codified in the manual grew out of patterns of symptoms that doctors noticed in their patients. In a 1943 paper, for example, psychiatrist Leo Kanner outlined the criteria for autism based on the traits he observed in 11 children. But it is hard to prove that a condition defined by difficulties with social interaction combined with repetitive and restricted behaviors really exists. “On what basis do we say there’s a syndrome, do those things go together?” Plomin says. “They don’t go together. The components of autism are less genetically correlated with each other than major disorders are correlated with each other.”
More evidence that the DSM’s boundaries are soft, if not fictitious, has to do with the observation that many, if not most, people with a mental illness—as many as 82 percent of those with schizophrenia, recent data suggest—exhibit symptoms that cross those boundaries, sometimes leading to multiple diagnoses. For instance, individuals with depression often have anxiety, and those with post-traumatic stress disorder (PTSD) are likely to have a substance use disorder.
Features of mental illness that span classic diagnoses may be most common in the earliest stages of distress, McGorry says. “It’s kind of an array of anxiety, depression, maybe a few little warning signs of psychosis in a significant proportion, a bit of mood instability in others, drug and alcohol [misuse] in a subset,” he says. “If you hadn’t ever read a DSM, it wouldn’t naturally occur to you” that the manual’s method of classifying disorders would be useful.
Yet another sign of the arbitrary quality of the DSM’s designations has to do with the way a diagnosis can shift over time. In one longitudinal study, scientists tracked the mental health of roughly 1,000 people born in Dunedin, New Zealand, in 1972 and 1973. In a 2020 report of the participants at age 45, Caspi and his colleagues found that people who are diagnosed with a mental health disorder often see that diagnosis changed some years later. A substance use disorder may remit and give way to depression, for example, only to later return as depression is replaced by severe anxiety, Caspi says.
A general predisposition to mental illness—what many see as the essence of the p factor—could explain the fluidity in diagnoses. Evidence from genetics supports this view. Studies show considerable genetic overlap among diagnoses made based on the DSM.
In a genome-wide association study (GWAS) of a psychiatric disorder, researchers compare the genomes of tens or even hundreds of thousands of people who have that specific mental health condition with those of an equal number of individuals without such a disorder. In so doing, they link small changes in DNA to the condition being studied. The first such studies of psychiatric disorders, published about 15 years ago, showed that many of the specific versions of genes associated with bipolar disorder were the same as those associated with schizophrenia. “These things that we think are distinct, most notably bipolar and schizophrenia, are not at all distinct,” Plomin says. “That was kind of mind-boggling.”
Since then researchers have followed suit for almost all the other major psychiatric disorders—including ADHD, PTSD, depression and obsessive-compulsive disorder (OCD)—and found a number of genes common to all of them. Together, the full complement of these genes has been termed “genetic p,” a kind of biological lever that controls psychiatric risk. “It is a statistical abstraction,” Plomin says, “but it’s very important because it points to the fact that many genes are general in their effects” on mental well-being.
Investigators are now trying to understand what the biological roles of these genes might be. In 2019 a team conducted a GWAS across psychiatric and neurodevelopmental disorders such as depression, anorexia, autism, bipolar disorder and schizophrenia in a nationwide Danish birth cohort of almost 1.5 million people. It found genetic variants that were common to these conditions that had roles in fetal neurodevelopment. “As the brain develops, it’s really a critical period of time where a lot of these biological processes are manifesting”—processes that may shape psychiatric risk, says Gandal, who is one of the authors of the study.
Gandal and his colleagues recently found genetic variants associated specifically with the p factor, which partially explained psychiatric symptoms in 10,000 adolescents in the Adolescent Brain Cognitive Development (ABCD) Study, the researchers reported at the 2023 meeting of the American Society for Human Genetics. The researchers have since been investigating the biological function of the specific versions of genes that this and other studies have identified. “We can now resolve [a molecular role for] about 60 percent of these genetic hits,” he says.
Others on the search for the meaning of p are examining brain structure. In a 2020 study, brain scans of more than 12,000 people who had one of six major psychiatric disorders revealed similar structural differences in four conditions—major depressive disorder, bipolar disorder, schizophrenia and OCD—and distinct ones in ADHD and autism. A similar 2023 study of 5,549 people who were autistic or had schizophrenia, bipolar disorder or major depression revealed a common pattern of thinning of the cerebral cortex in people with schizophrenia, bipolar disorder or depression.
The p factor could also arise from a shared mental trait. “Before I want to go biology on you, I want to understand [on the psychological level], ‘What is it that really binds a lot of these different conditions?’” Caspi says. Difficulty with emotional regulation is one possibility, but Caspi suspects the common thread is disordered thought: cognitive distortions that might take the form of delusions in psychosis, irrational fears in anxiety, knee-jerk negativity in depression or intrusive thoughts in OCD. It’s just conjecture for now, he says, but once people have pinpointed what’s central psychologically, “you have an actionable target that maybe you can start to think about biologically.”
The possible existence of a p factor has implications for how mental health research is conducted. For one thing, it casts doubt on the practice of restricting clinical trials of treatments to participants with a single disorder, such as depression. After all, many people with mental illness have more than one diagnosis. “The question that arises is: What am I studying when I study just depression among people who have no other disorders?” Caspi asks. “That’s not necessarily representative of that condition.” Instead scientists should target research and therapies to more general properties of mental illness, he says.
About a decade ago the National Institute of Mental Health (NIMH) initiated a program called Research Domain Criteria (RDoC) that is intended to do just that, or, in the jargon accorded to this goal by the agency, to encourage “research on dimensional constructs that cut across disorders.” The program is designed to fund studies geared toward the biological roots of human behavior, both typical and atypical, as a route to a better understanding of mental illness. But critics charge that the agency has not followed through on this mandate. When McGorry wanted to expand an NIMH-funded study of psychosis risk to a broader range of diagnoses, he was told to stick to the original idea of focusing on psychosis and schizophrenia. “So I don’t think they are funding things outside the DSM categories even though they say RDoC is their paradigm,” McGorry says.
The NIMH counters that it continues to encourage and support projects that use RDoC principles to study psychopathology, and estimates that several hundred such projects have been funded. Some of these projects, the agency says, focus on understanding differences between people diagnosed with the same disorder as defined by the DSM, and others attempt to identify biological and behavioral mechanisms that are observed across DSM diagnoses. NIMH, however, does not exclusively fund research that takes an RDoC approach, which it says is "more appropriate for some projects than for others.”
What a p factor should mean for clinical decisions is uncertain. No one is saying that everyone with a psychiatric condition is the same. Symptoms such as hallucinations, alcohol misuse, disordered eating and social anxiety clearly do not appear uniformly among people with psychiatric conditions. In fact, Plomin recently measured the extent to which genetic variants for any of 11 conditions are not associated with p. He termed this measurement for ADHD, for example, “non-p ADHD.” “No one’s saying it’s all p. It’s just saying there is p,” Plomin says.
Watts, who has her doubts about p, maintains that mental illnesses are more likely to fall into a few broad categories than to bear some common set of features. So-called internalizing disorders—depression, anxiety, PTSD and other conditions characterized by negative emotion—tend to overlap, she says, and respond to the same treatments. The same is true for “externalizing” behaviors such as hyperactivity, aggression and rule violation, which reflect a lack of impulse control. “I think the evidence pretty clearly supports the presence of these broader higher-order dimensions but not necessarily this overarching dimension of p,” Watts says.
Yet even the idea of broad categories is inconsistent with the DSM’s divisions. “We’ve drawn arbitrary borders around these diagnoses for far too long,” Watts says. And these borders reflect another fallacy: that mental illnesses are distinct from wellness. In reality, the severity of conditions such as depression, substance dependence and social anxiety varies across the population following the pattern of the classic bell curve. Most people experience moderate symptoms, reflected by the big blip in the middle of the curve, with smaller “tails” for mild cases to the left and severe cases on the right. There are no obvious cutoff points.
Kotov and others have put forward frameworks to try to capture degrees of severity for a particular disorder. In the Hierarchical Taxonomy of Psychopathology (HiTop), developed by Kotov and his colleagues, patients are rated according to the severity of each of more than 100 psychiatric symptoms. The result is a mix of ratings specific to an individual. “Each person is represented by a profile, so diagnosis becomes a profile as opposed to being a label or a list of labels,” Kotov says.
Some researchers aim to increase the specificity of diagnoses by observing patients over time. In a longitudinal study of 5,432 people who were diagnosed with schizophrenia, Gandal and his colleagues found five clusters of people with similar patterns of diagnoses over their lives. It’s a step toward identifying schizophrenia subtypes, which might be further defined genetically, Gandal says of the 2021 results.
McGorry favors a staging system that, like the one used for cancer, centers on the severity of a condition rather than its qualitative features. “The staging idea allows the fluidity of the symptoms,” McGorry says, “to be captured and validated in a transdiagnostic-type way.” It also could help soften the perception that there is a hard border between illness and wellness and thereby reduce the stigma of mental illness, McGorry maintains. That, in turn, could encourage more people in distress to seek care.