Not a member? Sign Up!
Enter Username or Email to reset.
For much of her life Anne Dalton battled depression. She seldom spoke with people. She stayed home a lot. The days dragged on with a sense of “why bother?” for the 61-year-old from New Jersey who used to work at a Wall Street investment firm. After trying more than a dozen combinations of antidepressant drugs to no avail, things got so bad two years ago that Dalton went in for electroconvulsive therapy — in which “basically they shock your brain,” as she put it.
Like Dalton, most of the estimated 16 million U.S. adults who have reported a major depressive episode in the past year find little relief even after several months on antidepressants — a problem that some researchers say may stem from the way mental illness is diagnosed.
New research published by Liston and colleagues in Nature Medicine seems to point toward a long-sought goal in psychiatry — biological markers to distinguish different kinds of depression. The researchers used a noninvasive technique called functional magnetic resonance imaging (fMRI) to measure the strength of connections between neural circuits in the brain. Analyzing fMRI scans from more than 1,000 people, of whom about 40 percent had been diagnosed as depressive, the team identified four subtypes of depression. If confirmed in additional studies, the findings could enable clearer diagnoses and pave the way for personalized therapies targeting brain networks found to be awry in individual patients.
The Nature Medicine study, published in December, grew out of a curious observation Liston made during his earlier work on stress. As an MD/Ph.D. student he conducted experiments in rats and found that stress reduced neural connections in a brain area called the prefrontal cortex, which controls mental flexibility — the ability to adapt one’s thinking to new situations, and to overcome habitual responses.
Liston also studied stress in students preparing for their medical licensing exams. Like the rats, stressed students displayed abnormal electrical activity in brain circuits that involve mental flexibility. (Luckily, getting a month off after the high-stakes test allowed their faulty networks to recover, suggesting the brain is more resilient than expected.) In a later study that Liston conducted with Weill Cornell psychiatrist Marc Dubin, the brain-imaging researchers detected similar network changes in people who are depressed — but only in a small subset of these patients.
That intrigued Liston. It seemed to him that stress, or something like it, throws off the flexibility circuits in certain depressed individuals — whereas other people become depressed for different reasons. That would be consistent with the view that depression “is not just one biological thing,” Liston said.
That idea aligns with a new set of priorities called Research Domain Criteria, launched in 2008 by the National Institute of Mental Health to encourage scientists studying mental illness to drill down to core mechanisms rather than placing disorders under blanket labels. This shift in thinking has invigorated the search for a range of biomarkers for depression — toxic free radicals, the stress hormone cortisol and even epigenetics (environmental triggers that switch genes on and off). “Depression is too complex to be reduced to a single biomarker,” said Nunzio Pomara, a professor of psychiatry and pathology at New York University School of Medicine who was not involved in Liston’s work. At this point no individual biomarker is considered good enough to use routinely as a clinical tool, but researchers hope the best ones could someday be combined to improve diagnosis and treatment of depression and other psychiatric conditions.
To look for new biomarkers, the Weill Cornell team used a method called resting-state fMRI to check for differences in brain connectivity between depressed and healthy people. The procedure scans the brain while a person lies on a bed for five minutes — but the resulting data is complex and messy. Brain fMRI measurements are sensitive to minuscule differences between subjects such as whether people look around the room or shut their eyes during the scan. To do a rigorous analysis, Liston knew he needed a mountain of data, far more than he could collect on his own. “I went around and begged a lot of people I knew, and some I didn’t know, who had collected data the same way we did,” he said. He ended up with brain scans from 1,188 individuals — some healthy, some depressed — studied at 17 research sites worldwide. Having that much data yielded enough statistical power that “we didn’t have to constrain ourselves to [analyzing] just a few regions” of the brain, Liston said. For each subject the team examined 258 brain areas, measuring how strongly each connects with other areas.
Using an approach called machine learning, in which a computer teaches itself to find patterns in the data, the analysis showed depressed people could be distinguished from healthy controls based on brain connectivity differences measured by fMRI in the limbic and frontostriatal areas. The limbic system controls emotions and frontostriatal networks help coordinate motor and cognitive functions. One brain area, called the subgenual cingulate cortex, has unusually strong connections with other regions of the brain in people who are depressed.
Prior imaging studies had implicated these areas in depression, and some of those analyses suggested connectivity measures could differentiate between depressed and healthy people. But the Weill Cornell team is believed to be the first to confirm findings in a separate population — an additional analysis that is seen as a mark of scientific rigor. “This represents an exciting approach,” Pomara said. “It sets the stage for future studies.” He noted, though, that brain connectivity data does not address the underlying biology of depression. It does not explain what is going on at the level of cells and chemical messengers — the sorts of discoveries that guide development of new drugs. Still, he said the new fMRI analysis “goes beyond what has been done with similar neuroimaging techniques” by identifying four kinds of depressed patients on the basis of connectivity problems. Most imaging analyses merely distinguished healthy and depressed people.
In the new study, the fMRI-based subdivisions could be linked to particular symptoms. Patients falling into the first two subtypes reported more fatigue whereas those in the other two reported more trouble feeling pleasure. This subtyping has implications not just for diagnosis but potentially for non-pharmaceutical treatment. Relative to groups two and four, people with depression subtype 1 were three times as likely benefit from a newer therapy known as transcranial magnetic stimulation, or TMS. This technology uses a magnet to produce small electric currents in brain areas affected by depression. Although the procedure is gaining popularity, it is generally reserved for patients who have not responded to conventional antidepressants — people like Dalton.
In 2015, when Dalton became suicidal but felt she could not stand the memory loss associated with electroconvulsive therapy, she came to Dubin’s office for TMS. Her sister drove her in for the half-hour procedure five days a week for four weeks. By the middle of the second week, “something lifted from my brain,” Dalton said. “I laughed easier. I didn’t have those thoughts of suicide. Everything wasn’t lost to me. I thought, okay, I can do this. I can get up every day and get going.”
Dubin says TMS could one day be tuned to treat patients with different depression subtypes. After scanning a patient’s brain with fMRI, as done in the recent study, a physician could adjust the TMS magnet so it aims directly at the brain areas with abnormal connectivity in that person. “In the next five years we could be doing that,” he said.
(This article was originally published in Scientific American)