Stanford Researchers Identify Six Subtypes of Depression Using Imagining Technology and Machine Learning

With better methods needed for matching patients with depression with appropriate treatments, researchers at Stanford University, reporting today in Nature Medicine, say they have identified six biological subtypes (biotypes) of depression. The team also identified existing treatments that could be more likely, or less likely, to be effective against three of those biotypes.

According to statistics from the National Institute of Mental Health’s 2021 National Survey on Drug Use and Health, roughly 21 million people (8.3% of the population) in the United States suffer from major depression. The rates are significantly higher for women (10.3%) than for males (6.2%). Major depression is defined as a period of at least two weeks when a person experienced a depressed mood or loss of interest or pleasure in daily activities, and had a majority of specified symptoms, such as problems with sleep, eating, energy, concentration, or self-worth.

Among people with depression, around 30% of them have treatment-resistant depression, meaning multiple kinds of medication or therapy fail to improve their symptoms. Further, as many as two-thirds of people with depression fail to fully reverse their symptoms through treatment. One reason is a lack of data showing which treatments could help which patients, with current practice relying on trial and error to find a treatment that works. In some cases, depression can worsen during the period of trying to find the correct treatment.

“The goal of our work is figuring out how we can get it right the first time,” said senior author Leanne Williams, PhD, a professor of psychiatry and behavioral sciences, and the director of Stanford Medicine’s Center for Precision Mental Health and Wellness. “It’s very frustrating to be in the field of depression and not have a better alternative to this one-size-fits-all approach.”

To gain a better understanding of the different biotypes of the disease, Williams and colleagues enrolled 801 people in their study, all of whom were previously diagnosed with depression or anxiety using functional magnetic resonance imaging (fMRI), to measure their brain activity. Brains scans were conducted both while the participants were at rest, doing, and while doing different tasks that test their cognitive and emotional functioning. The imaging was focused on areas of the brain that are already known to play a role in depression.

With these data, the investigators then applied the machine learning approach called cluster analysis to group the brain images of the patients and identified six distinctive patterns of brain activity in the regions studied.

The team also randomly assigned 250 patients to receive one of three commonly prescribed treatments for depression, either medication or behavior talk therapy. One subtype, characterized by overactivity in cognitive regions of the brain showed best results in participants when treated with venlafaxine (brand name Effexor) compared with those who have other biotypes.

A second subtype whose signature showed higher levels of activity in three regions of the brain associated with depression and problem solving in patients at rest showed better treatment outcomes when they participated in talk therapy. People with the third subtype, who had lower levels of activity at rest in the brain circuit that controls attention, were less likely to see improvement of their symptoms with talk therapy than those with other biotypes.

According to Jun Ma, MD, PhD, a professor of Medicine at the University of Illinois Chicago and one of the authors of the study, the biotypes and how they react to different forms of therapy correlate with what is already known about these regions of the brain. But this research also breaks new ground.

“To our knowledge, this is the first time we’ve been able to demonstrate that depression can be explained by different disruptions to the functioning of the brain,” Williams noted. “In essence, it’s a demonstration of a personalized medicine approach for mental health based on objective measures of brain function.”

The researchers also showed that the different biotypes identified also exhibit differences in symptoms and task performance by the study’s participant. For instance, those with overactive cognitive regions of the brain higher levels of the inability to feel please—referred to as anhedonia—than the other biotypes. People with anhedonia also performed worse in executive function tasks. The group of patients who responded best to talk therapy, also performed poorly on executive tasks, though they performed well on cognitive tasks.

An interesting finding of the study was that one of the six biotypes showed no noticeable differences in brain activity when compared with the brain activity of people without depression. Williams added that this is likely due to gaps in our understanding, and a lack of data on the full range of brain activity underlying depression and that other regions of the brain than those that were studied play also play a role in the disease.