Brain Data Could Read Moods, Potentially Treat Depression

Roughly 16 million American adults have major depression and 42 million American adults suffer from anxiety disorders, according to the National Alliance on Mental Illness, an advocacy group. About 20% of U.S. adults live with a mental illness, the equivalent of nearly 45 million people, according to 2016 government data. Mental health conditions are a leading cause of disability, resulting in productivity losses and lower quality of life, according to recent research.

“We need better treatments,” said Edward Chang, a neurosurgeon at the University of California, San Francisco’s Weill Institute for Neurosciences and a lead author of the study. “It’s a huge issue.”

In many cases, available medications and behavioral treatments don’t work. One early option for such patients was to create therapeutic brain lesions to alleviate symptoms. Eventually, neurosurgeons noticed stimulation could also improve symptoms for disorders like obsessive compulsive disorder, with some success. Scientists also noted these patients’ moods improved, which led to testing deep brain stimulation for depression.

Early studies of implantable devices that deliver electrical pulses to brain areas thought to be involved in mood and reward processing showed promise in helping patients with severe depression, but results from larger trials funded by

Abbott Laboratories



PLC were mixed, prompting psychiatrists and neuroscientists to try to develop more targeted ways of stimulating the brain.

So far, the U.S. Food and Drug Administration has approved only one deep brain stimulation device for obsessive compulsive disorder, according to the agency, though there are more ongoing clinical trials that seek to test whether deep brain stimulation can help patients with depression and anxiety.

Abbott is still “really committed to this space,” said Binith Cheeran, medical affairs director for Abbott’s neuromodulation business, but he declined to comment on specific ongoing work. Medtronic declined to comment.

To better target stimulation, experts are working to decode how the brain normally works and what goes wrong in disease, which includes figuring out what regions are involved in a particular patient’s symptoms. That can differ from person to person, even in patients with the same diagnosis, which complicates matters, psychiatrists said.

For the new study, which was published in Nature Biotechnology, researchers continuously recorded the activity of hundreds of neurons for multiple days in patients being monitored for epileptic seizures. They also had these patients rate their moods every couple of hours. They used these two sets of data to train software to understand what brain activity correlated with how a person was feeling. For each patient, the signature of brain activity—or the regions that lit up—that was predictive of mood was slightly different.

The new algorithm, developed by Maryam Shanechi, one of Dr. Chang’s collaborators at the University of Southern California with funding from the U.S. Department of Defense research arm Darpa, is a step toward reading and decoding mood-related brain activity reliably—a prerequisite for delivering more personalized care in the future, some psychiatrists said. They called the findings exciting, but preliminary.

“It sets the stage to think about how you move that mood around,” said Helen Mayberg, director of the Icahn School of Medicine at Mount Sinai’s Center for Advanced Circuit Therapeutics who wasn’t involved in the study. But first, she said, researchers need to figure out whether what they are measuring is pathological. General mood and depressive episodes may be related, but separate, signals, like two distinct instruments in an orchestra, she added.

Dr. Chang’s approach is one of several neuroscientists are taking to enable precision medicine for the brain. Some include uncovering, with the help of noninvasive imaging, what brain regions doctors should stimulate to increase the chances of improving patients’ health. Others seek to decode brain activity associated with behavioral deficits, including in learning and memory, impulsivity and emotion regulation that are common in various psychiatric disorders, including depression. These readouts are less subjective than mood and their brain circuitries better understood, according to Darin Dougherty, a Massachusetts General Hospital psychiatrist who is taking that approach and is also funded by Darpa.

Some researchers are making use of already available devices, while others like Dr. Chang’s and Dr. Dougherty’s groups, are aiming to develop novel algorithms for decoding brain activity, plus new hardware capable of recording and delivering stimulation. These would, in theory, read and decode brain signals in real-time with the help of machine learning and then deliver small electrical pulses to correct abnormalities associated with psychiatric conditions in an adaptive way.

Most commercially available devices can’t do that, which, some scientists say, may be why some of the early clinical trials failed.

For the foreseeable future, the type of machine-learning enabled brain interfaces these scientists are developing are unlikely to pose a big threat to privacy, said Hank Greely, a bioethicist at Stanford Law School.

“Until we get noninvasive ways to do this, it’s not going to be very common,” he said. “I don’t think we have to worry about Big Brother checking on our mood yet.”

Write to Daniela Hernandez at