AI uncovers biomarker to predict deep brain stimulation successes in depression: NIH study

In the last decade, several clinical studies have found that deep brain stimulation (DBS) therapy may help ease symptoms of major depressive disorder. While the technology has yet to earn regulatory clearance for that indication, and scientists are still understanding the mechanisms at play in applying DBS to depression, findings from a new study may improve its chances as a viable, effective treatment.

The study—backed by the National Institutes of Health (NIH) and led by researchers from Mount Sinai, Georgia Tech and Emory University—looked for common changes in the brain activity of people who had been treated with DBS for treatment-resistant depression.

That approach was a success: The researchers were able to pinpoint a shared signature among the patients’ neurological data that lined up with changes in their depression symptoms—and which could in turn be used to predict those changes ahead of time.

“This biomarker suggests that brain signals can be used to help understand a patient’s response to DBS treatment and adjust the treatment accordingly,” Joshua Gordon, M.D., Ph.D., director of the NIH’s National Institute of Mental Health, said in an agency release. “The findings mark a major advance in translating a therapy into practice.”

To find the biomarker, the scientists used artificial intelligence to analyze the mass of brain data collected from six of the study’s 10 participants throughout the study period; each of the patients underwent DBS therapy for six months, with the implanted devices’ electrodes specifically aimed at the subcallosal cingulate cortex region of the brain.

Almost all of the enrolled patients responded remarkably well to the treatment, with 90% showing significant improvements in their depression symptoms and 70% ending the six months in full remission from the condition. That high response rate made it easier for the AI to spot patterns among the patients’ neurological data, as the biomarker it ultimately discovered correlated with those significant changes in their depression symptoms.

The biomarker also lined up with the researchers’ analyses of MRI scans of the participants’ brains, which were collected before they were implanted with the DBS devices, as well as with changes in their facial expressions in recorded interviews throughout the study.

Once defined, the scientists were able to develop an AI algorithm that works in the opposite direction: homing in on the biomarker to detect changes that could serve as an “early warning signal” that a patient’s depression may be worsening, per the NIH release, and letting their doctors know that it’s time to increase their DBS dose and clinical care.

“We showed that by using a scalable procedure with single electrodes in the same brain region and informed clinical management, we can get people better,” said co-senior author Christopher Rozell, Ph.D., a professor of electrical and computer engineering at Georgia Tech. “This study also gives us an amazing scientific platform to understand the variation between patients, which is key to treating complex psychiatric disorders like treatment-resistant depression.”

The study’s authors are now working on confirming their findings in another cohort of patients. They also plan to continue delving into the DBS technology’s effects on depression, including by analyzing its real-time, continuous effects on mood alongside the more general, longer-term impacts.

Additionally, according to the release, the researchers noted that the study’s results potentially open the door for DBS to be used to treat a variety of other mental disorders—including obsessive-compulsive disorder, post-traumatic stress disorder, binge eating disorder and substance use disorder—if biomarkers can be uncovered to map the technology’s effects on those illnesses, too.

Source