A New Role for Diffusion MRI in Treating Anxiety and Depression

Gary Boas

Anxiety disorders and depression are widespread among adolescents in the U.S., affecting as many as one in four 13 to 18 year olds. Determining the best course of treatment can be difficult, though, as we still don’t fully understand the biology underlying them.

Now, using cutting-edge brain imaging technology, a study under way at MGH Martinos Center for Biomedical Imaging could offer new insights into this biology, and in doing so help to improve the ways we approach anxiety and depression. Ultimately, the work could also yield a quantitative means of diagnosing the disorders.

Examining Mental Disorders as Part of the Human Connectome

The study, the Boston Adolescent Neuroimaging of Depression and Anxiety (BANDA) study, emerged from the Human Connectome Project (HCP), a large-scale, multi-institutional collaboration including the Martinos Center. The HCP has demonstrated since its launch in 2010 an extraordinary ability to map the neural pathways in the healthy human brain. Using a range of MRI-based technologies, many of them developed in the Martinos Center, it has already helped to answer a range of seemingly intractable basic science questions.

Begun in late 2015, the BANDA study is now also applying these technologies to a population of adolescents with anxiety disorders and depression. In fact, it is among the first projects funded by the National Institutes of Health to look at a disease population using data collection protocols developed by the HCP.

The study is thus an important step forward, says Anastasia Yendiki, an Assistant Professor of Radiology at Harvard Medical School and principal investigator of the Martinos Center site of the study. It opens up new areas of inquiry for the HCP while also aiding a population very much in need of the insights it can provide.

“Our understanding of the biological mechanisms of mental illness is still limited,” she says. “This makes it very challenging to predict which treatment will work for which patient. We hope that, by mapping the brain signatures of depression and anxiety disorders at an age that is critical for brain development, we can discover reliable biomarkers that will allow doctors to perform accurate diagnoses and prescribe appropriate treatments for patients.”

The study has been recruiting patients from three different sites across Boston, including the Child Cognitive Behavioral Therapy program at MGH as well as sites at Boston University and McLean Hospital. It has also been recruiting patients from among those presenting to the general child outpatient psychiatry department at MGH. All of the scanning for the study is done at the Martinos Center using the state-of-the-art MRI instrumentation housed there.

Yendiki has played a dual role in both the development and the application of the methods used in the study—a role for which she has been lauded in the mainstream press, especially over the past year as she and colleagues have pressed forward with the data collection stage of the study. Last summer, the magazine Fast Company named her one of the 100 most creative people in business, citing her work in mapping “the back roads of our brains.” And in December, InStyle magazine included her on a list of “badass women” defying preconceptions of gender and making the world a better place.

Diffusion MRI and the Highways and Byways of the Brain

The researchers are exploring the brain mechanisms of anxiety disorders and depression by studying the wiring between different areas of the brain, and in particular by scrutinizing the white-matter fiber bundles that connect those areas using a technology known as diffusion MRI.

Introduced just over a decade ago, diffusion MRI has already yielded important insights into major pathways in the human brain: the superhighways of neural connectivity. Now, as the technology improves, researchers are seeing smaller pathways, one- or two-lane roads merging with the highways and then pulling away again, twisting and turning toward some other part of the brain.

Being able to image these smaller roads is especially important, Yendiki says, because the changes in connectivity associated with anxiety and depression are not likely to disrupt an entire highway. Rather, they are more likely to be subtle disruptions of specific offshoots. For this reason, the investigators have been exploring ways to further refine the reconstruction of white-matter fiber bundles using diffusion MRI.

In a paper recently published online in the journal NeuroImage, the researchers report an algorithm they developed to parse the hundreds of thousands of brain connections obtained from a high-resolution diffusion MRI scan, and group them into anatomically meaningful bundles. Typically, such algorithms will bundle connections based on their proximity to each other. This can be problematic, though, as fibers near one another do not necessarily belong to the same pathway. In the newly published paper, Yendiki and colleagues—including first author Viviana Siless, Ken Chang and Bruce Fischl—present an algorithm that bundles connections based on the surrounding anatomical structures that they pass through or pass near to. In doing so, the algorithm “behaves more like an anatomist,” Yendiki says. Indeed, in a validation study using healthy subject data from the Human Connectome Project, the landmark-based approach showed a 20% improvement in the overlap with manually defined pathways.

The researchers are already using the algorithm to analyze data from the first year of scanning in the BANDA project. Any number of other applications could also benefit from its use, including studies looking at large data sets with substantial anatomical variability, such as healthy subjects and disease populations like Alzheimer’s or epilepsy patients, or healthy subjects across a wide range of ages.