Functional MRI: The next tool for mental health?
Researchers develop an algorithm for predicting medication response in mental health patients
ATLANTA—A team of scientists at the Center for Translational Research in Neuroimaging and Data Science (TReNDS) has received a $875,110 grant from the National Institute of Mental Health to further the development of a tool to help psychiatrists treat mood disorders. The technology is based on research by Vince Calhoun, who is a Distinguished University Professor of Psychology, a Georgia Research Alliance Eminent Scholar and the founding director of TReNDS. He also holds appointments in electrical and computer engineering at Georgia Tech and neurology and psychiatry at Emory University.
TReNDS is a research effort supported by Georgia State University, the Georgia Institute of Technology and Emory University that is focused on research that makes better use of complex brain imaging data. The two-year grant was awarded to Advanced Biomedical Informatics Group, LLC (ABMIG), a startup company led by Jeremy Bockholt, who has collaborated with Calhoun for more than 15 years.
“ABMIG is our partner on this project, and they are supporting the development of web-based tools and dashboards as well as access to additional data. We have applied for many commercialization grants with ABMIG as the business partner,” note Calhoun, Bockholt and Eric Verner, associate director of Innovation at TReNDS and a co-investigator on the project. “Our algorithm will be used on resting functional magnetic resonance imaging (fMRI) scans to tell whether a particular patient has brain patterns that are more like those who have responded well to antidepressants or mood stabilizers. The psychiatrist will use this information, along with information obtained through other methods such as interviews and medical history, to prescribe medication. This will reduce the trial and error often involved in finding the right medication for a patient.”
“We have collected resting fMRI scans and treatment response from around 100 subjects in a previous study to build our initial algorithm and will add data from hundreds more subjects as part of our commercialization grant to refine the algorithm,” the researchers add.
There are currently no biologically based clinical tools to diagnose mental illness, and as a result, distinguishing between mood disorders like bipolar disorder and depression can be challenging. Finding a mental health treatment that works for each individual is often a process of trial and error.
“This tool could give clinicians an objective window into a patient’s brain, helping them make more tailored treatment recommendations. Regardless of the diagnosis, is the patient’s brain more similar to someone who responded better to mood stabilizers, or to someone who responded better to antidepressants?” Bockholt queried.
The team plans to use the grant money to refine the algorithm by incorporating additional data, including scans from a more diverse set of patients and scans from various types of fMRI machines.
“Initially, the algorithm will only use the resting fMRI scan as an input, but clinical and demographic data and additional data about brain volumes are also planned for the future,” state Calhoun, Bockholt and Verner.
“On the previous data set, our tool was over 90-percent accurate in predicting medication outcomes, so that shows us that we’re on the right track,” Verner pointed out. “By training the model on more data, it should perform better on a wider variety of patients.”
The researchers also want to complete interviews with psychiatrists to learn more about how and when a tool like this could be helpful in a clinical setting. They hope to submit the technology to the FDA for approval as a medical device.
“We use interviews to understand how psychiatrists would view a tool like this and how they view the application of artificial intelligence in psychiatry in general. We ask about how to build a tool that psychiatrists would feel comfortable using, is convenient for them, and gives them the right amount of information needed to make decisions,” the researchers say. “We will also study how easy it will be to provide referrals to an imaging facility, a process which is not common in the psychiatry field.”
“Once it reaches the market, this would be the first clinical decision support tool using fMRI for psychiatry, and we are excited about the potential for this technology to help patients suffering from bipolar disorder or depression,” they conclude. “This tool could reduce the amount of time patients and their doctors spend searching for a medication that works, possibly by months or years, while the patients continue to suffer from their illnesses.
“Although this technology would not be needed on patients with straightforward diagnoses, it could have a large impact on patients who have been labeled as complicated or treatment-resistant, as there are millions of people diagnosed with these conditions every year.”
