MIT and Massachusetts General Hospital Researchers Are Developing Machine Learning Algorithms to Provide Physicians and Patients with Early Recognition of Mental Health Changes

Physicians treating patients with major depressive disorder may soon be able to include personalized treatments designed specifically for each patient’s medical needs

Researchers from Massachusetts Institute of Technology (MIT) and Massachusetts General Hospital (MGH) are developing a machine learning algorithm designed to provide early recognition of mental health changes to both patients and physicians, enabling early intervention and personalized treatments.

This five-year study, funded by the National Institutes of Health (NIH), is nearing completion. The researchers anticipate that new technological tools developed as a result of their study will improve patients’ access to mental health care and enable healthcare givers to provide more effective psychiatric care, according to an MIT News article.

MIT professor Rosalind Picard, Sc.D., and Assistant in Psychology at MGH Paola Pedrelli, PhD, have partnered together to make this new artificial intelligence (AI) program a reality. Picard is a leader in the field of affective computing and a machine learning expert. Pedrelli is Director of Dual Diagnoses Studies and an Instructor in Psychology at Harvard Medical School.

Prof. Rosalind Picard, Sc.D., of Massachusetts Institute of Technology (MIT), and Paola Pedrelli, PhD, of Massachusetts General Hospital (MGH) and Harvard, are studying major depression using using their expertise in affective computing and dual diagnoses studies.
Rosalind Picard, Sc.D. (above left), of MIT, and Paola Pedrelli, PhD (above right), of MGH and Harvard, are developing AI algorithms that, they hope, will give patients struggling with mental health issues greater access to quality healthcare. “We’re trying to build sophisticated models that have the ability to not only learn what’s common across people, but to learn categories of what’s changing in an individual’s life,” Picard explained in an MIT News article. (Photo: MIT/MGH)

Leverages Person-Generated Wearable and Smartphone Device Data

Picard and Pedrelli are four years into their study, which follows patients with major depressive disorder, and who had recently changed their treatments. The patients wear specially designed wristbands that monitor biometric data. Additionally, the study participants download an app onto their smartphones that collects data on movement, calls, and text frequencies, as well as provides a biweekly depression survey.

Study participants then meet with clinicians weekly to have their depressive symptoms evaluated.

“We put all of that data we collected from the wearable and smartphone into our machine-learning algorithm, and we try to see how well the machine learning predicts the labels given by the doctors,” Picard said.

The researchers believe their machine learning algorithm could make healthcare more accessible for patients living with mental illness.

“It’s been very, very clear that there are a number of barriers for patients with mental health disorders to accessing and receiving adequate care,” Pedrelli said. One of the biggest of these barriers is knowing when to seek help, a barrier that Picard and Pedrelli believe their technology will help overcome.

Algorithm, Do No Harm

One of the more significant problems that the researchers are focused on solving is developing ways to clinically implement the data.

“The question we’re really focusing on now is, once you have the machine-learning algorithms, how is that going to help people?” Picard asked. “I think there’s a real compelling case to be made for technology helping people be smarter about people.”

The researchers acknowledge that, while there are definitive benefits to being able to accurately predict mental health changes, the way this information is used could actually cause harm. Warning someone with depression that their depression is beginning to worsen, for example, could actually cause depression symptoms to worsen more than they would have, the researchers noted in the MIT News article.

“What could be effective is a tool that could tell an individual ‘The reason you’re feeling down might be the data related to your sleep has changed, and the data related to your social activity, and you haven’t had any time with your friends, your physical activity has been cut down,” Picard explained. “The recommendation is that you find a way to increase those things.”

Forward looking healthcare leaders with an interest in treating psychiatric conditions will want to follow the results of this research as it enters its last months. The ability to use machine learning-based precision medicine to identify mental health changes as early as possible may enable clinicians to provide patients with personalized care for their depression. It could save lives in severe cases of major depressive disorder.

—Caleb Williams

Related Information:

Deploying Machine Learning to Improve Mental Health  

Rosalind Picard, Sc.D.

Paola Pedrelli, PhD

Interviews With Precision Medicine Movers

Browse All Briefings