Using Artificial Intelligence to Improve Mental Health?

Posted: May 3, 2018
Using Artificial Intelligence to Improve Mental Health

Foundation-supported researchers are featured in Wall Street Journal article

In a feature story in the April 29 edition of The Wall Street Journal, readers learned of exciting new efforts to apply powerful artificial intelligence (AI) technologies to help people with a variety of mental health conditions.  

Talk to a machine or screen instead of a doctor?  Apps – user-friendly computer applications – that run on personal computers and cell phones are already making it possible, on an experimental basis, for people with clinical depression to interact with AI-powered screen interfaces whenever they feel down or anxious or isolated.  One, called Woebot and featured in the article, asks the user a series of questions about “what is happening in his life this day and how he is feeling.” In an exchange that takes a few minutes per session, the program behind the app analyzes the way the patient frames his replies “and provides coping strategies drawn from cognitive behavior therapy.”  CBT is a widely practiced form of talk therapy that focuses on identifying and solving an individual’s current problems, in part by surfacing critical patterns of thought and behavior that tend to work against the individual. 

Two Foundation grantees figure prominently in the story, exemplars of different strategies to apply computing power and Big Data in the effort to help improve treatments for depression and other mental health conditions.  One is Conor Liston, M.D., Ph.D., of Weill Cornell Medical College, a 2013 Young Investigator and 2016 Freedman Prize honorable mention. Research led by Dr. Liston and published last year in Nature Medicine used artificial intelligence to find distinct patterns of depression – pulling these from a mountain of data gathered via brain scans of nearly 1200 people, only some of whom were depressed.  The four depression subtypes identified by Liston and colleagues are already helping clinicians determine which patients stand to benefit most from particular therapies, as we describe here and here.  

2002 Young Investigator David Fleck, Ph.D., of the University of Cincinnati, was also mentioned in the story.  He’s developing an AI system that would help determine which patients with bipolar disorder will be helped by lithium, widely prescribed but not effective in all patients. 

Although still in its infancy, AI technology as applied to improving mental health, and especially to our understanding of how to personalize treatments, has generated results that are “intriguing,” in the words of Dr. Joshua Gordon, the Director of the National Institute of Mental Health, a member of the foundation’s Scientific Council and in 2003 and 2001 the recipient of Young Investigator grants.  How helpful they will be will be determined in the period ahead, only with trials conducted with large numbers of patients over extended periods of time.