Machine learning may allow researchers to identify people suffering from suicidal thoughts, a new study in Nature Human Behavior reports.
In the study, a team of researchers from various U.S. universities looked at 34 subjects, half of which were experiencing suicidal thoughts. They then put the participants under functional magnetic resonance imaging (fMRI) and showed them words related to suicide, such as “death” and “distressed.” At the same time, they also flashed words related to positive and negative emotions.
Based on the subjects already identified as suicidal, the team managed to identify five regions of the brain and six different words that were connected to suicidal tendencies. They then used that information to train an algorithm to identify suicidal patients. Of the 34 participants, the algorithm correctly identified 15 of the 17 suicidal subjects and 16 of the 17 members in control group.
Researchers also ran a second experiment where they divided participants into two groups: one that had attempted suicide and one that had not. A separate algorithm correctly categorized 16 of those 17 patients as well, Gizmodo.
This research is important because it gives the first insight into how the brains of healthy and suicidal people differ. For example, the word “death” triggered a reaction in the region of the brain associated with shame.
“Suicidality isn’t that you can’t cope with life; it’s that you’ve somehow gotten into a pattern of thinking that leads you to consider suicide,” explained lead author Marcel Just, a professor of psychology at Carnegie Mellon University, according to Mashable.
This new study could provide a new tool to help diagnose and treat different mental health disorders. While there has been a lot of information gathered on psychiatric illnesses, many of the disorders have vague diagnoses. Understanding where in the brain those problems first occur could then lead to more targeted, and more effective, treatments.
Researchers plan to conduct more studies with the new technology in the near future.
“The most immediate need is to apply these findings to a much larger sample and then use it to predict future suicide attempts,” said study co-author David Brent, a researcher at the University of Pittsburgh, in a statement.