Training an artificial intelligence computer program to both recognize and identify emotions, a professor of psychology at Carnegie Mellon believes that how the brain processes certain words can be an important indicator of suicidal tendencies in an individual.

The potential to identify and possibly prevent suicidal behavior using such a system is valuable considering that the only current method to determine if an individual is suicidal is simply asking them, according to Professor Marcel Just. However, that method is typically unreliable.

“Nearly 80% of patients who die by suicide deny suicidal ideation in their last contact with a mental health care professional,” according to the study.

Using an MRI machine in conjunction with the computer program, Just observed the brain scan reactions of 34 participants (evenly split between those who said they were suicidal and those who said they were not) to a list of 30 different words. With a mixture of words both positive and negative, Just noted that those who previously attempted suicide responded to words that related to death with less sadness than those who hadn’t previously attempted suicide.

According to the research, the system was 91 percent accurate in identifying the suicidal participants from the non-suicidal participants. However, Just realizes that the system won’t prevent all suicides.

"This isn't a wild pie in the sky idea," Just said. "We can use machine learning to figure out the physicality of thought. We can help people.”

The study is published in the journal Nature Human Behavior.