Home Science They develop a ‘suicide detector’ with Artificial Intelligence

They develop a ‘suicide detector’ with Artificial Intelligence

US scientists use Artificial Intelligence and brain imaging techniques to identify people at risk of suicide.

They develop a ‘suicide detector’ with Artificial Intelligence

Often, suicides surprise their intimates. Contrary to what one might think, a high percentage of people who decide to take their own lives do not show that they are going to do it or preach it.

Hence the importance of the method developed by researchers at Carnegie Mellon University and the University of Pittsburgh, who have combined brain imaging techniques and Artificial Intelligence algorithms to discern which individuals present suicidal tendencies, with an amazing 94 percent success. The authors of the work, published in Nature, hope that their findings will help save lives.

The algorithm that anticipates disaster

Scientists at these two American institutions have discovered that the neuronal activity of those who are at risk of killing themselves is different from that of other people. How did you find out?

They analyzed 34 subjects. 17 who had had suicidal thoughts at some point (nine of them had tried to kill themselves) and 17 who did not have mental health problems. The brain of each individual was scanned while presented with a list of ten words related to death, ten related to positive concepts – such as “carefree”, for example- and ten related to negative feelings, such as “problem”.

Then, they applied an algorithm of machine learning – a method that allows computers to learn on the fly – to six concrete concepts: death, cruelty, problem, nonchalance, good and praise. Based on the type of neuronal activity that these words triggered in each person, the Artificial Intelligence software detected with 91 percent accuracy whether the subject in question belonged or not to the group of possible suicide bombers. And what is more surprising: the program identified the nine individuals in the sample who had gone so far as to attempt to take their own lives in the past, and did so with an accuracy of 94 percent.

Vital prevention

“Our research is unique because it identifies brain alterations associated with the ideas of taking our own lives. Auto-learning algorithms help us create a neural representation of specific concepts related to suicide. This opens a window to the brain and the mind and throws light on the way people with this problem think about suicide and the concepts related to this emotion, “says David Brent, professor of Psychiatry, Pediatrics, and Epidemiology at the University of Pittsburgh.

“Next he adds will apply this method to a much larger sample of people and use it to prevent suicide attempts.  Remember that, often, people at risk of killing yourself thinking present alterations detectable with this technique”.

Lethal statistics

To get an idea of the potential importance of this finding, it is enough to mention information from the Mental Health Foundation Spain:  “Suicide is the main cause of unnatural death in Spain, producing more than twice as many deaths as traffic accidents, 7 times more than work accidents and 70 times more than gender violence, being also the first absolute cause of death among men from 20 to 34 years and the second, after tumors, in both sexes of those ages.

Although the highest number of suicides in both sexes occurs between 40 and 49 years, the risk of suicide increases with age, especially in males, where it multiplies by 5 compared to the earliest ages”.


Please enter your comment!
Please enter your name here