Developed a program to identify people with suicidal desire with accuracy of 91%, utilize brain scanning and machine learning

A computer program capable of identifying people with suicidal desires based on brain scan was developed. Researchers involved in development say that at this time the size of the research is small, but this program will be available for future diagnosis of human mental health.

According to the World Health Organization (WHO)About 800,000 people commit suicide each yearIt is said that someone is committing suicide somewhere in the world every 40 seconds. However, predicting suicide is also a very difficult thing. A research that addressed such a suicide problem was published in Nature.

In this research, in order to ascertain whether there is a difference in the brain activities of those with suicidal aspirations and those without suicide desires, the subjects were divided into two groups according to the desire of suicide aspiration, and words such as "evil" or "praise" were given to the subjects We are considering brain activity at that time and are conducting experiments. Based on the data obtained in this experiment, we succeeded in developing "Algorithm to hit people with suicidal desire with accuracy of 91%". Furthermore, this algorithm seems to be able to apply whether it tried suicide with an accuracy of 94%.

The algorithm developed by the research group is not perfect, it seems that medical examination is necessary. Also, because brain scan costs a lot of money, this method will not become widespread, says The Verge, an international news media. However, Marcel Just, a psychologist at Carnegie Mellon University, says, "It is a pleasure to have such a method as an additional option."

For the experiments, 34 participants including 17 people with suicide desires and 17 people without suicide wish are participating as subjects. All subjects werefMRIWe scan the brain waves using the scan, and they read about 30 words at the time of scanning, such as positive words (eg bliss), negative words (eg cruel), words that lead to death (eg suicide) is.

According to Just, when humans think about something, brain neurons are said to fire with a specific pattern. For example, the term "hammer" always has the same pattern to fire a neuron, and for another word, the neuron ignites in another pattern. Measuring the firing pattern of neurons means that we can obtain much more accurate information than other brain studies that only look at the area of the activated brain.

In this study, we compare the neuron firing patterns of the two groups to identify neuron firing patterns of people with suicidal desire. Then, it seems that the response pattern to the six words such as "death" "trouble" "distraction" "good" "praise" "cruel" was found to be greatly different. Therefore, using a machine learning algorithm, we analyze what kind of neuron firing pattern the subject's brain causes for each word. And we succeeded in developing a program that can predict whether the subjects examined for electroencephalograms are "people with suicide wish" or "people who are not." The program has succeeded in identifying those with suicidal desires and 91% of the precision, and using similar machine learning algorithms, "a person who has attempted suicide" and "a person who has tried suicide" We have also developed an algorithm to identify, and we succeeded in identifying with an accuracy of 94%.

Blake Richards, a neuroscientist at the University of Toronto, said the experimental results are quite interesting, but "There is definitely a biological foundation on whether someone commits suicide, all in our spiritual life There is also a biological base on the aspect of the biological foundation, but in order to actually develop highly reliable tests that can be used in the clinical setting, it is necessary to question whether the biological basis is sufficiently available by fMRI There is it, and it is said that at present it is unknown whether it can be used for actual diagnosis.

Also, researchers who participated in the experiment also acknowledged that the number of subjects is too small. Nevertheless, I believe that we will be able to use the program to diagnose people with suicidal aspirations in the future and to check whether treatment of mental disorders are functioning, so more subjects It is said that they are planning to conduct a research that collects.