Identifying suicidal individuals by determining specific brain response to emotional words

Identifying suicidal individuals by determining specific brain response to emotional words

According to the Centers for Disease Control and Prevention (CDC), suicide emerged as the 10th leading cause of death in the United States in 2015 claiming the lives of over 44,000 people. Given the large number of casualties due to suicide, it is regarded as a public health crisis.

While suicide was the second leading cause of death among individuals between 15 and 34 years of age, it stood third among individuals between the ages of 10 and 14. Thus, suicide appears to be one of the biggest risk factors among adolescents and young adults. The trend is only increasing as noted by the National Institute of Mental Health (NIH).  Unfortunately, many people are still reluctant sharing their inner fears with others and continue to remain in the denial mode.

Moreover, the challenge increases when it comes to predicting or assessing those who are liable to commit suicide. In the absence of adequate awareness, one’s family or friends and even medical practitioner fail miserably at diagnosing his or her suicidal behavior. Moreover, people wrestling with such demonic feelings tend to hide them to avoid restrictive care.

In this direction, Marcel Just from Carnegie Mellon University and David Brent from the University of Pittsburgh have devised a potentially ingenious approach to identify suicidal individuals by using machine learning algorithms to evaluate the alterations in the brain. The primary ambition is to comprehend the way a suicidal individual thinks about suicide and related concepts by assessing the neural changes because of the specific suicide-related concepts.

Neural alterations on death-related topics predicts the risk of suicide

In order to make the clinical assessment of suicidal risk comparatively easier by integrating the above-mentioned biological measures, a study was conducted to determine suicidal tendencies by examining the neural changes on a set of death-related written words.

The study consisted of two groups of 17 people with the known suicidal tendencies and 17 healthy young adults. They were presented a list of 10 death-related words, 10 words associated with positive concepts like carefree and 10 words related to negative ideas like trouble.

The mechanism of machine learning algorithm was applied to six word concepts used for differentiating between the above-mentioned groups of participants. These words were death, cruelty, trouble, carefree, good and praise. By assessing the brain images of the participants exposed to these specific words, researchers were able to predict suicidal risk among them with 91 percent accuracy. Furthermore, they were able to differentiate participants who have attempted suicide earlier with 94 percent accuracy using the same methodology.

This approach has the potential to serve a useful purpose to clinicians in identifying, monitoring and intervening in the case of individuals having suicidal tendencies.  The only need of the hour is to apply the above findings on a larger sample.

Neural signatures for emotions

Often termed as explainable artificial intelligence, the machine learning approach was able to accurately detect suicidal risk among participants by 85 percent based on the changes in the emotions.

The emotions that people experience when they think about some of the test concepts are very different for people with suicidal tendencies. For instance, the concept of ‘death’ elicited more shame and sadness in the group than the concept of suicide. These understandings may assist in developing treatment that endeavors to change the emotional response of these individuals to certain concepts. According to the study authors, the immediate need is to apply these findings to a larger sample that will aid in predicting the future suicide attempts.

Ignoring mental disorders could prove fatal

Given the marked rise in the number of suicides, tools that help predict suicidal risks are necessary and long overdue. The advances in the field of neuroimaging and machine learning algorithms are making it possible to analyze the brain activity and predict the outcomes. Moreover, they can lead to better clinical decision-making and treatment approaches.

If you or your loved one is suffering from any mental health problem, contact the California Mental Health Helpline for more information about reputed mental health centers in California  offering comprehensive treatment programs. Call our 24/7 helpline number 855-559-3923 or chat online with one of the specialists to get advice on the best mental health treatment centers in California.