Official records say that in the UK in 2016, 4,508 men and 1,457 women died as a result of suicide, but some experts believe the true number may be double that. Men appear particularly vulnerable: in fact, suicide is the leading cause of death in men under 50 in the UK, claiming more lives than car accidents, heart disease or cancer.


The reasons so many men take their lives are a complex web of social, psychological, biological and cultural pressures. But new scientific approaches are presenting unexpected avenues for untangling the threads. Virtual reality experiments and artificial intelligence are revealing those most at risk, and presenting the prospect of predicting who is likely to try and take their life. Together, they present the prospect of better prevention.

Stepping into a virtual world

According to Prof Rory O’Connor, who runs the Suicidal Behaviour Research Lab at the University of Glasgow, changes in society are making men especially prone to feelings of entrapment that seem to be a key driver to suicide. His lab works with suicide survivors and conducts studies to find links between suicide and psychological and social characteristics.
Proving such links is not easy. But American psychologist Dr Joe Franklin, who heads the Technology and Psychopathology Lab at Florida State University believes he may have found an answer by probing the causes of suicide using virtual reality and a form of artificial intelligence called machine learning.

Franklin’s team exposed their test subjects to standard psychological scenarios designed to make them feel mildly socially rejected. Then they put them into a virtual reality scenario in which they’re standing on top of a high building. Some of those who had been socially rejected chose to jump.

There are potentially thousands of factors that contribute to suicide. But the only way of sorting out which of those factors is significant is by using machine learning.

China's Big Brother © Alamy

“You give the machine every bit of information you have,” says Franklin. “You say: we have 500 people who died of suicide, another 500 who didn’t. Here’s 2,000 bits of information about them all. Now sort out the best algorithm for pulling those groups apart.” This system could be plugged into national electronic health records to find patterns of contributing factors and to identify individuals’ suicide risk.

More like this

Choke Points

Amid the complexity, the data from virtual reality experiments and machine learning is likely to reveal psychological ‘choke points’, says Franklin, where preventative action may work on many fronts. There’s already evidence about the effectiveness of some public health choke point initiatives, effectively making it harder to commit suicide. The number of paracetamol overdoses in the UK fell significantly when a limit was placed on the number of tablets each customer was permitted to buy.

Investigating how the latest tech can be used to identify those most at risk will be at the top of the to-do list for Conservative MP Jackie Doyle-Price in 2019. As the newly appointed minister for suicide prevention, she is leading the government’s efforts to cut the number of suicides and overcome the stigma that stops people seeking help.

Suicide Statistics

  • 878 male students took their lives between 2001 and 2017
    (England and Wales)
  • 45% of trans young people have tried to take their lives (UK)
  • 12 men kill themselves every day (UK)
  • 3 in 4 suicides are by men (UK)

How to find help

  • Samaritans is a safe place for anyone to talk about difficult feelings, 24 hours a day. Phone free (UK/ROI) on 116 123 or email
  • The CALM helpline is for men in the UK who need to talk or find information and support. Open 5pm until midnight 0800 58 58 58


Follow Science Focus on Twitter, Facebook, Instagram and Flipboard


Simon is a freelance science and health writer