Preventing suicides with AI
Most people who think about killing themselves never reveal their plans to anyone, making suicidal behavior extremely difficult to predict or prevent. But a new study suggests artificial intelligence could address this problem, through brain scans that effectively predict who is most likely to attempt suicide. Researchers have developed a machine-learning algorithm that uses brain scans to identify those with suicidal thoughts. They tested the technology on 34 young adults, including 17 who had a history of suicidal thoughts or behavior. The participants were shown 30 words as their brains were scanned: Some were related to suicide, such as “death”; others were feelings, such as “desperate” and “carefree.” By analyzing brain patterns, the algorithm was able to distinguish between the people who had considered suicide and those who had not with 91 percent accuracy, NPR.org reports. Lead author Marcel Just, a cognitive neuroscientist at Carnegie Mellon University, says, “There really is a difference in the way [suicidal] people think about certain concepts.”
reports NBCNews.com. After considering variables such as the participants’ age, education, and whether they smoked, they found that the men who frequently ate alone were 45 percent more likely to be obese and 64 percent more likely to develop metabolic syndrome—a cluster of risk factors associated with heart disease, diabetes, and stroke. The women who often ate solo had a 29 percent greater risk for the condition. One possible explanation: loneliness. Those who ate alone tended to be single and live on their own. “Who wants to cook a whole meal for one?” says Andrew Abeyta, a psychology professor at Rutgers. “People who eat alone are more likely to eat unhealthy fast food or foods that, like frozen or boxed foods, are quick to prepare.” Social isolation also exacerbates the effects of stress, which increases the risk for heart disease and other health issues.
relieve chest pain. They then had an operation to have a stent inserted—but only some actually had the device fitted. When the researchers examined the patients six weeks later, both groups said they had less chest pain and showed similar levels of improvement on treadmill tests. Cardiologists said one reason stents might not be effective—except as placebos—is because blockages occur in many other blood vessels besides the artery. Some scientists downplayed the study, arguing that the assessment period should have been longer. But others said a procedure as invasive as stenting should be used only to clear extreme blockages during heart attacks. “For someone who puts in stents,” says Brahmajee Nallamothu, a cardiologist at the University of Michigan, “it’s a very humbling study.”