Russell Poldrack, director of the University of Texas, Austin's Imaging Research Center, and colleagues analyzed data from 108 subjects who sat in a magnetic resonance imaging scanner while playing a video game that simulated risk-taking.
The researchers used specialized software to look for patterns of activity across the whole brain that preceded a person's making a risky choice or a safe choice in one set of subjects. Then they asked the software to predict what other subjects would choose during the game based solely on their brain activity. The software accurately predicted people's choices 71 percent of the time, Poldrack said.
"These patterns are reliable enough that not only can we predict what will happen in an additional test on the same person, but on people we haven't seen before," Poldrack said in a statement.
When the researchers trained their software on much smaller regions of the brain, they found just analyzing the regions typically involved in executive functions such as control, working memory and attention was enough to predict a person's future choices.
The researchers concluded when people make risky choices it is primarily because of the failure of the body's control systems to stop.
"We all have these desires, but whether we act on them is a function of control," said lead author Sarah Helfinstein, a postdoctoral researcher at UT Austin.
The findings were published online the journal Proceedings of the National Academy of Sciences.