10.02.2005

Risky Business

Imagine a woman with little financial savvy, who after saving some cash is looking for a good investment. For a safe bet with small payoffs, she could buy government bonds. But if she’s willing to lose it all for a chance at fortune, she could instead buy shares of a volatile stock. Either way, her brain must somehow evaluate her economic circumstances and take action. As Daniel Bernoulli, the Dutch physicist and mathematician who is perhaps best known for his principles of fluid dynamics, explained in a 1738 paper on economic theory: “…there is no doubt that a gain of one thousand ducats is more significant to a pauper than to a rich man, though both gain the same amount.” The woman’s choice might depend upon her job security, for instance, or the number of children she has to support, or even the number of years she has left to enjoy those potential earnings. New research in monkey brains is coming closer to decoding the brain signals that ultimately drive such decisions. In a study published in this month’s Nature Neuroscience, researchers at Duke University found patterns of neural signaling that reliably indicate the riskiness of a choice. Their findings may lead to mathematical models of human risk behaviors and a better understanding of why we sometimes make irrational and even harmful decisions.

How do researchers watch and manipulate monkey decisions? Most begin by teaching them to play a simple game on a computer screen, called a visual gambling task. Two circles appear on the screen, and the monkeys are trained to pick one of by looking directly at it. Eye tracking sensors attached to their heads tell the researchers which circle they have chosen and connect to a feeding tube that instantly pays them a fruit juice reward. Old studies have focused on the objective factors that influence their decision, like the size of the reward, because they’re much easier to measure than subjective factors like how thirsty the monkey is or how much it enjoys the thrill of taking a risk. In 2003, for instance, neuroscientist Allison N. McCoy found that when given the choice between a circle that gave a small juice reward and another that gave a large reward, monkey brain responses were stronger when they chose larger rewards. These results are perhaps not so surprising—a $20 million jackpot is, after all, more exciting than a $20 scratch card.

But little had been done on the more difficult testing of subjective factors–until now. In the new Duke study, McCoy collaborated with Michael L. Platt to tackle the question of what choice would be made if they kept constant the objective value of the reward—i.e. the amount of juice the monkey would get after many trials—but varied the riskiness of its choice. By looking at one of the circles, the “certain target,” the monkey would always get a fixed amount of fruit juice. By looking at the other “risky target,” though, it had a 50:50 chance of getting a shot of juice that was either smaller or larger than the certain target. In other words, to look at the risky target was to take a gamble.

And McCoy and Platt were able to systematically change the size of the gamble. Say the certain target paid out a 50-ml reward and--because it’s certain-- this was paid every single time the monkey looked at it. But the risky target, remember, didn’t always give the same reward. On”high-risk” trials, a monkey choosing the risky target got either a 75-ml reward or a 25-ml reward. In contrast, in “low-risk” trials, looking at the risky target reaped either a 45-ml or a 55-ml dose. In all cases, the average pay-off over many trials was the same—using these hypothetical numbers, 50-ml. So you wouldn’t expect the monkeys to prefer one option over another. But here’s the kicker: McCoy and Platt’s monkeys overwhelmingly chose the risky target. What’s more, they chose it more often on the high-risk trials.

In the next part of the study, the neuroscientists looked for patterns of neuron firing that correlated with these preferences. They inserted an electrode—a tiny needle that records the number of electrical signals produced by a single neuron-- in a long and narrow brain region spanning the top of the skull to the top of the ear, called the posterior cingulate cortex. This area was chosen for recordings because it was known from previous studies that its cells were highly responsive during the activities—seeing shapes on a screen, moving the eyes quickly, receiving a reward—involved in the visual gambling task. McCoy and Platt recorded the electrical patterns generated while the monkeys performed the visual gambling game. Not only were the neuronal firings higher when the monkeys took risks, but they also fired in a predictable code: as risk was systematically increased, so too was the firing frequency. In short, they were exquisitely sensitive to risk.

These primate results could be easily translated into human terms: monkeys prefer risk, therefore humans prefer risk. But human behavior is complex, and the temptation to assume that these findings will lead to the ability to predict human behaviors should be squelched. As Daeyeol Lee put it in a Nature commentary, “an individual might insure a car used to drive to the casino.” Moreover, even given identical circumstances, every brain has a slightly different inclination toward risk; not every casino patron becomes a pathological gambler, and not every teenager who picks up a cigarette goes on to smoke a pack a day. Like most of today’s neurological research, these results are remarkable mostly because they expose systematic and surprising patterns at a microscopic level; though the authors did hint their study may have some human application “as an important model for probing the neural processes that underlie pathological risk taking in individuals with addictions to drugs, sex, food or gambling.” As for the woman looking to invest—nothing ventured, nothing gained?