Thursday, May 17, 2007

The gay science

Over the centuries since Adam Smith, economists have developed mathematical frameworks for maximizing economic success. However, despite the intellectual power of these theories and the often simple logic involved in their calculations, humans continue to amass credit card debt, default on loans, fail to save for retirement, and on the whole refuse to do what these rational, reward-maximizing equations tell them to do.

The irrationality of human decision-making attracts the fierce interest of two very different fields: neuroscience and economics. Economic theories of human decision-making are essentially based on two parameters: what something is worth and the probability of its occurrence. Neuroscientists, on the other hand, think of decision-making as a product of physical neural circuits: sensory information enters the brain, journeys through the brain where a decision is "made," and eventually exits the brain to evoke bodily responses. Economics ignores these biological, more proximal roots of behavior, whereas neuroscience ignores the economic goals that ultimately guide our decisions.

These two approaches have recently been integrated in the hybrid field of neuroeconomics. Neuroeconomics attempts to unify abstract economic variables with neuroanatomy, and thus understand the physical mechanisms by which our brains make decisions. The basic premise is that somewhere along the sensory-motor circuit are the neural substrates that represent "value" and "probability." These areas must interact and influence the flow of information along the circuit, thereby prompting a certain decision and its subsequent behavior. The most pressing questions, then, are
how and where these abstract variables are combined in the brain, and the dynamics of the neural computation which engenders a "decision."

Inherently, neuroeconomics is not a means to exploit the free market by, for example, scanning the brains of consumers to calculate the maximum price they will willingly pay for a good. Although such endeavors are opportune beneficiaries of this sort of research, I believe neuroeconomics to have grander, more noble intentions. As a neuroscientist, I view neuroeconomics with bright, hopeful eyes, eager for the insight that economics can lend the neurobiological study of human behaviors. Although the former "dismal science" is abstract and far removed from biological mechanisms, it offers one thing behavioral studies tend to lack: great mathematical beauty.

Because economists base their models on optimal behavior, they have the ability to develop a precise, unified framework for interpreting human behavior; the thesis is, essentially, that humans choose alternatives that maximize rewards. Neuroeconomics draws upon the precision and rigor of the formal models of economics to go beyond the sensory-motor circuit, allowing opportunities for understanding the neural basis of more abstract economic ideas, such as value and the profitabilities of outcomes (a bit more challenging to study than sensory and motor systems). Thus, the principles of economics allows neuroscientists to explore the physical mechanisms underlying high level cognitive processes.

Particularly intriguing subjects for these studies are human choices that violate simple logic,; those which are neither selfish nor generous but blatantly, unbiasedly, irrational. I've previously explored irrational behavior in my post on risk aversion; another interesting example is "time inconsistency." When people make decisions about the distant future, they tend to behave as rationally as economic equations dictate. In contrast, when faced with the same decision relating to the near future, they are reckless and impulsive, unwilling to delay gratification. For example, when people are offered the choice of $20 now or $22 in a month, they often choose to receive the smaller amount immediately. However, if given the choice between $20 in a year or $22 in a year and one month, they will choose the higher, delayed amount. This is irrational; in both situations, the time delay (1 month) and financial gain ($2) are equal, so the decision should be the same (the higher amount should always be chosen.)

Another example of irrational impulsivity is less quantitative than the above, but involves a more flagrant demonstration of vice versus virtue. If offered the choice of a chocolate bar now or an apple now, most people demand immediate gratification and will choose chocolate. But if offered to receive a chocolate bar in one week or an apple in one week, people will consider the long-term effects of each and prefer the apple.

Back in 2004, Jon Cohen, Director for the Study of Brain, Mind, and Behavior of Princeton University, teamed up with George Loewenstein of Carnegie Mellon to take a neuroeconomic approach to this perplexing behavior. Using fMRI, they searched for changes in brain activity as the subjects made decisions between small immediate rewards or larger delayed rewards, attempting to link irrational displays of time-inconsistency with brain activity. The results, published in
Science, suggested that decisions involved with the possibility of immediate reward activated the limbic system, which is associated with emotion, while both short- and long-term decisions activated the prefrontal cortex (PFC), associated with logical, abstract reasoning.

Interestingly, when students had the choice of an immediate reward but chose the larger, delayed option, the PFC was more strongly activated than the limbic system. In contrast, when they chose the immediate reward, the activity of the two regions was similar (with a trend toward more activity in the limbic system.) This data suggests that both systems are involved in the neural representation of "value," and that the decision-making process is guided by, as the authors state rather poetically, "a competition between the impetuous limbic grasshopper and the provident prefrontal ant within each of us."

Thus, by exploring the neural processes by which the brain generates economic decisions, the authors were able to gain insight into the circuit-level computations that may govern complex behaviors. The extent to which the computations of economic theory can truly be generalized to the computations performed by the brain (as well as to more complex decision tasks) is unknown, but the aims and progress of this field are promising. From the economist's point of view, neuroeconomics may be far "messier" than economics, but the theoretical analysis of what humans should do isn't, to me, nearly as fascinating as understanding what they actually do, and neuroeconomics brings us far closer to reality.

McClure SM, Laibson DI, Loewenstein G, Cohen JD. Separate neural systems value immediate and delayed monetary rewards. Science 306(5695):503-7 (2004).