Decision making and planning - review

One of the great unsolved mysteries of science is how the human brain can support the complex, coordinated, and intelligent behavior of which we are capable.

One of the great unsolved mysteries of science is how the human brain, built from a myriad of small processing units (neurons) with no single one in charge, can support the complex, coordinated, and intelligent behavior of which we humans are capable. How can this hodgepodge of hundreds of billions of noisy cells organize itself to play a game of chess, write a novel, or plan the course of a career? More remarkable still is that each individual neuron operates at about one tenmillionth the speed of an average desktop computer. As yet, we have relatively little insight into the answers to these questions. However, certain themes have begun to emerge, both from traditional behavioral and neuroscientific research and more recent studies using brain imaging technology. These findings are helping shape the way we think about the unique faculties of mind that make us human. First, it is clear that the capacity for virtually all higher cognitive functions, such as reasoning, problem solving, and language, as well as decision making and planning, rely heavily on two fundamental functions discussed our section on inhibition and control. Searching for the solution to a complex problem, interpreting the meaning of a conversation, or formulating a course of future action are all forms of goal-directed activity. They rely on a person’s ability to pursue an intended behavior even when more compelling but inappropriate alternatives are available. Classic examples of this include a teenager learning to study when he or she could be socializing, and a boss coaching employees gently rather than badgering them in order to create the most productive workforce. Our frontal lobes seem to play a critical role in this capacity for control and are therefore likely to play an equally important role in the higher faculties that depend on control. Indeed, damage to the frontal cortex dramatically impairs the capacity for sensible decision making and planning. Recall the personality changes in Phineas Gage, the unfortunate railroad foreman who suffered damage to his prefrontal cortex. Originally considered to be thoughtful, responsible, and of sound judgment, he became “capricious . . . and unable to settle on any of the plans he devised for future action.” Since the time of Phineas Gage, doctors have continued to observe similar changes in patients with damage to the frontal cortex, which often results in their inability to properly plan even the simplest of activities, such as cooking dinner. More recently, brain imaging has shown increased activity in the prefrontal cortex when people engage in tasks that rely heavily on decision making and planning, such as playing chess, or in solving problems that require them to evaluate the outcome of a complex set of alternatives. These studies have not yielded any specific insights yet into precisely how the frontal cortex carries out these functions. Nevertheless, the opportunity presented by imaging machines to study the brain’s normal workings noninvasively and with ever increasing detail promises, with time, to reveal critical information about our decision-making and planning processes. Decision-Making Difficulties Another important theme that has begun to emerge about our decision-making behavior is that we do not always—or even usually—conform to purely logical principles. This may not come as a surprise. We say with some pride that “we aren’t computers,” whose workings are based on mathematical logic. On the other hand, we humans also pride ourselves on our ability to reason. We believe that ability distinguishes us from our evolutionary ancestors, and we often hold rationality among our highest ideals. Indeed, the idea that human behavior is on average rational lies at the heart of traditional economic theory, which assumes that individuals seek to maximize their rewards in the best possible way. So ideally we should draw on those reasoning powers for every decision. It is becoming increasingly apparent, however, that ordinary human behavior is often far from optimally rational. Both psychologists and economists have identified many situations in which people’s thinking seems to be illogical or inconsistent— and consistently so! Furthermore, such deviations from optimality may not just be due to faulty reasoning, but may reflect specific adaptations of underlying neurobiological mechanisms. To explore how this might work, let us consider some examples of less-than-optimal human judgment and decision making. In the 1970s the psychologists Daniel Kahneman and Amos Tversky started a remarkable series of studies examining human decision making, thus creating some of the most influential research in modern psychology. Kahneman and Tversky showed that humans systematically misestimate the probabilities of certain types of events and therefore make logically incorrect decisions. A common example is the general sense many people share that airplane travel is more dangerous than car trips. In fact, the reverse is true. Yet even though the real statistics have been widely publicized (not least by the airline industry), many people are still more afraid of flying than of riding in a car. Kahneman and Tversky attributed this erroneous feeling of risk to the fact that people regularly overestimate the frequency of highly salient events like airplane crashes. That is, when an event is highly noticeable—it receives a lot of media attention, affects a large number of people all at once, and is emotionally arousing—it is more likely to “make an impression” on our memories. That event then influences people’s future judgments and decisions out of proportion to its actual likelihood of happening again. Another form of “wayward” reasoning occurs when people must decide between alternatives. Notwithstanding traditional economics, this can be very difficult for the human brain. For example, in one experiment researchers asked people to choose between two objects that they could keep (for example, two types of pens); if they preferred not to choose, they would instead receive a small reward worth significantly less than either object. When one of the objects was clearly more desirable than the other, the choice was easy; virtually everyone chose the more valuable object. However, when the two objects were roughly comparable in apparent worth, people often decided not to choose and took the reward instead—even though it was the least valuable alternative. Clearly such decision making was not optimal. Even if they could not tell which object was more valuable, completely rational humans would simply have chosen one or the other at random and come out ahead. Instead, people avoided the close decision. This behavior implies that the human brain is programmed to believe that deciding between apparently equally valuable alternatives has a cost. You may have experienced that phenomenon in your own life. Consider the difficulty of having to choose between two attractive plans for an evening, such as whether to go to a ball game or a movie. That situation seems better than having only one of the options, or neither. Yet the decision, paradoxically, can be more difficult. How often in such situations have you thought, “Oh, I’ll just stay home”? Evolution might explain that seemingly irrational form of decision making. Difficult decisions require time, and in the wild such time can be costly. An animal might do better to make a quick decision based on a simple rule, such as “A bird in the hand is worth two in the bush.” Our brains may thus have evolved mechanisms that avoid uncertainty and lead to more assured, though less than optimal, results. This approach may no longer be as useful to us today, and therefore can appear illogical in particular situations. However, in the circumstances under which such a response evolved, it may have been quite reasonable. But that explanation is only speculative. Both studies and everyday life also reveal the profound influence of emotions on our decisionmaking processes. Perhaps one of the most striking examples of this was the chess match between Garry Kasparov and the IBM computer Deep Blue. Kasparov was the human chess champion of the world, and Deep Blue was the computer chess champion. Many touted this as the ultimate intellectual showdown between human and machine: a contest to determine whether a device built from flesh and blood or from silicon could lay claim to the most powerful reasoning abilities on the planet. Deep Blue surprised everybody and won, which might have been taken as evidence that it possessed the superior capacity for the logic needed to play chess. However, chess pundits also reported that Kasparov was not playing at his best; in fact, he made several uncharacteristically impetuous moves. In interviews during the match, Kasparov seemed focused— even obsessed—with the fact that he was playing a computer. It appears that the champion may simply have “freaked out”: the unusual situation evoked emotionally disturbing reactions that interfered with his ability to think clearly and play chess at his best. Thus, even a task that would seem to tap only pure, specialized human reasoning in a person who had most honed that ability was subject to emotional factors. In other words, emotions can insinuate themselves into our every thought and affect our every decision. When Time Comes to Decide Connecting complex behavior with mechanisms in the brain is difficult, and most studies of human decision making have relied on two approaches. First, as with many other brain functions, abnormalities can reveal as much or more than normal functioning. Studies often start with what we see when a part of someone’s brain is damaged or the normal neurochemical balance is altered. Second, by posing a controlled, artificial challenge, such as a quiz or card game, to volunteers in the laboratory, researchers can observe their brains at work during the precise moment they are making up their minds. Studies of patients with damage to the prefrontal cortex confirm the importance of this region in human decision making. Neurologist Antonio Damasio has reported the behavior of people with damage to a particular region of the frontal cortex—the orbital frontal area, along the lower surface of the frontal lobes, just above the eye sockets, or “orbits”—in gambling tasks. His research team created a game that rewarded a conservative strategy over going for an immediate payoff. Most volunteers were able to play accordingly, but those with damage in the orbital frontal area had difficulty resisting the temptations of an immediate reward. Observations of the brain in action seem to confirm that the orbitofrontal cortex plays a critical role in evaluating potential rewards. The first of these studies used direct recordings of animals’ neurons in that area. Modern neuroimaging methods have begun to provide similar evidence in normal human brains as people play games very similar to those Damasio used to test his patients. Since parts of the prefrontal cortex are also crucial to our emotional processes, that suggests how emotions so easily affect the decisions we make. It also seems significant that the orbitofrontal cortex is largely guided by dopamine; the system for that neurochemical signal is hijacked when people abuse drugs, which obviously affects their decision-making abilities. Of course, the orbitofrontal cortex is not the only brain system involved in emotional processing. Both animal and brain imaging studies show that the amygdala—a specialized structure along the inner surface of the temporal lobes—plays a vital role in evaluating the emotional relevance of what a person’s body is sensing. This is especially so for fearful stimuli, but may also be true for other types. Studies of the amygdala suggest that this structure may be especially important in detecting emotionally significant events (for example, events that should cause us to prepare for a “fight or flight” response), while other structures, such as the orbitofrontal cortex, may be more important for the careful or deliberative consideration of such events. When those other brain structures are damaged, the influence of the amygdala on decision making may become stronger. Another brain structure that has attracted significant attention is the anterior cingulate cortex— a part of the frontal lobes that lies along their midline, just above the corpus callosum. Brain imaging studies have shown that in normal individuals this structure becomes active in response to pain and negative feedback, but it is abnormal in patients who suffer from anxiety or depression. Recent studies have revealed that the anterior cingulate also responds when peoplemake errors in simple cognitive tasks, and even when they respond correctly to a question but cannot be sure they are correct. More generally stated, it appears that the anterior cingulate responds in situations associated with poor performance: uncertain decision making in challenging tasks, committing an error, receiving negative feedback, and experiencing pain (the ultimate consequence of poor performance!). This suggests that the anterior cingulate may monitor our internal decision-making processes for signs of deteriorating performance and the dangers that could ensue, much as the amygdala is thought to monitor the external world for signs of threats. Intriguingly, the anterior cingulate could be the evolutionary adaptation proposed earlier: the mechanism that keeps our brains from spending too much time considering a difficult choice. But, as stated before, that is just speculation. A Glimpse of Higher-Level Reasoning The brain systems just mentioned provide some insight into the apparent idiosyncrasies of human decision making and how these systems interact with emotional processes. It is a big step from these simple processes to the more complex areas of social interactions and higher-level reasoning. Neuroimaging studies are giving us our first glimpse of the brain systems engaged in such higher-level cognitive processes. For example, recent studies have examined the patterns of brain activity of people during a game that involves cooperation and trust, comparing those playing against a computer with those playing against another human being. Comparison of the two groups shows significantly different areas of brain activity, and many of the differences involve the areas that process emotions. Moral reasoning is a field that we humans often consider our highest intellectual achievement. It is another area in which we exhibit patterns of behavior that have consistently puzzled philosophers and psychologists alike. It is not simply a matter of people holding different value systems, or of someone being “immoral.” Even the same individual can reason quite differently when presented with dilemmas that appear different but are in fact similar. For example, consider the following scenarios: ■ A trolley is headed toward five unsuspecting workers. A switch operator can avoid killing them all by moving the trolley to a sidetrack, on which there is only a single worker who will die. ■ A trolley is headed toward five unsuspecting workers. There is a large man on a footbridge just above the trolley, and pushing this man off the bridge onto the track will cause the trolley to run over him and derail, saving the five workers. Most people agree that it is morally justifiable, even imperative, to switch the train but impermissible to push the man off the bridge. Yet the logic of the two circumstances is the same: sacrificing one life will save five others. There are many possible explanations for inconsistent choices in these two scenarios, and others researchers have imagined, but philosophers have failed to identify any general logical principles that can account for people’s judgments under all the variants of those circumstances. This has led some to believe that people’s moral judgments rely on a mixture of logic and emotion that varies with the specific nature of the circumstances. Thus, people may react more emotionally to the thought of physically pushing a man off a bridge to his death (even if it is to save five others) than they do to the thought of flipping a switch, which is an impersonal object. As with simpler decision making, the paradox posed by these scenarios is reflected in the way the brain responds to each. A recent imaging study, described in the box to the right, has shown that the footbridge scenario activates emotional areas of the brain, whereas the switch scenario activates areas associated with more abstract processing. Again, our brains may have evolved to work the way they do because those adaptations were valuable in other circumstances. Perhaps our emotional reactions to directly harming people serve important social functions, prohibiting forms of aggressive or antisocial behavior that would challenge the species as a whole. Planning Ahead If deciding between options is an intricate brain action, planning is even more so. Planning is, in a way, an advanced form of decision making. It involves not merely making choices but also imagining what we want to do, what we need to accomplish that goal, and what might get in our way. It requires keeping those goals in mind over time. The function of consciously planning for the future instead of simply reacting to our environment according to instinct may be the most important ability we humans enjoy. Once again, we have relatively little knowledge about the specific ways in which a human brain manages to plan a vacation, an election campaign, or any other complicated endeavor. However, we do know that this process relies heavily on the prefrontal cortex, most probably in close interaction with structures responsible for long-term memory and storage, such as the hippocampus (an area along the inner sur-face of the temporal lobes, very close to the amygdala). These brain systems are responsible not only for allowing us to organize our behavior to deal with the complexity of daily life but for setting and pursuing long-term goals that can last a lifetime. Not surprisingly, these functions take a long time to develop to mature form. It is not simply an illusion that young children and adolescents have little sense of long-term planning; their brains are still developing that skill. In fact, with the prefrontal cortex not fully developed until late adolescence, children may not even be biologically ready for long-term planning. Perhaps the most important aspect of planning is that it involves the ability to carry out actions well into the future, long past when we first conceive of them. For instance, a person may decide to earn an advanced degree in order to enter a new field, with an eye to eventually running his or her own business. That plan covers years, and it would be impossible for anyone to be actively working toward those goals every minute of the day. Indeed, there are long stretches in which the person will be engrossed in immediate tasks and not be thinking about his or her career plan at all. The ability to stick to a plan therefore depends on our brain’s long-term memory. The prefrontal cortex maintains a mental representation of the goals of a person’s current behavior (to balance the checkbook, to serve dinner), while the hippocampus and other parts of the brain invoke mental processes and actions related to career pursuits only when they are needed. A simple example illustrates how the planning process may work in a much shorter time frame. Consider this situation: you wake up in the morning and go to the refrigerator to make breakfast, only to find that you are out of orange juice. You make a plan to stop at the grocery story on the way home from work. Obviously, you don’t keep this plan actively in mind (that is, in your prefrontal cortex) all day long, repeating to yourself over and over, “I have to go to the grocery store at six o’clock.” You devote your conscious thoughts to more immediate tasks, and to working on plans you’ve made as part of your work. Your ability to carry out Operation Orange Juice depends on its being stored in your mind and reactivated at the appropriate time. Here’s one model of how that reactivation may occur: when you make a mental note to stop at the grocery store on the way home from work, an association forms in your hippocampus between the steps of that plan (for example, turn left at the light rather than making the habitual right toward home, park in the store lot, and so on) and the circumstances under which it should take effect. Cues for activating the plan might include reading six o’clock on your watch, seeing the sun go down, and feeling your stomach growl. Whatever the association is, it allows you to put the plan “out of ind” (that is, to deactivate its representation in the prefrontal cortex). The day goes on, and you don’t think about it. Then six o’clock rolls around, the sun goes down, and your stomach growls. Because these cues are linked with your plan to buy orange juice, they activate the representation of that plan in your prefrontal cortex, and you think to yourself: “Gotta go to the grocery store.” Activation of this plan also ensures that you make the left turn at the light rather than the default right turn home. We do not know for sure that these interactions are what actually happen in the brain when you make a plan, nor whether similar interactions occur when you make and follow even longerterm plans, such as for a family vacation or a new career. However, we do know that complex behaviors like planning are likely to involve intimate interactions between different brain regions, each of which contributes critical and complementary functions. This model of interaction between the hippocampus and the prefrontal cortex provides an example of the hypotheses about brain function that are beginning to drive research in neuroscience and psychology. Using the tools now available to test such theories, such as brain imaging, we can develop new understanding of the brain mechanisms underlying some of our highest and most cherished faculties.