Neuroeconomics is currently one of the fastest moving bandwagons in systems neuroscience. If you are a neuroscientist trying to decide whether or not to clamber aboard, or, if on board already, whether to jump off, or if you are just trying to work out where the bandwagon is going and whether it might be about to hit you, then a couple of weeks with Neuroeconomics: Decision making and the Brain might enable you to make a decision. For neuroeconomists, behavior consists of decisions and learning how best to make decisions. Neuroeconomics attempts to describe not just why one choice might be the one to take but also to elucidate the neural mechanisms that bring about such decisions. Of course, as with any bandwagon, it always feels important to keep up with what everyone else is doing—and, by reading Neuroeconomics, you will start to understand why this may be so. The book has a major emphasis on understanding one of the most important influences on one's own decision making, namely the decisions taken by others. There is, in some quarters, confusion about what constitutes neuroeconomics. The inclusion of chapters by two Nobel laureates, V. Smith and Kahneman, not to mention the 29 other chapters, makes clear from the outset that this discipline is a serious branch of scholarly enquiry. Several degrees of separation distinguish it from neuromarketing—the attempt to use neuroimaging tools, such as functional magnetic resonance imaging (fMRI), to measure preferences at a deeper level than might be revealed by questionnaires alone in order to find the best way to sell things. For many neuroscientists, the appeal of neuroeconomics is that it provides formal frameworks to describe the behavioral output of several major brain divisions including parts of frontal, cingulate, and parietal cortex, striatum, and, for want of a better word, the limbic system. Visual and motor psychophysics provide a framework in which to understand the operation of the visual and motor systems, and in a sense the bold claim implicit in Neuroeconomics is that if you want to understand what approximately half of the forebrain is doing then you need an account of how and why decisions are made. Two further important claims are examined in Neuroeconomics. First, it is argued that because economics and social sciences have already been forced to think hard about how decisions should be made (so-called “normative” approaches to decision making) and to measure how decisions are actually made (“descriptive” approaches), these disciplines offer new perspectives on what goals brain processes might be trying to accomplish. Second, it is contended that since economic decisions are the product of neural processes then greater knowledge of such processes should endow economists with greater power to predict decisions. That the perspective offered by neuroeconomics provides new insights into neural data has been evident since publication of a study by Platt and Glimcher, 1999Platt M.L. Glimcher P.W. Nature. 1999; 400: 233-238Crossref PubMed Scopus (1218) Google Scholar that many consider a foundation for the field (reviewed in Platt and Padoa-Schioppa's chapter). For two decades there had been a debate about the function of brain regions such as the lateral intraparietal (LIP) area. Everyone agreed that LIP neuron activity increased when monkeys paid attention to stimuli to which they were about to make saccades. While one camp argued that the activity reflected attentional modulation of sensory processing another camp submitted that it reflected intentional processes related to motor preparation of the saccade. Platt and Glimcher showed that the firing of LIP neurons reflected both the probability that a simple event, the appearance of a visual target for an eye movement, would occur in their receptive field and the magnitude of gain associated with that event in terms of food reward. According to the seventeenth century philosopher Pascal, the importance or value of a possible event is a function of its probability of occurrence and the magnitude of the gain it entails. Monkeys therefore have to decide to where in the visual field it was most valuable to pay attention and parietal neurons encoded the expected value to be assigned to the sensorimotor transformations that would take the monkey's eyes to that location in space. Platt and Glimcher were careful to keep constant the magnitude of gain and the probability throughout fairly long periods of their experiments. Of course, in the real world the probability of one event or another may well change. Corrado and colleagues, in a thoughtful chapter discussing the pitfalls and advantages of various ways of describing behavior, review their suggestion of applying an exponential filter to the recent history of rewards associated with each decision option in order to estimate their value. In other words, the estimate of an option's value will depend heavily on the recent rewards associated with it and to a lesser degree with the more distant history of rewards. Behrens et al., 2007Behrens T.E. Woolrich M.W. Walton M.E. Rushworth M.F. Nat. Neurosci. 2007; 10: 1214-1221Crossref PubMed Scopus (1181) Google Scholar have demonstrated that the extent of the past reward history that should exert this influence is not fixed but depends instead on the volatility of the reward environment; when the environment is more volatile, the past is a less reliable guide to the future and people adjust their value estimates to be based on only the most recent history of rewards. When the value of spatial choices is estimated using Corrado and colleagues' method it correlates with the firing rates of LIP neurons encoding the positions of the possible choices the monkey might make. Volatility changes are also associated with activity in the anterior cingulate cortex as measured with functional magnetic resonance imaging (fMRI). It is clear from single neuron data that information about reward history is present in a number of frontal areas that may influence action selection (Lee's and Wang's chapter). Many of the chapters in Neuroeconomics are coauthored by researchers with different backgrounds. Lee's and Wang's combination of computational modeling and neurophysiology enables them to provide a review not just of neuron firing patterns but also an excellent summary of how such patterns may actually instantiate the making of a decision. A hallmark of Neuroeconomics is a preoccupation with bounded, unambiguous descriptions of probability, risk, and uncertainty. Perhaps nowhere is this clearer than in the chapter by Bossaerts and colleagues. These authors argue, from the foundations of financial analysis, that while activity in some insula regions is correlated with the probability of reward, activity in another insula area is correlated with the risk associated with reward—the expectation of reward variance or squared reward prediction error. Both reward probability and reward risk are also encoded in dopaminergic neurons in the ventral tegmental area (VTA) when cues associated with reward are shown to monkeys (Schultz's chapter). Such approaches are important, first, because they suggest alternative ways in which brain regions might be encoding the values of choices—regions such as the insula may encode a potential choice in terms of its mean return and variance/risk. Second, knowledge of the expected variance in reward may be important during learning in order to scale activity in response to outcomes, for example when there has been a prediction error. Such prediction errors appear to drive learning because they indicate that previous predictions about the world were incorrect but a learner might not update a prediction much even in response to a prediction error if risk, the expected squared reward prediction error, is high. There is now evidence that risk-based scaling of VTA neuron responses to rewards occurs. One implicit connecting strand in Neuroeconomics is the strong reliance on formal modeling. While a normative approach is central to economics, the theories that inform interpretations of neural data in Neuroeconomics are not always solely derived from this discipline. Indeed, it is to psychology or even computer science that neuroscientists have traditionally turned for accounts of behavior to then be related to neural activity. While it may have its detractors, the idea that the dopaminergic system encodes reward prediction errors is surely one of the most influential to have emerged recently in systems neuroscience and that influence is largely due to the careful exploitation of paradigms devised by behavioral learning theorists. This strand of research and the important contribution that has been made to it by computational modeling is illuminated by Niv and Montague and by Doya and Kimura, while Balleine and colleagues provide an excellent complimentary review of the origin of related ideas in psychology. While it is evident that thinking carefully about decision making has been of benefit to neuroscientists what has their involvement with neuroscience gained economists (an issue intelligently addressed in chapters from an economist's perspective by Bernheim, and from a psychologist's by Gallistel)? Economics has traditionally assumed that it has no useful access to internal variables, instead relying on observed behavior to impute the relationship between external factors and decision making. If the aim of economics is to predict choices then the question becomes whether or not knowledge of brain activity makes such predictions any easier or more accurate. In an important paper reviewed by both Fox and Poldrack and by Bossaerts, Berns et al., 2008Berns G.S. Capra C.M. Chappelow J. Moore S. Noussair C. Neuroimage. 2008; 39: 2047-2057Crossref PubMed Scopus (53) Google Scholar used fMRI to measure activity in several regions associated with valuation when people were presented with only one option. In subsequent testing, free choices made between options close in ranking were predicted significantly better by fMRI measurements in nonchoice trials than by previous behavioral measurements. Another test of whether neuroeconomics will be of assistance to economists is whether it can adjudicate between competing economic theories. Delgado and Phelps and colleagues, who coauthor and author chapters, respectively, have tried to do just that to explain why people are prepared to pay more for an item, given its value, in auctions (“overbidding,” Delgado et al., 2008Delgado M.R. Schotter A. Ozbay E.Y. Phelps E.A. Science. 2008; 321: 1849-1852Crossref PubMed Scopus (126) Google Scholar) . Interindividual variation in ventral striatal activity on occasions when bidders were unsuccessful was correlated with variation in overbidding, particularly when bidding was conducted in a social context. The authors argued that contrary to some accounts of overbidding it is fear of losing in a social context that drives overbidding. These findings also highlight an area where neuroscience and psychology have already had an influence on economic thinking, namely on the issue of how emotions influence choice behavior. One of the precursors to the neuroeconomic movement was the work of Damasio and colleagues who made the direct connection between the poor decision making and the social and emotional changes of patients with ventromedial prefrontal lesions (summarized in Damasio's chapter). Such findings have prompted some to favor a dual process model in which so-called cool, rational, deliberative thinking is in competition against automatic, emotional, impulsive urges (Loewenstein et al., 2008Loewenstein G. Rick S. Cohen J.D. Annu. Rev. Psychol. 2008; 59: 647-672Crossref PubMed Scopus (262) Google Scholar). The degree of separation between such systems, however, is a matter of contention. For instance, in two chapters, the function of anterior insula is described as mediating emotional representations which guide choice behavior (Fehr, and Sanfey, and Dorris), while another relates such activity to the encoding of risk (Bossaerts et al.). The difficulties in teasing apart what constitutes an emotional or an economic perspective, and how this affects interpretation of neuroimaging data, are patiently explained by Phelps. Ultimately, the argument that knowledge of brain processes might augment an understanding of economic decision making is based on the assumption that brain processes have a certain primacy over behavior. While many chapters in Neuroeconomics concentrate on correlative techniques such as fMRI or single-unit recording that can identify covariation in behavior and brain activity, there are relatively few examples of manipulation of brain activity in order to cause behavioral changes. In one exception, Fehr describes how transcranial magnetic stimulation (TMS) over the right but not left lateral prefrontal cortex, used to transiently disrupt brain activity, affects choices in a two-person decision making game; it increases “self-interested” choices which benefit the subject but do not punish the partner's unfair offers. Evaluating causality should be an important consideration for future neuroeconomic studies (see also Balleine et al.'s chapter for elegant examples of how lesion studies have helped elucidate dissociable aspects of value processing). The quest to uncover how decision making occurs in interactive or social contexts has generated considerable excitement in recent years and is a theme particularly well covered in Neuroeconomics. Fehr concentrates on how many decisions appear to be motivated not solely by a desire to maximize the value of the outcome for the decision maker but instead reflect an interest in how the outcome will affect others. Such “other regarding preferences” or “social preferences” concerning outcomes are associated with activation in brain regions such as the VTA and striatum in the same way that consideration of potential rewards for oneself activate such regions (Moll et al., 2006Moll J. Krueger F. Zahn R. Pardini M. de Oliveira-Souza R. Grafman J. Proc. Natl. Acad. Sci. USA. 2006; 103: 15623-15628Crossref PubMed Scopus (616) Google Scholar). It is not yet clear whether such activation reflects the consequence of having adopted another's perspective or a premium that is placed on fairness. An essential ingredient of the games used to study social preference is their “one-shot” nature. In such games, players play just a single round with any other player because the aim is to uncover the importance placed on fairness independently of beliefs about what the other player's actions might be; such beliefs might, in turn, affect one's own actions. It is, however, just such interactions and such beliefs about others' intentions that preoccupy other chapters (Camerer, Sanfey, and Dorris). A quite distinct set of brain regions, near the paracingulate and posterior superior temporal sulcus, are called into play when subjects attempt to adopt another person's perspective and estimate the decisions that they will take or the influence that their own decisions will have on the other person. An attempt to place social preferences in the broader context of the behavior of other primate species is made by Brosnan and by Silk. Like humans, some other primates show evidence of aversion to unfair decisions that leave themselves at a disadvantage. By contrast, evidence that other primates place a premium on fairness even when it leaves them relatively disadvantaged, is limited. There is, however, evidence that chimpanzees will assist others in obtaining desired outcomes in situations when such actions impose little cost on themselves (Warneken et al., 2007Warneken F. Hare B. Melis A.P. Hanus D. Tomasello M. PLoS Biol. 2007; 5: e184https://doi.org/10.1371/journal.pbio.0050184Crossref PubMed Scopus (373) Google Scholar). It has now been seven years since a special issue of Neuron allowed neuroscientists to present theoretical reviews which have become mainstays of the discipline. This issue introduced many to the entwined connections between dopamine release, motivation, formal learning theory, economic games, and ecology. It is therefore somewhat surprising that the last of these—behavioral ecology—a discipline that also investigates why animals choose what they do using normative modeling, but in the context of evolutionary constraints, is one of the few topics given less direct attention. As pointed out in the chapters by Gallistel and by Santos and Chen, there is a tacit assumption in Neuroeconomics that a nonhuman animal's individual economy—decisions concerning factors such as nutrition, safety, reproduction—can be understood by reference to human economics. However, the influence may also be in the opposite direction: given that the human brain is the product of evolution, might we gain clues to help inform neuroeconomic theorizing by considering how the brains of mammals have been shaped by the environments they are required to negotiate? For instance, risk preferences in animals are strongly shaped by motivational state and overall reward rate rather than consistently directed toward risk-seeking or -averse behaviors (cf. Platt and Huettel, 2008Platt M.L. Huettel S.A. Nat. Neurosci. 2008; 11: 398-403Crossref PubMed Scopus (370) Google Scholar) as assumed by some economic approaches (chapter by Weber and Johnson). Similarly, economic models often assume that costs discount benefits uniformly regardless of their source. Nonetheless, there is now good neural and behavioral evidence that costs such as the delay between a choice and its consequences or the amount of work required to obtain a goal are weighed up differently (Rudebeck et al., 2006Rudebeck P.H. Walton M.E. Smyth A.N. Bannerman D.M. Rushworth M.F. Nat. Neurosci. 2006; 9: 1161-1168Crossref PubMed Scopus (437) Google Scholar, Stevens et al., 2005Stevens J.R. Rosati A.G. Ross K.R. Hauser M.D. Curr. Biol. 2005; 15: 1855-1860Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar). Whether Neuroeconomics will work as a course textbook is unclear. Unlike most textbooks, several chapters cover similar material and the linear development of argument and idea is not always apparent in the ordering of chapters. For example, discussions of topics of broad interest, such as prospect theory, reinforcement learning theory, and social preference, appear in several chapters, and figures summarizing the same experiment, data, or idea appear in more than one place. For many readers, however, the differences in perspectives on these key issues will be illuminating. Moreover, the coauthored chapters, often written by contributors with quite different specializations, ensure that the accessibility of the material is exemplary (for example, Fox and Poldrack, with backgrounds in behavioral economics and neuroimaging, provide an outstanding introduction to Prospect Theory). Perhaps most importantly, Neuroeconomics manages to convey why neuroscientists in this field are excited by what their colleagues in anthropology and economics have to tell them. Regardless of whether or not the name “neuroeconomics” prevails into the future as the title of an autonomous discipline, it is clear that the approaches to decision making outlined in Neuroeconomics are likely to make their influence felt for some time to come.