Scholarly article on topic 'From behavioural economics to neuroeconomics to decision neuroscience: the ascent of biology in research on human decision making'

From behavioural economics to neuroeconomics to decision neuroscience: the ascent of biology in research on human decision making Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{}

Abstract of research paper on Clinical medicine, author of scientific article — Peter Bossaerts, Carsten Murawski

Here, we briefly review the evolution of research on human decision-making over the past few decades. We discern a trend whereby biology moves from subserving economics (neuroeconomics), to providing the data that advance our knowledge of the nature of human decision-making (decision neuroscience). Examples illustrate that the integration of behavioural and biological models is fruitful especially for understanding heterogeneity of choice in humans.

Academic research paper on topic "From behavioural economics to neuroeconomics to decision neuroscience: the ascent of biology in research on human decision making"

COBEHA 98 X-X ARTICLE IN PRESS

WÊÈtÊ^ i-i^'5''®» Available online at www.sciencedirect.com ScienceDirect Behavioral Sciences

ELSEVIER

From behavioural economics to neuroeconomics to decision neuroscience: the ascent of biology in research on human decision making§

Peter Bossaerts1,2,3 and Carsten Murawski1

Here, we briefly review the evolution of research on human decision-making over the past few decades. We discern a trend whereby biology moves from subserving economics (neuroeconomics), to providing the data that advance our knowledge of the nature of human decision-making (decision neuroscience). Examples illustrate that the integration of behavioural and biological models is fruitful especially for understanding heterogeneity of choice in humans. Addresses

1 Faculty of Business and Economics, The University of Melbourne, 198 Berkeley Street, Parkville, VIC 3010, Australia

2 Department of Finance, Eccles School of Business, University of Utah, 1655 E Campus Center Drive, Salt Lake City, UT 84112, USA

3 The Florey Institute of Neuroscience and Mental Health, 30 Royal Parade, Parkville, VIC 3052, Australia

Corresponding author: Bossaerts, Peter (peter.bossaerts@unimelb.edu.au)

Current Opinion in Behavioral Sciences 2015, 5:xx-yy

This review comes from a themed issue on Decision making/ neuroeconomics

Edited by John O'Doherty and Colin Camerer

http://dx.doi.org/10.1016/j.cobeha.2015.07.001

2352-1546/Published by Elsevier Ltd.

Economic theories of human choice

For a large part of the 20th century, research on human choice was dominated by economic theories, particularly rational choice and revealed preferences theory. This approach starts from a limited set of properties that are imposed on choices (rationality axioms). It then determines to what extent choices can be summarised (represented) by maximisation of some latent mathematical function, typically referred to as utility or value function. The form of the value function depends on the nature of the axioms [1]. The value function and

its maximisation merely constitute a compact way to summarise choices. In binary (pairwise) choice, for instance, the economist does not need a look-up table: to determine whether one option would be chosen over the alternative, the economist merely picks the option with the maximum value.

In economic theory, the value function does not necessarily reflect subjective preferences, or the agent's 'needs' or 'wants.' Preferences are formulated in a way that is independent of the type of agent (human, market, firm) whose choices the preferences describe. Thus, the economist's definition of the term 'preferences' is fundamentally different from the psychologist's. To economists, preferences are merely a description of choices, and preferences and choices are observationally equivalent.

Soon after the emergence of the first instances of axiomatic choice theories, it became apparent that they could not capture many key regularities of human choice. The two most famous examples are the Allais [2] and Ellsberg [3] paradoxes. In subsequent years, new value functions were proposed that improved the fit with the empirical data [4,5]. This development culminated in Prospect Theory [6], which summarised salient characteristics of actual human choice under uncertainty in terms of maximisation of a utility index that featured a reference point, a kink, probability weighting, and differential curvature in the gain and loss domains. Some of these features accommodated cognitive biases. Loss aversion, for instance, is not merely a tendency to avoid risk (which rational agents are allowed to do). Instead, it is a cognitive bias that makes an agent choose differently depending on whether a prospect is presented as losses or as gains [7].

Prospect Theory models capture human cognitive biases within a framework of utility maximisation. Thus, its approach is consistent with the approaches of earlier economic theories. The success of Prospect Theory was sealed when an axiomatic version of the theory emerged [8]. At the time, alternative (complementary or substitutable) theories were proposed such as Herbert Simon's 'satisficing' [9] or Gerd Gigerenzer's 'heuristics

Q1 § The authors gratefully acknowledge financial support from the finance department of the faculty of business and economics at the University of Melbourne and from the finance department of the David Eccles School of Business at the University of Utah. The contents of this article have been shaped by numerous discussions with colleagues in economics, psychology and neuroscience. Comments from two anonymous referees were particularly helpful. The views expressed here, however, are the authors' only.

www.sciencedirect.com Current Opinion in Behavioral Sciences 2015, 5:x-x

2 Decision making/neuroeconomics

toolbox' [10]. However, those theories cannot readily be translated into the language of traditional economic choice theory. Some have argued that Simon's theory could be translated into a value maximisation framework, by adding constraints to cognition [11]. Unfortunately, constrained optimisation often presupposes cognitive capabilities that contradict the bounded rationality that underlies satisficing behaviour. Indeed, constrained optimisation problems maybe very 'hard' [12]. Still, this is not a concern for traditional economics, where the agent would choose merely 'as if implementing constrained optimisation.

The strength of the axiomatic approach cannot be overestimated. It provides a disciplined way of modelling choice as utility maximisation. It avoids the pitfalls of other approaches that merely fit value functions to data. Indeed, a value function may fit data well but may be such that it violates rationality constraints that may be far less controversial than the observed cognitive biases that the value function was meant to capture in the first place. Such was the case with the original version of Prospect Theory [6], where the probability weighting function was at odds with the sure-thing principle — outcomes that would occur under any alternative prospect ended up influencing choice. (The subsequent, axiomatic version of Prospect Theory corrected this [13].)

The axiomatic approach and behavioural economics alike start and finish with choice data. The value or utility function that is maximised is just another way to describe choices. The maximisation process (which, as already mentioned, could be rather complex) is not to be taken literally: the agent chooses 'as if maximising utility. Importantly, the axiomatic approach does not provide a mechanistic account of how choice is implemented but only describes the properties of choices. Equally importantly, both approaches assume that preferences are exogenous, which unfortunately precludes an important type of intervention. 'Bad' choices (compulsive gambling, insufficient retirement savings, eating disorders, drug addiction, etc.) cannot be changed through a change of preferences, but only through a change of the available options or re-framing of the options [14], or through education [15].

From understanding choice to understanding neural circuitry: the advent of neuroeconomics

With the emergence of non-invasive human brain imaging techniques such as functional magnetic resonance imaging (fMRI), it was only a matter of time before economists and neuroscientists set out to determine if there was any biological foundation of economic theories of choice. Key aims were to determine how choices were implemented biologically, which neural circuitry was involved, and what algorithms were employed. A new

Current Opinion in Behavioral Sciences 2015, 5:x-x

field emerged, referred to as neuroeconomics, focusing on the description of algorithms underlying observed choice and their biophysical implementation. Human decision-making would thereby become understandable at a lower level of description than the traditional, abstract, axiomatic approach had done. It corrected a situation which actually was the opposite of that in vision research, where the biophysical took precedence over the abstract [16].

Very quickly, this research program led to some fascinating results, including the discovery of, and subsequently, ability to manipulate, the very value (utility) signals that constitute the core of the axiomatic theory [17—21][17— 19,20*,21]. More recently, it has provided more detail into how value maximisation is implemented at a neural level, borrowing ideas from drift-diffusion models in psycho-physics [22] and detailed neural networks with mutual inhibition [23], among others. This line of research also led to the discovery that some basic axioms of choice theory such as Irrelevance of Independent Alternatives (IIA) are violated due to fundamental properties of the central nervous system, namely, divisive normalisation [24]. Violations occur when the availability of a third, clearly inferior option, makes people choose the lower-valued option in a pair more frequently than in the absence of this third option. Under divisive normalisation, inputs (e.g., sources of light, auditory signals, values of available options) are re-scaled to fit a preset range. Biophysically, divisive normalisation happens because neuronal firing is affected by activation of nearby neurons. The discovery was particularly exciting, because divisive normalisation may predict behavioural features that economists had not detected yet. One small step in that direction is the prediction that independent alternatives may actually have the reverse effect on choice when the values of options are relatively close. The example is also important because it shows how biological data, hitherto outside the field of view of economists, can help to make sense of choice anomalies.

To date, neuroeconomic data have mainly been used to better distinguish between competing valuation models when choice data alone were not sufficient (given typical sample sizes). Neuroeconomics has shown, for example, that valuation based on Bayesian principles better explains neural activation and choices in a reversal learning task [25]. Similarly, neurobiology demonstrated that in certain settings, choice under uncertainty seems to be based on mean-variance analysis rather than more traditional expected utility theory [26]. Mean-variance analysis is popular in financial economics, yet unlike expected utility theory, can cause violations of simple rationality principles [27].

Despite all the successes of the neuroeconomic research program, economists may argue that it is of little relevance to economic theory, because of the perception that

www.sciencedirect.com

Biology in research on human decision making Bossaerts and Murawski 3

the levels at which one can understand human decision making are relatively independent, an opinion also voiced in vision research [16]. While it may be interesting to know which neural algorithms implement observed choice, and what biophysical constraints cause violations of the axioms of choice theory, such knowledge is deemed irrelevant for the future development of choice theory [28,29].

Curiously, economists do appeal to biological principles in other domains, in order to put discipline on the parameters of their choice models. An important example is the use of principles of evolutionary fitness to answers questions such as: What are acceptable risk aversion parameters? Will preferences feature inter-generational substitution? [30,31]. Such approaches referring to the theory of evolution to restrict preferences is not based on observation, however (there are no data in the cited work!). It is merely a device to restrict the parameter space when axioms of choice are too weak to constrain the theory sufficiently.

What if we started from biology? The emergence of decision neuroscience

So far, biology has only played a supporting role in the quest for a better understanding of human behaviour, helping to differentiate between existing valuation models, or elucidating the biophysical mechanics and implementation algorithms behind human economic decision making. However, in recent years evidence has emerged that there is significant biological variation that does not map into parametric variation of even the best economic models. These findings are part of a new field, decision neuroscience, focused on decision-making, but where biology no longer subserves economics and instead takes a central role [32].

Below, we give some examples of biological diversity that maps into variation in behaviour that is reflected in the error term of the most popular economic model that ties valuation to choice, namely, logit, or in the language of neuroeconomics, softmax. For economists, the error term of the softmax model captures 'unobserved heterogeneity' [33]. However, it appears that this 'error' actually contains useful information, because biological markers explain it. Hence, it ought to be modelled explicitly.

One important area of research investigates the relation between neurotransmitters and behaviour. Administration of Levadopa (L-dopa), a drug designed to increase levels of the neurotransmitter dopamine in the brain, has unintended effects on economic choice (unintended in the sense that it does not explain conjectured effects on parameters of existing choice models). In one study, L-dopa appeared to speed up learning in a two-armed reward bandit problem where subjects had to discover the option that was most rewarding on average [34*]. Closer

inspection, however, suggests that participants who received L-dopa were less erratic in their choices, which effectively meant that they were 'better optimisers.' Increases in estimated learning speed could merely be the consequence of better fit of the economic model. Thus, administration of L-dopa changed the properties of the error term of the softmax model that linked valuation with choice. More recently, it has been shown that, even absent learning, L-dopa intervention has no effect on Prospect Theory parameters, but instead shifts the error term of the softmax function [35*]. The researchers who discovered the effect offered a quintessentially biological explanation, namely, Pavlovian approach behaviour. This dimension of behaviour (phenotype) has yet to be captured by economic theories.

Other research has shown that while genetic variation explains differences in risk taking across humans, this genetic variation does not cause shifts in the parameters in Prospect Theory that are meant to capture risk attitudes [36*]. Genetic variation actually correlated with a tendency towards more or less optimising (relative to the predictions of Prospect Theory), but only when available

Figure 1

E ro ro

<D O O CO O

'¡Ô c <D CP

90 80 70 60 50 40 30 20 10 0

-4 -2 0 2 4 net expected utility of gamble

Current Opinion in Behavioral Sciences

Propensity to choose one gamble against another as a function of difference in value. Value is prospect theory expected utility as estimated from all other choices. Diamonds are MAOA-L carriers; dots are MAOA-H carriers. Estimated differences in values are no different across genotypes (observations across genotypes are equally distant on horizontal axis), yet choice propensities are very different (observations on vertical axis are not equidistant). Solid black line extending to black dashed line is best fit (softmax) for MAOA-H carriers; best fit for MAOA-L carriers (solid black line extending to grey dotted line) has a kink at zero (softmax with kink). Consequently, prospect theory cannot capture observed differences in choice propensities (phenotypes) for MAOA-H against MAOA-L carriers. Overall, MAOA-H carriers take less risk, yet prospect theory does not predict so. Parameter estimates (loss aversion, risk attitudes in gain and loss domain, probability weighting) were statistically indistinguishable (P = 0.05) across genotypes. Reproduced from [36*].

www.sciencedirect.com

Current Opinion in Behavioral Sciences 2015, 5:x-x

4 Decision making/neuroeconomics

gambles were more valuable than the risk-free alternative. There, too, the error term of the softmax model shifted (Figure 1).

This example is rather limited, because it is rare that a single gene predicts behaviour. There is more promise in polygenic or entire gene pathway analysis [37,38]. But the example demonstrates how prior research had been wrong to focus only on phenotypes that traditional economic analysis recognised. In this case, the phenotype concerned risk attitudes, and humans who were willing to accept risky gambles were categorised as 'more risk tolerant.' The genes correlating with this tendency to accept risky gambles, it was concluded, were the ones controlling risk aversion [39]. But the conclusion was wrong: those who accepted risky gambles more frequently were not more risk tolerant; they were actually merely 'better optimisers.' Quality of optimisation is a phenotype that is not captured by traditional economic analysis, but evidently very much present in human decision-making, and apparently has a simple genetic basis. (The gene, MAOA, regulates catabolism of, among others, dopamine;

Figure 2

hence, the genetic effect on optimisation quality is not unlike that of administrating L-Dopa [34*].)

The examples point to a potential weakness of neuroe-conomics: it often presupposes existing economic theories when analysing biological data, which has sometimes led to the mis-interpretation of the latter. A particularly pertinent example is related to the role of emotions in economic decision making. It has been known for some time that emotions and rational decision-making are not orthogonal. A key study [40] contrasted choices under uncertainty among patients with prefrontal and amygdala brain lesions, and discovered that emotional engagement during risk taking is crucial for 'reasoned' decision making. In contrast, economists had been modeling emotions as interfering with rational decision making, in the form of dual-self theory (e.g., [41]). Accounts of neuroeconomics, too, often tend to emphasise a sharp delineation between, among others, 'cognitive' and 'affective' processes

[42]. Such dual-self theories unfortunately have biased interpretation of neural signals on a number of occasions

Biological models Economic models Pharmacological intervention

L-DOPA intervention

Behaviour (choices)

Neuro-economics

u(c) = 1 - e-Yc

Biology

Current Opinion in Behavioral Sciences

The link between genetics, brain and behaviour (choices). Economic preferences revealed by choices are linked to brain and genes through the biological organism (green box). The neuroeconomic approach links preferences to properties of the brain and the genome through economic models (blue box). In this approach, an economic model is fitted to choice data and model parameters, such as risk aversion, are then correlated with properties of the brain or the genome. Recent research suggests that this approach has severe limitations. A study using an intervention at the biological level (administration of l-DOPA) [35*] to induce behavioural change showed that the change in behaviour was not captured by parameters of traditional economic models. Instead, the intervention changed the properties of the error term of economic models, suggesting that existing (neuro-)economic models are missing important dimensions of human behaviour.

Current Opinion in Behavioral Sciences 2015, 5:x-x www.sciencedirect.com

Biology in research on human decision making Bossaerts and Murawski 5

The future

Traditional economic approaches (including behavioural economics) as well as neuroeconomics have taken a choice-centred approach. If biology is appealed to at all, it is used as 'supporting' or 'converging' evidence for extant theories of choice. In contrast, we advocate an approach where biology takes a more central role. In such an approach, biological variation would be used to identify potential behavioural variation that would be missed (read: absorbed by the error term) if one were to follow economic theory alone. Behavioural scientists are to engage in a genuine dialogue with biologists, because biologists observe phenomena relevant to choice that traditional models do not capture, and they have research methods to manipulate these phenomena (e.g., pharmacological interventions) that behavioural scientists do not have. The contribution of biology to research on choice should not be limited to providing a mechanistic account of human decision-making. Its role should be extended to providing some of the foundations of theories of human behaviour (Figure 2).

In our discussion of the merits of decision neuroscience for understanding human decision-making, we have focused on choice heterogeneity, both between and within individuals. One cannot overestimate the practical relevance of an improved understanding of heterogeneity, not only for economic policy and welfare, but also clinical psychology, psychiatry and public health. An illustrative example is dopamine-replacement therapy in Parkinson's disease. While effective in alleviating many symptoms of the disease, about one fifth of patients receiving such therapy develop impulse-control disorders, the most frequent one being pathological gambling [one quarter of affected cases; [44]]. At present, the effect is not fully understood but it is conceivable that the pathway is similar to the one described in Section 'What if we started from biology? The emergence of decision neuroscience' and thus outside of current economic theory. Illumination of the pathway will help develop improved medication for Parkinson's disease, and in the process enhance our understanding of the factors influencing risk-taking.

A final remark concerns the scope of this novel research approach. It should be obvious that it only applies to individual human decision making. The discipline of Economics is more ambitious than that, however. Economists want to describe (and predict) economic decision making in general, whether the decision is made by an individual, or by an institution such as a financial market or a government. Although noble in its goal, this research program may be overly ambitious. We know now that the behaviours of individuals and institutions, such as markets, have very different properties. Traditional theories of choice evidently explain market (i.e., aggregate) phenomena better than the choices of the individuals who populate the market [45,46]. Only recently have we

begun to understand why [47]. Importantly, this does not mean that decision neuroscience cannot help explain phenomena at an aggregate level, such as in markets. Indeed, recent research in neuroscience has interesting things to say about how individuals behave in the face of, for example, institution-generated uncertainty, and why [48-50].

Conflict of interest statement

Nothing declared.

Uncited reference

[51*].

References and recommended reading

Papers of particular interest, published within the period of review,

have been highlighted as:

• of special interest

1. Von Neumann J, Morgenstern O: Theory of Games and Economic Behavior. Princeton, NJ: Princeton University Press; 1947.

2. Allais M: Le comportement de l'homme rationnel devant le risque: critique des postulats et axiomes de l'ecole americaine. Econometrica 1953, 21:503-546.

3. Ellsberg D: Risk, ambiguity, and the savage axioms. Q J Econ 1961:643-669.

4. Friedman M, Savage LJ: The utility analysis of choices involving risk. J Polit Econ 1948:279-304.

5. Markowitz H: The utility of wealth. J Polit Econ 1952:151-158.

6. Kahneman D, Tversky A: Prospect theory: an analysis of decision under risk. Econometrica 1979, 47:263-291.

7. De Martino B, Kumaran D, Seymour B, Dolan RJ: Frames, biases, and rational decision-making in the human brain. Science

2006, 313:684-687.

8. Wakker P, Tversky A: An axiomatization of cumulative prospect theory. J Risk Uncertainty 1993, 7:147-175.

9. Simon HA: Theories of decision-making in economics and behavioral science. Am Econ Rev 1959:253-283.

10. Gigerenzer G, Selten R: Bounded Rationality: The Adaptive Toolbox. MIT Press; 2002.

11. Rubinstein A: Modeling Bounded Rationality. MIT Press; 1998.

12. Meloso D, Copic J, Bossaerts P: Promoting intellectual discovery: patents versus markets. Science 2009, 323:1335-1339.

13. Tversky A, Kahneman D: Advances in prospect theory: cumulative representation of uncertainty. J Risk Uncertainty 1992, 5:297-323.

14. Thaler RH, Sunstein CR: Nudge. Yale University Press; 2008.

15. Lusardi A, Mitchell OS: The economic importance of financial literacy: theory and evidence. J Econ Lit 2014, 52:5-44.

16. Marr D, Poggio T: From understanding computation to understanding neural circuitry.

17. Shizgal P: On the neural computation of utility: implications from studies of brain stimulation reward. In Foundations of Hedonic Psychology: Scientific Perspectives on Enjoyment and Suffering.

18. Platt M, Glimcher P: Neural correlates of decision variables in parietal cortex. Nature 1999, 400:233-238.

19. Plassmann H, O'Doherty J, Rangel A: Orbitofrontal cortex encodes willingness to pay in everyday economic transactions. J Neurosci 2007, 27:9984-9988.

Q3 393 394

www.sciencedirect.com

Current Opinion in Behavioral Sciences 2015, 5:x-x

6 Decision making/neuroeconomics

20. Padoa-Schioppa C, Assad JA: Neurons in the orbitofrontal

• cortex encode economic value. Nature 2006, 441:223-226. Identifies neural signals in monkey orbitofrontal cortex that reflect the building blocks of choice based on utility maximisation, and hence, demonstrated that economic decision theory is not just 'as if' (agents choose as if they are maximising utility).

21. Figner B, Knoch D, Johnson EJ, Krosch AR, Lisanby SH, Fehr E, Weber EU: Lateral prefrontal cortex and self-control in intertemporal choice. Nat Neurosci 2010, 13:538-539.

22. Krajbich I, Rangel A: Multialternative drift-diffusion model predicts the relationship between visual fixations and choice in value-based decisions. Proc Natl Acad Sci USA 2011, 108:13852-13857.

23. Padoa-Schioppa C: Neuronal origins of choice variability in economic decisions. Neuron 2013, 80:1322-1336.

24. Louie K, Khaw MW, Glimcher PW: Normalization is a general neural mechanism for context-dependent decision making.

Proc Natl Acad SciU S A 2013, 110:6139-6144.

25. Hampton A, Bossaerts P, O'Doherty J: The role of the ventromedial prefrontal cortex in abstract state-based inference during decision making in humans. J Neurosci 2006, 26:8360-8367.

26. Christopoulos G, Tobler P, Bossaerts P, Dolan R, Schultz W: Neural correlates of value, risk, and risk aversion contributing to decision making under risk. J Neurosci 2009, 29:12574.

27. d'Acremont M, Bossaerts P: Neurobiological studies of risk assessment: a comparison of expected utility and mean-variance approaches. Cogn Affect Behav Neurosci 2008, 8:363-374.

28. Gul F, Pesendorfer W: The case for mindless economics. Found Pos Norm Econ 2008:3-42.

29. Bernheim BD: The psychology and neurobiology of judgment and decision making: What's in it for economists?

Neuroeconomics 2009:115-126.

30. Robson AJ, Samuelson L: The evolutionary foundations of preferences. Handb Social Econ 2010, 1:221-310.

31. Zhang R, Brennan TJ, Lo AW: The origin of risk aversion. Proc Natl Acad Sci U S A 2014, 111:17777-17782.

32. Shiv B, Bechara A, Levin I, Alba JW, Bettman JR, Dube L, Isen A, Mellers B, Smidts A, Grant SJ et al.: Decision neuroscience. Market Lett 2005, 16:375-386.

33. Wooldridge JM: Simple solutions to the initial conditions problem in dynamic, nonlinear panel data models with unobserved heterogeneity. J Appl Econ 2005, 20:39-54.

34. Pessiglione M, Seymour B, Flandin G, Dolan RJ, Frith CD:

• Dopamine-dependent prediction errors underpin reward-seeking behaviour in humans. Nature 2006, 442:1042-1045.

Dopamine regulation is claimed to impact learning rates, but appears to have a bigger effect on the error term of the learning model, thus presaging the finding in [35^].

35. Rutledge RB, Skandali N, Dayan P, Dolan RJ: Dopaminergic

• modulation of decision making and subjective well-being. J

Neurosci 2015. (in press). Shows how dopamine regulation correlates with the error term of a state-of-the art psycho-economic model based on prospect theory & softmax.

36. Frydman C, Camerer C, Bossaerts P, Rangel A: MAOA-L carriers • are better at making optimal financial decisions under risk.

Proc R Soc B: Biol Sci 2011, 278:2053-2059. Like [354, but here genetic variation rather than dopamine regulation is the basis for behavioural variation that is not capture by prospect theory & softmax; effectively, genes control the quality of optimisation.

37. Set E, Saez I, Zhu L, Houser DE, Myung N, Zhong S, Ebstein RP, Chew SH, Hsu M: Dissociable contribution of prefrontal and striatal dopaminergic genes to learning in economic games.

Proc Natl Acad Sci U S A 2014, 111:9615-9620.

38. Stice E, Yokum S, Burger K, Epstein L, Smolen A: Multilocus genetic composite reflecting dopamine signaling capacity predicts reward circuitry responsivity. J Neurosci 2012, 32:10093-10100.

39. Kuhnen CM, Chiao JY: Genetic determinants of financial risk taking. PLoS ONE 2009, 4:e4362.

40. Bechara A, Damasio H, Tranel D, Damasio AR: Deciding advantageously before knowing the advantageous strategy.

Science 1997, 275:1293-1295.

41. Fudenberg D, Levine DK: A dual-self model of impulse control.

Am Econ Rev 2006:1449-1476.

42. Camerer C, Loewenstein G, Prelec D: Neuroeconomics: how neuroscience can inform economics. J Econ Lit 2005:9-64.

43. Rustichini A: Neuroeconomics: what have we found, and what should we search for. Curr Opin Neurobiol 2009, 19:672-677.

44. Ambermoon P, Carter A, Hall WD, Dissanayaka NN, O'Sullivan JD: Impulse control disorders in patients with Parkinson's disease receiving dopamine replacement therapy: evidence and implications for the addictions field. Addiction 2011, 106:283-293.

45. Bossaerts P, Plott C: Basic principles of asset pricing theory: evidence from large-scale experimental financial markets. Rev

Finance 2004, 8:135-169.

46. Asparouhova E, Bossaerts P, Roy N, Zame W: 'Lucas' in the laboratory. J Finance 2015. (in press).

47. Bossaerts P, Plott C, Zame W: Prices and portfolio choices in financial markets: theory, econometrics, experiments.

Econometrica 2007, 75:993-1038.

48. Bruguier AJ, Quartz SR, Bossaerts P: Exploring the nature of 'trader intuition'. J Finance 2010, 65:1703-1723.

De Martino B, O'Doherty JP, Ray D, Bossaerts P, Camerer C: In the mind of the market: theory of mind biases value computation during financial bubbles. Neuron 2013, 79:1222-1231.

Smith A, Lohrenz T, King J, Montague PR, Camerer CF: Irrational exuberance and neural crash warning signals during endogenous experimental market bubbles. Proc Natl Acad Sci USA 2014, 111:10503-10508.

51. Boureau Y-L, Dayan P: Opponency revisited: competition and * cooperation between dopamine and serotonin.

Neuropsychopharmacology 2011, 36:74-97. Provides a biology-founded alternative approach to decision making under uncertainty, based on Pavlovian approach.

Current Opinion in Behavioral Sciences 2015, 5:x-x

www.sciencedirect.com