The Birth of a Field: From Utility Theory to Brain Scans

Every day, we make countless decisions, from the mundane to the life-altering. What to eat for lunch? Which job offer to accept? How to invest for retirement? Underlying these choices is a fundamental computational problem: how does the brain assign value to different options in a world full of uncertainty? For centuries, this question was the province of philosophy and, later, economics. But in the last few decades, a new field has emerged at the intersection of psychology, neuroscience, and economics—neuroeconomics—to tackle it head-on by peering directly into the brain. This chapter tells the story of how neuroeconomics was born, tracing the historical convergence of classical utility theory, behavioral challenges, and neuroscientific breakthroughs that together created a new science of decision-making.

The Rational Actor: Foundations and Cracks

The story begins in the 19th century with the marginal revolution in economics, spearheaded by William Stanley Jevons, Carl Menger, and Léon Walras. They introduced the concept of marginal utility—the idea that the value of a good is determined by the utility of the last unit consumed. This laid the groundwork for a formal, mathematical model of choice. The culmination came with expected utility theory, most famously axiomatized by John von Neumann and Oskar Morgenstern in their 1944 *Theory of Games and Economic Behavior*. The model is elegant in its simplicity: faced with a set of probabilistic outcomes, a rational actor assigns a subjective utility $u(x)$ to each outcome $x$, weights it by its probability $p_i$, and chooses the option with the highest expected utility.

$$EU = \sum_{i=1}^{n} p_i u(x_i)$$

This framework became the cornerstone of neoclassical economics, shaping fields from finance to public policy. It provided a clear prescription: to maximize well-being, one should consistently choose the option with the highest expected utility. The model’s power lies in its normative force—it defines what rational behavior *should* look like.

However, by the mid-20th century, cracks began to show. Economists and psychologists started documenting systematic deviations from expected utility predictions. The most famous was the Allais paradox (1953), demonstrated by Maurice Allais. Consider two choice pairs:

• **Pair 1:** Choose between:
• A: A sure gain of \$1 million.
• B: A 10% chance to gain \$5 million, an 89% chance to gain \$1 million, and a 1% chance to gain nothing.
• **Pair 2:** Choose between:
• C: An 11% chance to gain \$1 million and an 89% chance to gain nothing.
• D: A 10% chance to gain \$5 million and a 90% chance to gain nothing.

Logically, if someone prefers A over B (sure thing over risky), they should also prefer C over D, because the only difference is the common 89% chance of \$1 million. But in experiments, most people prefer A over B *and* D over C—violating the independence axiom of expected utility theory. This suggested that people’s valuations were not linear in probabilities; the certainty of a sure gain carried extra weight.

Other anomalies piled up: the preference reversal phenomenon (where people choose one option in a direct comparison but assign a higher price to the other), the disposition effect (holding losing stocks too long and selling winners too early), and the preference for status quo. These findings hinted that human decision-making was not a cold calculation but was influenced by context, framing, and emotion. Yet, psychology lacked a unified, quantitative theory to explain these patterns—until the arrival of behavioral economics.

The Behavioral Revolution: Heuristics and Biases

In the 1970s, psychologists Daniel Kahneman and Amos Tversky launched a systematic attack on the rational actor model. Their work, culminating in prospect theory (1979), provided a descriptive model of how people actually make choices under risk. Prospect theory replaced the expected utility formula with a more psychologically plausible one:

$$V = \sum_{i=1}^{n} \pi(p_i) v(x_i)$$

Here, $V$ is the overall value of a prospect, $\pi(p)$ is a probability weighting function that overweights small probabilities and underweights moderate to high probabilities, and $v(x)$ is a value function defined over gains and losses relative to a reference point (usually the status quo), not final wealth. Crucially, $v(x)$ is concave for gains (diminishing sensitivity) and convex for losses (diminishing sensitivity to losses), but steeper for losses than for gains—capturing loss aversion. Empirically, losses loom about twice as large as gains.

The theory was a masterstroke of synthesis. It explained why people buy both lottery tickets (overweighting tiny winning probabilities) and insurance (overweighting tiny losing probabilities). It explained the Allais paradox via the certainty effect: outcomes that are certain receive disproportionate weight. It also introduced the concept of framing effects: the same choice presented as gains or losses leads to different preferences. In the famous Asian disease problem, people preferred a sure saving of 200 lives over a probabilistic saving of 600, but when the same options were framed as losses (400 people will die), they preferred the probabilistic option. The reference-dependent nature of value was now formalized.

Kahneman and Tversky also catalogued a suite of heuristics—mental shortcuts that simplify judgment but introduce biases. The representativeness heuristic leads people to judge probability by how much an instance resembles a category, ignoring base rates (e.g., assuming a quiet person is more likely a librarian than a farmer, ignoring that farmers are far more numerous). The availability heuristic judges frequency by how easily examples come to mind, leading to overestimation of dramatic, media-covered risks (like tornadoes) and underestimation of common ones (like strokes). The anchoring and adjustment heuristic sees people start from an initial value (the anchor) and insufficiently adjust, explaining why the first price offered in a negotiation often sets the final range.

Behavioral economics thus provided a rich catalog of *what* went wrong with rationality, but it left open the *how*: what neural mechanisms give rise to these heuristics and biases? How does the brain implement a reference-dependent value function? To answer these questions, researchers needed to open the black box of the mind. The tool that would let them do just that was already being perfected in neuroscience: functional brain imaging.

The Neuroscientific Revolution: Peering Inside the Black Box

While Kahneman and Tversky were revolutionizing psychology, neuroscience was undergoing its own quiet revolution. Techniques like positron emission tomography (PET) and, later, functional magnetic resonance imaging (fMRI) allowed scientists to measure brain activity in awake, behaving humans. The principle behind fMRI is clever: neural activity increases local blood flow (the hemodynamic response), and deoxygenated hemoglobin is paramagnetic, affecting the MRI signal. By comparing brain images during a task to a baseline, researchers could map which regions “lit up” during decision-making.

But the most direct evidence for a neural value signal came from single-neuron recordings in animals, a technique that predates human fMRI by decades. The landmark study, published by Wolfram Schultz and colleagues in 1997, recorded from dopamine neurons in the brains of thirsty monkeys. The monkeys were trained to associate a visual cue (a light) with the delivery of fruit juice seconds later. Initially, the dopamine neurons fired strongly in response to the juice itself—a primary reward. But as learning progressed, something remarkable happened: the neurons stopped firing at the juice reward and instead began firing at the predictive cue. Conversely, if the expe