Let’s play a game. Imagine the pandemic is over and you take a vacation. You go somewhere warm with no plans other than to bask in the sun, swim in the ocean, and have lots of drinks garnished with flowers, fruit, and tiny umbrellas. When you arrive, the weather is excellent. The first few days are everything you wanted. But on the third day, there’s a sharp, strong breeze that makes it impossible to go swimming. On the last day, it rains, forcing you inside all day.
So will you remember this as a good vacation? Probably not, say neuroscientists and behavioral economists. In fact, you will likely remember this experience more negatively than a vacation during which it rained most of the time and only cleared up on the last day. It doesn’t matter that you had more warm weather fun during the first vacation. “It turns out that people want happy endings,” says Martin Vestergaard, a cognitive and behavioral neuroscientist at Cambridge University.
In a study published Monday in The Journal of Neuroscience, Vestergaard and coauthor Wolfram Schultz show that participants prefer experiences with happy endings to experiences that became slightly less enjoyable towards the end. Thanks to their work with fMRI imaging, Vestergaard and Schultz are also able to suggest some of the mechanical underpinnings of this preference by showing that different parts of the brain preserve and process different pieces of information from the same experience.
“The specific idea that we were interested in in this work is the disconnect between what people enjoy and what they want,” says Vestergaard. Although people may enjoy sunny beach vacations, if they don’t remember them fondly, they won’t opt for those experiences again when they are available in the future. “Now we have a problem,” he says. “If people don’t choose what they enjoy, then their choices don’t really serve their best interests.”
This phenomenon, known as the “happy ending effect,” has been recorded by researchers before. In one study, Daniel Kahneman, who won the Nobel Memorial Prize in Economic Sciences for his work on judgment and decisionmaking, had colonoscopy patients rank their pain at intervals throughout the procedure, then asked them to rank how painful the experience was after it was over. Kahneman found that the length of the procedure didn’t affect participants’ memories. Instead, the peak intensity of the pain, and the pain during the last three minutes of the procedure, influenced patients’ memories the most. The more painful those final minutes, the worse the memory.
Vestergaard and Schultz were curious about what was happening in people’s brains as they processed experiences that trended in either the positive or negative direction over time. But they faced some challenges. “We can’t expose the volunteers to holidays in a brain scanner,” says Vestergaard. “It’s not possible.”
Instead, he devised a virtual gambling scenario which participants could play on a screen while lying in an fMRI scanner, which measures brain activity by tracking changes in blood flow. In each round, 27 participants watched two sequences in which different-sized coins appeared and pulsated before falling into pots, first on the left and then on the right side of the screen. The bigger the image of the coin, the bigger its monetary value. Next, participants had to choose which pot held more money. At the end of the game, they were paid a fixed fee (in real money, not game coins), plus some additional prize money according to how well they chose.
While participants were choosing which pot of money they’d like, the fMRI scanner revealed which regions of their brains were being activated. What the researchers found was that all participants processed information about the gambling task in two areas. The first was the amygdala, two almond-shaped clusters in the right and left hemispheres of the brain. The second was the anterior insula, which is folded into the deep fissure between the parietal and frontal lobes.
When the researchers compared the scans of participants, they found that those who most often correctly chose the pot with the most money had more activation in the amygdala, while those who made suboptimal decisions had more activation in the anterior insula. “The amygdala encodes the actual money, the sum,” says Vestergaard. In their study, the authors concluded that the amygdala is responsible for processing information about the overall experience, not just whether or not it ends well. And that’s important, because that’s the thing you should be paying attention to if you want to optimize winning money.
In contrast, they wrote, the anterior insula “marks down,” or penalizes, sequences in which the coins’ sizes start to decrease over time. But rationally, temporary decreases are something you should ignore, not focus on. After all, a decrease at the end of an experience may be offset by increases at the beginning.
“The statistical strength of this neural code can separate efficient decision-makers from suboptimal decision-makers,” the authors wrote in their study. “Optimal decision-makers encode overall value more strongly, and suboptimal decision-makers encode the disincentive markdown more strongly.”
Vestergaard acknowledges there are limitations to this study. The subjects were all men ages 21 to 36. Could there be a difference between men and women, or between old and young people? What about differences between, say, artists or mathematicians, who might have different ideas about what makes them happy or about what’s valuable? “These are all good questions,” Vestergaard says. “This has to be next time, I’m afraid.”
Nathaniel Daw, a professor of computational and theoretical neuroscience at Princeton who was not involved in this study, says that while the happy ending effect has been known for a while, scientists still don’t know why people will consistently opt for bad or irrational choices. “We don't understand very well why they arise—what is the mechanism for how the brain evaluates options from experience, and why does it make these apparent mistakes?” he wrote in an email to WIRED. “The brain measurements provide additional information that can give us clues about that.”
One theory was that we may just be forgetting the good parts of the experience because they happened earlier. But this research shows that this isn’t the result of poor memory, Daw argues: “The fact that they find a distinct brain signal assessing how rewards are declining, separate from signals related to how good the options are, supports the idea that the brain is actively constructing this evaluation.”
Daw added that in other studies the anterior insula is often associated with processing negative experiences, disgust, and pain. “It really seems like people are actively repulsed by the declining options, which is not necessarily what I would have expected,” he wrote.
Vestergaard isn’t ready to speculate on what it means that the anterior insula is activated. What is important, he says, is that even though a negative or declining trend may not be relevant information, the brain is penalizing the experience because of it. “Your brain registers it, and it informs your decisions, whether you want it or not,” says Vestergaard.
There may be some reasons why an animal’s brain wants to keep track of upward or downward trends. Daw notes that it can be an important adaptation for foraging animals that need to keep track of whether a food source is getting more or less available. If an animal can make a quick decision about whether to continue looking for food in one place, or whether to change course because their options are dwindling, it could find food more efficiently.
When it comes to the way people respond to scarce resources, we might even try to game our own neurobiology. As far back as 1993, the behavioral economists George Loewenstein and Drazen Prelec published a paper in Psychological Review suggesting that people prefer an improving trend or sequence of experiences, rather than a decreasing trend, because they like reveling in the anticipation of a positive event. “Individuals are not helpless in the face of their urges,” Loewentstein and Prelec wrote. They pointed to real-life examples like participating in dieting clubs or the Social Security system, activities in which people put off the enjoyment of something now so that they can indulge in it later, whether it be dessert or the comfort of a healthy retirement savings.
Or, there are shorter-term scenarios in which someone might intentionally defer the happy ending effect; for example, if you’re at a birthday party and really looking forward to the cake, you may wait until the end of the party to eat it. “You want to delay and save a big reward for later so that you can have a lot of time to look forward to the happy ending,” says Kiyohito Iigaya, a postdoctoral neuroscience researcher at Caltech who was not involved in the new study.
Iigaya points out that what’s considered rational or irrational behavior is often subjective. He refers to another paper by Loewenstein that found that people will often put off meeting a celebrity so they can enjoy looking forward to the event. This may be irrational in some ways, but, Iiagaya says, from the point of view of maximizing the fun of anticipation, it does have a rationale.
The happy ending effect could also be used to steer people in positive ways. Kahneman’s study showed that people base their memories of colonoscopies on those final painful minutes—but what if doctors were able to improve them? Daw suggests that making the end of the exam more comfortable might mean people wouldn’t have as many negative memories—even if those changes make the procedure longer overall. “This would make people hate it less and more likely to return for their next one,” he wrote.
Vestergaard agrees that our preference for happy endings can be an evolutionary boon, but he points out that it can also lead us astray. It may make it easier to make decisions quickly, because it gives us an “intuitive” or “gut” feeling about which option is best. But that can be problematic, because our gut can be wrong or be easily misled. This is essentially what his study did, however harmlessly. Using different sequences of coins, the researchers were able to trick people into choosing the wrong pot—one that “felt” like it had more value, simply because the coins falling into it had gotten larger right before the sequence ended.
Similarly, Vestergaard gives the example of a political campaign. “Politicians always try to appear strong and successful prior to the end of their tenure,” he says. If a voter only considered that politician’s most recent actions, instead of the entirety of their time in office, they might end up voting counter to their own interests.
Vestergaard stresses that just because our brains process information about upward and downward trends doesn’t mean we have to privilege that information. “You can still stop and think,” he says. When it comes time to make a big decision, he suggests taking a step back, writing a pros and cons list, and analytically evaluating the options out there: “Just think one more time: Is that really in my best interest?”