When it comes to making choices, past decisions may play a surprisingly large role. The traditional view of decision-making is that our choices are guided by what we remember about the outcomes of previous choices we’ve made. But in recent years, a complementary idea has arisen: that the mere memory of the choices we make, whatever their outcome, may affect future decisions.
Essentially, this idea posits that “when you’re deciding between two options that are roughly equivalent, you tend to prefer the previously chosen option over the other non-chosen option,” explains Lennart Luettgau, a cognitive science graduate student at Heinrich Heine University in Düsseldorf, Germany. In a recent paper in Nature Communications, he and his colleagues set out to look for evidence in the brain that might support such two-way influence between choice and memory.
The group’s findings suggest that remembering a previous choice may lead us to favor that choice in the future, regardless of any outcome of the decision. “The choice behavior itself has an effect on how memories are represented at the neural level,” Luettgau says. The work could have implications for understanding seemingly irrational human tendencies, when for example choosing particular products or political candidates.
In the new study, participants ranked pictures of three desserts in order of preference. Then, in a classic Pavlovian learning setup, the researchers taught participants to associate neutral images—in this case, specific Japanese kanji characters (distinctive images with no real meaning for the subjects, none of whom knew Japanese)—with those desserts. In this way, each kanji was associated with a high, intermediate, or low-value sweet. The participants learned to associate two different kanji with each food item.
Next, the participants were asked to choose between two kanji associated with differently valued treats. In this step, called choice-induced revaluation, the pairwise choices included only one of the two kanji associated with each food item. In some of the experiments, participants were asked to select between kanji linked with an intermediate versus high-value dessert; in others, they chose between kanji linked with an intermediate versus low-value dessert.
In the final phase of the study, participants were shown all of the different kanji pairings and asked to select one of each pair. Given the choice between two kanji associated with the same dessert, participants now showed a preference for the one chosen during revaluation. What’s more, a kanji that was shown but not selected during revaluation became less favored than the unpresented item associated with the same food.
Some of the participants also underwent functional magnetic resonance imaging (fMRI) scans before and after revaluation to look for a neural representation of the value-based decisions. Luettgau’s team relied on the tendency of neural tissue to become less responsive immediately after being activated. If two stimuli—such as a kanji and a food item—are tightly associated, they should stimulate the same patterns of activity. The researchers assessed the strength of the kanji-food associations by showing two stimuli in rapid progression and measuring suppression of activity in brain areas involved in associative memory and reward responses.
The brain imaging results supported the behavioral findings. Neural representations of the learned associations in the hippocampus and lateral orbitofrontal cortex were altered based on what the participants chose. There were stronger associations for options selected during the revaluation phase and weakened associations for the non-selected items. “This happens even without being presented with the outcomes or the rewards that are associated to their choices,” Luettgau says. “This strongly suggests to us that decisions themselves are not only guided by the memory systems … but also that choices really influence or change the associative memory structures.”
Few people have looked for a neural basis of preference changes that occur in the absence of some kind of reward, says Bradley Love, a professor of cognitive and decision sciences at University College London who was not involved in the study. He commends the approach, noting that in real life “these kinds of decisions in which we don’t receive feedback are probably more common than the typical ones used in the lab… where we have immediate feedback.”
The involvement of prefrontal areas and the hippocampus “really makes sense” in learning and decision-making tasks, he adds. “To me, this all dovetails with the function of these areas in other paradigms and studies.”
The results could help explain some seemingly irrational consumer behavior, such as choosing which products to buy or which candidate to vote for. “We try to be consistent with what we did in the past,” Luettgau notes. “So even though we experience that our past choices or past votes might have turned out not so favorable, we might still continue voting for the same candidates.”
Love agrees that this is one interpretation, though not the only one. Through the memory recall needed to choose between the options, people may be strengthening prior associations or beliefs. Or perhaps, he notes, “we just come to like what we choose.”