The Role of the Posterior Medial Prefrontal Cortex in Confirmation Bias

Post by Shireen Parimoo 

What's the science?

Our beliefs can be influenced in many ways. For example, people often exhibit confirmation bias, which is the tendency to ignore information that is inconsistent with their existing beliefs while giving greater weight to information that confirms their beliefs. Similarly, people are more likely to be influenced by strongly expressed views than by weakly expressed views. It is important to understand how new information that could affect our decisions is processed. Activity in the posterior medial prefrontal cortex (pmPFC) has been implicated in monitoring and evaluating decisions, such as changing behavior after making a mistake. However, it is not known whether representations of beliefs in the pmPFC are sensitive to the strength (strong or weak) or type (consistent or inconsistent) of new evidence. This week in Nature Neuroscience, Kappes and colleagues used functional magnetic resonance imaging (fMRI) to investigate people’s sensitivity to the strength of new evidence and how this influences their judgments when the evidence confirms or contradicts existing beliefs.

How did they do it?

Participants completed the study in pairs across two testing sessions. In the first session, two participants individually played a real estate investment game, in which they were shown a property with a price and asked to (i) make a judgment about whether the true price was higher or lower than the price that was shown, and (ii) place a wager on their judgment (between 1 and 60 cents). They were informed that if they were correct, they would receive the amount of money that they wagered and if they were wrong, they would lose that amount. In the second session, the pairs of participants underwent fMRI scanning in adjacent rooms. They were shown the same properties as before along with their judgments and wagers. Importantly, they were also shown what were ostensibly their partner’s evaluations, which could be consistent or inconsistent with their own judgments, as well as the amount of money that their partner seemingly wagered on their judgment, which could be high (strong evidence) or low (weak evidence). Participants then had to decide whether they would change the amount of money that they wagered (but not the judgment itself) based on their partner’s evaluations. Thus, the partner’s judgment indicated whether participants saw confirmatory or contradictory evidence, and the partner’s wager amount indicated the strength of the evidence.

The authors first assessed whether participants showed confirmation bias. That is, were participants more likely to change their final wager when presented with consistent compared to contradictory evidence? Then, they determined whether participants were influenced by the strength of new evidence when it was consistent or inconsistent with their judgments. Finally, they performed a moderated mediation analysis to examine the sensitivity of the pmPFC to the type and strength of evidence. This allowed them to determine whether pmPFC activity mediated the effect of strong and weak evidence on participants’ final wagers, and if this varied based on whether the evidence was consistent or inconsistent with their own judgments.

What did they find?

Participants exhibited a confirmation bias, as they were more likely to increase their wager when their partner agreed with them. The strength of new evidence only affected behavior when the evidence was consistent with their existing beliefs. That is, participants were more likely to increase their wager when the evidence was strong (their partner made a large wager as well) than when the evidence was weak (their partner made a small wager). However, when their partner’s judgment did not align with their own judgments, the strength of the new evidence had no effect on their final wager.

The strength of new evidence was negatively correlated with pmPFC activity, but only for confirmatory evidence. Specifically, activity in the pmPFC was lower when the participant’s partner placed a higher wager on the same judgment, but pmPFC activation was not related to the partner’s wager when their judgment was inconsistent with the participants’ own judgments. Similarly, activity in the pmPFC mediated the effect of the strength of the new evidence and the participants’ final wager, but only when their partner’s judgments were consistent with their own. In other words, there was a reduction in pmPFC activity when participants were presented with strong confirmatory evidence, which was related to an increase in their final wager. Thus, the strength of new evidence only affects behavioral and neural responses when the evidence confirms existing beliefs.

What's the impact?

This study is the first to show that the pmPFC only tracks the strength of new evidence when it is consistent with prior beliefs, and that regardless of how strong the new evidence is, it does not influence behavior when it contradicts prior beliefs. These findings have important implications in numerous contexts, ranging from personal lifestyle to advocacy and policy making.

Kappes et al. Confirmation bias in the utilization of others’ opinion strength. Nature Neuroscience (2019). Access the original scientific publication here.