Can brain activity predict voting behavior? A new study in Scientific Reports suggests that the answer is more complicated than expected. By measuring the N400 brainwave during exposure to political statements, researchers found differences in how Democratic and Republican voters tend to process information that contradicts their political affiliations. But this brainwave was not a reliable predictor of actual voting behavior.
The researchers aimed to understand the gap between what people consciously express and the subconscious processes that influence their decisions, particularly in voting. To do this, they used a neural measure known as the N400, which is a specific type of event-related potential (ERP). ERPs are patterns of electrical activity in the brain that occur in response to specific stimuli, such as reading a sentence or seeing an image. The N400, in particular, is linked to processing meaning and detecting when something is unexpected or doesn’t align with what we already believe or know.
For example, when a person reads a statement that contradicts their deeply held beliefs, the brain generates a stronger N400 response. By comparing participants’ brain activity when they read political statements that either aligned or conflicted with their political preferences, the researchers sought to determine whether these implicit neural responses could reveal deeper, unconscious preferences and potentially predict how someone would vote.
“I was interested in this topic because we have limited insights into the individual-level processes that lead voters to make up their mind on who to vote for. These processes constitute, to a certain extent, a black box,” said study author Emmanuel Mahieux, who received a PhD in experimental psychology and neuroscience from University College London.
“Even recently, after the U.S. presidential debate, campaigning professionals were running focus groups with undecided voters to understand their reactions to the debate. Although common themes sometimes emerge from such focus groups, they often provide mixed or unclear insights on how people will vote at an upcoming election.”
“Given recent advances in neuroscience, I wanted to see if neural responses to political statements could make better predictions of how voters would vote rather than relying on their responses to political statements. Giulia Galli and her colleagues had developed a fascinating paradigm testing this and I wanted to see if it could be replicated in the U.S. electoral context.”
The study took place in the United States during a highly polarized political period — the 2022 Texas gubernatorial election. Participants were 55 undergraduate students from the University of Texas at San Antonio. These participants were split among those who had decided to vote for the Republican or Democratic candidate, and those who were undecided.
Beginning 25 days before the election, participants’ brain activity was monitored using electroencephalography (EEG) as they read a series of 184 political statements. These statements covered three major issue areas that were important in the election: the economy, immigration, and societal issues like abortion and gun rights. Each statement ended with a conclusion that was either pro-Republican or pro-Democrat.
For instance, one statement might say, “The ownership of automatic assault weapons like AR-15s needs to be restricted,” which would reflect a Democratic stance, while another might conclude, “The ownership of automatic assault weapons like AR-15s needs to be protected,” reflecting a Republican view.
The key was to measure participants’ N400 brain responses to see whether the brain reacted differently to political statements that aligned or conflicted with their stated preferences. Importantly, the participants indicated whether they agreed or disagreed with the political statements they saw, allowing the researchers to compare explicit political preferences with the N400 brainwave response.
As expected, the Democratic voters in the study showed larger N400 brain responses when reading statements that supported Republican views compared to those that supported Democratic views. This aligned with the hypothesis that when people encounter ideas that go against their beliefs, their brains show greater activity in the N400 region.
Unexpectedly, Republican voters showed no such distinction in their brain activity. Their N400 responses to both pro-Republican and pro-Democratic statements were similar, meaning their brain did not react more strongly to statements that opposed their political beliefs. This suggests that, for these voters, there might have been a disconnect between their neural processing and their explicit political beliefs.
“My expectation was that decided Democrats and decided Republicans would present opposite N400 patterns in their neural responses to political statements,” Mahieux told PsyPost. “The surprise was that decided Republicans did not show the expected pattern, as their N400 responses were ambivalent between pro-Democratic and pro-Republican statements. For undecided participants who voted Republican, their N400 response was more similar to the N400 response of decided Democrats than of decided Republicans, although their small sample size (N = 6) is too small to draw conclusions with certainty.”
Among the 55 participants in the study, 31 ended up voting for the Democratic candidate for governor, while 24 voted for the Republican candidate. When the researchers looked at the overall predictive power of the N400, they found it didn’t reliably forecast voter behavior. While it captured participants’ implicit reactions to political statements, it didn’t determine how they would vote in the election.
Instead, it appeared that voting behavior was more closely tied to partisan identity — whether a person identified as Republican or Democrat — rather than the deeper, subconscious preferences revealed by the N400 brainwave.
“Explicit political preferences — those that voters explicitly stated — were overall the best predictors of how people would vote,” Mahieux explained. “However, we found that what some participants said they believed about politics diverged from what brain measures suggested were their deeply-held preferences. Our brain measure of implicit preference — the N400 electrical potential of the brain — suggested that Republicans in our sample were overall less polarised and less aligned with their party than their explicitly stated preferences indicated.”
“Our neural measure suggests that some of these voters might have had conflicting or at the least divided views about some political questions. This shows us a new angle of political decision-making which surveys and polls cannot access.”
Like all research, this study had limitations. The sample size was relatively small, and all participants were young adults from a single university, which may limit the generalizability of the findings to broader, more diverse populations. Additionally, the study only captured a snapshot of brain activity at a particular moment in time, right before an election. Political views can evolve, and further research might investigate how consistent these brain responses are over time or whether they change with shifting political climates.
While the N400 did reflect implicit preferences, these preferences did not always align with the participants’ final voting choices. This opens up interesting questions for future research, particularly about the role of unconscious preferences in decision-making and how these interact with conscious beliefs and social identities.
“I would like to test which measures of implicit political preference improve existing models’ ability to predict how undecided voters will vote,” Mahieux said. “This is because we still have a limited understanding of the individual-level psychological factors that shape undecided voters’ voting choices. In general, I think it will be interesting to see which non-political measures -like brain potentials- are effective at predicting political behaviours like voting.”
The study, “The N400 effect captures nuances in implicit political preferences,” was authored by Emmanuel Mahieux, Lee de-Wit, Leun J. Otten, Joseph T. Devlin, and Nicole Y. Y. Wicha.
