A new study published in Scientific Reports, a prestigious journal from the Nature group, shows how specific electrical activity in the brain can predict voting choices. The research stems from the work of an interdisciplinary team of neuroscientists, psychologists, and political scientists, formed through a collaboration between Sapienza University of Rome, Kingston University London, Luiss Guido Carli, University of Rome “Tor Vergata,” and the University of Melbourne.

The study was conducted in the five weeks leading up to the 2019 European elections, a period when the political debate was dominated by the clash between populist and mainstream parties. Research participants took part in a pre-election survey with statements expressing varying degrees of populist positions on a wide range of issues, from immigration to the economy to the European Union. However, unlike a traditional pre-election survey, participants answered the survey while their brain activity was recorded in the laboratory led by Viviana Betti in the Department of Psychology at Sapienza University.

The authors used an electroencephalograph (EEG), a device that records the brain’s electrical activity. In particular, the focus was on the N400, an electrical brain activity that manifests as a wave the brain produces whenever we encounter information that contradicts our beliefs. In the week following the vote, all participants were contacted again and asked which party they had voted for. The researchers expected—and the results confirmed—that participants’ brains would respond with an N400 wave to any statement that contradicted their political beliefs, and that the amplitude of this N400 wave would predict their vote.

Indeed, the results showed that the brain responded differently based on the participants’ political beliefs: the N400 wave was clearly evident in response to political disagreement, for example, when participants who supported non-populist parties were shown populist content. Not only that, but the magnitude of this brain response could predict with high accuracy whether participants would later cast a populist vote or a vote for a mainstream party, outperforming traditional voting predictors typically used in election forecasts. Surprisingly, the N400 brain signal was evident in response to statements about economic issues, such as universal basic income. In essence, the brain responded differently to populist and non-populist content, but only when that content addressed economic issues rather than topics traditionally associated with populism, such as sovereignty or opposition to the establishment.

What do the results of this study tell us? First of all, recording brain activity, in addition to explicit responses to surveys, enhances the predictive power of the surveys themselves. And it's not hard to see why, considering that survey responses—regardless of the type—are often influenced by the so-called social desirability bias, which is the tendency to provide socially acceptable answers. In this way, recording brain activity allows us to “bypass” people’s explicit responses and gain access to their more genuine opinions or attitudes.

Another potential benefit of recording brain activity relates to undecided voters. In a previous study from 2016, conducted by one of the authors in the days leading up to the Brexit referendum, the brain responses of undecided voters already predicted whether they would later vote “Leave” or “Remain.” This demonstrates that undecided voters may have “embryonic” preferences they are not yet aware of, but which can still be detected through brain activity.

Traditional pre-election polls certainly have advantages over measuring brain activity, not least the ability to gather information from a large number of people with reduced costs and time. However, the results of this new study show that using neuroscientific methods allows for more accurate and detailed predictions. Building on neuroeconomics studies that have shown how an individual's brain responses can predict collective choices, this study could be seen as a first step toward “neuroprediction” of voting choices.