Can fNIRS predict future behavior?

When participants watch anti-smoking advertisements or hear short arguments for wearing sunscreen, the amount that their mPFC fires during that message can tell neuroscientists how much the participant will reduce their smoking or wear sunscreen. This brain activity is sometimes an even better predictor of future behavior than participants’ own self-report, which tends to be biased. This has immediate practical relevance – if marketers or policy teams can use neuroimaging to predict how effective their messaging campaigns will be, they could better predict how well those messages will perform in the general population and thus save a lot of money on potentially bad campaigns.

However, prior studies of this phenomenon have used a technology called functional magnetic resonance imaging, or fMRI for short. fMRI is good at finding brain activity in specific brain areas but is very expensive equipment to use. Most of these machines, therefore, are found in Academic hospitals. This is not ideal for studying the brain-behavior link anywhere outside of the academic context. Our goal with this study was to replicate prior persuasion neuroscience research in a new neuroimaging technology, called functional near infrared spectroscopy (fNIRS). fNIRS is a more mobile and affordable version of fMRI. If brain activity as recorded with fNIRS could also predict persuasion behavior, fNIRS could be used by non-academics for the practical applications described previously.

Approach

In the present study, we borrowed a paradigm from Vezich et al. (2017) that was originally used to study persuasion using fMRI. First, we recruited 84 participants – 37 people who regularly wore sunscreen, and 47 people who did not. Those participants filled out a pre-survey that provided information about how often they wore sunscreen per week, and how often they intended to use sunscreen but didn’t. Then, while wearing the fNIRS headcap to record brain activity, these participants listened to several arguments about why they should wear sunscreen. There were different kinds of arguments, such as gain-framed messages (“you will look younger if you wear sunscreen”) and loss-framed messages (“you are more likely to get skin cancer if you don’t wear sunscreen.”) The arguments were written out on a computer screen and also read aloud by a recorded voice. After watching all the messages, participants again answered a post-survey about how much they now intended to use sunscreen. Finally, a week later, those participants were re-contacted and asked to provide information about how much they had used sunscreen in the week since the experiment took place.      

Results

When analyzing the data, we first looked to see if a participant’s average amount of brain activity in the mPFC during the sunscreen messages would predict how much they would use sunscreen in the following week. This was our main hypothesis taken from previous fMRI studies, and this replicated successfully at a significance level of p < 0.001. Then, we also asked some more nuanced questions from prior fMRI work – does being an existing user or non-user of sunscreen modulate the brain-behavior link? Are there differences in brain activation to different kinds of persuasive messages? Our answer to the first question was that yes, non-users had a tighter brain-behavior link than did users. This is perhaps because sunscreen users have already thought about all of these persuasive arguments before, so their brain isn’t doing anything new when seeing the messages again. We did not find any differences in how the kind of persuasive message affects behavioral predictability from brain activity, which does not replicate previous fMRI results. Finally, we found a new kind of result in this data – activity in a different brain area called the right lateral prefrontal cortex (right lPFC) predicted resistance to the sunscreen messages, rather than persuasion.

Conclusions

The results of this study are largely encouraging – we can use fNIRS to predict when people will change their behavior in line with persuasive messages. Most of the results previously found in fMRI studies on this topic can be repeated with the fNIRS technology. The only qualification to this statement would be where we did not find brain activation differences in different kinds of persuasive messages. This may be because fNIRS is slightly less accurate than fMRI in how well it records brain activity. But this may also be because prior fMRI findings have not been retested before, so it is not actually replicable.

Therefore, it is unclear if fNIRS could be used as a good substitute for fMRI in academic studies. However, for practical applications of measuring brain activity to predict behavior (our main hypothesis), fNIRS has many uses. Indeed, fNIRS is more affordable to acquire and operate, as is more portable than fMRI. In fact, our team has already used it to study persuasion processes in the brain in samples of native Arab men in Amman, Jordan, a population that has not been studied before in persuasion neuroscience research (Burns et al., 2019). We are also currently using it to study the interpersonal persuasion process during board room meetings with real business teams, rather than just undergraduate college students. fNIRS would therefore seem to be an exciting technology to consider adding to one’s marketing arsenal.

REFERENCE
Burns SM, Barnes LN, Katzman PL, Ames DL, Falk EB, & Lieberman MD (2018), A Functional Near-Infrared Spectroscopy (fNIRS) Replication of The SunscreenPersuasion Paradigm. Social Cognitive and Affective Neuroscience, May 4, 628-636

CONTACT INFORMATION
University of California
Contact: Shannon M Burns, smburns47@ucla.edu

This article was originally published in the Neuromarketing Yearbook. Order your copy today