Access the full text.
Sign up today, get DeepDyve free for 14 days.
Activity in medial prefrontal cortex (mPFC) during persuasive messages predicts future message-consistent behavior change, but there are signiﬁcant limitations to the types of persuasion processes that can be invoked inside an MRI scanner. For instance, real world persuasion often involves multiple people in conversation. Functional near infrared spectroscopy (fNIRS) allows us to move out of the scanner and into more ecologically valid contexts. As a ﬁrst step, the current study used fNIRS to replicate an existing fMRI persuasion paradigm (i.e. the sunscreen paradigm) to determine if mPFC shows similar predictive value with this technology. Consistent with prior fMRI work, activity in mPFC was signiﬁcantly associated with message-consistent behavior change, above and beyond self-reported intentions. There was also a difference in this association between previous users and non-users of sunscreen. Activity differences based on messages characteristics were not observed. Finally, activity in a region of right dorsolateral PFC (dlPFC), which has been observed with counterargu- ing against persuasive messages, correlated negatively with future behavior. The current results suggest it is reasonable to use fNIRS to examine persuasion paradigms that go beyond what is possible in the MRI scanner environment. Key words: fNIRS; persuasion; replication; mPFC; dlPFC; health behavior functional near infrared spectroscopy (fNIRS), which allows for Introduction more portable, mobile and flexible neuroimaging than what is Persuasion is the art of soft power—using words and images to possible using functional magnetic resonance imaging (fMRI). convince others to change their beliefs and behavior. Although Our goal was to replicate a recent fMRI persuasion paradigm the study of persuasion has an incredibly long history, our using fNIRS to see whether the brain-as-predictor approach understanding of the brain’s role in persuasion processes is in- would succeed with this method as well. credibly short. It is only in the last decade that a concerted effort has begun to localize some of the brain regions that contribute Past fMRI persuasion research to persuasion and its behavioral consequences. Critically, Over the past decade, a number of studies have identified researchers have used a ‘brain-as-predictor’ approach (Falk medial prefrontal cortex (mPFC) as a central predictor of whether et al., 2015a) to predict whether an individual or even a mass persuasive messages will be successful in changing behavior. In a audience will respond in the desired way to a persuasive mes- sage. In the current research, we extend this work to the use of number of studies aimed at persuading behavior changes, Received: 17 July 2017; Revised: 17 April 2018; Accepted: 23 April 2018 V C The Author(s) (2018). Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact firstname.lastname@example.org Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 S. M. Burns et al. | 629 ventromedial prefrontal cortex (vmPFC) or mPFC activity during than laying in a dark and loud MRI scanner. Additionally, fNIRS message exposure was associated with smoking cessation (Chua is much less expensive to acquire and operate than fMRI, mak- et al.,2011;Falk et al., 2011; Cooper et al., 2015; Riddle et al.,2016), ing it a more feasible option for policy makers or community increased sunscreen use (Falk et al., 2010a; Vezich et al.,2017) researchers interested in applying neuroscience techniques to and decreased sedentary behavior (Falk et al., 2015a; Cooper et al., their work. 2017). In fact, when controlling for prior behavior and self- Given these advantages, fNIRS has the potential to be ex- reported intentions to follow the messages, mPFC activity still tremely useful in both the continued development of persuasion adds significantly more predictive ability to persuasion outcome neuroscience and in the application of these findings. However, models (Falk et al., 2011; Riddle et al., 2016), accounting for as there are downsides to using fNIRS for neuroimaging—namely, much as 23% more variability in behavior (Falk et al., 2010a). its signal cannot penetrate more than a few centimeters into the Additionally, the predictive utility of mPFC responses is not cortex, its spatial resolution is not as fine as fMRI (fNIRS signal restricted to the individual level in persuasion research. Other correlates best with MRI when an ROI radius of five voxels is studies have also investigated the neural correlates of persua- used, Cui et al.,2011), and there is no structural brain image gen- sion in focus groups, and then used that information to predict erated so the localization of activity must be inferred through the population-level behaviors. Activity in mPFC areas of partici- 10–20 external positioning system. Thus, before fNIRS can ser- pant samples has been used to predict the performance of iously be used to further our understanding of the brain during large-scale phone (Falk et al., 2012) and e-mail anti-smoking persuasion in naturalistic contexts, it is important to empirically campaigns (Falk et al., 2016), the future sales of new music test whether or not the brain-as-predictor findings from fMRI can (Berns and Moore, 2012) and amount of message propagation replicate with fNIRS as well. between people (Falk et al., 2013). However, all of the above results come from fMRI research. Current study There are many benefits to using fMRI technology for studying the neural correlates of persuasion, such as differentiating be- Given the need for testing the potential of fNIRS in persuasion tween the roles of prefrontal subdivisions in persuasion (Cooper neuroscience, the main goal of this study was to replicate our et al., 2017) or analysing the interaction between cortical and prior fMRI research. Specifically, we replicated the procedure of subcortical areas in message processing (Ramsay et al., 2013). Vezich et al. (2017). In this work, participants viewed persuasive But fMRI is also a highly expensive apparatus that restricts the paragraphs about using sunscreen everyday, and then were sur- movement of research participants and is not transportable. veyed about their intentions for future sunscreen use. A week These constraints limit the types of real world persuasion that later, they were re-contacted about their sunscreen use follow- can be studied with fMRI. For instance, populations in rural ing the imaging session. They found a significant relationship areas or who are mobility impaired cannot be easily recruited to between activity in the mPFC during message exposure and university laboratories, and fMRI cannot be brought to them. later sunscreen use after the scan. Our main goal then was to Persuasion is also a highly contextualized and social phenom- replicate these results in this a priori region of interest. While enon, but fMRI requires that participants be removed from the this particular study design does not capitalize on the advan- stimulus-rich environment they normally experience and dis- tages fNIRS poses for naturalistic experiments, a close replica- connected from real-time interaction with other people. Some tion of an fMRI paradigm will hopefully bolster confidence in studies have studied real-time social influence with fMRI by vir- using fNIRS for persuasion work in these other contexts. tually connecting participants to an outside agent or another The Vezich et al. study also probed the kind of persuasive participant in a separate MRI scanner (Montague et al., 2002; messaging that elicited the best predictive neural activity, and Schilbach et al., 2006; Saito et al., 2010; Wilms et al., 2010), but whether or not this predictive ability varied as a function of indi- these paradigms are still much more artificial than everyday vidual differences across participants. These questions help clar- interpersonal interaction where influence often occurs. ify why the mPFC works as a predictor, and in what contexts its In contrast, functional near infrared spectroscopy (fNIRS) is activity is a reliable indicator of later behavior. Specifically, the a similar neuroimaging method to fMRI that has not been used researchers tested the extent to which ‘gain’ vs ‘loss’ framing of a to study persuasion neuroscience before, but has important message would elicit differential mPFC activity, considering that advantages over fMRI for this purpose (Cui et al., 2011; Ferrari factually equivalent information presented as a gain can increase and Quaresima, 2012). Both imaging modalities indirectly meas- motivation and self-relevant valuation more than loss framing ure brain activity via the concentration of oxygenated and (Kahneman and Tversky, 1979; Salovey and Wegener, 2003). In deoxygenated hemoglobin (HbO and HbR) in the cortex. But other words, it was predicted that describing the advantages of while fMRI uses the magnetic properties of HbO to make these using sunscreen would elicit greater persuasion-related brain ac- measurements, fNIRS relies on the optical properties. Because tivity than describing the consequences of not using sunscreen. skin and bone are relatively transparent to near infrared light, They found that mPFC activity in response to gain messages was while HbO is not, researchers can measure HbO concentration significantly greater than activity during loss messages, and that changes in the cortex by affixing light emitters and detectors to the extent of this difference in each individual predicted amount the scalp, measuring the intensity of light that propagates of sunscreen use. through the head, and then calculating HbO concentration In addition, the activation difference between previous users changes via the Modified Beer-Lambert Law (for more informa- and non-users of sunscreen was investigated (i.e. participants tion on the biophysics of fNIRS, see Ferrari et al., 2004; who either did or did not use sunscreen prior to participation in Scholkmann et al., 2014). Because of this, a simple cap or head- the experiment). Psychological models of persuasion and health band can be worn to hold the optodes to the head. This means behavior make a distinction between the processes necessary participants are free to sit up and move around during experi- ments. This considerably widens the range of experimental for initial behavior enactment and subsequent behavior main- designs researchers may use, enabling them to test how neural tenance (Weinstein et al., 1998; Miilunpalo et al., 2000; Fogg, activity during persuasion occurs in more naturalistic situations 2009), and Vezich and colleagues indeed found that mPFC Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 630 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 6 activity was associated with future sunscreen use in previous Experimental persuasion task non-users of sunscreen, but not in existing sunscreen users. The persuasion task consisted of 32 different persuasive mes- Finally, Vezich et al. tested messages about ‘why’ to use sun- sages about using sunscreen. These messages were divided into screen vs ‘how’ to use sunscreen in order to investigate the dif- four conditions, such that the messages were framed to de- ferential effects of intent and action planning, but the results scribe either the benefits of using sunscreen (‘Gain’ messages), for this contrast were located in the rostral inferior parietal lob- the risks of not using sunscreen (‘Loss’), how to use sunscreen ule and posterior inferior frontal gyrus, which are areas that our appropriately (‘How’) and neutral facts about sunscreen formu- fNIRS set up was not designed to measure. So in the current lation (‘Fact’). Examples of messages in each of these conditions work, we repeated the contrasts of gain vs loss and users vs can be found in Table 1. Participants read each message one at a non-users to test whether the fMRI results in the mPFC general- time on a computer screen, while a pre-recorded voice read the ize to fNIRS. It is important to note that while the general MPFC- message aloud at the same time. Each message was presented behavior effects have been observed numerous times, the mPFC for 16–20 s, and appeared in a pseudo-random order such that effects due to gain vs loss were novel to the Vezich study and one message from each condition was presented in random user vs non-user investigations are also few in number (Weber order within a block, and eight blocks total of that same order- et al., 2015). ing was shown to the participant. A jittered rest period sepa- This work will also investigate one more hypothesis that rated each message, showing just a small cross hair on the was not examined in Vezich et al. (2017)—whether or not the screen. After four blocks, participants were given a brief break right dorsolateral prefrontal cortex (right dlPFC) might also be before continuing on to the final four blocks when they were ready. The entire task lasted 18 min. associated with behavior change after persuasive messaging— specifically, negatively associated. In Falk et al. (2010b), an area Post-study questionnaire. After the experimental task, partici- of the right dlPFC activated for unpersuasive messages more pants again reported how many days and how many times over than persuasive messages, when viewing arguments about the next 7 days they intended to use sunscreen. Participants activities that people would have few strong opinions about. In also answered questions about how confident they were in their other studies, dlPFC activity was associated with arguing ability to use sunscreen more often, their beliefs on the poten- against previously held attitudes (Kato et al., 2009; Ramsay et al., tial benefits of increasing sunscreen use and the persuasiveness 2013). Additionally, in pilot work, our group has found that of the messages they viewed as well as their openness to per- counterarguing against persuasive messaging was associated suasive messaging in general. These ratings were provided on a with both increased right dlPFC and decreased message consist- sliding scale between 0 and 100. ent behavior change (Falk et al., unpublished data). Based on these findings, increased right dlPFC activity during persuasive One-week follow-up questionnaire. Eight days after completing messaging might reflect counterarguing and rejection of per- the in-lab session, participants were emailed a follow-up ques- suasive messages. If paired with a measure of mPFC activity, tionnaire and were asked to report on their sunscreen use over data from these two regions might give a more robust predic- the past 7 days as well as their intention to use sunscreen over tion of future behavior after persuasive message exposure. We the next 7 days, in terms of number of days they used sunscreen therefore investigated whether decreased activity in the right over the previous week and number of days per week they now dlPFC will also be a predictor of persuasion, and if combining intend to use sunscreen in general. Before leaving the lab after the right dlPFC and mPFC into one model will be an even better the scanning session, participants consented to being contacted predictor than either region alone. again for general research inquiries, but were not expecting to be re-contacted about their sunscreen use specifically. Materials and methods fNIRS neural data acquisition Participants Neural data during the persuasion task was measured using A total of 84 participants were recruited for participation in this near infrared spectroscopy (fNIRS). The fNIRS system used in study (37 sunscreen users, 47 non-users). This is roughly double this study was the fNIR Imager 1000 from fNIR Devices (fnirdevi- the sample size of Vezich et al., because it was unknown how ces.com). This system measured the relative changes in oxygen- large the effect size would be for this study in the fNIRS modal- ated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) in ity. Of these participants, 3 did not complete the study, and the the prefrontal cortex while participants proceeded through the data from 12 more participants was deemed unusable due to experimental persuasion task, using near infrared wavelengths poor neuroimaging data quality. The final sample consisted of of 730 and 850 nm. A diagram of the source and detector optode 69 participants—29 sunscreen users and 40 non-users. All were layout can be seen in Figure 1. This layout is fixed within a female right-handed undergraduate students in university semi-rigid headband device, so the optode spacing of 3 cm was (M age¼ 21.25, s.d. age¼ 2.66). Written informed consent was standardized across all participants. To ensure that the head- obtained and all participants were paid for their participation. band was positioned in the same place for each participant, a The study protocol was approved by the University of small indicator line was located in the middle of the bottom California—Los Angeles Institutional Review Board. edge of the headband, and the headband was secured to the head so that this indicator line rested just above the nasion Pre-study questionnaire. In this questionnaire prior to completing point (Figure 1A). This provided ample coverage of the medial, the experimental task, participants were asked to report how ventromedial and lateral prefrontal cortices (Figure 1B). many days in the previous 7 days they wore sunscreen, in add- After securing the fNIRS headband to participants’ heads, ition to six distractor questions such as how many days they experimenters checked the signal quality coming from each flossed or exercised. Finally, participants reported how many data channel by running a test data acquisition session in COBI days over the next 7 days they intended to use sunscreen. Studio acquisition software (BIOPAC Systems, Inc.). If the Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 S. M. Burns et al. | 631 Table 1. Examples of the four types of messages shown to participants about using sunscreen Message type Example Fact ‘In USA, sunscreen products are regulated as over-the-counter (OTC) drugs by the U.S. Food and Drug Administration (FDA). The FDA has several safety and effectiveness regulations in place that govern the manufacture and marketing of all sunscreen products, including safety data on its ingredients.’ How ‘Apply liberally and evenly to all exposed skin. The average adult in a bathing suit should use approximately one ounce of sunscreen per application. Not using enough will reduce the product’s SPF and the protection you get. Be sure to cover often-missed spots: lips, ears, around eyes, neck, scalp if hair is thinning, hands and feet.’ Gain ‘Daily application of broad spectrum sunscreen with SPF 15 or higher has been clinically demonstrated to keep skin looking younger, more elastic and healthier. Maintaining good habits about using sunscreen is crucial for having beautiful skin for years to come, that not only looks better but is more likely to remain healthy.’ Loss ‘Studies have found that inconsistent use of sunscreen is associated with a number of skin issues. These include, but are not limited to, wrinkling, sagging, splotchy, leathery, uneven skin. To avoid these issues, you should apply broad- spectrum sunscreen with SPF 15 or higher to any and all skin that will be exposed to the sun.’ Fig. 1. (A) Placement of the fNIRS headband device on the participant’s head. Positioning across participants was standardized using the 10–20 external landmark sys- tem. (B) Approximate location of each channel of data (numbered), projected downward onto the surface of the MNI standard brain cortex from the 10–20 positions. measured intensity of light signal received by detector optodes indicated that they were finished with this questionnaire, they was >3500 V or <200 V, efforts were made to bring the signal were paid for their participation in this session and allowed to into the acceptable range. This includes changing the gain level leave. of the detectors, tightening or loosening the headband on the One week after the laboratory session, participants were participant’s head and pinning any hair out of the way of the e-mailed a link to the follow-up questionnaire in order to gauge optodes. Once all possible improvements were made to the sig- their sunscreen use and intentions over time. If participants nal strength, the experimenters then initiated the data record- completed this survey, they were invited to return to the lab to ing session in the acquisition software and turned off the lights receive payment for this final part. in the experimenting room to reduce ambient light noise. fNIRS data analysis Procedure Neural fNIRS data were first analysed for signal-to-noise ratio to The procedure for this study involved two sessions—one in the determine its usability. If the amplitude of the light intensity in lab and one a week later online. In the first week, upon arrival, a channel fell <500 V or >4200 V, indicating that a good connec- participants were given information about the study and asked tion between the scalp and optode was not achieved, then that to provide informed consent. Then, participants filled out the channel of data was removed. If more than half of a partici- pre-study questionnaire in privacy. After completing the ques- pant’s data channels were over or under saturated in this way, tionnaire, experimenters attached the fNIRS headband to their then that entire participant was removed from analysis. head, checked and adjusted signal levels, initiated fNIRS data Thirteen participants were removed from analysis for this rea- acquisition and then began the experimental persuasion task. son, and the majority of other participants had at least one At this point experimenters then turned out the lights and left channel removed. the room, so that participants could read the experimental mes- Raw light intensity fNIRS data were then preprocessed in nirsLAB (http://nirx.net/nirslab-1/) using a band-pass sages in private for the entirety of the task. Once this task was finished, experimenters returned to the room to turn on the filter of 0.01–0.2 Hz to remove cardiac fluctuation and slow lights, stop neural data acquisition, remove the fNIRS headband signal drift. Motion artifacts were also identified and removed and open the post-study questionnaire for participants to fill if they took the form of spikes or discontinuities that out before again leaving the room. As soon as participants exceeded 5 s.d. from the variance of the rest of the data. Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 632 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 6 Fig. 2. Image showing strength of correlation between activity in each signiﬁcant data channel and sunscreen use behavior 1 week after the imaging session. Right and left mPFC were correlated r¼ 0.314 and 0.296, respectively, and right dlPFC was correlated r¼0.205. In a binomial regression model, right and left mPFC showed a signiﬁcant positive association with future behavior over and 2 2 2 above behavioral intentions [v (2, n¼ 51)¼ 31.093, P< 0.001, DR ¼ 0.344; and v (2, n¼ 53)¼ 16.502, P< 0.001, DR ¼ 0.261, respectively]. Right dlPFC was also signiﬁ- 2 2 cantly associated with future behavior [v (41)¼ 8.318, P¼ 0.0039, DR ¼ 0.348]. Including both mPFC and right dlPFC in the model resulted in a better prediction than either region alone (Ddeviance P<0.001). Fig. 3. There was an interaction effect in the right mPFC, such that previous Light intensity values were then converted to HbO and HbR non-users of sunscreen showed a signiﬁcant relationship between neural activ- hemoglobin concentration using the Modified Beer Lambert ity and future behavior while users did not. (A) The spatial location of this inter- Law. action. (B) Right mPFC parameter plotted with number of days sunscreen was After converting the data to hemoglobin concentration, just used post-scan, separated into user/non-user distinction. Linear trend lines in the HbO chromophore was chosen for analysis. While HbR each group are included for ease of visualizing the group difference. changes tend to be more spatially specific (Franceschini et al., 2000), HbO seems to have a relatively stronger correlation with our sample means without the assumption of normality. the fMRI BOLD signal in the prefrontal cortex during task en- Table 2 reports the sample means and the significance of the gagement (Cui et al., 2011), a comparatively stronger signal amp- pre-post differences. There was a significant increase in self- litude (Hoge et al., 2005) and slightly better signal-to-noise ratio reported sunscreen use between pre-scan and 1 week after the (Strangman et al., 2002). Because the neural areas of interest in scan for all participants (P< 0.001, Cohen’s d¼ 0.540). There was this study are larger cortical divisions, and the experimental also a significant increase in intention to use sunscreen, both goal is to investigate the ease of replicating prior fMRI research, between pre-scan and post-scan (P< 0.001, d¼ 0.825) and be- HbO was the preferred signal to investigate. tween pre-scan and 1 week after the scan (P< 0.001, d¼ 0.688). Finally, first level statistical analysis was conducted by mod- Within user and non-user groups, these changes were also sig- eling the recorded brain data with a convolution of the message nificant. Comparing pre-post sunscreen use difference scores condition design matrix and a canonical HRF. Pre-coloring was between previous users and non-users of sunscreen revealed used to account for serial correlations, as pre-coloring tends to that non-users increased their sunscreen use significantly more estimate temporal correlation in fNIRS data better than pre- than users (P¼ 0.0196, d¼ 0.586), although users still used sun- whitening (Ye et al., 2009). The resulting betas from this estima- screen significantly more than non-users did at the 1 week tion step were then used to predict sunscreen behavior 1 week follow-up (P< 0.001, d¼ 1.574). Non-users also had a larger in- after the scan. Spatial localization of these effects were deter- crease in intention pre-post (P¼ 0.001, d¼ 1.078) and pre-week mined through a built-in nirsLAB process to convert 10–20 co- later (P¼ 0.001, d¼ 1.487). Users still intended to use sunscreen ordinate locations to MNI space. more than non-users did, both at post-scan (P¼ 0.001, d¼ 1.320) and the follow-up (P< 0.001, d¼ 1.359). Further breakdown of Results the behavioral data can be found in the Supplementary data. Behavioral results fNIRS results Sunscreen use behavior and intentions were recorded in num- Relationship between mPFC activity and sunscreen use. To analyse ber of days per week that sunscreen was used/intended to be whether or not mPFC activity predicted sunscreen behavior, used. The distribution of this data was non-normal with cluster- ing around both response boundaries, so a 10 000 iteration per- Vezich et al. (2017) contrasted gain with fact messages as well as mutation test was used to shuffle the pre/post-label of paired gain with loss messages and then used the resulting contrasts scores or the user/non-user label of difference scores, then gen- to predict sunscreen behavior (minus intention) in a correlation erate an experimental null distribution against which we tested analysis. In this study, we chose to perform a binomial Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 S. M. Burns et al. | 633 Table 2. Means and standard deviations (in parentheses) of behavioral results Sunscreen use Intention to use sunscreen Pre-scan 1 week later Pre-scan Post-scan 1 week later Everyone 2.000 (2.797) 3.203*** (2.837) 2.884 (3.183) 5.087*** (2.672) 4.478*** (2.769) Users 4.759 (2.325) 5.241 (2.132) 5.966 (2.009) 6.724* (0.702) 6.241 (1.504) Non-users 0 (0) 1.725*** (2.331) 0.650 (1.610) 3.900*** (2.942) 3.200*** (2.785) Scores represent number of times a week participants intended to or engaged in sunscreen use. Bolded numbers signify a signiﬁcant difference between scores relative to an experimental null created via 10 000 iteration permutation test. *P< 0.05, **P< 0.01, ***P< 0.001. regression analysis controlling for intentions instead because that sunscreen non-users had a stronger relationship between that would better serve the bimodality of the behavior variable activity in the right mPFC and behavior than sunscreen users distribution. When we performed this test for the NIRS data did (Table 4; Figure 3). This was revealed by a significant inter- channels in the mPFC region (lower half of Brodmann’s area 10), action term between user status and NIRS data channel in a bi- removing outlier data points for having extreme influence on nomial regression model [v (4, n¼ 53)¼ 4.679, P¼ 0.0305]. the test results (DF Beta> 1), the results were similar to those Investigating the interaction showed that non-users had a sig- seen in Vezich et al. Specifically, the gain> fact parameters in nificant association between mPFC activity and future behavior 2 2 the channels over the right and left mPFC significantly [v (2, n¼ 30)¼ 11.36, P< 0.001, McFadden’s R ¼ 0.281] while 2 2 predicted sunscreen use behavior, over and above the effect of users did not [v (2, n¼ 23)¼ 1.22, P¼ 0.270, McFadden’s R self-reported intention, and accounted for significantly more ¼ 0.179], replicating Vezich et al.’s findings. variance than post-scan intentions alone [right mPFC: v (2, n¼ 51)¼ 10.705, P¼ 0.001, deviance change test D¼ 95.32, Relationship between right dlPFC and persuasion. To evaluate how P< 0.001; left mPFC: v (2, n¼ 52)¼ 11.658, P< 0.001, deviance well right dlPFC could predict behavior change, we again ran a change test D¼ 94.03, P< 0.001] (Figure 2). McFadden’s R calcu- binomial regression predicting sunscreen use behavior from lations showed that right mPFC activity accounted for 28.5% of activity in the right dlPFC for gain, loss and how messages, the remaining variance in future behavior and left mPFC incorporating post-scan sunscreen intentions as a regressor. accounted for 28.1% of the remaining variance, over and above There was a significant negative association between right what participants themselves predicted in their self-reported dlPFC activity and sunscreen use behavior 1 week after the ex- 2 2 intentions. For the gain> loss parameter, the right mPFC signifi- periment [v (2, n¼ 44)¼ 8.318,P¼ 0.004, McFadden’s R ¼ 0.348] cantly predicted sunscreen use behavior over and above inten- (Table 5; Figure 2). A deviance change test showed that includ- tion [v (2, n¼ 53)¼ 8.215, P¼ 0.004, deviance change test ing this region also significantly improved model fit over and D¼ 71.93, P< 0.001, McFadden’s R ¼ 0.215]. above the effect of self-reported intentions alone (D¼ 178.814, Alternatively, if we contrast all persuasive messages (gain/ P< 0.001). loss/how conditions) with neural baseline like some other fMRI Further, incorporating bilateral mPFC and right dlPFC into persuasion studies do (Falk et al., 2010, 2011), we also get signifi- the model all at once created a better model fit than either of cant results for both the right and left mPFC [right mPFC: v (2, those regions alone. The deviance change test over and above n¼ 51)¼ 31.093, P< 0.001, deviance change test D¼ 115.05, just mPFC was D¼ 151.144, P< 0.001, and over and above right 2 2 2 P<0.001, McFadden’s R ¼ 0.344; left mPFC: v (2, n¼ 53)¼ 16.502, dlPFC was D¼ 121.404, P< 0.001. McFadden’s R showed that P< 0.001, deviance change test D¼ 87.37, P<0.001, McFadden’s both regions accounted for an additional 58.46% of the variance R ¼ 0.261]. In both mPFC channels, the standardized coeffi- beyond self-reported intentions alone. cients for this contrast were greater than those for either gain- It is worth noting, however, that because this model was > fact or gain> loss (Table 3), indicating that for the present composed of three separate data channels, missing data in any dataset this contrast is the best predictor of future behavior. of these channels due to poor optode-scalp contact would result None of these results were affected by number of times rain or in a list-wise deletion of that participant’s data. Thus, out of 81 fog was recorded in the area in the week after the scan, which subjects that completed the study, only 27 had complete might be expected to influence sunscreen use. enough data for model inclusion. Thus, this reported effect size may be an inflated estimate. Message and participant characteristics results. We proceeded with the following analyses using only the baseline contrast, as Discussion it produced the strongest effect in the first analysis and the message/participant characteristic and the right dlPFC hypothe- The immediate goal of this investigation was to determine ses are less well replicated. Based on an analysis of variance of whether previously observed persuasion effects examined with average activity levels in the mPFC channels for gain, loss and fMRI would replicate with fNIRS. The larger goal was to deter- fact messages, there were no significant differences. This is a mine whether fNIRS is a useful tool for the study of persuasion notable deviation from the original study. There were also no that in the future can be used in novel contexts that cannot eas- significant differences between average activity for gain and ily be studied with fMRI. fact messages, or loss and fact messages. Two out of three of the original fMRI results were repeated in Consistent with the prior study, there was a difference be- this study. Specifically, greater mPFC activity during persuasive tween sunscreen users and non-users in the right mPFC, such messages about sunscreen use positively predicted future Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 634 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 6 Table 3. Binomial regression results for models testing relationship between mPFC activity and future behavior, with self-reported intention included as a regressor Model Parameter Unstandardized Standard Standardized Wald’s v P value coefficient error coefficient Gain > fact Behavior ¼ intention þ right mPFC Intercept 3.3891 0.4681 52.410 <0.001 intention 0.5980 0.0771 0.8917 63.439 <0.001 Channel 8 6363.2 1944.8 0.2375 10.705 0.001 Behavior ¼ intention þ left mPFC Intercept 3.2799 0.4583 51.227 <0.001 intention 0.5948 0.0735 0.8561 65.529 <0.001 Channel 10 8085.3 2368.0 0.2571 11.658 <0.001 Gain > loss Behavior ¼ intention þ right mPFC Intercept 3.0960 0.4107 56.834 <0.001 intention 0.5486 0.0655 0.8106 70.119 <0.001 Channel 8 5840.0 2037.6 0.2047 8.215 0.004 Behavior ¼ intention þ left mPFC Intercept 3.0650 0.4256 51.870 <0.001 intention 0.5374 0.0666 0.7620 65.175 <0.001 Channel 10 3142.8 1649.8 0.1273 3.629 0.057 Gain/loss/how > baseline Behavior ¼ intention þ right mPFC Intercept 3.6552 0.4855 56.674 <0.001 intention 0.5761 0.0747 0.8591 59.543 <0.001 Channel 8 12220.3 2191.5 0.5610 31.093 <0.001 Behavior ¼ intention þ left mPFC Intercept 3.0077 0.4430 46.100 <0.001 intention 0.5261 0.0698 0.7567 56.843 <0.001 Channel 10 7430.7 1829.2 0.3753 16.502 <0.001 Coefﬁcient values are reported in log odds, with unstandardized (relative to unit increase in predictors) and standardized (relative to standard deviation increase in predictors) values included. Table 4. Binomial regression results for models in users and non-users of sunscreen, in terms of the relationship between mPFC activity and future behavior Model Parameter Unstandardized Standard Standardized Wald’s v P value coefficient error coefficient Gain/loss/how > baseline Behavior ¼ intention þ user right mPFC Intercept 0.0304 2.1650 0.0002 0.989 intention 0.1578 0.3156 0.0492 0.250 0.617 channel 8 2162.1 1959.2 0.1133 1.218 0.270 Behavior ¼ Intention þ non-user right mPFC Intercept 3.2719 0.4901 44.564 <0.001 intention 0.4032 0.0794 0.6405 25.809 <0.001 channel 8 10249.0 3040.5 0.5064 11.363 <0.001 Coefﬁcient values are reported in log odds, with unstandardized (relative to unit increase in predictors) and standardized (relative to standard deviation increase in predictors) values included. Table 5. Binomial regression results for model testing relationship between right dlPFC activity and future behavior, as well as full model including both right dlPFC and bilateral mPFC Model Parameter Unstandardized Standard Standardized Wald’s v P value coefficient error coefficient Gain/loss/how > baseline Behavior ¼ intention þright dlPFC Intercept 2.976 0.4691 40.242 <0.001 Intention 0.4806 0.0756 0.7076 40.414 <0.001 Channel 1 3741.0 1297.1 0.246 8.318 0.004 Behavior ¼ intention þ right dlPFC þ right mPFC þ left mPFC Intercept 2.967 0.5763 26.499 <0.001 Intention 0.446 0.0909 0.650 24.110 <0.001 Channel 1 7877.9 2556.4 0.486 9.496 0.002 Channel 8 28 130.6 5178.2 1.396 29.512 <0.001 Channel 10 17 290.9 4607.0 0.914 14.087 <0.001 Coefﬁcient values are reported in log odds, with unstandardized (relative to unit increase in predictors) and standardized (relative to standard deviation increase in predictors) values included. Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 S. M. Burns et al. | 635 sunscreen use over and above self-reported intentions. There this analysis if they were missing data in any of three different was also a difference between users and non-users of sun- channels. This data loss is also a concern for possible applica- screen, such that previous non-users’ mPFC activity significant- tions of this approach that would aim to design personalized ly predicted future behavior, but users’ mPFC activity did not. persuasive messages based on one person’s neural activity. Both of these findings were reported in Vezich et al., and our The cause of this data loss is largely due to the specific fNIRS mPFC results had a similar effect size to previous fMRI findings technology that we used. While the Biopac unit is one of the (Falk et al., 2010, 2011). most affordable fNIRS units available, the rigidity of its head- We also observed that right dlPFC activity during persuasive band device meant that accommodating variation in partici- messages about sunscreen use was negatively associated with pants’ head shapes was difficult. Additionally, the flat and wide future sunscreen use, and information about the activity in this design of the optodes meant that light could not reach the scalp area paired with information from the mPFC offers better be- if there was any hair in the way. More advanced fNIRS units are havioral prediction than either area alone. The inspiration for now available that better address these issues, so data loss examining this region of interest came from past studies where should be much less of a problem in the future. Further direct right dlPFC activity was associated with persuasion resistance tests of the new technology’s capabilities to replicate fMRI find- and explicit counterarguing. While this study did not test ings should be performed to ensure that design improvements whether counterarguing was the specific psychological phe- do solve the data loss issue and to identify what, if any, differ- nomenon responsible for the dlPFC-behavior link, the results in- ences in effect sizes are possible to detect. dicate that activity in this area is likely associated with some Taken together, these results provide evidence that fNIRS process that inhibits persuasion. The results also suggest that can replicate the reliable finding that signal from the mPFC dur- this process is somewhat independent from the self-integration ing passive viewing of persuasive messages predicts later be- occurring with mPFC activity, due to the model improvement havior. It can also distinguish between types of audience that occurred when both areas were taken into account. Future members, though not types of persuasive content framing in research can thus examine in more detail the specific psycho- this particular study. This suggests that, while perhaps not as logical mechanism responsible for activity in the right dlPFC reliable for detecting small neurophysiological effects as fMRI, and build a comprehensive account of the real time internal fNIRS can provide meaningful predictions about behavioral out- processes leading to persuasion and behavior change. comes in response to persuasive messaging. This is especially However, we did not find any activation differences between valuable for use cases where fMRI is not practical or possible— different kinds of persuasive messages. This is in contrast with for example when equipment must be transported to access re- the original study, which identified greater neural activation on mote populations; when large study samples are required with an average for gain over fact and loss messages. In addition, the limited research budget or in applied research designs where successful replication results above are the result of using a dif- participants need to move in and interact with a more ecologic- ferent neural activity parameter than in Vezich et al. (all persua- ally valid environment than the fMRI scanner. Thus, the results sive messages over baseline, vs gain> fact and gain> loss). The here do indicate that fNIRS can be used toward the service of original contrasts, gain> loss and gain> fact, did not replicate persuasion research, so long as research designs account for the the user/non-user interaction and were weaker predictors of identified limitations. mPFC activity for the whole group than the baseline contrast was. It is not immediately clear why the different types of per- Supplementary data suasive messages in this study did not produce differential neural responses. Perhaps it is because the original study had Supplementary data are available at SCAN online. less statistical power than this one, making the effect estimates more unstable, and that any true within-subject difference be- Funding tween these message types is a relatively small effect that can- not be identified with fNIRS, which has a lower signal-to-noise This work was supported by the Minerva Initiative from the ratio than fMRI. Our alternative approach of contrasting average US Department of Defense (13RSA281, PI: M.D.L.). message activation with baseline has precedent in previous neuroimaging studies of persuasive messaging (Falk et al., 2010, Conflict of interest. None declared. 2011, 2016; Cooper et al., 2015), so we believe our results still show a meaningful relationship between neural processing of References persuasive information and downstream behavior. Further, it is useful to know what analysis approach is more appropriate for Berns, G.S., Moore, S.E. (2012). A neural predictor of cultural fNIRS specifically, for the purpose of future persuasion research popularity. Journal of Consumer Psychology, 22(1), 154–60. efforts in this modality. Chua, H.F., Ho, S.S., Jasinska, A.J., et al. (2011). Self-related neural At the same time, a number of issues in this study limit how response to tailored smoking-cessation messages predicts extensively the results can be interpreted. Firstly, there was quitting. Nature Neuroscience, 14(4), 426–7. quite a bit of neural data that were removed from analysis be- Cooper, N., Bassett, D.S., Falk, E.B. (2017). Coherent activity be- cause of poor signal-to-noise. Twelve participants-worth of tween brain regions that code for value is linked to the malle- data had to be removed out of 81 total, and the majority of those ability of human behavior. Scientiﬁc Reports, 7, 43250. remaining participants had at least one channel of data Cooper, N., Tompson, S., O’Donnell, M.B., Falk, E.B. (2015). Brain removed for poor quality. This did not severely limit our power activity in self- and value-related regions in response to online in the analysis of mPFC activity, because the relative affordabil- antismoking messages predicts behavior change. Journal of ity of fNIRS allowed us to recruit many more subjects than in a Media Psychology, 27(3), 93–108. typical fMRI experiment. It did however limit our power in the Cui, X., Bray, S., Bryant, D.M., Glover, G.H., Reiss, A.L. (2011). A analysis of our more complex model that included both the quantitative comparison of NIRS and fMRI across multiple cog- mPFC and rdlPFC, because participants had to be excluded from nitive tasks. NeuroImage, 54(4), 2808–21. Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018 636 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 6 Falk, E.B., Berkman, E.T., Lieberman, M.D. (2012). From neural Miilunpalo, S., Nupponen, R., Laitakari, J., Marttila, J., Paronen, O. responses to population behavior: neural focus group predicts (2000). Stages of change in two modes of health-enhancing population-level media effects. Psychological Science, 23(5), physical activity: methodological aspects and promotional implications. Health Education Research, 15(4), 435–48. 439–45. Falk, E.B., Berkman, E.T., Mann, T., Harrison, B., Lieberman, M.D. Montague, P.R., Berns, G.S., Cohen, J.D., et al. (2002). Hyperscanning: simultaneous fMRI during linked social inter- (2010a). Predicting persuasion-induced behavior change from actions. NeuroImage, 16(4), 1159–64. the brain. The Journal of Neuroscience, 30(25), 8421–4. Ramsay, I.S., Yzer, M.C., Luciana, M., Vohs, K.D., MacDonald, Falk, E.B., Berkman, E.T., Whalen, D., Lieberman, M.D. (2011). A.W.I.I.I. (2013). Affective and executive network processing Neural activity during health messaging predicts reudctions in associated with persuasive antidrug messages. Journal of smoking above and beyond self-report. Health Psychology, 30(2), Cognitive Neuroscience, 25(7), 1136–47. 177–85. Riddle, P.J., Jr, Newman-Norlund, R.D., Baer, J., Thrasher, J.F. Falk, E.B., Cascio, C.N., Coronel, J.C. (2015a). Neural prediction of (2016). Neural response to pictorial health warning labels can communication-relevant outcomes. Communication Methods predict smoking behavioral change. Social Cognitive and and Measures, 9(1-2), 30–54. Affective Neuroscience, 11(11), 1802–11. Falk, E.B., Morelli, S.A., Welborn, B.L., Dambacher, K., Lieberman, Saito, D.N., Tanabe, H.C., Izuma, K., et al. (2010). “Stay tuned”: M.D. (2013). Creating buzz: the neural correlates of effective Inter-individual neural synchronization during mutual gaze message propagation. Psychological Science, 24(7), 1234–42. and joint attention. Frontiers in Integrative Neuroscience, 4, 127. Falk, E.B., O’Donnell, M.B., Cascio, C.N., et al. (2015b). Salovey, P., Wegener, D.T. (2003). Communicating about health: Self-afﬁrmation alters the brain’s response to health messages message framing, persuasion, and health behavior. In: Suls, J., and subsequent behavior change. Proceedings of the National Wallston, K.A., editors. Social Psychological Foundations of Health Academy of Sciences, 112(7), 1977–82. and Illness. Malden, MA: Blackwell Publishing, 54–81. Falk, E.B., O’Donnell, M.B., Tompson, S., et al. (2016). Functional Schilbach, L., Wohlschlaeger, A.M., Kraemer, N.C., et al. (2006). brain imaging predicts public health campaign success. Social Being with virtual others: neural correlates of social inter- Cognitive and Affective Neuroscience, 11(2), 204–14. action. Neuropsychologia, 44(5), 718–30. Falk, E.B., Rameson, L., Berkman, E.T., et al. (2010b). The neural Scholkmann, F., Kleiser, S., Metz, A.J., et al. (2014). A review on correlates of persuasion: a common network across cultures continuous wave functional near-infrared spectroscopy and and media. Journal of Cognitive Neuroscience, 22(11), 2447–59. imaging instrumentation and methodology. NeuroImage, 85, Ferrari, M., Mottola, L., Quaresima, V. (2004). Principles, techni- 6–27. ques, and limitations of near infrared spectroscopy. Canadian Strangman, G., Culver, J.P., Thompson, J.H., Boas, D.A. (2002). A Journal of Applied Physiology, 29(4), 463–87. quantitative comparison of simultaneous BOLD fMRI and NIRS Ferrari, M., Quaresima, V. (2012). A brief review on the history of recordings during functional brain activation. NeuroImage, human functional near-infrared spectroscopy (fNIRS) develop- 17(2), 719–31. ment and ﬁelds of application. NeuroImage, 63(2), 921–35. Vezich, I.S., Katzman, P.L., Ames, D.L., Falk, E.B., Lieberman, M.D. Fogg, B.J. (2009). The behavior grid: 35 ways behavior can change. (2017). Modulating the neural bases of persuasion: why/how, In: Chatterjee, S., Dev, P., editors. Proceedings of the 4th gain/loss, and users/non-users. Social Cognitive and Affective International Conference on Persuasive Technology. Claremont, CA, Neuroscience, 12(2), 283–97. 26–29 April, New York: ACM, 42–6. Weber, R., Huskey, R., Mangus, M., Westcott-Baker, A., Turner, Franceschini, M.A., Toronov, V., Filiaci, M.E., Gratton, E., Fantini, B.O. (2015). Neural predictors of message effectiveness during S. (2000). On-line optical imaging of the human brain with counterarguing in antidrug campaigns. Communication 160-ms temporal resolution. Optics Express, 6(3), 49–57. Monographs, 82(1), 4–30. Hoge, R.D., Franceschini, M.A., Covolan, R.J., Huppert, T., Weinstein, N.D., Rothman, A.J., Sutton, S.R. (1998). Stage theories Mandeville, J.B., Boas, D.A. (2005). Simultaneous recording of of health behavior: conceptual and methodological issues. task-induced changes in blood oxygenation, volume, and ﬂow Health Psychology, 17(3), 290–9. using diffuse optical imaging and arterial spin-labeling MRI. Wilms, M., Schilbach, L., Pfeiffer, U., Bente, G., Fink, G.R., NeuroImage, 25(3), 701–7. Vogeley, K. (2010). It’s in your eyes – using gaze-contingent Kahneman, D., Tversky, A. (1979). Prospect theory: an analysis of stimuli to create truly interactive paradigms for social cogni- decision under risk. Econometrica, 47(2), 263–91. tive and affective neuroscience. Social Cognitive and Affective Kato, J., Ide, H., Kabashima, I., Kadota, H., Takano, K., Kansaku, Neuroscience, 5(1), 98–107. K. (2009). Neural correlates of attitude change following posi- Ye, J., C., Tak, S., Jang, K., Jung, J., E., Jang, J. (2009). NIRS-SPM: tive and negative advertisements. Frontiers in Behavioral statistical parametric mapping for near-infrared spectroscopy. Neuroscience, 3,6. NeuroImage, 44(2), 428–47. Downloaded from https://academic.oup.com/scan/article-abstract/13/6/628/4992588 by Ed 'DeepDyve' Gillespie user on 03 July 2018
Social Cognitive and Affective Neuroscience – Oxford University Press
Published: May 4, 2018
Access the full text.
Sign up today, get DeepDyve free for 14 days.