Journal of Consumer Behaviour, J. Consumer Behav.

, 15: 538–548 (2016)

Published online 30 June 2016 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/cb.1591

An fMRI study of advertising appeals and their relationship to product

attractiveness and buying intentions
HYO JUNG (JULIE) CHANG*, MICHAEL O’BOYLE, RONALD C. ANDERSON and CHOMPOONUT SUTTIKUN
Texas Tech University, Lubbock, TX USA

ABSTRACT

This study examines the effects of various visual apparel advertisements on consumers’ brain activation during exposure to different types
of advertising appeals (i.e., celebrity, non-celebrity, and rational). The influence on consumer perceptions of products and their subsequent
buying intentions are also measured. A repeated measures experimental design was employed, and the total of 27 right-handed female
subjects participated in the study. The results of the quantitative data showed significant differences in perceived product attractiveness
for each of the three types of advertising appeals, but not in buying intentions. Regarding the fMRI results, our findings support the notion
that celebrity advertising appeals are associated with heightened brain activation of memory-related/retrieval regions, reflecting how consumers
remember the ad and are influenced by the attractiveness of the source. Non-celebrity advertising appeals were more closely associated with
brain activation of regions thought to mediate self-reflection and also engaging executive functions. For rational advertising appeals, our findings
showed significant activation in brain areas associated with logical evaluative decision making reflecting more logical processing value assess-
ments and reward potential. Therefore, retailers/marketers must be particularly mindful to match the appropriate advertising appeal to the specific
purpose of the advertisement. This study also provides brain-based insight into the effectiveness of different types of advertising appeals and
whether or not they have the desired impact on the consumer. Copyright © 2016 John Wiley & Sons, Ltd.

INTRODUCTION LITERATURE REVIEW

Esthetic design and attractiveness are important factors Consumer neuroscience

influencing consumers’ intention to purchase fashion prod- Consumer neuroscience is defined as the field of research seek-
ucts, such as apparel and home décor items (Eckman and ing brain-based explanations for consumer behaviors using
Wagner, 1994). Thus, it is important to investigate how to current neuroscientific methods (Kenning and Plassmann,
capture consumer attention in positive ways, thereby increas- 2008). Consumer neuroscience is unique in that it focuses on
ing perceived product attractiveness and enhancing their in- the consumer and how various advertising appeals affect their
tention to buy. One essential tool for doing so is individual preferences and purchasing behavior at the neuro-
conducting research to discover how various types of adver- physiological level (Plassmann et al., 2012). The application
tising appeals affect consumer behavior in both experimental of neuroimaging techniques to marketing and consumer
and real-life settings (Chandy et al., 2001). Advertising ap- behavior has gained significant popularity and researchers
peals not only are thought to affect actual buying behaviors, have become interested in gaining an in-depth understanding
but also can influence the consumers’ point of view by initi- of the neural and cognitive processes underlying consumers’
ating a comparison between themselves (usually favorably) purchasing decisions (Jai et al., 2014).
with the models appearing in an advertisement (Sadeghi Although much research has focused on the behavioral
et al., 2015). Thus, it is important to understand how differ- effects of visual merchandising strategies and their impact
ent types of advertising appeals impact consumer percep- on consumer decision making (Mower et al., 2012), there is
tions, emotions, and behaviors. a lack of empirical research regarding brain activation patterns
In addition to the traditional approaches, an increasingly and their correlation to marketing stimuli and their ultimate
popular method for assessing these factors is the study of connection to consumers’ perceptions, attitudes, and behav-
the underlying brain activation that accompanies different ioral intentions. Therefore, this study attempts to fill part of
types of advertising appeals designed to promote a given the current gap in the literature by using functional magnetic
product. The data derived from neuroimaging studies is im- resonance imaging (fMRI) techniques to investigate brain
portant as it can provide brain-based insights into consumer activation in response to apparel advertisements using three
behavior beyond those revealed by more traditional research different types of advertising: celebrity endorsement, non-
methods. Thus, the present study examines the effects of var- celebrity endorsement, and rational appeals.
ious visual apparel advertisements on consumers’ brain acti-
vation during exposure to different types of advertising Advertising appeals and consumer attitudes toward
appeals. The influence on consumer perceptions of products products
and their subsequent buying intentions are also measured. Advertisements play an important role in the business world
Furthermore, the present research investigates the study of as they are known to affect consumer attitudes toward a
the neural effects of different types of advertising appeals given brand or product (Connell et al., 2014). Additionally,
on U.S. female consumers’ perceptions and their buying advertisements play an important role in driving actual sales
intentions. (Jones, 2006). According to Taylor et al. (2013), exposure

Copyright © 2016 John Wiley & Sons, Ltd.

 An fMRI study of advertising appeals 539

via television and online advertising are particularly effective celebrity–product associations lead to deeper memory
methods of increasing brand awareness and product sales. encodings, thus increasing the consumer’s ability to retrieve
Marketers seek effective communication through various information about the product being advertised (Stallen
types of advertisements because consumers are known to et al., 2010). Based on these findings, the current study exam-
be influenced by attractive advertising appeals (Ruiz and ines the brain activation correlates of product attractiveness
Sicilia, 2004). In light of this fact, marketing and information and buying intentions when using celebrity advertising ap-
systems pay close attention to research on consumer attitudes peals. The latter are anticipated to produce heightened activa-
(Tsang et al., 2004), being careful to choose the right type of tion of the brain areas associated with the formulation and
advertisements to reach an intended target market. When retrieval of short- and long-term memories, regions that in-
consumers have favorable feelings toward an advertisement, clude (but are not limited to) the prefrontal and frontal cortex.
they are more likely to accept and even “enjoy” the informa-
tion provided about an advertised product. Non-celebrity advertising appeals and buying intentions
Interestingly, the degree to which consumers feel con- Non-celebrity endorsers have also been used in advertise-
nected to an advertisement is affected by the specific type ments, not only as a cost cutting measure, but also because
of advertising appeal employed (Bush et al., 1999). Specifi- they are thought to influence consumers by increasing the at-
cally, the overall attitude of the consumer is influenced by tractiveness of products and thereby enhancing sales (Wang
product endorsers and/or the messages contained in a given et al., 2009). Previous research has shown that retail store
advertisement (Bush et al., 1999). For example, advertising image, the brand name of products, and their price are the
researchers report that consumers are much more likely to three most important components when consumers are decid-
purchase products featured in a given advertisement if they ing to purchase (Grewal et al., 1998). Clearly, using non-
have an overall positive attitude toward the advertising ap- celebrity endorsers can be beneficial from an operating ex-
peal (Phelps and Thorson, 1991). Furthermore, consumers pense point of view by reducing the price of advertisements,
are thought to be more readily persuaded by an advertise- given that it costs significantly less to employ non-celebrity
ment if it tends to “match” them in some way—either in ap- than celebrity endorsers, while retaining a positive effect on
pearance, by personality or both (Fabrigar and Petty, 1999). brand/product image.
Recently, Gaied and Rached (2010) found that non-
Celebrity advertising appeals and buying intentions celebrity endorsements sometimes have more persuasive
Celebrities in advertisements create positive images of the power in the advertisement of some product types compared
brand/product (Tom et al., 1992). Therefore, marketers often to celebrity endorsements. This finding highlights the impor-
use famous (but not always expert) endorsers to send a mes- tance of source credibility compared to the attractiveness or
sage about their product, attempting to generate positive “fame” of the endorser (Gaied and Rached, 2010). In the cur-
brand reputation and encourage consumers to buy (Van der rent study, we examine how the brain activation generated by
Waldt et al., 2009). For example, successful tennis players non-celebrity advertisements may differ from celebrity ad-
often endorse tennis-related brands/products, a ploy that is vertisements (and in contrast to rational advertising appeals,
believed to instill positive attitudes in consumers regarding e.g., linguistic descriptions of products) when judging prod-
these products. Advertisers also tend to select endorsers uct attractiveness and generating buying intentions. It is hy-
who have personalities close to or matching the attributes pothesized that because of a smaller gap in the perceptions
of a given target audience. Use of the latter is thought to of consumers between themselves and non-celebrity en-
create trust in a given brand as a result of presentation by dorsers, the latter will produce a strong connection with the
endorsers who have personalities or images that can be products being advertised. We anticipate that such “self-
positively related to a given brand/product (e.g., If Sarah comparisons” will be reflected in greater activation of the
Jessica Parker, a famous and fashionable celebrity, endorses insula cortex for non-celebrity as compared to celebrity and
a particular fashion product or brand, consumers may be rational appeals.
induced to trust the product/brand as a result of the “fashionable”
(positive) image associated with her as a celebrity) (Choi and Rational advertising appeals and buying intentions
Rifon, 2007). In our study, we examined the effect of celebrity Persuasive messages in advertisements are known to affect
status of the model (who may or may not be “expert” in the consumer cognition and impact their buying intentions.
product they are endorsing) and anticipate that the celebrity Thus, it is necessary to examine other factors (rather than
status of the model will instill a positive connection with the simply the celebrity versus non-celebrity status of the en-
product, resulting in a beneficial change in consumer attitudes dorser) that are also known to influence consumer decisions
and behavioral intentions. to purchase products. Information on these factors will help
Accordingly, Klucharev et al. (2008) believe that con- retailers determine the right combination of advertising fac-
sumers’ attitudes and behavioral intentions may be revealed tors to maximize consumer engagement (Nettelhorst and
and potentially explained by neuroimaging data. For exam- Brannon, 2012). In this way, retailers would benefit from a
ple, they note the increased long-lasting positive effects on more complete understanding of consumers’ attitude and
memory that have been reported for celebrity as compared buying decisions, thus enhancing product profitability.
to non-celebrity product endorsements (Klucharev et al., According to Rath et al. (2015), consumer rational deci-
2008). Also, the investigation of the brain mechanisms under- sion making is a byproduct of their reasoning and logic and
lying the impact of celebrity endorsement suggests that strong is directly related to the information advertisers place in their

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
540 H. Chang et al.

advertisements (Wang et al., 2013). For example, Sadeghi H1: Perceived product attractiveness will be rated differ-
et al. (2015) found that rational advertising appeals are more ently among the three advertising appeals (i.e., celebrity,
influential compared to emotional appeals when consumers non-celebrity, rational advertising appeals).
are contemplating a purchase decision. Thus, the present
H2: There will be different brain activation patterns for
study examines the brain activation patterns associated with
perceived product attractiveness that will be related to
rational appeals (e.g., informational descriptions) compared
the type of advertising appeal presented: (i) celebrity ap-
to both celebrity and non-celebrity endorsements, and how
peals will evoke more activation of memory-related brain
they differ in terms of their influence on product attractive-
regions; (ii) non-celebrity appeals will evoke more activa-
ness and consumer buying intention. It is hypothesized that
tion in self-referencing related brain regions; (iii) rational
rational appeals will engage several brain regions, particu-
appeals will evoke more activation in the reasoning-
larly the left frontal lobe which mediates language processing
related brain regions compared to the other two types of
as well as evaluative judgements.
advertising appeals.
Product attractiveness, price, and buying intentions H3: Consumer intentions to buy will be different depend-
The attractiveness of a product is defined by the esthetic ap- ing upon the nature of the three advertising appeals pre-
peal inherent in its design (Eckman and Wagner, 1994). sented (i.e., celebrity, non-celebrity, rational).
Companies are interested in developing more attractive
H4: There will be different brain activation patterns associ-
products in an effort to enhance consumer satisfaction and
ated with consumer buying intentions as a function of adver-
to influence their buying intentions. Growing competition
tisement type: (i) celebrity appeals will evoke more
in the business world encourages marketers and retailers to
activation of memory-related brain regions; (ii) non-celebrity
develop the best products and advertising promotions to
appeals will evoke more activation in the self-referencing
maximize impact on sales (Johnson and Fornell, 1991).
related regions; (iii) rational appeals will evoke greater
Therefore, it is important to understand how visual attractive-
activation in the reasoning-decision making regions of the
ness of fashion products can be improved through the use of
brain as compared to the other two advertisement types.
various advertising cues. The latter help retailers compete
against price challenges of competitors by enhancing their
own product attractiveness and thus ensuring consumer METHOD
satisfaction and loyalty (Johnson and Fornell, 1991).
In addition to product attractiveness, price is a very im- Participants
portant factor when it comes to influencing consumer pur- Participants were recruited through a general posting placed
chasing decisions (Kim et al., 2012). Perceptions of on the on-line announcement page of a large Southwestern
fairness of price are known to be strongly related to brand University (USA). Twenty-seven right-handed female sub-
image and product attractiveness, which in turn have a signif- jects (average age = 20.8 years) participated in the study (see
icant impact on purchasing decisions (Jacoby and Olson, Table 1 for demographic details). Desmond and Glover
1977). Brain responses to specific types of advertisements (2002) have suggested that in fMRI experiments, for a liberal
should provide a more complete understanding of consumer
behavior and its influence on product attractiveness and price Table 1. Demographic characteristics of the participants
evaluations. For example, in the present study, it is antici-
pated that product attractiveness will be correlated with acti- Percentage
Characteristics N (%) Mean
vation of the nucleus accumbens in the brain (i.e., a reward
center), while price perceptions and the subsequent decision Gender
processes will be reflected in significant activation of the pre- Female 27 100%
frontal, frontal, and insula cortex involved in decision mak- Age (Mean) 21
Ethnicity
ing (Knutson et al., 2007). Caucasian/White 14 52%
In the present study, we examine the underlying processes African–American 03 11%
by which visual characteristics found in different types of Hispanic/Latino 4 15%
advertisements differentially affect consumer buying inten- Asian 05 18%
tions. Through the use of fMRI, we also directly investigate Other 01 4%
Class
the brain activation patterns induced by differing advertising Freshmen 07 26%
appeals, thus providing brain-based insights into the pro- Sophomore 04 15%
cesses used by consumers in their evaluations about product Junior 05 18%
attractiveness and their buying intentions. Senior 07 26%
It was predicted that there would be significantly different Graduate 04 15%
Household income
brain activation patterns related to each of the aforemen- $24,999 or less 08 30%
tioned advertising appeals. Also, participants will perceive $25,000–34,999 01 4%
increased product attractiveness from the apparel advertise- $35,000–49,999 03 11%
ments employing either celebrities or non-celebrities com- $50,000–74,999 05 18%
pared to those ads using rational appeals. From these $75,000–99,999 02 7%
$100,000 or above 08 30%
assumptions, four hypotheses were developed:

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
 An fMRI study of advertising appeals 541

threshold of significance (p < 0.05), approximately 12 partic- Research protocol

ipants per group are required to reach 80 per cent power at There were three different stimulus blocks used in the present
the voxel level of analysis. And while the number of partici- study: (i) celebrity advertisement; (ii) non-celebrity adver-
pants in our study is smaller than that used in more tradi- tisement; and (iii) rational advertisement. For celebrity
tional behavioral research, our sample size is comparable to advertisements, six famous female celebrities (i.e., Rachel
that of other fMRI investigations (Button et al., 2013). Note McAdams, Jennifer Lopez, Scarlett Johansson, Kristen
that pregnant women and individuals having metal in or per- Stewart, Sarah Jessica Parker, and Jennifer Lawrence) were
manently attached to their bodies were excluded from the presented one at a time, wearing dresses; for non-celebrity
study as these factors represent contraindications for expo- ads, six female non-celebrities wearing the same dresses
sure to fMRI scanning. This research was conducted with were presented, while for the rational block, one of the
the approval of the Human Research Protection Program at six linguistic statements were presented that described a
said University. Participation was voluntary and subjects re- given garment in detail, including the dress name, color,
ceived written informed consent prior to the study. size, and material. The three ad blocks were presented in
Participants read and signed a consent form upon arrival a counterbalanced order to prevent carryover effects.
at the neuroimaging institute. They completed an fMRI Participants viewed a fixation crosshair presented before
safety checklist to ensure that they had no metal in or at- each trial. This fixation crosshair was provided to alert par-
tached to their body. This was also monitored by the scanner ticipants that a new ad was forthcoming and to focus their
technician before participants were placed in the neuroimag- attention on the center of the screen. Following each trial
ing equipment. Participants practiced a simulated version of (i.e., viewing a given advertisement) questions about prod-
the task on a laptop in an adjacent waiting area. After prac- uct attractiveness and buying intention using a 7 point-
tice, participants entered the fMRI scanner and when com- Likert type scale were presented (i.e., Was the dress
fortably positioned, the acquisition of the brain images attractive? Would you buy this dress?).
commenced. Task stimuli were randomly projected through
a viewing screen attached to the head coil of the scanner
using E-Prime presentation software. Participants also an- fMRI acquisition
swered a post-scanning survey about the products presented. Neuroimaging data were collected for each of the three
blocks and consisted of 236 whole brain images. All images
Stimuli materials were acquired using a Siemens 3T Skyra scanner. For the
A 3 (advertising appeals) × 2 (prices) experimental design functional (EPI) scan the following acquisition parameters
was employed in the current study. Three different advertis- were used: Scan Time = 8 min; TR = 2000; TE = 20; Field of
ing appeals (i.e., celebrity, non-celebrity, and rational) were View (FOV): 240; Flip Angle = 80; voxel size:
presented in the context of two different product prices 3 × 3 × 3 mm; 44 slices acquired in the axial orientation. A
(i.e., low: $50 and high: $250). On a given trial an ad 3D high resolution T1 anatomical scan followed the func-
depicting an item (i.e., a dress) was presented for a total of tional scan and employed the following parameters: Scan
30 s, including fixation (2 s), ad presentation (4 s), rating of Time: 4:26; Repetition Time (TR): 1900; Flip Angle: 9; Echo
product attractiveness (10 s), reappearance of the ad with a Time (TE): 2.49; .9 × .9 × .9 voxel size; 192 slices in the sag-
price (4 s), and subsequent buying intention decision (10 s). ittal orientation. Concomitant behavioral responses (i.e.,
A total of 36 trials were presented to each participant. speed of attractiveness ratings and buying intention) while
Figure 1 illustrates an example of the sequence of events being scanned were recorded using E-prime presentation
on one trial. software.

Figure 1. Sequence of events on a given trial (e.g., rational advertising appeal).

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
542 H. Chang et al.

Figure 2. Activation epochs analyzed for the attractiveness and buying intention segments of each trial.

fMRI data analysis 46.71) = 10.41, p < 0.001). Perceived attractiveness of the
fMRI mixed effects analysis was performed using FSL product was rated significantly higher for non-celebrity
(FMRIB’s software library). Pre-processing steps were ap- (M = 4.90, SD = 0.87) compared to celebrity appeals
plied to all acquired functional images including skull re- (M = 3.99, SD = 0.82, p < .001) Also, perceived attractiveness
moval (BET), motion correction, spatial smoothing (5-mm was significantly higher for non-celebrity compared to ratio-
full-width/half-maximum Gaussian kernel), intensity nor- nal appeals (M = 4.18, SD = 1.04; p < .01). There was no sig-
malization, and high-pass temporal filtering. For attractive- nificant difference in product attractiveness between the
ness evaluation, functional images acquired during the first celebrity and rational appeals. Thus, H1 was partially
6 s of each of the three blocks were analyzed. To evaluate at- supported.
tractiveness the 4 s of ad presentation and the first 2 s of the
attractiveness rating component were analyzed; for the buy- Neuroimaging data
ing intentions the 4 s of the ad reappearance and 2 s of the Figure 3 presents a comparison of the brain activation maps
buying intention component were analyzed (see Figures 1 illustrating where significant differences were found during
and 2). Note that the fMRI data were collapsed across price attractiveness evaluation as a function of the specific type
when evaluating buying intentions. Statistical significance of advertising appeal. The exact coordinates for the activated
was thresholded at Z > 2.3 (cluster size of 10) with an alpha brain regions can be found in Table 2.
of Pcorr > 0.01 (Worsley et al., 1992) (see Figures 1 and 2). The product attractiveness contrast comparing celebrity to
non-celebrity appeals revealed significantly greater activa-
tion for celebrity appeals in the left and right lingual gyrus
RESULTS AND DISCUSSION (Brodmann Area (BA) 17), which is primary visual cortex
(Banich and Compton, 2010). This finding suggests that ce-
Product attractiveness data lebrity appeals instantiate more visual information process-
Attractiveness ratings ing (and potentially more visual mental imagery) than do
The results of a repeated-measures analysis of variance non-celebrity advertising appeals. The product attractiveness
(ANOVA) showed significant differences in perceived prod- contrast comparing celebrity to rational advertising appeals
uct attractiveness for each of the three types of advertising also revealed significantly greater activation in the left lin-
appeals: celebrity, non-celebrity, and rational (F(1.80, gual gyrus (BA 17). This too suggests celebrity advertising

Figure 3. Significant contrasts for the attractiveness evaluation process. Note: F = front; B = back; R = right; L = left.

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
 An fMRI study of advertising appeals 543

Table 2. fMRI table depicting each advertising appeal contrast during the attractiveness evaluation process
MNI coordinates
Region BA Side x y z Z-score Cluster Voxel size Mental function of Interest
Celeb > Non-Celeb (p < 0.01)
Lingual gyrus 17 L 8 82 3 6.84 1 101,499 Visual mental imagery, visual perception,
Lingual gyrus 17 R 2 85 6 6.05 memorization
Celeb > Rational (p < 0.01)
Lingual gyrus 18 L 2 88 0 6.83 1 207,925
Lingual gyrus 18 L 7 81 3 6.65
Lingual gyrus 18 L 3 88 1 6.57
Non-Celeb > Celeb (p < 0.01)
Cingulate gyrus 31 L 1 34 38 3.86 5 49,952 Self-reflection, emotion
Postcentral gyrus 3 L 15 29 76 3.82 Imagined actions
Postcentral gyrus 3 L 20 33 72 3.7
Middle frontal gyrus 10 L 39 48 20 3.56 4 23,291 Recognition, recall, recognize emotions
Transverse temporal gyrus 41 L 49 21 10 3.53 3 13,714
Angular gyrus 39 L 30 59 39 3.31
Superior parietal lobule 7 R 45 52 59 3.32 2 7320
Inferior parietal lobule 40 R 45 51 45 3.19
Supramarginal gyrus 40 R 48 43 38 3.05
Postcentral gyrus 40 R 68 16 15 3.84 1 6212
Superior temporal gyrus 42 R 70 28 10 3.65
Transverse temporal gyrus 42 R 67 7 6 3.48
Non-Celeb > Rational (p < 0.01)
Middle temporal gyrus 39 R 57 63 13 5.31 3 70,701 Self-reflection, goals
Middle occipital gyrus 19 L 51 74 14 4.89 Processing of visual motion
Precuneus 7 R 4 57 42 4.02 2 19,459 Self-referencing
Posterior cingulate 30 L 2 50 23 3.85 Emotion, motivations
Posterior cingulate 23 R 4 52 21 3.53
Cerebellar tonsil L 28 57 38 3.45 1 9002

appeals induce more visual processing than rational appeals. non-celebrity as compared to celebrity advertising appeals.
Therefore, H2a was supported. These findings suggest that non-celebrity endorsers induce
The product attractiveness contrast comparing non- more activation in brain areas thought to mediate attention
celebrity to celebrity appeals revealed five significant clusters and the process of “self-reflection” (i.e., participants
of brain activity. Specifically, activation of the left cingulate comparing/relating themselves more to non-celebrity than
gyrus, left angular gyrus, and right transverse temporal gyrus the celebrity endorsers, and thus potentially better able to
(BA 31, 39, and 42) were all found to be greater for imagine themselves in the outfit when worn by a non-

Figure 4. Significant contrasts for the buying intention process. Note: F = front; B = back; R = right; L = left.

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
544 H. Chang et al.

Table 3. fMRI result table depicting each advertising appeal contrast during the buying intention process
MNI coordinates
Region BA Side x y z Z-score Cluster Voxel size Mental function of Interest
Celeb > Non-Celeb (p < 0.01)
Middle occipital gyrus 18 L 12 91 25 3.95 1 40,965 Visual mental imagery, visual perception
Lingual gyrus 19 L 20 63 4 3.95 Memorization
Celeb > Rational (p < 0.01)
Middle occipital gyrus 18 R 15 96 19 4.66 2 26,988
Lingual gyrus 18 R 8 92 5 4.64
Fusiform gyrus 37 R 49 42 17 3.65 1 5514 Visual analysis, face recognition
Non-Celeb > Celeb (p < 0.01)
Superior frontal gyrus 10 R 42 56 10 3.71 6 25,158 Recognition, recall, emotions,
Middle frontal gyrus 6 L 33 25 52 3.55 perceptual priming, imagery,
Superior frontal gyrus 9 L 21 52 29 3.51 decision making
Superior frontal gyrus 6 L 11 22 64 3.37 5 7408 Action planning, self-reflection
Superior frontal gyrus 8 R 3 35 47 3.31
Precentral gyrus 9 R 50 28 34 3.51 4 6911
Thalamus (ventral lateral nucleus) R 11 7 7 3.32 3 6117
Caudate (NAcc) R 12 7 9 3.32 2 6013 Reward anticipation, pleasure
Inferior parietal lobule 40 R 50 40 54 3.45 1 5748 Emotion, goals
Non-Celeb > Rational (p < 0.01)
Thalamus (pulvinar) R 9 26 8 3.21 1 7997
Posterior cingulate 29 L 5 45 19 3.05 Recall, self/other distinction
Rational > Celeb (p < 0.01)
Anterior cingulate 24 L 1 36 14 3.64 2 12,040 Value of reward, reasoning, attention
Anterior cingulate 24 L 2 30 14 3.6
Medial frontal gyrus 10 L 6 65 1 3.3 Language, calculation
Middle frontal gyrus 10 R 34 54 13 3.59 1 7699
Rational > Non-Celeb (p < 0.01)
Cuneus 18 L 4 81 28 3.5 1 8122 Attention, reward expectation
Cuneus 18 L 9 82 29 3.33

celebrity endorser). Brain activation of the left transverse tem- Neuroimaging data
poral gyrus and right inferior parietal lobule (BA 41, and 40) Figure 4 represents a comparison of the brain maps where
was also found to be greater for non-celebrity compared to there were significant differences in activation during the
celebrity appeals. The latter may reflect increased emotional buying intention stage of the process for each advertising ap-
engagement (Banich and Compton, 2010) of the participants peal. The areas differentially active during buying intentions
when viewing non-celebrity endorsements. Similarly, the corresponding to each of the three types of advertising ap-
contrast of the non-celebrity with rational appeals revealed peals are listed in Table 3.
three significant clusters of brain activation, including the right For buying intentions, the celebrity to non-celebrity con-
middle temporal gyrus, right precuneus, and left cerebellar trast showed two significantly increased clusters of brain ac-
tonsil (BA 39 and 30). These areas are associated with the tivation during celebrity advertising, which also included
aforementioned “self-referencing” process. Finally, greater ac- price information: the left middle occipital gyrus and the left
tivation of the left and right posterior cingulate (BA30 and 23) lingual gyrus (BA 18 and 19). This finding suggests that ce-
was found for non-celebrity compared to rational appeals, lebrity appeals evoke more visual processing and potentially
which is suggestive of heightened attention and increased emo- more mental imagery than non-celebrity advertising. This re-
tional response to the non-celebrity endorsers as compared to sult is similar to the pattern of activity found for the attrac-
rational appeals (linguistic descriptions) of the product. tiveness component of the process. The contrast of
Therefore, H2b was supported. Note that there were no in- celebrity to rational appeals also revealed heightened activa-
stances in which rational advertising appeals produced greater tion of the right middle occipital gyrus, lingual gyrus, and fu-
activation in any brain region when compared to either non- siform gyrus (BA 18 and 37) when viewing celebrity ads.
celebrity or celebrity appeals. Thus, H2c was not supported. These regions are thought to mediate various visual pro-
cesses, including object perception and face recognition
Buying intention data (Schacter et al., 2010). Thus, celebrity advertising appeals
Buying intention ratings seem to evoke more visual/mental imagery than rational ad-
The repeated-measures ANOVA showed no significant dif- vertising appeals. Therefore, H4a was supported.
ferences in buying intention for the three types of advertising Also, the contrast of non-celebrity to celebrity appeals re-
appeals and the two different prices ($50 vs. $250): celebrity vealed six significantly greater clusters of activation. Specifi-
(M = 3.41, SD = 0.98), non-celebrity (M = 3.16, SD = 0.76), cally, greater activation of the right and left superior frontal
and rational (M = 3.48, SD = 0.77). Therefore, H3 was not gyrus (BA 10 and 6) was found for non-celebrity appeals,
supported. which again may be related to the process of self-reflection,

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
 An fMRI study of advertising appeals 545

as well as the engagement of several executive functions re- Our data also revealed that different types of advertising
lied upon during product evaluation and buying intention pro- appeals impact product attractiveness ratings. Similar to the
cesses (see Burgess et al., 2007). Additionally, the contrast of finding of Ruiz and Sicilia (2004), our results suggest that at-
non-celebrity to celebrity appeals showed significantly tractive advertising appeals influence consumer attitudes pos-
greater activations in the caudate nucleus, which suggests that itively. However, participants rated the products advertised
non-celebrity advertising engages more reward anticipation by celebrities as least attractive of the three appeals, a finding
centers as compared to celebrity advertising (Grahn et al., that stands in direct contrast to what is often reported in the
2009). Last, the contrast of non-celebrity appeals to rational literature (e.g., Burgess et al., 2007). This result led us to al-
appeals revealed significantly greater brain activation in the ternative considerations regarding the actual attractiveness of
right thalamus and left posterior cingulate (BA29) for non- the celebrity endorsers. Perhaps, rather than adding appeal to
celebrity appeals, suggesting that the latter engage greater at- the product, the attributes of celebrity endorsers distract con-
tention and likely enhance memory for product information sumer attention away from the quality and potential value of
relative to rational appeals. Therefore, H4b was supported. the products advertised, thus failing to produce the intended
The contrast of rational to celebrity advertisements result. Interestingly, non-celebrity advertising appeals had
showed significantly greater clusters of activation for rational the most positive effect on perceived product attractiveness
appeals in the left anterior cingulate, left medial frontal gy- and on consumer buying intention, which were both corre-
rus, and right middle frontal gyrus (BA 24 and 10). The acti- lated with activation of brain regions important for self-
vation of these frontal regions, particularly the anterior reflection. In keeping with the findings of Wang et al.
cingulate, reflects increased attentional processing as well (2009), our result suggests that when viewing non-celebrity
heightened mediation of reward/value assessments. This appeals, consumers pay more attention to the product than
finding supports the notion that rational appeals are associ- the celebrity endorsing it.
ated with more “logical” buying intentions as compared to The fact that participants rated the dresses modeled by ce-
celebrity advertising appeals (the latter perhaps being based lebrities as least attractive was surprising. In fact the dresses
more on emotion than logic). The contrast of the brain maps rated as most attractive were those modeled by non-
associated with rational appeals compared to non-celebrity celebrities. This suggests that celebrity appeals may not al-
appeal revealed a significantly greater cluster of activation ways be the most effective advertising strategy for enhancing
in the left cuneus (BA 18). And because this region is purchase intentions or sales, although this conclusion must
thought to play a role in reward expectation (Jai et al., be tempered by the fact that certain types of products (e.g.,
2014), rational appeals may involve a more logical assess- sporting apparel) and the intended branding of a product
ment of reward expectations compared to non-celebrity ap- (i.e., increasing brand notoriety) may still benefit from celeb-
peals. Thus, H4c was supported. rity endorsement. Therefore, it is important to create unique
advertisements that are specifically configured to connect
with a particular product/brand category if one is to enhance
GENERAL DISCUSSION AND CONCLUSIONS marketing effectiveness (Hartnett et al., 2015). Note, how-
ever, it still may be that the best sales strategy for most prod-
The present study examined the effects of three different ad- ucts is to use non-celebrity appeals, those that feature
vertising appeals (i.e., celebrity, non-celebrity, rational) on advertisements depicting non-celebrity models with whom
both behavioral ratings and brain responses of female con- the consumers can most readily identify and connect.
sumers during product attractiveness and buying intention Regarding the fMRI results, celebrity advertising appeals
evaluations. Generally, our findings support the notion that evoked more activation in memory-related brain regions dur-
celebrity advertising appeals are associated with heightened ing the attractiveness evaluation process than non-celebrity
brain activation of memory-related/retrieval regions, perhaps ads. Although there is a diverse range of celebrity familiarity,
reflecting how consumers remember the ad and are influ- this finding suggests that celebrity endorsements may evoke
enced by the attractiveness of the source. It also suggests that stronger mental images (e.g., associations with/about the
consumers who have pre-existing memories associated celebrity) which aid in enhancing memory for a given
with/about the celebrities portrayed in the ads are having product/brand.
those memory traces “refreshed” which subsequently pro- Notably, as compared to the other advertising appeals,
duces a stronger and more complex memory trace, one that non-celebrity advertising engaged a greater number of brain
may have the collateral effect of enhancing remembrance regions associated with emotion. As previously mentioned,
for information about the product being depicted in the ad). this may be related to the fact that non-celebrity advertising
Non-celebrity advertising appeals were closely associated appeals induce self-reflection (i.e., “I can imagine how I
with brain activation of regions thought to mediate self- would look in this dress”) as compared to the dresses
reflection (i.e., “how this dress might look on me”) and also modeled by celebrity endorsers (i.e., “of course it looks great
engaging executive functions (e.g., paying attention). For ra- on Sarah Jessica Parker, but it may not on me”). The activa-
tional advertising appeals, our findings revealed significant tion of these regions points to a more realistic evaluation of
activation in brain areas (frontal) associated with logical the product when modeled by non-celebrities.
evaluative decision making (i.e., perhaps verbally mediated) As for buying intentions, there were no significant differ-
reflecting more logical (rather than emotional) processing ences related to the three different types of advertising ap-
value assessments and reward potential. peals based on our rating data. This finding may reflect the

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
546 H. Chang et al.

fact that product price (rather than strictly advertising type combine to influence consumer behavior. Further research of
per se) has a larger effect on buying intentions (Kim et al., this type may prove to be particularly useful when examining
2012). Regarding the fMRI results for buying intentions, ce- the neural correlates and subsequent impact of other types of
lebrity advertising appeals evoked more activation in brain advertising appeals not specifically employed here (e.g., deep
areas associated with visual processing and mental imagery, discounts). With regard to practical implications, the current
a result also found during the attractiveness evaluation por- study provides brain-based insight into the effectiveness of
tion of the study. This finding is congruent with that of different types of advertising appeals and whether or not they
Knutson et al. (2007) suggesting that celebrity advertising have the desired impact on the consumer. The present results
appeals produce effective memory retrieval cues. In contrast, support the hypothesis that different advertising appeals trig-
for buying intentions, non-celebrity advertising appeals acti- ger different patterns of brain activation that subsequently
vated more regions associated with self-reflection, emotion, (and meaningfully) influence evaluations of product attrac-
and reward anticipation (Fransson and Marrelec, 2008), each tiveness and buying intentions and do so in somewhat differ-
of which contributes to a more self-referenced buying inten- ent ways. An interesting real-world application of such
tion formulated by consumers when viewing non-celebrity research would be to use this technology as a pre-testing tool
appeals. for the creation and development of specific advertising ap-
peals prior to their release to the public (Northover, 2012).

IMPLICATIONS

In summary, the present study suggests that celebrity en- LIMITATION AND FUTURE RESEARCH
dorsements may be most influential when attempting to es-
tablish name recognition for a new brand or product. And One limitation of our study is that the celebrities employed
while the use of celebrity endorsements is thought to increase may not have been “familiar” enough (and thus less influen-
product attractiveness, it does not always do so and is not tial), a factor that is of course, dependent upon the level of
necessarily the most effective means of inducing buying in- fashion awareness of the participant. Given that our subjects
tention or actual buying behavior. However, researchers need were young female college students, our expectation would
to be cautious when applying this principle as the effective- be that their fashion awareness would be relatively high.
ness of using celebrity endorsers may vary depending on Also, the inherent importance of the product advertised in
the product/brand categories being marketed (e.g., male ver- the study (in this case apparel) is not likely to be the same
sus female consumer responses, endorsing sporting goods for all participants, but again one would anticipate that cloth-
versus clothing/apparel). On the other hand, as found in our ing would be of high interest to our sample. However, future
study, non-celebrity endorsements were most influential in research using other types of celebrities (e.g., sports stars or
terms of producing positive buying intentions, perhaps be- popular musicians) and a wider range of products is required
cause they induce self-reflection and thus generate a connect- to provide a more complete picture of the brain mechanisms
edness between the consumer and the product being underlying consumer evaluations and decisions.
advertised. Thus, if the retailing goal is to increase positive It is important to note that our study only measured buy-
buying intentions and presumably actual buying behaviors, ing intentions, not buying behavior per se, as participants
then non-celebrity endorsement may be the most cost effec- did not actually purchase the items being displayed. As such,
tive option. Note that even though rational advertising ap- it remains an open question as to how well buying intentions
peals did not appear to significantly increase attractiveness or the related consumer attitudes being formed about a given
of a given product, these types of ad appeals do have an im- product actually reflect genuine buying behaviors (Wright
pact on buying intentions, and thus may best be employed as and Klÿn, 1998). Future studies of the influence of different
supplementary input to accompany non-celebrity advertising types of advertising appeals are needed to confirm such a
appeals. This dual approach may prove to be particularly ef- connection (e.g., one could provide the participant with an
fective for online retailers (i.e., providing combined rational on-line account credited with real money, ask them to “save”
and non-celebrity advertisements), where consumer connect- outfits they are interested to their “shopping cart” and then
edness is often difficult to establish. Again, one must be par- monitor how many items they actually buy before their dollar
ticularly mindful of matching the appropriate advertising allotment is depleted). As previously mentioned, our subjects
appeal to the specific purpose of the advertisement (i.e., in- were young and females and so further research is needed to
creasing brand awareness versus increasing the volume of understand how different age groups and the sex of the con-
products sold). sumer relate to the effectiveness of different types of adver-
The present research also used fMRI technology to help tising appeals. Finally, because all our participants were
understand the neural correlates of various types of advertis- female college students, the perceived prices of the products
ing appeals on consumer evaluations of attractiveness and may have masked important decision factors as college stu-
buying intentions, thus bridging a potential gap between dents (compared to older career women) are known to have
what consumers say (do) and what their brains are actually less disposable income. These shortcomings notwithstand-
processing. The academic implication of the study resides ing, research involving both rating and brain-based method-
in the successful demonstration of how neuroimaging can as- ologies appears to hold great promise for the study of
sist in providing new insights in to the myriad of factors that consumer decision making processes and hopefully

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
 An fMRI study of advertising appeals 547

providing important insights into their subsequent buying Gaied AM, Rached KSB. 2010. The persuasive effectiveness of fa-
behaviors. mous and non famous endorsers in advertising. IBIMA Business
Review 2010: 1–14.
Grahn JA, Parkinson JA, Owen AM. 2009. The role of the basal
ganglia in learning and memory: neuropsychological studies.
BIOGRAPHICAL NOTES Behavioural Brain Research 199(1): 53–60.
Grewal D, Monroe KB, Krishnan R. 1998. The effects of price
Hyo Jung (Julie) Chang, PhD is an Assistant Professor in the comparison advertising on buyers’ perceptions of acquisition
Department of Hospitality and Retail Management at Texas Tech value, transaction value, and behavioral intentions. The Journal
University, USA. Her research interests include social psychological of Marketing 62: 46–59.
aspects of apparel and consumers, consumer behavior, fashion Hartnett N, Kennedy R, Sharp B, Greenacre L. 2015. Creative that
merchandising, and sustainable consumption. sells: how advertising execution affects sales. Journal of
Michael O’Boyle, PhD is the Associate Dean of Research in the Advertising 45(1): 102–112.
College of Human Sciences and Professor of Human Development Jacoby J, Olson JC. 1977. Consumer response to price: an attitudi-
and Family Studies at Texas Tech University. He is a leading expert nal, information processing perspective. Moving Ahead with
in cognitive neuroscience using fMRI techniques to investigate Attitude Research 39(1): 73–97.
individual differences in brain function. Jai TMC, O’Boyle MW, Fang D. 2014. Neural correlates of
sensory-enabling presentation: an fMRI study of image zooming
Ronald C. Anderson is a doctoral student in Electrical and Computer and rotation video effects on online apparel shopping. Journal of
Engineering at Texas Tech University. His research interests include Consumer Behaviour 13(5): 442–350.
medical image processing, pattern recognition, and data mining. Johnson MD, Fornell C. 1991. A framework for comparing cus-
Chompoonut Suttikun is a doctoral student in Hospitality Administra- tomer satisfaction across individuals and product categories.
tion at Texas Tech University, USA. Her research interests include Journal of Economic Psychology 12(2): 267–286.
consumer behavior, marketing, service management, customer per- Jones JP. 2006. When Ads Work: New Proof that Advertising
ception and satisfaction, and motivation. Triggers Sales. ME Sharpe, Inc: New York, NY.
Kenning PH, Plassmann H. 2008. How neuroscience can inform
consumer research. IEEE Transactions on Neural Systems and
Rehabilitation Engineering 16(6): 532–538.
Kim HW, Xu Y, Gupta S. 2012. Which is more important in Inter-
REFERENCES net shopping, perceived price or trust? Electronic Commerce
Research and Applications 11(3): 241–252.
Klucharev V, Smidts A, Fernández G. 2008. Brain mechanisms of
Banich MT, Compton R. 2010. Cognitive Neuroscience3rd edn.
persuasion: how “expert power” modulates memory and atti-
Wadsworth Cengage Learning Belmont: Cengage Learning.,
tudes. Social Cognitive and Affective Neuroscience 3(4):
Belmont, CA.
353–366.
Burgess PW, Dumontheil I, Gilbert SJ. 2007. The gateway hypoth-
esis of rostral prefrontal cortex (area 10) function. Trends in Knutson B, Rick S, Wimmer GE, Prelec D, Loewenstein G. 2007.
Cognitive Sciences 11(7): 290–298. Neural predictors of purchases. Neuron 53(1): 147–156.
Bush AJ, Smith R, Martin C. 1999. The influence of consumer so- Mower JM, Kim M, Childs ML. 2012. Exterior atmospherics and
cialization variables on attitude toward advertising: a compari- consumer behavior: influence of landscaping and window dis-
son of African–Americans and Caucasians. Journal of play. Journal of Fashion Marketing and Management: An Inter-
Advertising 28(3): 13–24. national Journal 16(4): 442–453.
Button KS, Ioannidis JP, Mokrysz C, Nosek BA, Flint J, Robinson Nettelhorst SC, Brannon LA. 2012. The effect of advertisement
ES, Munafò MR. 2013. Power failure: why small sample size choice on attention. Computers in Human Behavior 28(2):
undermines the reliability of neuroscience. Nature Reviews 683–687.
Neuroscience 14(5): 365–376. Northover, HJ. 2012. Assessing the value of neurophysiological
Chandy RK, Tellis GJ, MacInnis DJ, Thaivanich P. 2001. What to measurement for advertising pre-testing, (Ph.D. Dissertation).
say when: advertising appeals in evolving markets. Journal of University of South Australia.
Marketing Research 38(4): 399–414. Phelps J, Thorson E. 1991. Brand familiarity and product involve-
Choi S, Rifon NJ. 2007. Who is the celebrity in advertising? Under- ment effects on the attitude toward an ad-brand attitude relation-
standing dimensions of celebrity images. The Journal of ship. Advances in Consumer Research 18(1): 202–209.
Popular Culture 40(2): 304–324. Plassmann H, Ramsoy TZ, Milosavljevic M. 2012. Branding the
Connell PM, Brucks M, Nielsen JH. 2014. How childhood adver- brain: a critical review and outlook. Journal of Consumer
tising exposure can create biased product evaluations that per- Psychology 22(1): 18–36.
sist into adulthood. Journal of Consumer Research 41(1): Rath T, Baker K, Dumont JA, Peters RT, Jiang H, Qiao SW,
119–134. Lencer WI, Pierce GF, Blumberg RS. 2015. Fc-fusion proteins
Desmond JE, Glover GH. 2002. Estimating sample size in func- and FcRn: structural insights for longer-lasting and more effective
tional MRI (fMRI) neuroimaging studies: statistical power therapeutics. Critical Reviews in Biotechnology 35(2): 235–254.
analyses. Journal of Neuroscience Methods 118(2): 115–128. Ruiz S, Sicilia M. 2004. The impact of cognitive and/or affective
Eckman M, Wagner J. 1994. Judging the attractiveness of product processing styles on consumer response to advertising appeals.
design: the effect of visual attributes and consumer characteris- Journal of Business Research 57(6): 657–664.
tics. Advances in Consumer Research 21: 560–560. Sadeghi M, Fakharyan M, Dadkhah R, Reza M, Vosta SN, Jafari M.
Ehrenberg AS, Chakrapani C. 1974. The pattern of consumer attitudes 2015. Investigating the effect of rational and emotional advertis-
in. AAPOR Conference Lake George: New York 14(25–34). ing appeals of Hamrahe Aval mobile operator on attitude to-
Fabrigar LR, Petty RE. 1999. The role of the affective and cognitive wards advertising and brand attitude (case study: student users
bases of attitudes in susceptibility to affectively. Personality and of mobile in the area of Tehran). International Journal of Asian
Social Psychology Bulletin 25: 363–381. Social Science 5(4): 233–244.
Fransson P, Marrelec G. 2008. The precuneus/posterior cingulate Schacter DL, Gilbert SD, Wegner DM. 2010. Psychology2nd edn.
cortex plays a pivotal role in the default mode network: evidence Worth Publishers: New York; Solomon MR. 2008. Consumer
from a partial correlation network analysis. NeuroImage 42(3): Behavior: Buying, Having, and Being8th edn. Pearson Prentice
1178–1184. Hall: Upper Saddle River, NJ.

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
548 H. Chang et al.

Stallen M, Smidts A, Rijpkema M, Smit G, Klucharev V, Fernández Van der Waldt DLR, Van Loggerenberg M, Wehmeyer L. 2009.
G. 2010. Celebrities and shoes on the female brain: the neural Celebrity endorsements versus created spokespersons in
correlates of product evaluation in the context of fame. Journal advertising: a survey among students. South African Journal of
of Economic Psychology 31(5): 802–811. Economic and Management Sciences 12(1): 100–114.
Taylor J, Kennedy R, McDonald C, Larguinat L, El Ouarzazi Y, Wang Y, Sun LW, Toncar M. 2009. Examining beliefs and attitudes
Haddad N. 2013. Is the multi-platform whole more powerful toward online advertising among Chinese consumers. Direct
than its separate parts? Measuring the sales effects of cross- Marketing: An International Journal 3(1): 52–66.
media advertising. Journal of Advertising Research 53(2): Wang JS, Cheng YF, Chu YL. 2013. Effect of celebrity en-
200–211. dorsements on consumer purchase intentions: advertising ef-
Tom G, Clark R, Elmer L, Grech E, Masetti J, Sandhar H. 1992. fect and advertising appeal as mediators. Human Factors
The use of created versus celebrity spokespersons in advertise- and Ergonomics in Manufacturing & Service Industries 23
ments. Journal of Consumer Marketing 9(4): 45–51. (5): 357–367.
Tsang MM, Ho SC, Liang TP. 2004. Consumer attitudes toward Worsley KJ, Evans AC, Marrett S, Neelin P. 1992. A three dimen-
mobile advertising: an empirical study. International Journal sional statistical analysis for CBF activation studies in human
of Electronic Commerce 8(3): 65–78. brain. Journal of Cerebral Blood Flow and Metabolism 12: 900.
Tse DK, Wilton PC. 1988. Models of consumer satisfaction Wright M, Klÿn B. 1998. Environmental attitude: behaviour corre-
formation: an extension. Journal of Marketing Research 25: lations in 21 countries. Journal of Empirical Generalisations in
204–212. Marketing Science 3(3): 42–60.

Copyright © 2016 John Wiley & Sons, Ltd. J. Consumer Behav., 15: 538–548 (2016)
DOI: 10.1002/cb
Copyright of Journal of Consumer Behaviour is the property of John Wiley & Sons, Inc. and
its content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.