erps_brand
TRANSCRIPT
![Page 1: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/1.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
Late Positive Potential in brand logo recognition: relationship between emotional response and self-‐reported preference
Christa Villari, Nina Underman, Maya Reiselbach, Julia Gomez
Abstract
Previous market research suggests that many factors, like familiarity and emotional response,
influence consumer decision-making. The present study looked at a neural correlate of emotional
response, the Late Positive Potential (LPP), and surveys of self-reported brand preference in
order to better understand how advertising efforts may influence emotional processing and
potentially affect purchasing behavior. The hypothesis was that self-reported preference of a
brand would correlate with its LPP response, and that familiarity would increase the LPP
amplitude and strength of self-reported ratings. Participants were shown images of familiar and
unfamiliar brand logos as ERPs were recorded, then completed a survey rating their preference
for the brands. Survey ratings of familiar logos were rated significantly more positively and
strongly than unfamiliar logos. There was no significant difference in mean amplitude of LPP
response to familiar versus unfamiliar logos in the centroparietal region. However, there was a
trending, positive correlation between difference in strength of preference and difference in mean
LPP amplitude when comparing responses to familiar versus unfamiliar logos. The findings
support the principle that advertising is an important influence on consumer psychology,
increasing brand familiarity in ways that may strengthen a consumer’s emotional response to a
brand and possibly affect purchase behavior.
![Page 2: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/2.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
2
Introduction
Consumers are confronted every day with the decision to purchase just one product over
a large number of equally viable alternative products that perform the same task. Understanding
the consumer decision process that leads to the purchase of a certain brand over others is
valuable information for marketers and companies that allocate millions of dollars towards
advertising each year. The emphasis of the present study was to explore how advertising efforts
that increase a consumer’s familiarity with a brand may influence emotional processing, which
plays a crucial role in consumer decision-making.
The generalized and widely accepted model for decision making involves the process of
examining possible options, comparing the possible outcomes, and then choosing a course of
action, such as purchasing a certain product (Perreau, 2014). However, this method can be
inefficient. If a consumer follows this model every time he or she is bombarded by, for example,
the many different laundry detergents at the supermarket, then this shopper could spend hours
simply weighing the options of which detergent to buy. Given the number of options and the
similar functions of the products, the cognitive load of this basic decision would be tremendous.
Researchers have therefore come to support a simplified model of decision making for this basic
consumer choice task that minimizes cognitive demand and maximizes the potential for a
satisfying outcome of this purchase decision (Hoyer, 1984). This model is based upon the
understanding that decision makers apply simple heuristics, adjusting the rules to these heuristics
based on the feedback they receive from their environment (Hogarth, 1981). In the context of
consumer decision making, past interactions with a brand/product can result in a positive,
negative, or neutral experience. These interactions can occur in various ways, such as by first-
hand use of the product, or through advertising and marketing efforts. Upon evaluating this
![Page 3: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/3.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
3
experience, the consumer begins to form opinions about whether this brand satisfies (or does not
satisfy) individual needs better than existing alternatives. This phenomenon, experts believe, may
drive “brand loyalty,” the tendency for some consumers to continue to purchase the same brand
of items rather than competing brands (Bettman et. al, 1991).
This simplified decision-making, heuristic driving product choice is validated by a
particular study in which participants’ most “beloved” brands resulted in reduced activation of a
region within the dorsolateral prefrontal cortex (Schaefer et. al, 2007). Because the dorsolateral
prefrontal cortex is associated with decision-making, this study shows that the beloved brand has
the ability to reduce strategic reasoning and therefore result in more “impulsive” decision
making during the path to purchase. This simplified decision-making heuristic and the concept of
brand loyalty built on past experience are also consistent with Thorndike’s theory of
reinforcement learning. This theory, which is implicated in many decision making paradigms,
states that human beings learn from the consequences of their actions (Thorndike, 1970),
therefore suggesting that the previous interactions a consumer has with a brand’s product play a
crucial role in associating positive or negative emotions and attributes to the product.
Many studies on consumer behavior have attributed increased purchase behavior to this
heuristic model: consumers are more likely to purchase a product from a brand that they are
familiar with and feel positively about as opposed to an unfamiliar brand that they feel neutrally
towards (Whan et. al, 2010). This suggests that the decision whether to buy a product or not may
be influenced by emotional associations made as soon as the consumer views a characteristic
logo or brand name and before a consumer consciously strategizes about the features or functions
of the product. While the overall consumer decision-making process does involve other factors
besides brand loyalty, the general path-to-purchase for many typical consumers is believed to be
![Page 4: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/4.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
4
a stepwise process that integrates primary sensory information with emotional associations as
well as cognitive decision-making (Perreau, 2014).
Consumer neuroscience, or “neuromarketing,” is a developing field that uses knowledge
about neuroscience and human psychology as a market research strategy aiming to understand
the source and strength of consumer behavior via neural mechanisms. Within this field, the use
of electroencephalogram recording (EEG) has often been used to study and analyze the cognitive
processes underlying the consumer experience (Morin, 2011). Consumer-research EEG studies
are often concerned with studying the Late Positive Potential (LPP) event-related potential (ERP)
waveform. The LPP is a positive wave that arises about 300 ms after a stimulus is presented and
generally peaks about 400-600 ms post-stimulus over the centroparietal scalp region (Cuthbert
et. al, 2000). The LPP is thought to be a component strongly modulated by the emotional
intensity of a stimulus, regardless of the valence of this emotion (Brown et. al, 2012). This is
further confirmed by studies, for example, by Liu et. al (2012) that have shown a larger LPP
component for emotional images versus neutral images. Therefore, the LPP is a component of
crucial interest for exploring emotional processing in response to visual perception of brand
logos, a part of the consumer decision-making heuristic described above.
The purpose of this present study was to examine the emotional responses elicited by
various stimuli representative of specific brands (i.e. brand logos) to better understand the forces
that influence a consumer to remain loyal to or trust one brand over another. The perception of
brands and their logos, the identifying images tied to a specific company or product, are crucially
important in distinguishing equivalent products. Brands are often marketed in ways that try to
elicit specific emotional reactions in potential customers or cause a certain target demographic to
self-identify with specific qualities the brand claims to represent (de Azevedo, 2010). Given the
![Page 5: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/5.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
5
importance of establishing this connection between brand and consumer, marketing of the brand
logo is often repetitive or themed during advertising campaigns to make sure consumers are
aware of and familiar with the feelings or image the company is trying to convey. This fits in
with the simplified decision-making model described above. From a marketing perspective,
repetition of an advertising campaign that emphasizes a specific quality of a brand builds
familiarity and a specific emotional association that consumers hopefully recall as soon as they
see the brand logo in the store and use to influence purchase behavior. From a consumer
perspective, the emotional ties established between the consumer and the brand allow for more
efficient decision-making.
This study used self-report preference surveys and LPP ERP waveforms to explore the
influence that advertising efforts may have in creating emotional associations to brand images.
The self-report preference surveys in this study were used as a representation of the valence that
a consumer might feel towards a particular brand logo he or she sees on the store shelves or in a
television commercial. LPP ERP waveforms were averaged from EEG recordings to provide
insight into the neural basis of emotional intensity elicited by certain brand logos. We combined
these two experimental measures in an attempt to examine LPP responses to both familiar and
unfamiliar brand logos and to determine whether such responses were correlated with self-
reported ratings of these logos. Based on the assumption that familiarity and past experience with
a product create stronger emotions for consumers than an unfamiliar product with which they
have little or no experience with, the overall hypothesis of the present study was that familiar
brands would receive more positive LPP mean amplitudes and stronger self-reported preference
ratings, regardless of valence, and that differences in the emotional strength of these ratings
would correlate with differences of mean LPP response for unfamiliar and familiar stimuli.
![Page 6: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/6.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
6
Methods
Participants
Twenty participants (nineteen undergraduate students and one professor) provided EEG
and behavioral data for this study as part of course credit. Participants were 35% (7 out of 20)
male and 65% (13 out of 20) female, ranging from ages 20 to 34. Mean age of participants was
21.6. All participants were right-handed. The majority of participants were native English
speakers and free of neurological conditions or major head injuries; 19 out of 20 participants
qualified for each factor.
Stimuli
Eighty images, each a unique brand logo without letters or numbers, were used as target
stimuli in the experiment. Stimuli were pre-divided by experimenters into familiar and
unfamiliar categories, and these categories were confirmed by 25 peers, all of whom were
Bowdoin students not participating in the course. Stimuli were sized to fill the whole screen
during testing and were given white backgrounds on a computer monitor’s white screen.
Presentation of stimuli was created using E-Studio from E-Prime 2.0 for Windows, and then
presented to participants using E-Run from the same software suite.
Design
During the EEG-recorded task portion of the experiment, each participant was presented
with all 80 stimuli, with the order of stimuli and length of interstimulus screen presentation
randomized across the participants. During the self-report preference survey portion of the
experiment, participants received one of four versions of the survey that presented the images in
![Page 7: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/7.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
7
different orders. This was done to prevent bias of response based on how early or late in the
survey a participant was asked to rate a specific brand image.
Procedure
Each two-hour session was run through the neuroscience program in the psychology
department at Bowdoin College. Participants consented to the study and provided anonymous
demographic information prior to beginning EEG and survey data collection. Participants
secured the EEG cap onto their heads and two experimenters filled the electrodes with gel, while
checking that impedances were sufficiently low (25 kΩ). Participants provided data for 5 studies,
always in the same order, with the present study being 2nd of these 5.
In the present study, all participants viewed a randomized order of 80 stimuli: 40
unfamiliar and 40 familiar. Prior to viewing stimuli, participants were asked to be attentive to the
presentation. Stimuli were presented for 1500 ms each. In between stimuli, a white screen with a
fixation cross at the center was presented for a variable length of time between 900-1500ms. This
screen was intended to provide a baseline to viewing stimuli, as well as to prevent habituation to
the presentation. No behavioral response was required from participants during the EEG
recording. After finishing the 5th and final study, participants filled out a paper survey in which
they rated each brand on a valence scale of 1 to 5, with 1 being “strongly dislike”, 3 being
“neutral”, and 5 being “strongly like.” In subsequent analysis, we re-coded survey scores to
reflect strength of preference. A score of 3 remained “3”, whereas the average of the 2 and 4 was
coded as “4” and the average of 1 and 5 was coded as “5.” Thus, the higher the strength score,
the stronger the preference of the brand.
![Page 8: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/8.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
8
EEG/ERP Recording and Analysis
EEG was recorded from 64 electrode channels using Brain Amp. Impedances were
checked to be below 25 kΩ before testing and in between studies. Recordings were filtered to be
between 0.1-100 Hz and sampling was conducted at a 250 Hz rate. Recordings were referenced
to the Cz electrode site and pre-processed using EEG Lab with MatLab. Data was imported from
the vhdr file, and stimulus events were defined and binned as either familiar or unfamiliar. Any
bad channels were identified and corrected using spherical interpolation. A low pass filter of 40
Hz was applied. Recordings were then segmented from -200 ms pre-stimulus to 1000 ms post-
stimulus, with a baseline correction using the -200 ms pre-stimulus interval. Artifacts were
detected with a simple voltage threshold from -100 to 100 uV, and a sample-to-sample threshold
difference of 50 uV. Data was re-referenced to the class average of all channels. ERP waveforms
for each channel were averaged and plotted.
The LPP waveform was generated from all subjects’ centroparietal channel recordings:
Pz, CPz, FCz, CP1, and CP2 (Figure 1). Specifically, LPP was measured as the mean amplitude
at these sites from 400 to 1000 ms post-stimulus. These channel locations and time frame were
selected based on previous literature (Weinberg & Hajcak, 2010). In follow-up analysis, LPP
was re-measured as the mean amplitude from the same sites occurring 400 to 800 ms post-
stimulus.
Statistics
Statistical analysis was conducted on Microsoft Excel, SPSS, and GraphPad Prism 6.
Two-tailed, paired t-tests were used to assess any differences between familiar and unfamiliar
logos with regards to LPP mean amplitude, mean survey scores coded by strength, and mean
![Page 9: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/9.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
9
survey scores coded by valence. Significance was assigned at a p-value less than 0.05, and
significance was considered “trending” at a p-value less than 0.1. Average differences between
unfamiliar and familiar logo images were calculated for LPP mean amplitude, survey scores
coded by strength, and survey scores coded by valence. Two correlations of differences were
conducted, comparing LPP mean amplitude to survey scores coded by strength or valence.
Results
Survey Results
Survey results were coded either by valence or strength of the ratings (Figure 2). The
original hypothesis focused on looking at the overall strength of self-preference ratings based on
previous evidence that both positive and negative emotional images show a larger LPP than
neutral stimuli (Brown et. al, 2012). When coded by strength, a paired, two-tailed t-test showed,
as expected, that familiar logos were rated significantly higher (i.e. more strongly) than
unfamiliar ones (t(19) = 7.34, SEM = 0.06, p < 0.01). Additionally, a paired, two-tailed t-test for
mean survey scores coded by valence showed that familiar logos were rated significantly higher
(i.e. more favorably) than unfamiliar ones (t(19) = 17.59, SEM = 0.09, p < 0.01). Survey results
and statistics are summarized in Figure 3.
ERP Results
LPP was measured from five centroparietal channels Pz, CPz, FCz, CP1, and CP2, which
are circled in red in Figure 1. First, mean LPP amplitude from 400-1000ms was compared for
unfamiliar and familiar brand logos. It was expected that familiar logos would have an increased
LPP mean amplitude compared to that for unfamiliar ones. However, a paired, two-tailed t-test
![Page 10: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/10.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
10
showed no significant difference in mean LPP amplitude from 400-1000ms when comparing
unfamiliar and familiar logos (t(19) = 0.81, SEM = 0.35, p = 0.43). Based upon the average LPP
waveform (Figure 4) for this time frame, it was thought that mean LPP amplitude from 400-800
ms might significantly differ based on familiarity of logos. Yet, a paired, two-tailed t-test showed
no significant difference in mean LPP amplitude from 400-800 ms between familiar and
unfamiliar logos (t(19) =1.48, SEM = 0.36, p = 0.16). Mean LPP amplitude results are
summarized in Figure 5.
Correlation Results
It was predicted that, when comparing familiar to unfamiliar logos, differences in mean
survey scores coded by strength and LPP mean amplitude would be positively correlated with
each other. Average difference between unfamiliar and familiar LPP ERP amplitudes was
correlated with average difference between unfamiliar and familiar self-reported surveys.
Overall, survey results coded by strength produced stronger correlation values and a p-value of
the correlation that was closer to statistical significance than survey results coded by valence
(Figure 6). The correlation of differences in unfamiliar and familiar logo preference strength and
LPP response showed a weak, positive correlation (r(18) = 0.30, p = 0.10). The correlation of
differences in unfamiliar and familiar logo preference valence and LPP response showed a weak,
negative correlation (r(18) = -0.22, p = 0.17).
![Page 11: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/11.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
11
Discussion
Implications of results
These findings have interesting and relevant implications for the field of neuromarketing.
Whether coded by valence or strength, self-reported responses rated familiar brands significantly
higher (i.e. participants felt more positively and strongly about familiar brands) than unfamiliar
brands. There are two possible explanations for this result. Firstly, this could indicate that
familiar logos are inherently more preferred than unfamiliar logos; if one recognizes the brand,
one tends to rate it more highly.
The second possible conclusion is that the selection of brand stimuli content was
unintentionally skewed, such that there was not an equal distribution of stimuli for which it was
expected that subjects would feel either negatively or positively about. If this study were to be
replicated, conscious effort should be made to vary the brands represented so that more logos
would be rated negatively, creating an equal distribution of positive, negative and neutral brand
logos, especially in the familiar category. The current study used only 80 logos as stimuli, so
perhaps simply increasing this number could help eliminate a potential bias in selecting
positively regarded logos. Further studies could include a category of response on the self-report
survey to distinguish between mixed feelings that consumers may feel about specific brands and
neutral feelings. Additionally, further studies could include both images of brand logos and
branded products in an attempt to distinguish potential differences in feelings consumers may
have towards a brand as a whole versus the products it includes.
Though there was no statistical significance for the difference in the average LPP mean
amplitudes elicited by familiar and unfamiliar stimuli from 400-800 ms, it is worth noting that
the p-value for this comparison was closer to significance when compared to the results for mean
![Page 12: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/12.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
12
LPP from 400-1000 ms. The standard error of the mean for this analysis was large, indicating
that variability was so high that the difference between mean LPP amplitude in familiar and
unfamiliar logos was not significant. These results support the hypothesis that there is a greater
neural correlate for emotional response to familiar brands than unfamiliar ones. In the context of
consumer decision-making and brand loyalty, this result makes sense: because subjects likely
have no prior experience with the unfamiliar brands, it would not be expected for them to have
any kind of emotional association with these brand images.
There was also a weak, positive correlation between differences in self-reported strength
of preference for familiar and unfamiliar logos and differences in LPP responses to familiar and
unfamiliar logos. This finding indicates that, consistent with the original hypothesis, if
participants have a stronger self-reported response to a brand, regardless of response valence,
they also elicit a greater LPP in response to that brand. Since the LPP reflects a neural correlate
of emotional response, a higher self-reported response to a logo appears to be associated with a
greater emotional response. This correlation is not enough to predict actual consumer purchase
behavior, but confirms previous market research assumptions that emotional strength may be
involved in consumer decision-making and brand loyalty (Perreau, 2014).
Limitations of experimental design
Although the results of this study do support original hypotheses and are relatively
consistent with past research, there were a number of limitations in the study design. A larger
sample size could have produced more significant results. Additionally, 25% of the subjects (5
out of 20) had prior exposure to the brand stimuli used; for those 5 subjects, unfamiliar logos
may have become familiar due to the time spent collecting the stimuli and designing the study.
![Page 13: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/13.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
13
Because this study was presented in a block with other experiments, it is possible that the
experiment that occurred immediately prior to the present one could have affected fatigue and
attention during this study. Lack of attention could also have occurred due to the fact that the
experimental design required no motor response, leading to disengagement with the stimuli.
Despite these potential opportunities for fatigue/boredom, raw EEG data did not have many
alpha waves, suggesting the data was not too affected by potential participants’ boredom.
A major obstacle in regards to experimental stimuli was the basis on which the
experimenters determined what constituted a “familiar” versus an “unfamiliar” logo. We did
make efforts to accurately categorize the stimuli by first dividing the logos based on prior
knowledge and experience with them, then presenting the lists to a sample of individuals who did
not participate in the study. Feedback from this focus group was used to alter final
categorizations of stimuli. However, there was no way of determining whether the categorization
of stimuli was accurate for the specific participants. For instance, participants may have had
geographic differences in exposure to certain brand logos depending on where they grew up.
Another obstacle in regards to the post-experiment survey is whether the survey
measurement is an accurate reflection of preference. Although the survey was anonymous, there
is always a potential bias in self-perception that may not reflect neural correlates as strongly.
There also may have been differences in subjects’ interpretation of how the survey prompt was
worded. Participants rated brand logos based on a valence scale, but perhaps their responses
would have differed if the scale was based on strength or included an option for mixed feelings
towards a brand. Experimental design could have asked participants to indicate familiarity and
preference of each image immediately after presentation, though this would have led to
![Page 14: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/14.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
14
additional complexities in averaging ERP data and running statistical tests that this study sought
to avoid.
One major limitation concerns interpretation of the ERP waveform data. While we
assumed the waveform we analyzed from 400-1000 ms in the centroparietal region was the LPP,
a neural correlate of emotional intensity, there are other possibilities that could lead to different
interpretations and applications of these results. For example, previous research showed that an
increased LPP response to visual stimuli corresponded to increased attention for those stimuli,
suggesting that the LPP may also be a neural correlate for attention or recognition memory
(Pastor et al., 2008; Finnigan et. al, 2002). While it logically makes sense that familiarity,
attention, and recognition memory are needed to produce emotional responses to brand logos, it
is possible that the ERP results reflect one of these other neural correlates instead of emotional
response. Our weak, positive correlation of observed ERP waveforms and emotional strength of
preference provides loose support that the neural component analyzed was a reflection of
emotional intensity of a brand logo, as originally hypothesized. However, these alternative
possibilities should be considered, especially since the time interval and channel locations of the
averaged LPP mean amplitude waveform analyzed by this study could instead be the LPP of the
parietal old/new effect, which reflects recognition memory.
For the parietal old/new effect, old/repeated items elicit more positive LPPs than
new/unrepeated items (Finnigan et al., 2002). Studies examining the LPP in the context of the
parietal old/new effect have demonstrated that LPP waveforms taken from centroparietal
channels 500-800 ms after stimulus presentation represent decision making accuracy; correct
decisions in a word discrimination task produced greater LPP amplitudes than inaccurate
decisions (Finnigan et al., 2002). Additionally, this study did not find that the LPP corresponded
![Page 15: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/15.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
15
to emotional strength; at certain centroparietal electrodes, the peak LPP amplitudes were
equivalent for new/unfamiliar words and emotionally strong words (Finnigan et al., 2002).
Taken in the context of marketing, one could imagine that when individuals are faced
with numerous brand choices, they will be more likely to attend to a brand that is familiar or
triggers a recognition memory, without necessarily increasing strength of emotional response.
However, like the factor of emotional strength, factors of increased attention or awareness of
increased familiarity when a brand logo is viewed may not necessarily dictate consumer decision
to buy that specific brand’s product, though they may be more inclined to heavily weigh it as an
option.
Future directions
In addition to the suggestions made above to address specific limitations of this study,
different data analysis could be used to further explore the application of these results on
consumer neuroscience. Correlating differences of LPP amplitudes and survey responses was
done in this study for general simplicity of analysis. However, correlating by individual LPP
amplitude subcategories (familiar and unfamiliar brand logos) and survey response subcategories
(strong and neutral emotional response for familiar and unfamiliar brand logos) would provide
additional detail that could help confirm or alter interpretations of the results presented in this
study. Additionally, it would be interesting to code the data so that it could be analyzed by
individual brand image and by gender. In scoring the surveys, certain brand images, specifically
Playboy and Nike, produced opposite valence self-report preference ratings within the
participants. It would be interesting to see, for example, what the correlations between mean LPP
amplitude for the Playboy image and various categories of self-report preference ratings would
![Page 16: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/16.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
16
look like and if these correlations are significantly different for male participants versus female
participants.
Future data analysis could also use mean LPP averages from different time windows.
While significance was not seen in the mean LPP averages for familiar and unfamiliar logos
when these averages were taken from 400-1000 ms or from 400-800 ms in the centroparietal
region, the p-value did get smaller as the time interval shortened. Perhaps significance would
have been seen if mean LPP averages were obtained from 400-600 ms in the same location.
Previous studies have shown the LPP to begin roughly 300-400 ms after the presentation of an
emotional picture and last through the duration of the picture presentation (Liu et. al., 2012).
This background information and the 1500 ms presentation of each image in this study explains
the reasoning for the original calculation of average mean LPP amplitudes from 400-1000 ms
after image onset. Though calculating average mean LPP amplitude from 400-600 ms may seem
like a small time window, other studies, including the study whose experimental design inspired
this one, have demonstrated significance between brand images that subjects indicated they
preferred or felt indifferent towards using this 400-600 ms time window, even when each image
was presented for 6 seconds (de Azevedoa, 2010). It is possible that this study would produce a
shorter LPP wave in response to a presented stimulus because brand images were not reinforced
via a motor response. Additionally, unlike images used in other ERP studies of emotion, brand
images in this study were not purposely selected from a database of pictures with established
valence values like the International Affective Picture System, which may have also contributed
to a shorter LPP waveform (Cuthbert et. al, 2000).
Finally, ERP significance for mean LPP amplitude averages of familiar and unfamiliar
brand logos may have been seen if a different region of channels was used. While previous
![Page 17: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/17.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
17
research indicates that the LPP is maximal over centroparietal areas, other regions of the brain
have also shown an LPP response to emotional images, including images of brands (Cuthbert et.
al, 2000; de Azevedoa, 2010; Pastor et. al, 2008). Specifically, previous research using EEG to
explore perceptions of consumer brands showed that, while participants’ most preferred brands
had stronger LPP activations in the centroparietal regions from 400-600 ms compared to
indifferent and unknown brands, preferred brands were also more active in the frontal cortex
than indifferent brands during the same time window (de Azevedoa, 2010). In looking at the
grand average waveforms from this study, it does appear that the greatest LPP effect could have
been more frontal, though actual statistical analysis would need to be performed from an average
of frontal channels to confirm or deny this hypothesis (Figure 7).
Clearly, there are a plethora of future investigations necessary to fully and accurately
understand the results of this study. Overall, the results show that there may be a positive
correlation between LPP response and strength of self-reported preference, indicating that
marketing targeting consumer emotion should strive to build brand familiarity. However, the
relationship suggested by this correlation, that LPP response to a brand logo can impact actual
consumer decision-making, should be tested. If we assume, as our original hypothesis did, that
the LPP is a neural correlate reflecting emotion intensity, a logical future study could examine
the relationship between LPP response strength and likelihood of choosing to buy such a brand’s
product in a behavioral testing paradigm. This study could answer the question of whether, when
faced with choosing between an unfamiliar brand and a familiar but negatively perceived brand,
consumers will choose the brand with a greater emotional response, regardless of the valence of
that response, or be more likely to base purchase decision off of economic factors, for example.
The results of such a study would help to provide a complete picture of the consumer decision-
![Page 18: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/18.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
18
making process, allowing companies to market products more effectively and providing
additional evidence to support the importance of the consumer neuroscience field.
Conclusion
Overall, these results suggest that if marketers are aiming to target consumer emotion,
familiarity matters. This interpretation of the results provides better understanding as to why
companies spend millions of dollars every year to advertise their brands. By using advertising
repetition to boost consumer exposure to a brand, marketers can not only increase consumers’
overall familiarity with a brand but also dictate specific emotional associations or feelings of
identity that consumers perceive when viewing a brand logo. However, because it was not
possible to conduct EEG recording in an actual store as real consumers were making purchases,
we cannot predict how, if at all, the weak, positive correlation between LPP mean amplitudes
and self-report preference demonstrated in this study would actually translate into purchasing
behavior. Nevertheless, this study is a crucial first step for investigating the neural processes
behind consumer decision-making. Though the simplicity of the experimental design was
prioritized, leading to limitations discussed above, these findings are novel and relevant for
studying neuromarketing, providing a stepping stone for future exploration of the field.
![Page 19: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/19.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
19
References
Bettman, J. R., Johnson, E. J., & Payne, J. W. (1991). Consumer decision making. Handbook of
Consumer Behavior, 44(2), 50-84.
Brown, S. B., van Steenbergen, H., Band, G. P., de Rover, M., & Nieuwenhuis, S. (2012).
Functional significance of the emotion-related late positive potential. Frontiers in Human
Neuroscience, 6.
Cuthbert, B. N., Schupp, H. T., Bradley, M. M., Birbaumer, N., & Lang, P. J. (2000). Brain
potentials in affective picture processing: covariation with autonomic arousal and
affective report. Biological Psychology, 52(2), 95-111.
de Azevedoa, P. C. B. S. (2010). Perception of commercial brands and the emotional and social
value: A spatiotemporal EEG analysis. Technical University of Lisbon, Portugal.
Finnigan, S., Humphreys, M. S., Dennis, S., & Geffen, G. (2002). ERP ‘old/new’effects:
memory strength and decisional factor(s). Neuropsychologia, 40(13), 2288-2304.
Hogarth, R. M. (1981). Beyond discrete biases: Functional and dysfunctional aspects of
judgmental heuristics. Psychological Bulletin, 90(2), 197.
Hoyer, W. D. (1984). An examination of consumer decision making for a common repeat
purchase product. Journal of consumer research, 11(3), 822-829.
Liu, Y., Huang, H., McGinnis-Deweese, M., Keil, A., & Ding, M. (2012). Neural substrate of the
late positive potential in emotional processing. The Journal of Neuroscience, 32(42), 14563-
14572.
Morin, C. (2011). Neuromarketing: the new science of consumer behavior. Society, 48(2), 131-
135.
![Page 20: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/20.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
20
Pastor, M. C., Bradley, M. M., Löw, A., Versace, F., Moltó, J., & Lang, P. J. (2008). Affective
picture perception: emotion, context, and the late positive potential. Brain Research,
1189, 145-151.
Perreau, F. (2014). The 5 stages of Consumer Buying Decision Process. Consumer Behavior
Factors and Variables. Retrieved from http://theconsumerfactor.com/en/5-stages-
consumer-buying-decision-process/
Thorndike, E. L. (1970). Laws and hypotheses for behavior. Animal intelligence (pp. 241–281).
Darien, CT: Hafner Publishing Co.
Schaefer, M., & Rotte, M. (2007). Favorite brands as cultural objects modulate reward circuit.
Neuroreport, 18(2), 141-145.
Weinberg, A., & Hajcak, G. (2010). Beyond good and evil: the time-course of neural activity
elicited by specific picture content. Emotion, 10(6), 767-782.
Whan Park, C., MacInnis, Deborah J., Priester, Joseph, Eisingerich, Andreas B., & Iacobucci,
Dawn. (2010). Brand Attachment and Brand Attitude Strength: Conceptual and Empirical
Differentiation of Two Critical Brand Equity Drivers. Journal of Marketing, 74(6), 1-17.
![Page 21: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/21.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
21
Figures
Figure 1: Electrode channel locations used during EEG recording. Cz (blue circle) served as the reference during EEG recording for all subjects. The red circles indicate the five centroparietal electrode channels averaged for ERP analysis of the LPP from 400-1000 ms, and also 400-800ms.
Figure 2: Coding of self-report surveys done by both valence (left) and strength (right) of preference response. For valence coding, surveys were scored on a scale of 1-5, with 1 representing strong dislike, 5 representing strong like, and 3 representing neutral preference for the brand image shown during experimental testing. Survey responses were also coded by strength. In this scale of 3-5, 3 represents a neutral response, 4 represents a medium-strength response (4’s and 2’s grouped together), and 5 represents a high-strength response (5’s and 1’s grouped together).
![Page 22: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/22.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
22
Type of Coding (Scale) Average Survey Score: Unfamiliar Logos (SEM)
Average Survey Score: Familiar Logos (SEM)
Average Difference in Scores (p-value*)
Valence Coding (1-5)
3.01 (0.010)
3.65 (0.081)
0.64 (p < 0.001)
Strength Coding (3-5)
3.02 (0.009)
4.09 (0.061)
1.07 (p < 0.001)
*p-value from paired, two-tailed t-test Figure 3: Self-report survey results. Means and differences shown for unfamiliar and familiar logos coded by both strength and valence. Differences between unfamiliar and familiar logos were statistically significant in both coding conditions. Familiar logos were rated more positively and strongly compared unfamiliar ones.
Figure 4: ERP results. Average ERP mean waveforms for the LPP (centroparietal region, 400-1000 ms) for visualization of familiar and unfamiliar logos. Centroparietal region represented by an average of FCz, CP1, CP2, CPz, and Pz channels.
![Page 23: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/23.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
23
LPP AMPLITUDES FOR CENTROPARIETAL REGION
400-1000 ms 400 - 800 ms
Average LPP amplitude: Unfamiliar Logos (SEM)
0.5853 µV (0.3707)
0.6761 µV (0.4165)
Average LPP Amplitude: Familiar Logos (SEM)
0.8658 µV (0.4695)
1.2089 µV (0.5061)
P-Value (paired 2-tailed t-test)
0.43 0.16
Figure 5: Average LPP mean amplitudes for centroparietal region from 400 -1000 and 400- 800 ms. Average mean amplitudes were higher for familiar compared to unfamiliar logos in both time frames. However, while both p-values were not statistically significant, measuring from 400-800 ms produced a p-value that was closer to statistically significance than in the 400-1000 ms time frame. 6A
CORRELATION OF LPP AND SELF-REPORT SURVEY DIFFERENCES
Survey Responses Coded by Strength Survey Responses Coded by Valence
Average LPP amplitude difference (Familiar - Unfamiliar)
1.25 1.25
Average self-report survey response difference (Familiar - unfamiliar)
1.07 0.64
Correlation (R value) between LPP and Self-report survey differences
0.2992 -0.2218
R2 Value of correlation 0.0895 0.0491
P- Value of correlation 0.0999* 0.1738
*no significant correlation, but trending towards significance
![Page 24: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/24.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
24
6B
6C
![Page 25: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/25.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
25
Figure 6: Correlation results. Average difference between unfamiliar and familiar LPP ERP amplitudes was correlated with average difference between unfamiliar and familiar self-report surveys. Survey results coded by strength produced stronger correlation values and a p-value of the correlation that was closer to statistical significance than survey results coded by valence. Statistical results (A) are shown graphically for strength coding (B) and for strength and valence coding (C). Figure 7: Further analysis frontal channel locations (left, circled in blue) and example average ERP waveforms from 400-1000 ms (right).
Familiar Unfamiliar
![Page 26: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/26.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
26
Appendix 1: Stimuli Categorization Familiar:
Unfamiliar:
![Page 27: ERPs_Brand](https://reader033.vdocuments.us/reader033/viewer/2022051315/55cbe3acbb61eb841a8b4753/html5/thumbnails/27.jpg)
COGNITIVE NEUROSCIENCE OF BRAND LOGO RECOGNITION
27
Appendix 2: Survey Example
: