4.1 Neuromarketing Academic Production

Figure 1 shows scientific production associated with neuromarketing in the Scopus and WoS databases from 2004 to 2019 (736 documents). Until 2009, 38 documents had been published, making it the period with the lowest productivity; however, from this year onwards the interest of the scientific community in this area grew significantly. During 2019 (the period of highest production) 128 articles were published, which constitutes 17% of total publications. The trend lines show an increase of interest in the field, which is reflected in an annual growth rate of 31%.

Source: own elaboration.

Figure 1 Neuromarketing scientific production 

Regarding the analysis of scientific production by country, the United States leads the way in both databases, with 43 publications in WoS and 72 in Scopus. Figure 2 provides a list of the 10 countries that lead academic publications in neuromarketing; 8 of them (United States, Spain, Italy, Germany, United Kingdom, China, Japan, and Turkey) are in the top 10 in both databases.

The countries that populate these lists carry out 70% of global research on the subject of neuromarketing. The United States produces 24% of these publications and European countries produce 52% (Figure 2).

Source: own elaboration.

Figure 2 10 Countries with the highest production 

Table 2 provides a list of the most relevant authors and classifies them by the number of publications in each database, their h-index (an indicator proposed by ^{Hirsch (2005)} used to characterize the level of researchers’ scientific production) is also associated. Fabio Babiloni is the researcher with the highest number of publications in both databases, followed by Giovanni Vecchiato and Laura Astolfi; all three are co-authors of multiple articles (^{Astolfi et al., 2007b}; Babiloni et al., 2005; ^{Borghini, Astolfi, Vecchiato, Mattia, & Babiloni, 2014}). Furthermore, Babiloni has an h-index of 66, the best within the list of the most relevant authors.

Figure 3 shows the authors’ co-citation network. This analysis is a bibliometric technique based on the co-occurrence of article references, which allows the characterization of the structure of a field of study from the co-cited authors (^{Culnan, 1986}; ^{White, 2003}; ^{White & Griffith, 1981}). Authors with conceptual affinities in their research tend to cite each other. Of course, some are more relevant than others, allowing them to stand out (^{McCain, 1990}; ^{Sircar, Nerur, & Mahapatra, 2001}). Within the network, authors such as Plassmann, Knutson, Ariely, Lee, Kenning, Mcclure and Vecchiato are the most cited, and even the four most relevant documents (with respect to the number of citations) in this field of study are those in which Knutson, Ariely, Lee and Mcclure participate as authors (Figure 3). Two clusters can also be identified (only the most relevant researchers are mentioned): the largest consisting of Lee, Knutson, Plassmann, Kenning, Mcclure; and the second one including Ariely, Vecchiato, Zaltman and Ohm.

Source: own elaboration.

Figure 3 Network 

Figure 3 shows the network with a strong collaboration among authors such as Babiloni, Vecchiato, Astolfi, Cincotti, Mattia, Fallani and Toppi, which creates a community among these researchers. Another group of co-authors represented by Ma Q., Wang XY. and Wang CC. show another community, but not as important in relation to the number of publications and impact of their authors as the first one.

Table 3 lists the 10 journals (indexed in WoS and Scopus) with the largest number of published articles on neuromarketing. It also shows their impact factor and their quartile according to the database. The most important journal in this field is Frontiers in Neuroscience, a Swedish publication with an h-index of 71 and classified in Q1 by Scopus and Q2 by WoS.

4.2 Network analysis

The search equation produced an initial network composed of 9,162 nodes and 12,556 links; after processing them a final network of 260 nodes and 4031 links was established. The main inclusion criteria for choosing which documents to analyze were: In-degree (degree of entry, number of times a document is cited by others that are part of the network), Betweenness (degree of intermediation, citation of and by other documents in the network), and Out-degree (degree of exit, citations made to other documents in the network). A total of 40 documents were selected for this review: 10 articles with the highest level of In-degree, 15 with the highest level of Betweenness, and 15 with the highest level of Out-degree that involve relevant information over an extended period of time. The classical (hegemonic) documents have high Out-degree and zero In-degree, the current (recent) documents have high In-degree and zero Out-degree, and the structural documents have high Intermediation, which implies that they connect classical studies to current ones. The documents are presented through a tree structure, in which the classical documents are the roots, structural ones form the trunk and current ones are the leaves.

The Neuromarketing network (Figure 4), composed of 260 documents, was obtained through the use of the Modularity Class algorithm that allows the selection of different communities (clusters) within the network by separating the nodes that are more densely connected (^{Blondel, Guillaume, Lambiotte, & Lefebvre, 2008}); in this case 3 perspectives were identified. The size of the nodes (documents) represents the degree to which they are cited; the largest ones are those which have been cited most frequently within the network, which implies that they are the most relevant in this field of study. The four most important nodes are highlighted.

Source: own elaboration.

Figure 4 Neuromarketing network 

4.3 Classical documents

The documents considered to be hegemonic (Table 4) can be classified into two groups: the first includes theoretical research (^{Murphy, Illes, & Reiner, 2008}; ^{Fugate 2007}; ^{Plassmann et al., 2012}) and the second includes experimental research (^{Deppe, Schwindt, Kugel, Plasmann, & Kenning, 2005}; ^{Erk, Spitzer, Wunderlich, Galley, & Walter, 2002}; ^{Knutson, Rick, Wimmer, Prelec, & Loewenstein, 2007}; ^{McClure et al, 2004}; ^{Plassmann, O'Doherty, Shiv, & Rangel, 2008}; ^{Astolfi et al., 2009}; ^{Yoon, Gutchess, Feinberg, & Polk, 2006}).

4.4 Theoretical research

The following authors, who analyze conceptual components of applications, are located in this group. ^{Fugate (2007)} describes the origins of neuromarketing and explains it in simple terms, lists some of the findings of various studies to date, and suggests future directions for research on consumer behavior. This author concludes that the use of neuroscience techniques such as neural activity imaging can support conventional marketing techniques and improve their effectiveness. ^{Hubert and Kenning (2008)} agree with the views discussed by Fugate, however, they state a difference between consumer neuroscience and neuromarketing; for these authors the former is a scientific procedure, but the latter is the application of the findings of neuroscience to management practices. They conclude that consumer neuroscience is an emerging field of study that can meaningfully complement consumer studies. Finally, ^{Murphy et al. (2008)} propose a preliminary code of ethics that promotes research and development of this field of study, considering business profitability without putting consumers at risk.

4.5 Experimental research

Research in this category is done to evaluate brain activity by using techniques from neuroscience, such as fMRI (Functional Magnetic Resonance Imaging), NMR (Nuclear Magnetic Resonance), HR (Heart Rate), EEG (Electroencephalography) scans and GSR (Galvanic Skin Response). Linear regression models are the most used quantitative technique to process information.

Two of the most relevant documents within the hegemonic category are "Neural Correlates of Behavioral Preference for Culturally Familiar Drinks" and "Neural Predictors of Purchases." The former attempts to identify the influence of brand image on choices, behaviors and brain responses of Coca-Cola and Pepsi consumers; it concludes that the Coca-Cola brand has a significant influence on the preferences expressed by consumers and, more importantly, on their brain responses. The latter determines the response of different neuronal circuits to the preference of a product against excessive prices; it suggests that the activation of different circuits related to the anticipatory effect precedes and supports consumers’ purchase decisions.

^{Erk et al. (2002)} showed that artificial cultural objects (vehicles with different characteristics), categorized by the income and social classification of each segment, activate areas of the brain related to reward. Likewise, ^{Deppe et al. (2005)} studied through fMRI the way in which economic decisions are influenced by implicit memory. In the case of products that are more differentiated by the information that the consumer has about the brand, the brain shows less activity in the areas related to working memory and reasoning when comparing favorite brands with others. In contrast, the areas related to emotions demonstrate a higher degree of activity.

Later, ^{Yoon et al. (2006)} used fMRI methods to study if implicit (and often explicit) assumptions regarding product and brand qualities are processed in a similar manner to those about people; and found evidence to the contrary. ^{Kenning and Plassmann (2008)} verified how consumer satisfaction in neural terms is affected by the types of marketing and price changes of a product, with the authors suggesting that the brain calculates the satisfaction experienced through the integration of actual sensory properties and product expectations. Finally, ^{Vecchiato et al. (2010)} analyzed brain activity during the observation of commercials; they found that advertising considered pleasant increased activity in the left hemisphere of the brain.

Despite the use of tools from neuroscience, it is evident that development achieved in the field is not conclusive, and the experiments carried out thus far are only superficial approximations that do not yet allow us to understand what happens in the brain of the consumer.

4.6 Structural documents

Structural documents have the highest rate of centrality (they quote hegemonic documents and at the same time they are quoted by recent publications), which implies that they connect the whole network. The articles in this category deal mostly with experimental research that seeks to understand and identify how the brain and its different areas are affected by stimuli (Table 5).

The first paper in this category was authored by ^{Lee et al. (2007)} who mapped neuroscience techniques that can be used as a basis for research of marketing issues such as trust, pricing, and society-related marketing (ethics). The authors suggest that neuroscience can help us to understand how humans create, store, remember, and relate to brands. Within this theoretical avenue of enquiry, ^{Kenning and Linzmajer (2011)} do not use the concept of neuromarketing, they rather refer to consumer neuroscience and consider it a sub-discipline of neuroeconomics. They also believe that this area of knowledge is still an incipient discipline that requires more exhaustive tests and robust theoretical debate.

^{Wilson et al. (2008)} focus on analyzing the ethical issues involving consumer awareness, consent and understanding, and discuss the models of persuasion employed by marketing. They believe that neuromarketing should be used to help people understand what they really want and not against them, what they refer to as “ethical marketing strategies”.

Later, ^{Astolfi et al. (2008)} determined through various experiments using EEG tests that the parietal areas of the brain play an important role in relation to the flow of information during the observation of television commercials. These areas show greater activity when the subjects are exposed to commercials that have been recalled sometime after their observation. Likewise, ^{Ohme et al. (2010)} propose that the measurement of activity in the frontal lobe can be a diagnostic tool to help determine the potential of television commercials. ^{Guixeres et al. (2017)} established that aspects related to the impact of advertising and its propensity to be recalled can be detected. In their experiment they found a significant correlation between neuroscience metrics and the effectiveness of advertisements on a YouTube channel.

Subsequently, ^{Ariely and Berns (2010)} and ^{Ohme et al. (2010)} addressed the importance of using neuroscience tools and techniques in marketing processes, especially in market research, to evaluate or design products or as part of the development of advertising campaigns. In addition, ^{Plassmann et al. (2012)} proposed that consumer neuroscience can make a substantial contribution to brand positioning research with respect to psychology as a tool for observing the mental process in greater depth.

^{Kenning and Linzmajer (2011)} presented an overview of consumer neuroscience studies, verified methods, findings and implications, and categorized research in relation to its purpose (Product Policy, Pricing Policy, Communication Policy, Distribution Policy, Brand Research). Other authors, such as ^{Venkatraman et al. (2012)} and ^{Solnais et al. (2013)} consider that neuroscience techniques and tools do not replace current methods applied in marketing practice, and propose that these new procedures be seen as complementary to understanding consumer behavior.

4.7 Recent documents

These documents make reference to both classical and structural studies to give them support, structure and form. They have the least number of entries, or no entries at all. However, their contribution is fundamental to the development of new theories that demonstrate the advancement and evolution in this field, as well as playing an indispensable role in the current and future application of new non-conventional methods explained in three different perspectives (Table 6).

The authors (^{Vecchiato et al., 2011a}; ^{Vecchiato et al., 2014a}; ^{Cartocci et al., 2017}) focus their studies on the application of neuromarketing in the advertising industry, considering cultural and gender factors and reactions to public service content through the EEG method. Other authors incline their research towards a more cognitive perspective linked to psychology in order to better understand the functioning of the brain with branded and even latent stimuli which are influenced by gender differences (^{Kim, Kim, Han, Lim, & Im, 2016}; ^{Fehse, Simmank, Gutyrchik, & Sztrókay-Gaul, 2017}; ^{Hsu & Cheng, 2018}; ^{Ramsøy, Skov, Christensen, & Stahlhut, 2018}; ^{Ma, Zhang, & Wang, 2018}).

There is also a study by ^{Crespo-Pereira and Legerén- Lago (2018)} that reviews theories developed from the application of neuromarketing in the audiovisual industry, and the impact that it could have for the development of new content with these characteristics.

^{Harris et al. (2018)}, ^{Jiménez and Zambrano (2018)}, ^{Gani et al. (2018)}, and ^{Lim (2018)} explore the existing alternatives provided by neuromarketing that contribute to current and future theory and practice in different business processes, which also represent lower costs and are more advanced tools.

The perspectives that delve deeper into the three categories identified through the application of the clustering algorithm (Figure 4) are notable. Within these groups, the most outstanding authors in the field of research are still the same, i.e., Plassman, Cartocci, Vecchiato, Astolfi and Ma Qg. This will be later discussed in detail.

4.8 Perspective 1: brain and brand

Initial studies in the field of neuroscience highlight that activation of the ventromedial prefrontal cortex is critical for preference trials (^{Paulus & Frank, 2003}). In the same line of research, ^{McClure et al. (2004)} and ^{Stoll, Baecke, and Kenning (2008)} state "what you see is what you get", for which a functional MRI study on neuronal correlates focused on attractive packaging was conducted. In contrast, ^{Plassmann et al. (2008)} conducted a study on how marketing actions can modulate the neuronal representations of pleasure. These researchers also focus on more cognitive-oriented studies, such as a paper by ^{Klucharev, Smidts and Fernández (2008)} that mentions the "expert power" that marketing can have as a persuasive mechanism. ^{Tusche, Bode and Haynes (2010)} suggest that consumer choices can be predicted through certain neural responses.

Authors like ^{Reimann, Zaichkowsky, Neuhaus, Bender, and Weber (2010)} present the design of a package that gathers behavioral, neuronal and psychological research, which they call an "aesthetic package". Regarding the fMRI method, ^{Venkatraman et al. (2012)} present their perspectives on branding, propose new data for marketers of branding through scanning, and suggest that neuroscience is an effective tool for understanding brand preferences. An exploratory study by ^{Al-Kwifi (2016)} that uses high-tech products suggests that fMRI can play a key role in detecting attitudes towards brand change by focusing on new brands versus known brands analyzed in the field of neurophysiology, response latency, and choice.

With respect to the most current research, we found one paper by ^{Fehse et al. (2017)} which, by means of the fMRI method, studies the perception of the elements involving different functional pathways on organic or popular brands. Furthermore, ^{Goto et al. (2017)} show that the neural signals of selective attention are modulated by subjective preferences and purchase decisions in a virtual shopping task. In a more practical study, ^{Ma et al. (2018)} analyze how the effect of continuous gain influences consumer price perception, an ERP (Event- Related Potential) study.

4.9 Perspective 2: tools and techniques

Authors such as ^{Davidson and Irwin (1999)} indicate that the neuronal responses of the prefrontal cortex, more specifically of the ventromedial and dorsolateral zones, are the representation of positive and negative elemental states. ^{Aftanas, Reva, Varlamov, Pavlov, and Makhnev (2004)} used the EEG method with an experimental sample of 20 right-handed subjects to evaluate the (emotional) brain activity that occurs during the application of a stimulus; the results showed that there is synchronization of the right hemisphere in posterior areas and desynchronization in the left hemisphere in anterior areas. Likewise, ^{Astolfi et al. (2007a)} reaffirm that the use of the EEG and fMRI methods allow us to observe the cortical activity of the brain; they argue that with these tools and techniques, it is possible to identify patterns of functional connectivity that are given by the experimental conditions.

In a subsequent study, ^{Astolfi et al. (2008)} focused their research on cortical activity in order to obtain information on recall through the EEG method. They exposed a group of people to TV commercials, evaluating after a couple of days how many were recalled and how many were forgotten.

A similar experiment was devised by ^{Vecchiato et al. (2011b)}, who analyzed changes in brain activity during the observation of video clips, with the objective of observing asymmetry in frontal EEG activity with respect to the distribution of information from pleasant and unpleasant commercials. Later, ^{Vecchiato et al. (2014b)} proposed a new study in which they administered two tests to a group of people during the observation of television commercials. In both tests, the researchers used three neuronal measurement tools: EEG, GSR and HR; in the first test the information presented pleasant stimuli, while the second test showed results of emotional responses. They concluded that this technology could be useful for marketing specialists. ^{Guixeres et al. (2017)} carried out an experiment to evaluate the neurophysiological responses (variability of heart activity, neuronal responses and eye tracking) of thirty-five participants who watched eight commercials. The results obtained can be applied in the creation, improvement, and development of digital advertising content.

Gupta et al. (2016) performed an analysis of different EEG tests and suggested a method different from the traditional one, which they consider more effective in the evaluation of three emotions: valence, dominance, and excitement. These authors also presented a subjective classification of taste and found a better characterization of human active states during the observation of a video clip.

Additionally, ^{Kim et al. (2016)} proposed the use of different tools to measure physiological changes in consumers. These authors evaluated changes in oxyhemoglobin concentrations by means of an NIR spectroscopy system. Based on a set of Public Service Announcements (PSA), ^{Cartocci et al. (2017)} developed a seventeen-year study in the USA and Europe on anti-smoking advertising campaigns, proposing a method based on EEG, GSR and HR tests to evaluate the effectiveness of these campaigns. Later, ^{Borghini et al. (2017)} conducted a study along similar lines in order to bring the analysis of mental and emotional states closer to reality, moving from BCI (brain/computer interface) technology to passive BCI.

4.10 Perspective 3: consumer behavior

This perspective analyzes the emergence of neuromarketing as a complementary technique to traditional marketing, with the objective of understanding consumer behavior. Some authors (^{Telpaz, Webb, & Levy, 2015}; ^{Plassmann et al., 2015}) consider that although traditional methods have limitations in understanding consumers, this does not mean that some methods should replace others in the case of those used by neuroscience; on the contrary, they can complement each other.

^{Yoon et al. (2012)} indicate that neuroscience will enable the shaping of new models to explain how consumers make decisions, involving variables that are often not considered, for example physiological (hunger, stress, social influence) and neural. Meanwhile, ^{Plassmann et al. (2015)} propose that the usefulness of consumer neuroscience research (applied in the field of marketing) will be more widely accepted when it provides valid and reliable information that cannot be obtained using traditional methods to understand consumer behavior. Finally, based on previous studies, ^{Hsu (2017)} states that the methods applied in neuroscience allow for a deeper approach to thoughts, feelings and intentions, something that traditional methods cannot do. However, he agrees with the complementarity between both methods rather than with substitution.

Consumer neuroscience also allows for the evaluation of advertising content, such as a study by ^{Utkutug- Bozoklu and Alki´bay (2016)} in which, through an experiment, they analyzed the effectiveness of the strategies used and their impact on consumer memory. They confirmed that elements such as music, humor and emotion are specific to advertisements that are key to creating advertising memory.

^{Braeutigam, Lee and Senior (2019)} propose that consumer neuroscience should move away from the reactive view they currently offer of consumer behavior, as this approach presents a partial understanding of the functioning of the brain which is somewhat removed from reality. It should seek a more dynamic view of neuronal behavior as brain activity is more spontaneous and not always the result of stimuli (^{Lee et al., 2018}). Current methods denote a more marked inclination towards non-conventional procedures to make more in-depth evaluations, therefore, it is worth investing time in more rigorous research to strengthen knowledge in this area.

As stated in the previous paragraphs, one of the most relevant avenues of research in consumer neuroscience is the analysis of consumer behavior. However, it is clear that it still requires extensive exploration to allow for the construction of theory on the subject. As ^{Yoon et al. (2012, p.485)} indicate "There is no magic: one cannot look inside the head of a decision maker and predict the individual selection of toothpaste or tomorrow's visit to the grocery store."