1. Introduction
Generally, branded products with brand equity are traded at premium prices, compared with unbranded products [
1]. As for brand equity, it is one of the crucial profitable sources for enterprises and one of their greatest assets. Aaker [
2] segregated the elements of brand equity into five factors: brand name awareness, brand associations, brand loyalty, perceived brand quality, and other proprietary brand assets, such as patents, trademarks, and channel relationships. Thus, unlike factories and office buildings, brand equity is an asset attributed to consumers’ minds.
Numerous brand equity studies have been conducted over the past two decades in order to clarify how to build brand equity in consumers’ minds. Such research included empirical studies, theoretical studies, and practical cases. Meanwhile, the number of neuroscience studies to understand brand equity-related mental processes has been increasing. For example, McClure et al. [
3] showed that brain activations on beverage products with brand equity were observed in the hippocampus (HP) and dorsolateral prefrontal cortex (DLPFC), whereas both the ventral medial prefrontal cortex (VMPFC) and ventral striatum (VS) were activated in low-brand equity products. However, the activations of both the VMPFC and VS were observed in several brand equity studies [
4,
5,
6,
7,
8]. These regions are known as the “neural currency network” [
9]. The activations in these regions were also reported in several studies on unbranded objects [
10,
11,
12].
Barring these regions, the activations in the medial prefrontal cortex (MPFC) were observed in both branded and unbranded research. For example, in a comparison between familiar automobile brands and unfamiliar ones, the MPFC was activated [
13]. In related studies, Schaefer and Rotte [
14] confirmed the activations in the MPFC when comparing luxury automobile brands and unfamiliar brands, whereas Chen et al. [
15] investigated the brain activations associated with brand personality, which is an element of brand association composed of brand equity. In the latter study, they reported the activations in the MPFC, the cingulate cortex, and the caudate. Meanwhile, the activations in the MPFC were reported in numerous studies on unbranded objects [
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27]. Besides these brain regions, other brain activated regions have been found in studies on branded objects with brand equity, e.g., the insula [
28], the inferior frontal gyrus [
29], and the superior frontal gyrus [
30,
31]. It has also been reported that these brain regions are solely or multi-regionally activated.
Based on the aforementioned findings, brand equity-related brain regions are highly diversified. Even though brand equity has an influence on consumers’ decision-making, such as purchases, preferences, and attitudes, the watershed brain regions between brand equity and unbranded-related brain regions remain unknown. Therefore, the purpose of this study is to assess the unique characteristics of the mental processes associated with brand equity by identifying the watershed brain regions through a comparison between the brain regions related to brand equity and such regions related to unbranded objects without brand equity.
4. Discussion
Although the characteristic brain regions were not observed regarding the unbranded-related brain regions through the three assessments (i.e., the ALE method, the statistical hypothesis test, and machine learning), this study revealed that the brain regions around both the PHG and the left LG were characteristic brain regions to brand equity and anatomically close to one another. In the right LG, two assessment methods (the statistical hypothesis test and machine learning) were passed. Accordingly, the PHG and LG can be thought of as watershed brain regions for distinguishing mental processes of branded and unbranded objects. Therefore, when metabolic alternations in these regions are observed in a magnetic resonance spectroscopy (MRS) research hereafter, it will imply that the clustered brain regions around both the PHG and LG can be biomarkers for whether brand equity has been established in consumers’ minds. Specifically, the PHG, which corresponds to the centroid of Cluster 9, is associated with recognition [
45], episodic memory [
46,
47,
48], and visual and spatial scene processing and navigation [
49,
50]. In the function of recognition, the anterior part of the PHG is engaged in familiarity-based recognition [
51,
52]. Meanwhile, the posterior part of the PHG is engaged in recollection-based recognition [
53,
54]. During recollection, the activations of both the PHG and the posterior parts of the PHG (or single activations of hippocampus) were observed in many cases [
45]. In addition, the PHG has a tendency to activate in association with various elements, such as memory sources, and remembering targets when engaging functions of episodic memory [
46,
47,
48]. Thus, episodic memory engaged with the PHG can be thought of as “associative memory” or what Aminoff et al. [
55] described as “contextual association”. As for the LG, it is associated with mental imagery [
56], visual and spatial scene processing and navigation [
49], episodic memory [
57], divergent thinking [
58], predictive inferences [
59], and recognition [
60]. These functions, in which the LG is associated in elements of visual processing are required. For example, when generating predictive inferences or performing divergent thinking, visual images must be internally generated. Moreover, the LG also plays a crucial role in language processing, such as in visual recognition of words [
61,
62] and semantic processing of words [
63,
64,
65]. According to Zhang et al. [
65], the LG is involved in language processing and supramodal organization in patients who are not congenitally blind but lost their sight in their early teens. Musz and Thompson [
66] demonstrated that the LG plays a role in the semantic hub across the modalities of words. Thus, considering that consumers may recognize a brand as a type of word, the LG is believed to serve as a link connecting modalities and meanings of a brand. Interestingly, these regions are associated with the default mode network (DMN) [
67,
68]. The PHG is the core region of the DMN [
69], and the LG has functional connection with brain regions constituting the DMN [
68]. Given that the DMN is engaged in self-referential processing (e.g., episodic memory, autobiographical memory) [
67] and associative memory-based autopilot behavior [
70], mental processing of branded objects can be thought of as automated mental processes based on associative memories and effortless decision-making based on these mental processes.
Meanwhile, as described earlier, consistent results from the three assessments (i.e., the ALE method, statistical hypothesis test, and machine learning) were not observed in unbranded objects relative to that in branded ones. However, the IPL (BA40) was commonly observed as characteristic brain regions via two assessments (the ALE method and statistical hypothesis test). The IPL associates with calculation [
71,
72] and decision-making under uncertainty [
73,
74,
75]. Interestingly, connections between the IPL and insula were recorded in metacognition under uncertainty [
75]. The insula plays a crucial role in monitoring situations when decision-making under uncertainty. The IPL is involved in controlling and managing mental resources for problem solving under uncertainty. The insula was the brain region revealed in the assessment by the ALE method. In the consumer contexts, the insula detects and evaluates the social risks in purchase decision-making [
23]. Besides these regions, the medial frontal gyrus was revealed by the ALE method. The machine learning approach demonstrated that the superior frontal gyrus (Cluster 25) has an influence on unbranded flags. These brain regions are so close that they are placed in a dorsal and medial part of the prefrontal cortex (hereafter, the dorsomedial prefrontal cortex). The dorsomedial prefrontal cortex (DMPFC) is associated with action control, conflict monitoring [
76,
77], and decision uncertainty [
78]. The DMPFC performs these cognitive control-related functions by connecting with the executive control network [
78,
79]. Additionally, the DMPFC associates with the DMN and is involved in social cognition through a connection with brain regions of the DMN [
69]. The DMPFC plays a role in inferring others’ thoughts in complex social relationships [
80]. In this way, this region is engaged in organizing and adjusting information to solve problems in complex situations, such as a preference on options with equal values, and unstable situations [
81]. Thus, in mental processes of unbranded objects, cognitive control and deliberative aspects may be dominated to handle unknown objects, such as unbranded products and services. In other words, it can be presumed that consumers carefully behave while purchasing unbranded objects.
Cognitive decoding in Neurosynth (
https://neurosynth.org/, accessed on 7 September 2021) was also conducted to more rigorously decode the functions of these clustered brain regions. The decoding analysis was performed for the results of branded objects and unbranded objects. Additionally, the region of interest (ROI) was determined via the Mango software (Version 4.1;
http://ric.uthscsa.edu/mango/, accessed on 1 April 2021). In this regard, three ROIs (Cluster 9, 15, and 20) were established, and the shape of each ROI was set as a cube in branded objects. In unbranded objects, two ROIs (Cluster 8, and 25) were established. The length, width, and height in each cube were determined in accordance with the standard deviations of the coordinates in each cluster. Concretely, each measurement of the cube was set at 18 mm. The calculation procedures are as follows. First, the standard deviations (1 sigma) of each coordinate (x, y and z) in each cluster (Clusters 8, 9, 15, 20, and 25) were calculated. Second, the maximum and minimum values of the coordinates in each cluster were determined. For example, in Cluster 9, the maximum value of the x coordinate was determined by calculating 30 (x; centroid of Cluster 9) plus 13 (1 sigma of the x coordinate), while the minimum value of the x coordinate was determined by calculating 30 (x; centroid of Cluster 9) minus 13 (1 sigma of the x coordinate). As for the ranges of the ROI of Cluster 9, the x coordinate ranged from 17 to 43, the y coordinate ranged from −16 to 6, and the z coordinate ranged from −22 to −1. Regarding the ranges of the ROI of Cluster 15, the x coordinate ranged from −24 to −6, the y coordinate ranged from −96 to −77, and the z coordinate ranged from −11 to 11. Regarding the ranges of the ROI of Cluster 20, the x coordinate ranged from 10 to 31, the y coordinate ranged from −94 to −77, and the z coordinate ranged from −4 to 15. Third, each measurement of the cube was adjusted in accordance with these ranges calculated in the second step using the Colin27-T1 template in the Mango software. It was determined that 18 mm was appropriate for the measurement of the cube. Finally, these three ROIs were united into a single ROI (see
Figure 7A) in branded objects. Similarly, the two ROIs were united into a single ROI (see
Figure 7C) in unbranded objects. After determining the ROI, they were registered in the Neurovault database (
https://neurovault.org, accessed on 1 April 2021) for decoding. Subsequently, cognitive decoding was performed for the ROI through the Neurovault database, which is internally connected with Neurosynth. The results of the decoding are shown in
Table 8 and
Figure 7B,D. In this case, we adopted the top 40 terms, excluding both anatomical terms, disease and experimental task-related terms. The word cloud was created using Python. The higher the correlation values a term had, the larger the font size was set, and vice versa. Accordingly, the font size in the word cloud of branded objects is larger than that of unbranded objects as correlation values in decoded results of branded objects are relatively larger than those in decoded results of unbranded objects. The colors were randomly allocated to each term. In branded objects, the results show that both memory- and emotion-related terms are primarily dominant. Especially, emotion-related terms were ranked as the top 10 terms. This indicates that the emotional episodic memories of objects in consumers’ minds play a crucial role in differentiating between branded objects and unbranded objects. In contrast, in unbranded objects, many executive control-related terms (e.g., “competing”, “judgment”, “reasoning”, “switching”, “control network”, “conflict”, “executive control”, “cognitively”, and “monitoring”) were ranked. Besides this term category, language-related terms (“fluency”, “verbal fluency”, “lexical decision”) and social cognition related terms (“pain”, “default network”, “empathy”) were ranked. Although the decoded terms of unbranded object-related brain regions were not converged into specific categories as were the results of brain regions related to branded objects, the executive control-related terms were characteristic in the decoded results of unbranded objects’ related brain regions.
Overall, the findings of this study are consistent with previous theoretical and empirical brand equity studies. Specifically, the emotional and positive experiences of consumers influence their attitudes toward brands [
82], and vice versa. Similarly, it has been revealed that emotional aspects influence value-based decision-making in neuroeconomics and neurofinance studies [
83]. These emotional experiences are stored in consumers’ minds along with multimodal sensory information [
84]. In addition, the link between emotional episodes and brands help form brand associations, which is one of the crucial elements in brand equity [
2]. Hence, a strong brand association is created by episodic memories that are based on emotional experiences [
82,
85]. In collaboration, this study indicates that the PHG may be involved in emotional aspects of brand associations and the LG may function as a semantic hub connecting various multimodal elements of brand associations. Meanwhile, given that terms related to the executive control network were decoded in analysis of the IPL and the DMPFC, it is presumed that making decisions about unbranded objects may be effortfully executed based on rational mental processes. Therefore, regarding mental processes of branded objects, emotional aspects may be relatively dominant in decision-making. In contrast, cognitive and deliberative aspects may be relatively dominant in mental processes of unbranded objects.
The results of this study also provide useful implications for practitioners. First, when launching a new brand, managers should prioritize the creation of emotional brand associations, aside from other marketing practices. In this regard, they should carefully observe the emotional brand associations and related scores like “familiarity”, in addition to other indices, for tracking brand equity and managing an established brand. Second, when conducting qualitative research, such as in-depth interviews and focus groups, researchers should focus on eliciting emotional episodes on a brand from consumers. In this case, episodes that are visually vivid, spatially concrete, and positively presented can be core factors that strengthen brand associations.
Although the present study provided useful findings to both academicians and practitioners, there are several limitations that should be noted. First, the analyses were conducted using data with stimuli from B2C products and services. In other words, the data in this study included cross product and services data among B2C categories. Depending on these categories, it is possible that different results can be obtained when using data that focus on a specific product/category. In unbranded objects, inconsistent results among the three assessments may be induced owing to these reasons. Further, research on a specific product/category is required in near future. In addition, controlling the attributes and facets in both branded and unbranded objects will be required to overcome the inconsistencies of results for unbranded objects during analysis. Second, the analyses were conducted without considering the heterogeneous sample profiles such as sex, age, occupations, personalities, attitudes toward a brand, and brand usages. In marketing, segmented groups of consumers play a crucial role in setting a strategy and evaluating an outcome. However, in this study, both demographic and psychographic factors were not considered in the analyses. Consequently, it is possible that different results can be derived from these factors. Finally, regarding the analysis by machine learning, it is possible that the performance of the model was improved by conducting the more precise feature engineering. For example, the latter approach added other variables such as raw coordinate data, a task factor, and product categories. Therefore, the results of this study should be carefully interpreted before drawing any conclusions. In this way, although the study has several limitations with this approach, this is the first study in which the watershed brain regions between the branded and unbranded objects were comprehensively revealed based on the enormous brain regions that activated imaging data. However, additional work is required for more precisely identifying a neural mechanism of brand equity and mental processes in it.