Elsevier

Appetite

Volume 147, 1 April 2020, 104538
Appetite

An event-related potential study of consumers’ responses to food bundles

https://doi.org/10.1016/j.appet.2019.104538Get rights and content

Abstract

We conducted two event-related potentials (ERP) experiments to investigate consumers' responses to different types of food bundles. In Experiment 1, the participants were instructed to indicate their wanting of a three-yogurt bundle when their neural activity was recorded. The results of self-report wanting scores revealed that the participants wanted bundles consisting of their favorite yogurt products more than those of disliked products. Such a difference in self-report scores was also indexed by the N2 in frontal brain and the P1 in the left hemisphere. By contrast, bundles consisting of three different yogurt products elicited a smaller amplitude of the N2 than bundles consisting of two favorite products and one disliked product, but these two types of bundles received comparable wanting scores. Moreover, we asked the participants in Experiment 2 to perform a visual discrimination task on these bundles, and did not found these effects on the N2 or the P1. Collectively, these results revealed neural activities underlying consumers' responses to food rewards, and demonstrated the role of individuals’ variety-seeking tendency in wanting process.

Introduction

Bundling is a prevalent promotion strategy, and many factors have been shown to influence consumers' responses to product bundles (Hansen & Martin, 1987; Janiszewski & Cunha, 2004). For example, consumers prefer a bundle more when a separate price is presented for each component than when a single price is presented for the whole bundle (Chakravarti, Krish, Paul, & Srivastava, 2002). Making a selection among different bundles is different from choosing one single product. When choosing one single product, consumers may switch to less-preferred options even when they obtained less pleasure from switching than from repeating more-preferred options (Ratner, Kahn, & Kahneman, 1999), thus demonstrating a variety-seeking tendency (McAlister, 1982; van Trijp, Hoyer, & Inman, 1996). When choosing a product bundle, consumers are less likely to seek variety if they consider these products as a bundle rather than parallel single items (Mittelman, Andrade, Chattopadhyap, & Brendl, 2014). Manufacturers and retailers like to provide food bundles with variety (Howell, 2000), but very few studies have been conducted to systematically examine the component variety of these bundles and consumers’ responses to food rewards, such as their wanting for these foods.

Food wanting refers to people's disposition to eat or appetite, and can be triggered by mere exposure to a food or even imagination of this food's taste, smell, and texture (Pelchat, Johnson, Chan, Valdez, & Ragland, 2004). Food wanting has both explicit and implicit forms (Berridge, 2009). The explicit component of food wanting can be directly measured via self-report rating scales, whereas the implicit component of food wanting has been indirectly assessed via participants' performance in a variety of behavioral tasks, including the approach-avoidance task, the implicit association test, and the affective Simon task etc. (for a review, Tibboel, De Houwer, & van Bockstaele, 2015). Kraus and Piqueras-Fiszman (2016) have shown that direct and indirect measures of food wanting can provide consistent findings. However, indirect measures are often assumed to be superior to direct measures, presumably because they can capture unconscious processes and are less susceptible to social desirability concerns (Tibboel et al., 2015; Wiers & Stacy, 2006).

However, it may be difficult to make sure whether the indirect measures are valid. For example, Finlayson, King, and Blundell (2007) used the frequency of a food being chosen in a forced-choice task as an index for food wanting; whereas Finlayson, King, and Blundell (2008) proposed that the frequency of being chosen in a forced-choice task might assess a combination of liking and wanting, and used the participants' reaction times of making a choice to index implicit wanting. Piqueras-Fiszman, Kraus, and Spence (2014) asked their participants to complete an approach-avoidance task in which they pulled or pushed a joystick upon seeing food images, and used their reaction times to index implicit wanting. They obtained dissociations between the participants' behavioral responses in the approach-avoidance task and self-report ratings, but such indirect measure only appears to be more sensitive than the direct measure in testing participants’ disgust and rejection responses.

In contrast, a potentially better way to test consumers’ responses to food rewards may be to link them to food-related event-related potentials (ERPs; Carbine et al., 2018; Lin, Cross, Jones, & Childers, 2018). Characterized by a good time resolution, the ERP technology can detect rapid changes that may not be easily captured by behavioral tasks or self-report ratings (Plassmann & Karmarkar, 2015). In recent years, the ERP technique has been increasingly used to study food-related cognitive processes, including attention (Schwab, Giraldo, Spiegl, & Schienle, 2017), inhibitory control (Lapenta, Dierve, de Macedo, Fregni, & Boggio, 2014), and working memory (Rutters, Kumar, Higgs, & Humphreys, 2015).

Notably, Telpaz, Webb, and Levy (2015) had their participants view pictures of single products during the electroencephalography (EEG) recording session, and linked the amplitude of the N2 to the binary choices that the participants subsequently made between pairs of these products. Their results revealed that the N2 amplitude was smaller for the products that participants chose more often and liked more, compared to the unselected and disliked products. Considering that the chosen frequency in such a forced-choice task has been linked to wanting (Finlayson et al., 2007, 2008), Telpaz et al.’s (2015) findings suggested that the N2 may be used to index wanting. Moreover, Goto et al. (2017) recorded their participants' scalp EEGs while they were rating the pleasantness of and wanting for products in a virtual shopping task. Specifically, their participants were instructed to look at images of products while avoiding eye movements, and then rated their liking for and wanting of these products. Their results also revealed that the N2 amplitude was smaller for the products with high wanting scores than those having low or medium wanting scores.

Therefore, the present study was conducted to assess consumers’ responses to food rewards, and we chose the N2 peaking approximately at 160–250 ms post-stimulus as an ERP of interests. It should be noted that food wanting and liking are distinctive from each other (Stevenson, Francis, Attuquayefio, & Ockert, 2017), but it is challenging to dissociate them in laboratory-based studies because these two psychological constructs are often intertwined to each other (Havermans, 2011). Moreover, we also chose the P1 peaking approximately at 80–140 ms post-stimulus as a second ERP of interest, as previous research has shown that the P1 is sensitive to the influence of motivation (Cunningham, van Bavel, Arbuckle, Packer, & Waggoner, 2012; Hammerschmidt, Sennhenn-Reulen, & Schacht, 2017). Importantly, reward-related stimuli can elicit a larger P1 (Becker, Flaisch, Renner, & Schupp, 2016; MacLean & Giesbrecht, 2015).

In the present study, we presented fruit-flavored yogurt products in the form of three-product bundles as experimental stimuli, and manipulated the level of variety of each bundle. Menon and Kahn (1995) defined a “switch” as occurring when a chosen product was different from any of the previous chosen items in sequential choices. Similarly, we used the total number of switches within each three-yogurt bundle to index the variety of this bundle, and created four different types of bundles with varied levels of variety: (1) 3F bundles consisting of three cups of the same yogurt which was a participant's favorite flavor (number of switch = 0), (2) 3D bundles consisting of three cups of yogurt of the same flavor that the participant disliked (number of switch = 0), (3) 2F1D bundles consisting of two cups of yogurt of the same favorite flavor and one cup of yogurt of the disliked flavor (number of switch = 2), and (4) 3M bundles consisting of three different cups of yogurt in flavors with a medium level of popularity across participants (number of switch = 3).

In Experiment 1, we asked our participants to rate their wanting for these four types of bundles during the EEG recording session. First, we hypothesized that the 3F would receive higher wanting scores than the 3D bundles, so we expected to observe significant differences in the N2 and the P1 elicited by these two types of bundles. Specifically, we hypothesized that the N2 amplitude would be smaller for the 3F bundles than for the 3D bundles, and the P1 amplitude would be larger for the 3F bundles than for the 3D bundles. Second, we hypothesized that the 3M and 2F1D bundles would receive comparable wanting scores, presumably because both of these two types of bundles had pros and cons. That is, the 3M bundles provided the highest level of component variety that a three-product bundle could provide, but they did not contain the participants' favorite product; whereas the 2F1D bundles contained the participants' favorite products, but they also contained a disliked product and only provided a medium level of component variety. Therefore, comparing the N2 and the P1 elicited by the 3M and 2F1D bundles might reveal differences in consumers’ responses to these two types of food rewards that might not be easily detected by self-report ratings.

In order to rule out the confounding of visual differences between different types of bundles, we also conducted Experiment 2 as a control experiment in which the participants only performed a discrimination task on the bundles. Collectively, the findings of this study can provide more empirical evidence about the associations between the N2 and food wanting, and extend the scope from wanting for a single product to wanting for a bundle of products. Our findings about the P1 would also provide novel findings to link wanting with an earlier ERP than the N2. Moreover, our findings may demonstrate certain associations and/or dissociations between neural activities underlying the explicit and implicit components of food wanting.

Section snippets

Participants

Twenty-five Chinese participants (21.05 ± 1.57 years, ranging from 19 to 24 years; 11 females) took part in this experiment. We used the G*Power software to estimate the sample size, and the results revealed that a sample of 25 participants can detect the effects with ηp2 ≥ 0.24 (statistic power = 0.80).

In the present and the following experiments, all participants reported that they were right-handed, had normal or corrected-to-normal vision, and had no history of neurological or psychiatric

Methods

Twenty-five new Chinese participants (23.2 ± 2.48 years, ranging from 19 to 27 years; 16 females) took part in this experiment. All aspects of the methods in this experiment were the same as those in Experiment 1 except for the following differences. First, we randomly chose one flavor from the strawberry and yellow peach flavors and one flavor from seaberry and water chestnut flavors to present in the 3F and 3D bundles, respectively. Second, the participants were asked to perform a

General discussion

In summary, two major findings emerged from the present study. First, our results associated the N2 with wanting for food bundles, though this wanting effect was intertwined with liking (also see Goto et al., 2017; Telpaz et al., 2015). This result not only extended the scope from wanting for a single product to wanting for a bundle of multiple products, but also specifically linked people's wanting for food bundles to a smaller amplitude of the N2 in frontal region of the brain. Our results

Declaration of competing interest

None.

Acknowledgements

This research was supported by the National Natural Science Foundation of China (Grant No. 71872097 & 71472106) awarded to Xiaoang Wan. The authors would like to thank Prof. Shimin Fu for generously allowing the authors to use his lab, and Ms. Yihang Ouyang for her assistance in the preparation of the experimental stimuli.

References (43)

  • W.J. Jones et al.

    The shopping brain: Math anxiety modulates brain responses to buying decisions

    Biological Psychology

    (2012)
  • B.E. Kahn

    Consumer variety-seeking among goods and services: An integrative review

    Journal of Retailing and Consumer Services

    (1995)
  • N. Koranyi et al.

    Dissociating implicit wanting from implicit liking: Development and validation of the wanting implicit association test (W-IAT)

    Journal of Behavior Therapy and Experimental Psychiatry

    (2017)
  • A.A. Kraus et al.

    Sandwich or sweets? An assessment of two novel implicit association tasks to capture dynamic motivational tendencies and stable evaluations towards foods

    Food Quality and Preference

    (2016)
  • O.M. Lapenta et al.

    Transcranial direct current stimulation modulated ERP-indexed inhibitory control and reduces food consumption

    Appetite

    (2014)
  • M.H. MacLean et al.

    Neural evidence reveals the rapid effects of reward history on selective attention

    Brain Research

    (2015)
  • M.L. Pelchat et al.

    Images of desire: Food craving activation during fMRI

    NeuroImage

    (2004)
  • B. Piqueras-Fiszman et al.

    Yummy” versus ‘‘Yucky”! Explicit and implicit approach-avoidance motivations toward appealing and disgusting foods

    Appetite

    (2014)
  • D. Schwab et al.

    Disgust evoked by strong wormwood bitterness influences the processing of visual food cues in women: An ERP study

    Appetite

    (2017)
  • R.J. Stevenson et al.

    Explicit wanting and liking for palatable snacks are differentially affected by change in physiological state, and differentially related to salivation and hunger

    Physiology & Behavior

    (2017)
  • H. Tibboel et al.

    Implicit measures of “wanting” and “liking” in humans

    Neuroscience & Biobehavioral Reviews

    (2015)
  • Cited by (6)

    • Emerging biometric methodologies for human behaviour measurement in applied sensory and consumer science

      2023, Digital Sensory Science: Applications in New Product Development
    • Influence of the crossmodal congruency between color and flavor on product evaluations: Evidence from behavioral and oscillatory brain responses

      2022, Food Quality and Preference
      Citation Excerpt :

      In order to motivate the participants to actually process this image of the colored drink, we then asked them to rate to which extent they wanted to consume this beverage. Similar to Qiu, Qi, and Wan (2020) study, we asked our participants to use a mouse to choose a number from 0 to 100 in 7 s, with a greater number indicating a higher degree of wanting. After another 1-second fixation screen, 2 ml colored drink was delivered to the participant’s mouth during the next 4 s, followed by another 1-second fixation screen.

    • Influence of flavor variety and color variety on consumer preference and choices of yogurt bundles

      2022, Food Quality and Preference
      Citation Excerpt :

      As also shown in Fig. 1, we created three types of three-yogurt bundles, including the 2P1N, 1P2DN, and 3DN bundles. The three products were aligned horizontally, so a “switch” was defined when a yogurt had a different flavor from any of the items on its left side (Qiu et al., 2020), so the 2P1N, 1P2DN, and 3DN bundles had 2, 3, and 3 color switches, respectively. The popular flavor was always placed on the left, and the locations of new flavors were randomly determined.

    • A Study for ERP Classification of Food Preference Based on CSP and SVM

      2022, Chinese Journal of Biomedical Engineering
    View full text