Neural correlates of product attachment to cosmetics

The neurobiological basis of brand and product attachment has received much attention in consumer neuroscience research, although it remains unclear. In this study, we conducted functional MRI experiments involving female users of famous luxury brand cosmetics as participants, based on the regions of interest involved in human attachment and object attachment. The results showed that the left ventral pallidum (VP), which is involved in positive reward, and the right posterior cingulate cortex (PCC), which is involved in self-concept, a key concept in object attachment, are the core regions in cosmetic attachment. Moreover, the performed psychophysiological interaction analyses showed that VP-temporoparietal junction connectivity positively correlated with activity in the dorsal raphe nucleus, and PCC–anterior hippocampus (aHC) connectivity positively correlated with subjective evaluation of attachment. The former suggests that object attachment is a human-like attachment and a stronger tendency of anthropomorphism is associated with stronger feelings of security. The latter suggests that the individual’s concept of attachment as well as the relationships with the attached cosmetics are represented in the aHC, and the PCC–aHC associations produce subjective awareness of the attachment relationships. These associations between memory and reward systems have been shown to play critical roles in cosmetic attachment.

. Summary of the previous fMRI studies of human attachment relationships.

fMRI study
Numbers and age of subjects
In the visual with tactile session, the ROI analyses showed that the right PCC (Montreal Neurological Institute (MNI) coordinates: 3, − 22, 29; family-wise error rate [pFWE] = 0.018 18,20 30 ) were significantly activated in the attached cosmetic compared to that in the non-attached cosmetics, as shown in Table 2. In contrast, there were no significant differences in activity during the visual session. The ANOVA showed that there were significant main effects of attachment in the right PCC (6, − 16, 32; F = 10.443, p = 0.004), left PCC (− 4, − 22, 32; F = 5.791, p = 0.026), and left VP (− 9, 0, 6; F = 9.10; p = 0.007), all of which showed significantly greater activity in the attached than in non-attached cosmetics (Fig. 2). In addition, the PAG (3, − 35, − 28) was found to have a significant interactive effect between attachment and touch (F = 5.6; p = 0.029), and also had significantly greater activity in the attached than in the non-attached cosmetics, in the visual with tactile session (df = 9, t = 2.7; p = 0.014 < 0.025 = 0.05/2) (Fig. 2), while there was no significant difference in the visual session.
The   (Table 3). There was no significant

Discussion
Analysis of the subjective evaluations showed that the score of "attachment" was significantly higher for the attached cosmetics than for the non-attached ones, and it was confirmed that our participants felt a stronger attachment to their attached cosmetics than to the others. In addition, positive emotion or motivation including "security" was significantly higher for the attached cosmetics than for the non-attached ones. As for the brain regions involved in attachment to the cosmetics, the left VP and the left and right PCC were significantly activated in the main effect of attachment. Specifically, the left VP and right PCC showed significant functional connectivity to several other brain regions that are important for object attachment. These results show that these two regions play a major role in cosmetic attachment. Each of the VP and PCC is well known to be a core region in the reward system 46 and self-referential processing system 47 , respectively. The VP is a central convergent region for input from the orbitofrontal, prefrontal, and infralimbic cortex, amygdala, lateral hypothalamus, VTA, parabrachial nucleus, subthalamic nucleus, and other structures related to reward 48 . Conversely, the VP projects back to nearly  www.nature.com/scientificreports/ all of its input sources, including the nucleus accumbens for reciprocal information exchange 49 . Based on such limbic-related anatomical connectivity, the VP mediates reward and motivation functions at many levels in the brain 50,51 , such as social affiliation and pair bonding [52][53][54][55] . Indeed, activity in the VP is well known to be linked with pair bonding and attachment behaviors in monogamous prairie voles 25 . Moreover, securely attached children show greater VP activation than children with poor attachment 27 . In addition, the VP shows significant activities in both maternal love 28 and romantic love 29 . Fisher et al. 56 speculated that human activity in this region is likely related to feelings of attachment. Furthermore, a recent neuroimaging study related to decision making in hoarding disorders showed that the VP is activated in association with attachment to object possessions 30 . These findings and our present results show that the VP plays an important role not only in attachment in individuals, but also in object attachment, as we hypothesized. In addition, the PCC is the core region not only in the selfreferential processing system 32 , but also in the processing systems of autobiographical memory 57 and personal semantic memory (PSM) 47,58 . PSM is an intermediate entity between semantic and episodic memory 59 . Moreover, the PCC is well known as the strongest hub, with the highest number of functional connections 60 , and the core region of the default mode network (DMN) which is involved in self-referential processing 57 . Furthermore, this region plays a central role in supporting internally directed attention and cognition 60 . Based on these facts, the PCC region may be highly involved in self-concept, which is a key concept of object attachment 15,16 . Moreover,  www.nature.com/scientificreports/ the PCC plays a critical role for craving 61 , and its activity represents the situations that the individual is "caught up" or "attached to" his/her experiences 62 . In addition, it has been recently shown that a substantial part of the serotonergic influence on this core region in the DMN is mediated by different 5-HT1A binding sites 63 . This fact suggests the relationship between the PCC activity and the subjective feelings of "security" which is one of the most critical emotions in brand/product attachment [18][19][20] . Moreover, the left VP showed significant connectivity to several regions in the right PCC, and the left TPJ, and the PCC showed significant connectivity to the anterior HC and the RSC in the main effects of attachment. All these brain regions are included in neural networks for processing autobiographical and personal semantic memories 64 . Personal semantics are autobiographical knowledge or information extracted from repeated autobiographical events and is thought to be an intermediate entity between semantic memory and episodic memory 64 . Accordingly, such memory processing-related networks are considered to play a critical role in cosmetic attachment. Furthermore, the VP-PCC connectivity suggests that the association between neural networks for processing reward-related information and memory-and self-related information is the core mechanism for cosmetic attachment. Interestingly, such memory and reward associations have also been shown to be core mechanisms in nostalgic experiences 65,66 . Moreover, it has been shown that greater nostalgic connection between customers and products link to greater intensity of product attachment in marketing literature 13,[67][68][69][70] . In addition, there was a significant positive connectivity between the left VP and left TPJ in the main effect of attachment, and its individual strength was positively correlated with individual DRN activity (Fig. 4). The TPJ is well known to be involved in social cognition and mentalizing 71 . In particular, the left TPJ is involved in processing different subjective perspectives 72 . This suggests that the participant might view her attached cosmetic as possessing human characteristics. Indeed, it has been shown that there is an association between the structure of the left TPJ and anthropomorphism 35 . Fournier 14 claimed that individuals experience little difficulty in assigning personality features to brands, and it has been shown that consumers easily view brands as possessing human characteristics 21,22 . Moreover, Kwok et al. 23 showed that a stronger anthropomorphism tendency was associated with enhanced perception of the objects' sentimental and instrumental value, and this enhanced value mediated the relationship between anthropomorphism tendency and object attachment. In the present study, it was shown that a stronger anthropomorphism tendency represented in the left TPJ was associated with a higher reward value of the attached cosmetic represented in the left VP. Therefore, anthropomorphism was shown to play a critical role in cosmetic attachment, as hypothesized. Furthermore, our present results showed that stronger connectivity between the left VP and TPJ regions induces greater activity in the DRN. This result suggests that individuals who showed stronger connectivity between these regions feel more secure feelings because the DRN is a serotonin-rich site. Furthermore, our previous studies 42,43 showed that DRN activity is negatively correlated with attachment-related anxiety. Therefore, it is suggested that individuals with less attachment-related anxiety, that is, those who have better attachment relationships with the objects, show stronger neural communication between the VP and the TPJ, and assign personality features to the object (anthropomorphism). This result showed that the left VP and the left TPJ might coproduce a sense of security, which is one of the critical factors in product/brand attachment [18][19][20] , as hypothesized. In addition, our results of the left TPJ-VP connectivity may be based on similar mechanisms in a recent neuroimaging study that showed a similar relationship between functional connectivity between the left TPJ and ventral striatum and the positive feelings such as happiness 67,68 .
Moreover, there was positive connectivity between the right PCC and the right anterior HC, and the individual strength of this connectivity positively correlated with the individual subjective score of "attachment" (Fig. 5). The anterior HC region is involved in categorical/conceptual representations, while the posterior region is associated with the recovery of fine-grain perceptual detail 73 . Accordingly, the anterior HC is considered to be involved in establishing one's concept of attachment relationships with cosmetics by combining multiple forms of information, such as semantic and emotional information 74 , or integrating distinct experiences on a conceptual scale 75 . In addition, genetic and pharmacological studies have revealed that OXT receptors in the anterior dentate www.nature.com/scientificreports/ gyrus and CA2/CA3 play a critical role in the discrimination of social stimuli, and OXT receptors in the anterior CA2/CA3 neurons recruit a population-based coding mechanism to mediate social stimuli discrimination 76 . It has been suggested that there are indexical (personalized) and affect-based representations of the brand as part of the consumer's self-concept in strong connections between customers and products. Thus, such representations are highly salient and automatically retrieved when the consumer activates his or her self-concept 16 . Such mental representations may be schemata that encode past experiences of the interactions with the attached object, and of its availability and capacity to respond to the user's needs. The present results suggest that such schemata are represented in the anterior HC and make it possible to discriminate one's attached cosmetic as the special entity for her from the other cosmetics. Furthermore, from the perspective of social memory, it might discriminate one's attached cosmetic, which is like the person close to her (anthropomorphism) from the others. Accordingly, the PCC-anterior HC connectivity suggests that the sensory stimulation cues induce the automatic retrieval of such mental representations in the anterior HC, and the individual re-experiences the autobiographical episodes associated with such representations, mediated via the PCC-anterior HC connectivity, and these neural processes lead to one's subjective awareness of attachment to the cosmetic. In addition, there was a significant interaction in the PAG activities, and the PAG showed greater activation for attached cosmetics in the visual with tactile session, while it showed a greater deactivation for non-attached cosmetics. This structure is heavily connected to various limbic regions and contains a high density of OXT receptors 77 . OXT is implicated in regulating positive social interactions, social bonding, and maternal responsiveness in several mammalian species, including humans 78 . It has been shown that intranasal delivery of synthetic OXT motivated pair-bonded men to maintain a larger social distance from an unknown female experimenter and inhibited approach toward attractive women 79 , suggesting that OXT is also important for the maintenance of an already established pair bond 80 . Accordingly, this interactive effect found in PAG activities may be explained by a similar mechanism of approaching or avoiding behavior, based on the effects of OXT on the PAG. In addition, the PAG receives oxytocinergic fibers from the supraoptic nucleus in the hypothalamus [81][82][83][84] . Our additional analysis showed significant functional connectivity between the PAG and supraoptic nucleus in the hypothalamus (Fig. 6). Moreover, upregulation of OXT expression in the hypothalamus is known to be activated by somatosensory stimulation, such as massage, which is mediated via the spinothalamic pathway 39 . The thinly myelinated or unmyelinated afferent fibers activated by touching or rubbing are carried by the contralateral spinothalamic pathway. These impulses are sent to the thalamus and then the primary somatosensory cortex. These impulses are further sent to other brain regions, including the PAG, hypothalamus, and brainstem, via collateral connections 85 . Accordingly, the interactive effect observed in the PAG activity in the present study may also be explained by such a synergetic effect between direct somatosensory stimulation and OXT release via hypothalamic activity. In addition, several ROI regions were significantly activated in the visual with tactile session. However, there was no significant activity in these regions during the visual session. These brain regions include the memory-, reward-, emotion-related regions, and brainstem regions such as the DRN, PAG, and SN/VTA, which are the primary sites of neurotransmitters that modulate the basic functions of survival, such as reward, motivation, emotion, and security 42,43 . These results may also highlight the importance of tactile information processing in attachment to cosmetics. Skin-to-skin contact is one of the earliest communication channels that promotes attachment between infants and caregivers 84 . The OXT release associated with tactile stimulation, via hypothalamic activity, was considered to have some effects on these several brain regions involved in object attachment.
In the present study, the results showed that the left VP involved in positive rewards and the right PCC involved in the self-concept which is a key concept in object attachment, are the core regions in cosmetic attachment. Moreover, the associations between the reward (VP) and memory (PCC) systems were shown to play critical roles in cosmetic attachment. Furthermore, the PPI analyses showed that the VP-TPJ connectivity positively correlated with activity in the DRN, and the PCC-anterior HC connectivity positively correlated with subjective evaluation of attachment. The former suggests that object attachment is a human-like attachment and a stronger tendency of anthropomorphism is associated with stronger feelings of security. The latter suggests that the individual's concept of attachment as well as the relationships with the attached cosmetics is represented in the aHC, and the PCC-aHC associations produce the subjective awareness of the attachment relationships. However, some limitations of the present study should be noted. First, the research focuses on only one product category (i.e., face care products). Here, we showed the important concepts and characteristics of object attachment are based on the neural network centered on the VP and PCC and their functional connectivity. However, there may be the other neural factors which are related to other object-specific attachment, as well as general object attachment. Further research on different categories of products would be helpful to achieve the generalizability of the present findings. Moreover, it is unclear, in the present study, as to whether or how the other factors or the emotions that construct "attachment" are involved in the neural activity and connectivity related to object attachment. Further research on this topic should be performed for knowing more about the meanings of the neural networks related to object attachment.

Participants.
A total of 20 healthy right-handed women (age [mean ± standard deviation]: 33.4 ± 3.5 y) participated in this study. All recruited participants were regular users of one of the three face seram of famous luxury brands (A, B, and C), consuming them more than three times per week. These cosmetics were all within the same price range (A: \13,500/50 mL; B: \13,500/60 mL; C: \13,000/50 mL). The number of users for each serum was similar (A: n = 7; B: n = 7; C: n = 6), and more than five bottles of serum (5.2 ± 2.0 bottles) were used at the time of the fMRI experiment. Participants did not have any history of neurological or psychiatric disorders and provided written and oral informed consent to participate in the study. The Research Ethics Committee of Experimental stimuli and procedure. We used three types of face serums (A, B, and C) that had different textures and were in bottles with different visual appearances. Each participant could clearly discriminate the serum that she regularly used from the others. The fMRI experiment consisted of two sessions (with or without tactile cues) for each participant, with two types of stimuli (attached or non-attached cosmetic) per session: 1) a visual session consisting of the serum regularly used (attached cosmetic in visual session: AV) and the control (non-attached cosmetics in visual session: nAV), repeated four times, and 2) a visual with tactile session consisting of the attached cosmetic in the visual with tactile session (AVT) and the non-attached cosmetics in the visual with tactile session (nAVT), repeated four times. Each stimulus was presented for 30 s (task block) with a 30 s interval (rest block). The stimulus presentation order was counterbalanced across participants. In the first visual session, participants viewed the photo of a face serum bottle as a stimulus in the MRI scanner using goggles that allowed the photos to be projected. In the second visual with tactile session, participants were applied with the face serum on the back of their left hands while they viewed the photo of the face serum bottle (Fig. 7). The amount of serum applied was 0.2 mL per task block. A beauty specialist applied the serum on the back of the participant's hand using the fingers and palm of her right hand, moving her hand slowly and softly in a circle during application. The speed of movement was approximately 2.5 s per cycle. The application procedure was the same for all the task blocks. During the rest block, the other staff removed the serum from the participant's hand using a warm wet towel. The participants were instructed to pay attention to and experience or feel the stimuli without thinking anything during the experiment.
Subjective evaluation. After the fMRI experiment, participants evaluated all the face serums on a visual analog scale (from 0 to 100 points) of six items: "positive feeling of texture, " "expectation for skincare effect, " "want to buy, " "security, " "satisfaction, " and "attachment" after being applied with the serum in the same manner as they had in the scanner. The average scores of each subjective evaluation were compared between the attached and non-attached cosmetics using a paired t-test with a significance level of p = 0.05.
Functional MRI data analysis. Scanning was conducted using a 3.0 T MRI system (Achieva Quasar Dual; Philips Medical Systems, Best, the Netherlands  18,19,30 at p < 0.05. The additional ROIs were as follows: the right accumbens/ventral striatum (0, 10, 0) 18 , the left VP (− 34, − 2, − 6) and right VP (20, − 6, − 8) 19 , and the left VP (− 9, 0, 6) and right VP (18, − 9, 3) 30 . When there was a significant interaction between attachment and touch, we compared their activities between the attached and non-attached conditions using a paired t-test based on a Bonferroni correction (p < 0.025 = 0.05/2). Furthermore, we performed PPI analyses to identify brain regions whose activity depends on an interaction in the main effect of attachment ([AV + AVT] vs. [nAV + nAVT]). In addition, we performed PPI analysis on brain regions whose activity depends on interactions in the interactive effect of ([AVT + nAV] vs. [nAVT + AV]). For www.nature.com/scientificreports/ each seed ROI that showed a significant main effect of attachment or a significant interactive effect of attachment and touch in the ANOVA, the PPI procedure was performed at the single-subject level. For each participant, the seed region (5 mm radius sphere) was localized around the local maxima of the ROI. We then extracted the time course of activity in the ROI for each subject. The PPI analysis employed three regressors as follows: the deconvolved activation time course in the seed (Physiological), each of the contrasts of the main and interactive effect (Psychological), and their interaction (PPI). These regressors were entered into a first-level general linear model after deconvolution of the HRF, and contrast images of the PPI effects for each participant were entered into a random effects analysis at the second-level analysis (peak p = 0.005 and cluster-level pFWE < 0.05; peak p = 0.001 and cluster-level FWE < 0.05). For each region that showed significant connectivity with the seed, multiple regression analyses were conducted, with individual strength of connectivity (beta value of PPI variable) as the dependent variable, and each individual subjective evaluation and individual brainstem ROI activities as the independent variables, in the main effect contrast ([AV + AVT] vs. [nAV + nAVT]. The analyses were based on a stepwise method. Furthermore, we checked the residuals for all regression analyses by performing a Kolmogorov-Smirnov test of normality and calculated the D-W statistic for the null hypothesis of no autocorrelation. The significance level was set at p = 0.05, for all analyses.

Data availability
Due to confidentiality agreements with the participants, the data in this study are available only at the Shiseido Global Innovation Center and Tokyo Metropolitan University.