Changes in emotional response towards different facial application gestures following skincare application: An electroencephalogram approach

Abstract

Background

Amidst increasing global levels of stress and a growing concern for well‐being, facial massage gestures have emerged as an option to improve the comfort and well‐being of women.

Objective

To characterize four emotional indices (stress, comfort, valence, and arousal) and their evolution when induced from two experimental and standardized facial gestures (Q and S).

Methods

This prospective, monocentric, intra‐subject comparison study was conducted in Japan where 17 healthy female subjects aged 38–55 years were enrolled. Two experimental standardized gestures (Q and S) were administered in pairs to the subjects over 2 days. Gesture Q was a 60 s smoothing and hand press gesture while Gesture S was a smoothing gesture alone. Prior to each application, subjects cleansed their faces and completed a self‐assessment questionnaire. During the procedure, subjects were fitted with a portable electroencephalogram (EEG). A beautician applied a cosmetic product to the eye and cheek area based on specific gestures (P/R control gesture followed by Q, or S). EEG measurements were taken before, during, and after the application. Subsequently, participants filled out another self‐assessment questionnaire.

Important Results

Gesture Q was identified as less stressful, more comfortable during application, and higher positive valence after application which was aligned with self‐perception questionnaire results. The portable EEG analyser further revealed the evolution of emotions during application whereby Gesture Q induced a constant and positive evolution of emotion after 20 s, while Gesture S showed a slight decrease of stress only after 40 s and induced a dominant feeling of lower valence after application, associated with a negative feeling.

Conclusion

The portable EEG analyser showed high potential to capture and differentiate between emotions during the beauty experience. Gesture Q positively influenced the emotional state of subjects, which was complimentary to what subjects reported. Emotions captured during the evolution of gestures helped identify key moments of application. Combined with the self‐questionnaire, these findings contribute to the development of user experience.

Tracking emotional states (stress, comfort, valence, and arousal) in 17 Japanese women before and after standardized massage gestures using EEG and self‐assessments was evaluated. Gesture Q was least stressful and most comfortable, showing positive emotional evolution after 20 s. By combining instrumental multisensorial data with self‐assessments, it shifts from traditional questionnaires to implicit emotional insights, contributing to user experience research in emotions and cosmetics.

INTRODUCTION

In today's contemporary society, the relentless pace of urban life has cultivated a search for comfort, rituals of well‐being, where individuals seek happiness from the escalating stresses of life. Over the years, this has caused a paradigm shift in public interest towards happiness and well‐being [1]. Happiness reports generated over the last 10 years by the United Nations indicated that levels of stress, worry, and sadness have risen globally [1].

The study of emotions, intricately tied to sensory experiences, significantly influences a subject's well‐being, decision‐making, and overall happiness. In the past, assessing the emotional state conventionally relied on subjective methods such as questionnaires or focus groups [2, 3]. However, these methods were less reliable for gauging a consumer's emotions in real time, as verbal or written feedback can be distorted during the response process. Moreover, questionnaires fail to accurately encompass and mirror the entirety of a subject's feelings, posing a challenge in linking emotions to specific elements [3]. This limitation stems from the importance of grasping the ‘untold,’ which is crucial for understanding an experience that resonated and was embraced by consumers.

As a component of wellness, facial self‐massages are recognized for their morphological, biological, and psychological benefits [4]. A previous study investigated the impact of a daily 3‐min self‐massage routine on visible signs of stress on facial skin and well‐being in women using a lubricant made of botanical oils. After 2 weeks, all women reported the self‐massage routine as significantly improving their feelings of well‐being and felt more relaxed [5]. While these studies highlighted the benefits of facial massages on well‐being, the data reported were collected from questionnaires which do not grasp the full range of emotion felt by these women during their massage. Therefore, a more nuanced exploration of the emotional state is necessary.

Studying emotions relies on the principles that electrical signals emitted as a result of the brain's response to a stimulus are objectively and scientifically measurable [6]. This approach was previously studied through electroencephalogram (EEG), which is a neurological device that measures the electrical signals generated during brain activity using electrodes placed on the scalp [3, 7, 8, 9]. An exploratory study previously compared a consumer's emotions during application of two cosmetic products (CP1 and CP2) in 15 female subjects by using EEG. Neural activity recorded during application revealed that consumers were happier and more excited with CP2 than CP1 as CP2 induced more positive valence and arousal reading [1]. This study, among many others [3, 7, 8, 10] demonstrates the insight that EEG devices bring towards understanding emotional states of subjects vis‐à‐vis cosmetic experiences. Despite these past efforts to monitor emotions via EEG, assessing the multiple different emotions evoked by targeted facial gestures during application of cosmetic products is currently under‐researched, for which EEG may prove very insightful as an adjunct.

As such, the present study has two goals. The first is to characterize various emotional states in 17 Japanese women aged 38–55 years before and after different standardized facial gestural applications. Based on these beauty ritual gestures performed by beauticians, four emotional indices, such as stress, comfort, valence, and arousal were monitored via an EEG device [11] over the course of 2 days. Second, the evolution of these emotional states of each subject during the beauty step was monitored.

This combination of instrumental multisensorial data and self‐assessment questionnaires in this study presents a fresh perspective as the emphasis shifts from traditional feedback in questionnaires to focusing primarily on the subjects' authentic emotional experiences in an implicit manner, creating a distinct and unprecedented approach.

MATERIALS AND METHODS

Ethical considerations

The study protocol was submitted and approved in Tokyo, Japan on April 25th, 2022 by an ethical committee (Dentsu Science Jam, Inc. Reception number: 2022007). It was conducted according to the World Medical Association Declaration of Helsinki, local regulations of Japan (APPI) and ICH Guidelines for Good Clinical Practice. All subjects received verbal and written information concerning the study in accordance with the applicable local regulations. This information explained the nature, purpose and risks of the study and emphasized that participation in the study was voluntary and that subjects may withdraw from the study at any given time and for any reason. All subjects were given the opportunity to inquire about the study and were given sufficient time to consider their participation before consenting. The subject's written informed consent to participate in the study was obtained prior to any study related procedure being performed.

Study participants

This prospective, monocentric, intra‐subject comparison study was conducted in May 2022 on a cohort of female cosmetic users in Japan. Seventeen healthy Japanese women aged between 38 and 55 years of age were selected for the duration of the study which lasted 2 days. Subjects were excluded if they had any systemic or skin disease, concomitant medical treatment, and recent aesthetic procedure. Other exclusion criteria were the patients with diagnosed mental illness and allergy to skincare products. Women who were pregnant and lactating or intending to become pregnant during the study period were also excluded.

Material: Portable EEG analyser, EEG recording, and preprocessing

The portable analyser was a single channel EEG device developed by Keio University and Dentsu Science Jam, Inc. in Japan and comprises an emotion analyser kit that analyses the sensitivity from brain waves. The headset contains an electrode which is placed at both the participant's left temple and left earlobe, and a brain function is detected. This is converted into a signal and sent via Bluetooth to a measuring electronic device. The signals are analysed, and emotions (stress, valence, arousal, and comfort) are measured as points (pts) on a scale of 1–100 [12]. The preprocessing, data mining, and pattern recognition method used for developing the emotional indices were previously referenced [13, 14, 15].

Experimental procedure

The application gestures were performed with a product, Takami Skin Peel Essence, twice daily on the face. A control gesture (P/R) and two different experimental and standardized gestures (Q, S) were randomized, following which two gestures were utilized at a time on the subject each day (Control and Q or Control and S) (Figure 1a,b). Gesture P/R stood for the control gestures which comprised a 30 s smoothing gesture which was meant to simulate the initial product spreading phase during a typical skincare application which involved a shorter duration. Gesture Q and S were set as experimental gestures, based on two major self‐application gestures observed in the usage instruction of cosmetic skincare products in the Japanese market. Gesture Q was applied for 60 s which comprised a 20 s smoothing gesture and a 40 s hand press gesture. As for gesture S, it lasted for 60 s and comprised a smoothing gesture alone. Both experimental Q and S were the focus of the study and were therefore applied for a longer standardized period of 60 s to assess their specific effects.

FIGURE 1.

(a) The test‐flow that the subjects underwent during the study and (b) The different application gestures (P/R, Q, and S) performed by the beautician on the subject volunteers twice a day for 2 days. [Colour figure can be viewed at wileyonlinelibrary.com]

On both Day 1 and Day 2, the subjects came to the investigation centre and were instructed to clean their face with a face wash prior to answering a self‐assessment mood questionnaire (Figure 1a). The subjects were then instructed to lie down with their eyes closed before being fitted with the portable EEG Analyser (Dentsu Science Jam, Inc., Japan). A beautician then applied the Takami Skin Peel Essence on the whole face following specific gestures (P/R, Q and S) (Figure 1b). EEG measurements were recorded before, during and after the application of the Takami Skin Peel Essence. The subjects were asked once more to fill in the self‐assessment mood questionnaire.

Statistical analysis of EEG and self‐assessment data

Statistical analyses were performed using the R statistical software (R_VERSION 4.4.3). For EEG emotion indexes, the Shapiro–Wilk test was used to examine whether all datasets are normally distributed. Gesture comparison and time evolution of each gesture were analysed by one‐way analysis of variance, and Bonferroni test was performed post hoc. Comparison of before and after each gesture was analysed by student t‐test. The significance level was set at 5%, and Cohen's d indicated the effect size. For the results of self‐assessment questionnaire, Wilcoxon signed‐rank test was used to compare the self‐perceived mood before and after massage. The significance level was set at 5%.

RESULTS

Comparison of standardized gestures during application

The portable EEG analyser differentiated during product application the different emotion states caused by gesture Q, gesture S, and control gesture P/R. Four emotion indices (stress, comfort, valence, and arousal) were analysed by one‐way analysis of variance using gesture conditions (Q, S, P/R) with the score difference between baseline and during application. The Bonferroni test was performed post hoc. The significance level was set at 5%, and Cohen's d indicated the effect size.

Figure 2 showed a significant difference between the conditions Stress (F[2,45] = 9.259, p < 0.001), Comfort (F[2,45] = 9.683, p < 0.001), and Valence (F[2,45] = 9.966, p < 0.001). Gesture Q was significantly lower in stress score compared to Gesture S (p < 0.001, d = 1.28) and Control gesture P/R (p = 0.005, d = 1.17). Gesture Q was also higher in comfort compared to Gesture S (p = 0.001, d = 1.35) and Control gesture P/R (p = 0.001, d = 1.35) and higher in valence score than Gesture S (p < 0.001, d = 1.56). Gesture S was significantly lower in valence score than Control gesture P/R (p = 0.026, d = 0.97).

FIGURE 2.

The quantitative value of the emotional states (stress, comfort, valance and arousal) caused by the application gestures captured by the portable EEG analyser. Dark blue bars represent the average of Gesture Q; the pink bars represent the average of gesture S and the pale blue bars represent the average of the control gesture (P/R). *p < 0.05. [Colour figure can be viewed at wileyonlinelibrary.com]

Evolution of emotional states in standardized gesture application

In Figure 3, the portable EEG device provided insight on the evolution of the emotional state of subjects during product application for Gesture Q (Top) and Gesture S (bottom). Four emotion indexes were analysed by one‐way analysis of variance using time conditions (0–20 s/21–40 s/41–60 s) for each gesture, and the Bonferroni test was performed post hoc. The significance level was set at 5%, and Cohen's d indicated the effect size.

FIGURE 3.

Shows the evolution of the gestures Q (top) and S (bottom) from 0 to 60 s. Blue bars denote 0–20 s, pink bars denote 21–40 s, and pale blue bars denote 41–60 s. *p < 0.05. [Colour figure can be viewed at wileyonlinelibrary.com]

The results showed significant differences among the time condition of Gesture Q for Stress (F[2,48] = 23.853, p < 0.001); Comfort (F[2,48] = 16.379, p < 0.001) and Arousal (F[2,48] = 5.221, p = 0.009). After 20 s of massage, the stress score of gesture Q was significantly decreased compared to the beginning of massage (p < 0.001, d = 2.32). Both comfort scores (p < 0.001, d = 1.96) and the arousal scores were increased (p = 0.01, d = 1.06) (Figure 3).

Comparison before and after application

The emotional benefits of gestures Q or S were also assessed before and after application (Figure 4). The four indices were analysed using Student t‐test for paired samples (before/after application of gesture) for each gesture. The significance level was set at 5% and additionally at 10% to capture the weak tendencies, and Cohen's d indicated the effect size. Comparison of the before and after massage for gesture Q showed a weak tendency for the valence score to increase (p = 0.077, d = 0.458). For Gesture S, however, comparison of the before and after massage revealed an increase in stress (p = 0.022, d = 0.615) followed by a decrease in comfort (p < 0.01, d = 0.813), and valence (p = 0.043, d = 0.532). Self‐assessment questionnaire before and after application was also reported. A comparison of before and after the application of gesture Q and S was performed on n = 17 subjects (Table 1). Both gestures showed arousal decreased after application.

FIGURE 4.

Shows the difference between before versus after for gesture Q (top) and gesture S (bottom) Blue bars show the score before, and pink shows the score after. *p < 0.10. [Colour figure can be viewed at wileyonlinelibrary.com]

TABLE 1.

The SAM (Self‐Assessment Manikin) questionnaire with a 9‐point scale filled by volunteers comparing their perception before and after the gesture Q or S applications.

Average score	Comfort	Arousal
Before	6.12	4.06
After gesture Q (Hand press)	6.41 (ns)	3.00 (s)

Average score	Comfort	Arousal
Before	6.18	4.53
After gesture S (Smoothing)	6.35 (ns)	3.35 (s)

Average score	Comfort	Arousal
After gesture Q	6.41	3.00
After gesture S	6.35 (ns)	3.35 (ns)

Note: Wilcoxon signed‐rank test was used to compare the self‐perceived mood before and after massage. The significance level was set at 5%. (s), statistically significant and (ns), statistically non‐significant. Statistically significant differences (p < 0.05) after application of the hand gesture.

Self‐assessment questionnaire of different mood attributes

A comparison of before and after of gesture Q and S was performed on n = 17 subjects regarding various mood attributes. Table 2 showed that subjects self‐reported feeling significantly less nervous and feeling cosier after gesture Q. On the other hand, following gesture S, subjects reported feeling significantly less nervous but did not report improvement on other mood attributes (Table 2).

TABLE 2.

Score of higher or lower (p < 0.05) after the application of the hand gesture.

Average score	Nervous	Energetic	Afraid	Relaxed	Annoyed	Satisfied	Tensed	Feel cosy	Refreshed	Lively	Focused	Feel dim
Before	2.35	3.06	1.76	3.00	1.59	2.82	1.76	2.88	2.88	2.71	2.71	2.06
After gesture Q (Hand press)	1.76 (s)	2.94 (ns)	1.47 (ns)	3.24 (ns)	1.35 (ns)	3.12 (ns)	1.50 (ns)	3.24 (ns)	3.00 (ns)	2.59 (ns)	2.47 (ns)	2.18 (ns)

Average score	Nervous	Energetic	Afraid	Relaxed	Annoyed	Satisfied	Tensed	Feel cosy	Refreshed	Lively	Focused	Feel dim
Before	2.18	2.94	1.82	2.94	1.24	2.76	1.71	2.53	2.71	2.59	2.47	2.06
After gesture S (Smoothing)	1.71 (s)	2.94 (ns)	1.53 (ns)	3.29 (ns)	1.18 (ns)	3.06 (ns)	1.53 (ns)	2.88 (ns)	2.88 (ns)	2.71 (ns)	2.59 (ns)	2.35 (ns)

Average score	Nervous	Energetic	Afraid	Relaxed	Annoyed	Satisfied	Tensed	Feel cosy	Refreshed	Lively	Focused	Feel dim
After Q	1.76	2.94	1.47	3.24	1.35	3.12	1.50	3.24	3.00	2.59	2.47	2.18
After S	1.71	2.94	1.53	3.29	1.18	3.06	1.53	2.88	2.88	2.71	2.59	2.35

Note: Wilcoxon signed‐rank test was used to compare the self‐perceived mood before and after massage. The significance level was set at 5%. For gesture Q comparing Before and After, volunteers (n = 17) perceived significantly less nervous, less afraid of, less annoyed, more satisfied and felt cosy after application. (s), statistically significant and (ns), statistically non‐significant. Statistically significant differences (p < 0.05) after application of the hand gesture.

DISCUSSION

The use of EEG in the measurement of emotions has a long history [16, 17, 18], although not in the field of cosmetics. Despite these past attempts, a comprehensive evaluation of the multifaceted emotions felt by consumers was lacking. Specifically, the use of EEG in the measurement of emotions triggered by targeted facial gestures when applying cosmetics is under‐reported in literature. In this investigation, the instrumental approach with a multiparametric portable EEG analyser used together with self‐assessment questionnaire from subjects presented a novel perspective to current literature while previous studies have only explored clinical scoring and self‐assessment questionnaires [5].

In this study, valence and arousal were analysed separately to measure the emotions with EEG [19, 20]. Emotional valence describes the extent to which an emotion is positive and negative whereas arousal refers to its intensity or strength associated with the emotional state [19, 21]. The results obtained further validated the notion that specific facial massage gestures affected the consumer positively or negatively. The recorded underlying emotions from the EEG (Figures 2 and 4) were complementary to the subjective self‐assessment questionnaire completed by the subjects before and after application (Table 1) whereby the subjects experienced less arousal after. They further reported being less nervous (Table 2), which are ‘negative emotions’ associated with stress [4, 16]. Instead, they felt significantly cosier after application implying a more ‘positive’ feeling in line with what the EEG data generated. This insight helped discern the specific physical facial gestures that influenced the ‘positive’ emotions subjects felt during application. These data were corroborated by Davidson's frontal EEG asymmetry model which supported the view that an increase in the left prefrontal activity was related to positive emotions while an increase in the right prefrontal activity of the brain was related to negative emotions [22, 23]. Therefore, this study added new evidence to the potential use of certain facial gestures to improve the well‐being and happiness of subjects.

A previous large‐scale systematic review of 137 studies involving 12 966 individuals explored the efficacy of touch interventions on well‐being. It was reported that touch interventions were particularly effective in regulating cortisol levels and decreased pain, depression, and anxiety in adults. Owing to the large‐scale approach of the analysis, the results obtained were deemed robust with high statistical power. In line with this hereby study, ‘negative emotions’ such as anxiety were also reported having decreased following specific facial gesture further corroborating that certain gestures were capable of improving the well‐being of subjects [24]. However, to further increase the robustness of this claim, a larger sample could be used for future investigations.

The evolution of the subject's emotional response during the massage technique was also monitored. The subjects felt more comfortable and less stressed after 20 s of smoothing and hand press gesture, which would not have been perceivable on a self‐administered questionnaire alone. This revealed that emotions felt during the facial massages were not constant throughout. A 60 s massage window was set for this exploratory study as it reflects the maximum realistic time that users typically spend applying skincare products during their personal routines at home. This allowed for an evaluation under practical conditions. Being a complex phenomenon with multiple manifestations, measurements of brain responses can be a useful adjunct to other methods, for example, surveys, focus groups. As a future direction to this study, other investigative methods could also be included, such as the measurement of salivary oxytocin [5]. It was previously suggested that the level of oxytocin (responsible for reducing stress and stabilizing emotions) within the body increases when pleasant feelings are created by tactile stimulation during skincare, for example, when the face is covered with the palms [19, 25]. It would therefore benefit this study to have similar techniques included.

A previous study conducted by Arakawa et al. in 2021 on 34 US participants and 22 Japanese participants suggested behavioural differences between the two groups over various representative words expressing skin tactile conditions after skincare application. For instance, the 22 Japanese participants associated moisturized and skin softness to a pushing gesture while US participants associated similar words to a stroking action or both a stroking and pushing action. This implied a link between what participants felt and cultural elements [26]. Further studies by Arakawa in 2024 also looked at the verbal descriptions of 1545 participants from five countries and observed their haptic exploratory procedures (HEPs) to uncover cultural differences. They found that although participants used similar words to describe tactile perceptions, the HEPs varied across different cultures [27]. In this hereby study, it was therefore hypothesized that the feeling of comfort and reduced stress felt by the Japanese subjects was likely due to social and cultural factors that the gestures were associated with. As a future directive, it would be pertinent to extend the application of these gestures to a broader and more culturally diverse group of subjects. Studies focusing on the cross‐cultural elements are crucial to verify the differences in perception and preference [28]. A previous study on Brazilian and French women revealed important differences in the perceived sensations of different product textures which led to the conclusion that culture was a key factor that affected sensory properties in different countries [28]. In the context of this study, this would help understand how the emotions associated with different facial gestures vary in different countries.

McGlone et al. proposed that pleasant touch applied to hairy skin is processed in limbic‐related regions of the brain via large, myelinated mechanoreceptors (LMTs), suggesting it constitutes an innate, non‐learned affective response. In contrast, the perception of pleasantness from touch on glabrous skin (such as the palms or face) is thought to arise from learned associations or secondary reinforcement mechanisms [29]. In this connection, it was hypothesized that the sensations reported by participants during the tactile facial massage in this study would primarily involve glabrous skin and the associated perception of pleasantness might not stem from an innate affective system but rather from learned or conditioned responses. This raises the question of the extent to which cultural factors or social conditioning shape the emotional valence attributed to such tactile stimuli. Ultimately, understanding this concept clearly could help open up an avenue for exploring personalized beauty or skin care gestures.

When comparing the objective EEG response with the subjective self‐assessment data, a pattern of partial convergence for gesture Q is observed. In the objective EEG data, a decrease in stress and an increase in comfort and valance were reported for gesture Q, which was aligned with the self‐reported feelings of increased cosiness and reduced nervousness after the gesture Q application. This suggested that Gesture Q was consistently coherent both objectively and subjectively. However, for gesture S, the findings diverged, whereby the EEG data indicated an increase in stress and a decrease in comfort following the massage. Subjective data gathered through the questionnaire only reported lowered nervousness but showed participants were more tense and less cosy compared to gesture Q. This divergence suggested that although participants did not perceive the gesture as ‘negative’, the EEG captured subtle increases in negative emotional states that were not reflected in the self‐assessment scores. This reflected the higher sensitivity of EEG to subtle emotional changes or the tendency of self‐report to under‐represent mild or evolving discomfort.

The limitations of this study existed in two folds. First, a small number of subjects (n = 17) were recruited to participate in this study. While the number of subjects is suitable for a preliminary study, it would be interesting to conduct a follow‐up study with a scaled‐up number of subjects, varied gender and cultures to increase the robustness of the trends observed [30]. Second, limitation due to the EEG methodology should also be factored in whereby EEG signals are affected by external factors such as muscle movement and other environmental noises which may have complicated the interpretation of the emotional responses leading to less accurate results. To further expand the reach of this study, it would be advantageous to explore whether the positive gestures identified here can be expanded towards a self‐application process by the subjects themselves. Furthermore, future research could dive into the specific mechanisms by which the facial gestures exert their positive effects.

The outcome of this study sets the scene for exploring novel avenues pertaining to designing comprehensive and holistic beauty experiences for consumers. Through gestures based on individual preferences, the EEG and its associated outcome can be adapted seamlessly to devise a personalized skin routine adapted to the morning or evening from the comfort of a subject's home. It can also prove valuable to professional service providers such as hairstylists and make‐up artists to attain the best user experience by achieving an optimal combination of formula, gesture, and applicators. Ultimately, data gathered from this investigation paves the way towards improving the overall female beauty experience in the near future.

CONCLUSION

The impact of different facial application gestures (P/R, Q, or S) on the emotional state of subjects and the evolution of these emotional states were captured. Paired with a self‐assessment questionnaire, it was shown that smoothing and hand press facial application gesture Q positively influenced the emotional state of subjects and the evolution of these emotional states helped identify key moments during the beauty ritual. This study was limited by its small sample size (n = 17) and the inherent sensitivity of EEG measurements to external interferences such as environmental noise, which may have affected result accuracy. Future research with larger, more culturally diverse cohorts and self‐application protocols would add robustness. Combined with the self‐questionnaire, this study gives an unprecedented insight into user experience and their evolution in real time.

FUNDING INFORMATION

All costing of the present work was entirely met by L'Oréal Research and Innovation Department.

CONFLICT OF INTEREST STATEMENT

SY, DV, DL, and FF are employees of L'Oréal group.

ETHICS STATEMENT

The study is according to the guidelines of the Declaration of Helsinki, OECD Good clinical Practices (GCP) and APPI (Act of Protection of Personal Information). This study protocol was submitted and approved by an Ethical Committee located in Japan (Dentsu Science Jam, Inc. Reception number: 2022007). Informed consent was acquired from all patients involved.

Supporting information

Figure S1.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.