Measurement and Evaluations of Kawaii Products by Biological Signals

Abstract

To evaluate the affective value of industrial products, such subjective evaluation methods as questionnaires are commonly used, even though they have some demerits such as linguistic ambiguity and interfusion of experimenter and/or participant intention to the results. We began our research to objectively evaluate interactive systems by quantifying sensations using biological signals to supplement the above questionnaire demerits. We utilize biological signals to estimate participant feelings of relaxation, comfort, and excitement, which are considered positive sensations. First, we focus on a positive and dynamic feeling called “wakuwaku.” We construct various systems to evaluate affective values to derive wakuwaku feeling using biological signals and clarify the relation between the wakuwaku feeling and biological signals. Then we applied the obtained methods for wakuwaku feeling to evaluate kawaii feeling derived from various stimuli. Finally, we found that kawaii stimuli can be classified into exciting kawaii and relaxing kawaii.

1 Introduction

This chapter introduces our research on kawaii feeling using biological signals mainly based on the Refs. [1,2,3,4,5,6] and their extended work.

As we have already described in Chap. 1, Kansei/affective value has become very important for industrial products. To evaluate the affective value of these products, subjective evaluation methods such as questionnaires are commonly used. In previous chapters (Chaps. 2, and 3), we introduced research on kawaii feelings of objects employing questionnaires. Questionnaires are known for the established methods for subjective evaluation and have various merits. However, at the same time, they suffer from the following demerits:

• Linguistic ambiguity.

• Interfusion of experimenter and/or participant intention to the results.

• Interruption of the system’s stream of information input/output.

Solving these problems is crucial to evaluate the degree of interest and/or excitement of an industrial product or an interactive system, such as whether the system is really interesting, and to identify the moment of excitement. Evaluating the affective value of an industrial product or an interactive system only by such subjective evaluation methods as questionnaires is almost impossible.

We began our research [1] to objectively evaluate interactive systems by quantifying sensations using biological signals that offer the following merits and can supplement the above questionnaire demerits:

• Can be measured by physical quantities.

• Avoids influence from the intentions of experimenter and participants.

• Can be measured continuously.

Much previous research has measured mental sensations using biological signals. Ohsuga et al. used biological signals to measure mental stress or simulator sickness [7, 8], which are considered negative sensations. On the other hand, Omori et al. measured ECG and EEG to evaluate autonomous and central nerve activities evoked by color stimuli, where they treated relaxation or comfort [9]. We previously utilized the alpha waves of EEG to estimate participant feelings of relaxation [10]. Compared with negative sensations, relaxation and comfort are considered nonnegative sensations.

In this chapter, we first focus on a feeling called “wakuwaku,” which is a Japanese word for a positive sensation derived when someone feels something exciting or captivating. The word means thrilling or exhilarating in English. A wakuwaku feeling is also considered a non-negative sensation, as are relaxation and comfort. However, a big difference exists between those sensations: a wakuwaku feeling is considered dynamic, especially compared to the static sensations of relaxation and comfort. Now in 2018, much research exists on such positive and dynamic sensations as the wakuwaku feeling. However, little previous research existed in 2007 when we performed this research. In Russell’s circumplex model of emotion [11] (Fig. 5.1), such negative dynamic emotions as stress is in the second quadrant, and such positive static emotions as relaxing and comfort are in the fourth quadrant. “Wakuwaku” feeling is in the first quadrant (Fig. 5.2).

Fig. 5.1

Russell’s circumplex model and locations of feelings

Fig. 5.2

Russell’s circumplex model and locations of two types of kawaii

The purposes of the next section include to clarify the relation between such dynamic, positive sensation as wakuwaku feeling and biological signals. Then, we introduce various kinds of research applying the obtained methods to measure kawaii feelings using biological signals, especially ECG. Finally, we found that kawaii feelings are classified into excited kawaii and relaxed kawaii. In other word, kawaii stimuli are classified into exciting kawaii and relaxing kawaii.

2 Measurement of Wakuwaku Feeling

2.1 Background

In this section, we focused on a feeling called “wakuwaku,” a positive and dynamic sensation. The purposes of this section include to clarify the relation between wakuwaku feeling and biological signals [1].

2.2 Construction of Systems

We constructed various systems based on a treasure chest game to evaluate the degrees of wakuwaku feeling. We employed virtual treasure boxes by computer graphics instead of real treasure boxes because creating various real treasure boxes mentioned below would be costly and time-consuming. These constructed systems have various complicated components to promote wakuwaku feeling, such as the appearances of figures, their combination, and the actions of the combined figures. The parameters of these systems were the design of the boxes and the sound, the BGM and the effect as shown in Table 5.1. The three box designs are shown in Fig. 5.3. The constructed systems are shown in Table 5.2.

Table 5.1 System parameters

Fig. 5.3

Three types of boxes

Table 5.2 Constructed systems

The procedures of the game were as follows:

• 1. Confirm the figures in the boxes (Fig. 5.4a).

  Fig. 5.4

  

  System flow

• 2. Choose one of the boxes (Fig. 5.4b).

• 3. Watch the figure in the chosen box (Fig. 5.4c).

• 4. Repeat the above procedures (Fig. 5.4d).

• 5. Watch the combinations of the two figures (Fig. 5.4e).

• 6. Watch the combined figure (Fig. 5.4f).

These procedures were designed to promote wakuwaku feeling when expecting a figure’s appearance from the chosen box and combining two figures. Questionnaires and biological signals were employed to evaluate the degree of wakuwaku feeling of each system.

Figure 5.5 shows the system setup. The input device was a keypad, and the output devices were a 17-inch LCD display and a pair of speakers. The biological signals were measured by sensors and BIOPAC measurement equipment (Biopac Systems Inc.). Two PCs were employed for system display and to measure the biological signals.

Fig. 5.5

System setup

2.3 Experiments to Evaluate the Systems

2.3.1 Method

The participants randomly played the four games shown in Table 5.2 and answered questionnaires for each system.

The questionnaire about the wakuwaku feeling consisted of 23 items of paired 7-point evaluations such as “fun boring,” and five items of unpaired 5-scale evaluation such as “pounding.” For the paired 7-point evaluation items, four indicates neutral, seven indicates the best, and one indicates the worst. For the unpaired 5-point evaluation items, five indicates the best and one the worst. In addition, participants were asked some free description questions after playing four games.

The following biological signals were measured constantly during the experiments to detect the degree of wakuwaku feeling: Galvanic Skin Reflex (GSR), Electrocardiogram (ECG), and breathing rate. GSR, which is affected by states of emotion, was used as a physiological index to detect such emotions as anxiety and mental stress. ECG is changed not only by physical exercise but also by mental factors such as anxiety and stress. In addition, breathing rates and patterns are also indexes of stress and anxiety.

2.3.2 Results

Experiments were performed with 12 male students in their 20s who served as volunteers.

From the results of the analysis of variance for each questionnaire item with parameters in Table 5.1, the main effect of sound was significant for almost all questionnaire items including exciting and enjoyable. On the other hand, the main effect of the box design was not significant for almost all items. Table 5.3 shows the result of the analysis of variance for enjoyable. In the free description answers, some participants pointed out that the BGM and the sound effects were good points of the system, suggesting that sound is effective for wakuwaku feelings.

Table 5.3 Analysis of variance (enjoyable)

As for biological signals, we selected various physiological indexes as shown in Table 5.4. The RR interval, defined as the time interval between the two R waves of the ECG, is the inverse of the heart rate, the number of heart beats per minute. All indexes were normalized by the values in the quiet state for each participant. Since we designed the game flow with various events to promote wakuwaku feelings, we chose the following three moments for analysis:

Table 5.4 Physiological indexes

• Moment I: When the first box opened.

• Moment II: When the second box opened.

• Moment III: Just after combining the two figures.

The first and second moments were in the first half of the game, while the third moment was in the second part. By a paired difference test, the heart rate at each moment of the first and the second choices of boxes was significantly different between the systems with different box designs. On the other hand, the heart rate at Moments I or II was not significantly different between the systems with sound and without sound. However, the averages of GSR at Moment III were significantly different only between the systems with sound and without sound. Table 5.5 summarizes the results of all tests.

Table 5.5 Results of difference tests

2.4 Discussion

The above experimental results suggest that heart rate and GSR averages may show the wakuwaku feeling of the users of interactive systems. Moreover, the heart rate results are related to the system’s former part, and the results of GSR averages are related to its latter part. Since the questionnaire results agreed with the results of the GSR averages and disagreed with the heart rate results, they might reflect the wakuwaku feeling of the latter part of the systems. The questionnaire may reflect the wakuwaku feeling of the system’s last part because the participants can only remember it.

2.5 Summary

To evaluate the affective value of industrial products, we constructed various systems based on a treasure chest game to evaluate their affective values especially generated wakuwaku feeling using biological signals.

We performed experiments to measure the degree of wakuwaku feeling by using the constructed systems. From analysis of the experimental results, we obtained the following useful knowledge:

• The degree of wakuwaku feeling may vary depending on such parameters as object design and sound effects.

• The degree of wakuwaku feeling may be measured by such biological signals as GSR and ECG.

This work is the first step to measure wakuwaku feeling that occurred by interactive systems. We performed many researches on wakuwaku feeling after that such as [12,13,14,15,16].

3 Measurement of Kawaii Feeling and Biological Signals—Kawaii Color and Size

3.1 Background

From 2007, we focused on the kawaii attributes of industrial products, because we considered kawaii one of the most important affective values. Our aim is to clarify a method for constructing kawaii products from the research results. We previously performed experiments and obtained valuable knowledge on kawaii attributes (Chap. 2). For example, curved shapes such as a torus and a sphere are generally evaluated as more kawaii than straight-lined shapes. Brightness and saturation are effective for kawaii colors.

Meanwhile, only questionnaires were employed to evaluate kawaii shapes and colors in all the above experiments. However, questionnaires suffer from some demerits as described in Sect. 5.1. Thus, to compensate for questionnaire demerits, we examined the possibility of using biological signals.

This section describes our trials to clarify the relation between kawaii colors and biological signals and the relation between kawaii sizes and biological signals [2].

3.2 Kawaii Colors and Biological Signals

3.2.1 Experimental Method

The first experiment addressed kawaii colors. We performed a preliminary experiment to select color candidates from 381 colors in the color table [17] employing four male and two female students in their 20s as participants. As the experimental results, pink and colors basically identical to pink were selected as kawaii colors, while dark brown and dark green were selected as non-kawaii colors. Thus, we selected pink (5R 7/10 from the Munsell Color System [18], blue (10B 7/6), brown (5R 4/6), and green (between 2G 4/4 and 3G 4/4) for the following evaluation experiment, where green represented a middle color between kawaii and non-kawaii colors.

Figure 5.6 shows the experimental setup. Participants watched a large, 100-inch screen whose surface was covered by one of the four above colors and evaluated its kawaii degree on a 7-scale evaluation.

Fig. 5.6

Experimental setup

The experimental procedures were as follows:

• 1. Participants sat on chairs.

• 2. The experimenter explained the experiment.

• 3. Participants wore electrodes.

• 4. Participants remained quiet for 30 s.

• 5. Participants watched the screen displayed by a color for 30 s.

• 6. Participants answered the questionnaire.

Steps 5 and 6 were repeated four times for the four colors, which were displayed in random order. Such biological signals as heart rate, Galvanic Skin Reflex (GSR), breathing rate, and Electroencephalogram (EEG) were measured both before and while watching. The biological signals were measured by BIOPAC Student Lab (BIOPAC Systems, Inc.), except for EEG which was measured by the Brain Builder Unit (Brain Function Research Center).

3.2.2 Experimental Results

Experiments were performed with eight female and eight male student volunteers in their 20s. Based on our previous experiments [19,20,21,22], we selected the following physiological indexes for analysis:

• Average heart rate,

• Variance of heart rate,

• Average RR interval,

• Variance of RR interval,

• Average GSR,

• Variance of GSR,

• Number of breaths,

• Variance of number of breaths,

• Average breath magnitude,

• Variance of breath magnitude,

• Ratio of power spectrum of Theta, Slow alpha, Mid alpha, Fast alpha, and Beta waves,

• Ratio of dominant duration of Theta, Slow alpha, Mid alpha, Fast alpha, and Beta waves.

All indexes described above were normalized by the values in the quiet state for each participant.

Figure 5.7 shows the questionnaire results, where the horizontal axis shows the participants (male: a–h, female: i–p) and the vertical axis shows the kawaii degree of each color.

Fig. 5.7

Questionnaire results of the first experiment

From the results of a two-factor analysis of variance, the main effect of color is significant at the 1% level, the main effect of gender is significant at the 5% level, and a significant cross-effect exists between color and gender at the 5% level. Thus, we successfully analyzed the biological signals by dividing the participants into two groups by kawaii scores.

The data of the biological indexes were divided into the following two groups: the kawaii group where scores were above 0, and the non-kawaii group where the kawaii scores were below 0. The data with 0 score were omitted from analysis. From the unpaired t-test results of the difference of the mean value of the two groups, the heart rates, the numbers of heart beats per minute, showed a significant difference (Table 5.6). The heart rate of the kawaii group was significantly faster than that of the non-kawaii group. This result suggests that watching a kawaii color is more exciting than watching a non-kawaii color, because a decrease in the heart rate is considered an index of relaxation.

Table 5.6 Unpaired t-test of difference of mean value of heart rate

Moreover, from the unpaired t-test results of the difference of the mean value of the two groups by gender, the ratio of the dominant duration of the mid-alpha wave showed a significant difference. The ratio of the dominant duration of the mid-alpha wave of the kawaii group was significantly larger than that of the non-kawaii group. This result suggests that watching a kawaii color is more exciting than watching a non-kawaii color, because a decrease of the ratio of the dominant duration of the mid-alpha wave is also considered as an index of relaxation.

3.3 Kawaii Sizes and Biological Signals

The second experiment addressed the kawaii sizes of objects. The experimental setup resembled that shown in Fig. 5.4, except the display used two projectors with polarized filters and the participants wore polarized glasses to watch the objects on the screen stereoscopically. Participants watched an object on the large screen and evaluated its kawaii degree. The shape and the color of each object were set as torus and yellow, 5Y8/14 in MCS, as shown in Fig. 5.8, based on the results of our previous experiment described above. The size, which means visual angle, of each object was set as one of the four sizes shown in Table 5.7 based on our preliminary experiment.

Fig. 5.8

Objects employed for second experiment

Table 5.7 Analysis of variance

Experiments were conducted with 12 female and 12 male student volunteers in their 20s. Figure 5.9 shows the questionnaire results, where the horizontal axis shows the participants (male: m1–m12, female: f1–f12) and the vertical axis shows the kawaii degree for each object size. The results of a two-factor analysis of variance show that the main effect of size is significant at the 1% level and the main effect of gender is not significant. The biological signal data were divided into two groups similar to the first experiment. From the unpaired t-test results of the difference of the mean value of the two groups, the heart rates showed a significant difference (Table 5.8). Since a higher heart rate shows the unrelaxed state, the mental state when feeling kawaii is considered more exciting than not feeling kawaii. In addition, the results of the similar difference test of the heart rate for object C showed a significant difference at the 5% level, and the results of the similar difference test of the heart rate for object D showed a significant difference at the 1% level. These results show that the mental state when feeling kawaii is probably more exciting than not feeling kawaii even if the size of the object being watched is the same.

Fig. 5.9

Questionnaire results of second experiment

Table 5.8 Unpaired t-test of difference of mean value of heart rate

3.4 Summary

In this section, we focused on the relation between two kawaii attributes and biological signals. We performed two experiments to clarify the relation between kawaii colors and heart rates and the relation between kawaii sizes and heart rates. In both experiments, heart rates increased when the participants felt kawaii.

4 Evaluation of Kawaii Size Using Augmented Reality

4.1 Background

We performed experiments and obtained interesting tendencies about such kawaii attributes as sizes in the previous section (Sect. 5.3). In this section, we refocus on kawaii size to obtain the tendency of kawaii sizes using Augmented Reality. In our first experiment, we compared the results of virtual objects between virtual and real environments. In the second experiment, we classified the most kawaii size.

4.2 The First Experiment

4.2.1 Experimental Method

In our experimental setup, we employed a 42-inch LCD 3D monitor (Hyundai Corp.) and polarized glasses to stereoscopically show virtual objects in a virtual environment (VR) (Fig. 5.10). The distance between the monitor and each participant was 1.0 m. In a real environment, we employed 3D see-through glasses (Wrap920AR, Vizix Corporation) to stereoscopically show the virtual objects (AR) (Fig. 5.11). The marker for the AR display was set in the left hand of each participant. The distance between the participants and their hands was 0.5 m.

Fig. 5.10

VR setup

Fig. 5.11

AR setup

Participants watched an object in VR or AR and evaluated its kawaii degree. Each object’s shape was set as a cube and its color was set as red-yellow (orange) on the basis of the results of our previous experiments (Chap. 2). Each object’s size, which implies a visual angle, was set to one of four (Table 5.9) both in VR and AR based on our previous experiment (Sect. 5.3). We randomly showed these four objects to the participants, who evaluated the kawaii degree for each of them on a 7-point Likert scale:-3: extremely not kawaii, 0: neutral, and +3: extremely kawaii. We also measured the ECGs of the participants during the experiment.

Table 5.9 Visual angles of virtual objects both in VR and AR

4.3 Results and Discussion

Our experiments were performed with eight female and eight male student volunteers in their 20s.

Figure 5.12 exhibits the questionnaire results, where the horizontal axis shows the ratios of the kawaii scores and the vertical axis shows each object size in VR. The averaged scores for the kawaii degrees show that smaller objects tend to be more kawaii in VR, which is the same result as in a previous experiment (Sect. 5.3). The results of a two-factor analysis of variance for VR show that the main effect of size is significant at the 1% level, but the main effect of gender is not.

Fig. 5.12

Questionnaire results for VR

Figure 5.13 displays the virtual objects in AR, and Fig. 5.14 shows the questionnaire results for AR. The averaged scores for the kawaii degrees in AR show the same results in VR. The results of a two-factor analysis of variance for AR show that the main effect of size is significant at the 1% level, but the main effect of gender is not.

Fig. 5.13

Displays of virtual objects in AR

Fig. 5.14

Questionnaire results for AR

Figure 5.15 compares the averaged kawaii degrees for each size between VR and AR, where the vertical axis shows the averaged kawaii degree. The correlation coefficients between the sizes of the objects and the kawaii degrees are higher in AR than VR, implying that the AR result is stronger than that of VR.

Fig. 5.15

Comparison of averaged kawaii degrees between VR and AR

We calculated heart rates from ECG signals. Heart rates were normalized by the values in the quiet state for each participant. The data of the biological indexes were divided into the following two groups: the kawaii group where scores were above 0 and the non-kawaii group where the kawaii scores were below 0. The data with 0 score were omitted from analysis. From the unpaired t-test results of the difference of the mean value of the two groups, the heart rates, the numbers of heart beats per minute, showed a significant difference only in case of AR. The heart rate of the kawaii group was significantly faster than that of the non-kawaii group in case of AR. This result reconfirmed that kawaii feeling increases the heart rate.

Table 5.10 compared the averaged heart rates between feeling kawaii and non-kawaii in VR and AR. Figure 5.16 showed the averaged heart rates for each size in VR and AR. This figure shows the strong correlation between size and heart rate in AR.

Table 5.10 Comparison of the averaged heart rate between feeling kawaii and non-kawaii in VR and AR

Fig. 5.16

Averaged heart rate for each size in VR and AR

4.4 The Second Experiment

4.4.1 Experimental Method

The results of the first experiment showed that participants tended to judge smaller objects more kawaii. However, since the most kawaii size was not clarified, we did the second experiment.

Both for AR and VR, we employed the same setup as in the first experiment. The initial size of the virtual object was set as the smallest, as in Table 5.1, and participants could increase or decrease its size using the arrow key of the keyboard until they felt the object to be the most kawaii. The pitch of the change of the object’s size was 0.23°, which was 10% of the initial size.

4.4.2 Results and Discussion

The second experiment was performed with a female and four male student volunteers in their 20s.

Figure 5.17 shows the results, where the initial size was set to 1 on the vertical axis. Participant A should be considered as an outlier, because the most kawaii size for the other participants almost equaled double the object’s initial size. This result shows that the most kawaii size has a lower limit.

Fig. 5.17

The most kawaii size

The results of our first experiment suggested that smaller objects tend to be more kawaii. However, our second experiment results showed that the tendency of size of the first experiment was not always, and that an optimum size for kawaii objects was determined.

4.5 Summary

We performed two experiments to determine the most kawaii size using both VR and AR. In the first experiment, our results suggest that smaller objects are more kawaii in both environments. In addition, the AR result is stronger than that of VR from both questionnaires and heart rates. In the second experiment, we determined for kawaii objects an optimum size, implying a limit on the tendency obtained in our first experiment.

5 Evaluation of Kawaii Objects in Augmented Reality by ECG

5.1 Background

In this section, we employed ECGs in our evaluation to clarify the differences of kawaii feelings between different objects using augmented reality (AR) based on the previous section (Sect. 5.4).

5.2 Experimental Method

In our experimental setup, we employed 3D see-through glasses to stereoscopically show virtual objects. We set markers for the AR display in the left hands of the participants, who observed three objects in AR. One was an orange ball employed in our previous research (Fig. 5.18a). Another was chosen by each participant as the most ordinary object from the five candidates shown in Fig. 5.18b. The last object was also chosen by each participant as the most kawaii object from the five candidates in Fig. 5.18c.

Fig. 5.18

Objects observed by participants

We measured the ECGs before and while observing the objects. Because of the individual differences of heart rates (HRs), we used the difference between the HR while and before observing the objects as normalized HRs.

5.3 Experimental Results and Discussion

We performed our experiments with 5 female (1–5) and 5 male (6–10) student volunteers in their 20s. Figure 5.19 compares their averaged normalized HRs.

Fig. 5.19

Normalized HRs for all participants

When they observed the kawaii objects, the HRs were high for all of the females and most of the males. When observing the common objects, the HRs resembled the HRs before observing the objects for all the females and most of the males.

5.4 Summary

We experimentally clarified the relation between objects and ECGs. We observed kawaii objects in AR related to high heart rates and got the same results as from the elderly who observed kawaii spoons (Chap. 8).

6 Differences in Heartbeat Modulation Between Excited and Relaxed Kawaii Feelings During Photograph Observation

6.1 Background

A person experiences kawaii feelings when observing kawaii objects or attributes in everyday sceneries. Specifically, watching photographs elicits subjective, intuitive, and strong preferences depending on personal emotional feelings, suggesting that investigation of kawaii feeling that occurs while watching photographs provides valuable insights. Therefore, kawaii-related physiological changes while performing this task were measured. In a previous study, we detected heartbeat modulation when participants experienced kawaii feeling watching photographs [23, 24]. A pronounced modulation was observed when the subjects themselves selected kawaii photos that act as stimuli. Moreover, physiological brain responses were measured in event-related potentials when participants experienced kawaii feeling when watching photographs [25, 26].

We further hypothesized that the kawaii feeling evoked while watching photographs could be classified into two types [27]. The first type, called “excited kawaii” in this section, corresponds to a feeling of cuteness accompanied by excitation. The second type, or “relaxed kawaii,” is a feeling of cuteness accompanied by relaxation. To test this hypothesis, in this study, heartbeat modulations were measured when participants experienced both types of kawaii feeling while watching photographs and potential differences in these heartbeat modulations were examined with respect to both types.

This section is described based on Ref. [4].

6.2 Material and Methods

6.2.1 Participants

Seven healthy students in their 20s participated in the experiment. All participants gave informed consent prior to participation. The experimental procedure was approved by the ethics committee of Shibaura Institute of Technology.

6.2.2 Stimuli

Each participant subjectively selected (1) an exciting kawaii photograph, which evoked excited kawaii feeling, (2) a relaxing kawaii photograph, which evoked relaxed kawaii feeling, and (3) an uninteresting photograph, which evoked no specific feeling. Exciting kawaii photographs were expected to make the participants’ hearts palpitate and in turn enhance their heart rates. Relaxing kawaii photographs were assumed to make the participants unwind, resulting in lower heart rate increase compared with their exciting kawaii counterparts. Uninteresting photographs were not expected to cause any heartbeat modulation and acted as control stimuli. Heartbeat modulation differences according to the types of photographs would physiologically verify the hypothesis that the kawaii feeling resulting from watching photographs can be categorized as excited and relaxed.

Participants were instructed to select photographs that induced the desired feelings in a stable and permanent manner from the Internet. Typical exciting kawaii photographs showed animals and portraits across the subjects while their relaxing kawaii counterparts usually displayed animals. Uninteresting photographs generally presented tools, furniture, buildings, or sceneries. Interestingly, one participant selected an animal picture as an exciting kawaii photograph that was chosen as a relaxing kawaii one by another participant. This suggested that the kawaii classification strongly depended on subjective feelings. Selected photographs were converted to monochrome images and edited to fit a 192 pixel-high by 256 pixel-wide frame.

6.2.3 Heart Rate Measurement

Visual stimuli were displayed on a laptop computer screen placed in front of the participants (Fig. 5.20). During each trial, participants quietly faced a black screen for 30 s (rest period) before watching a picture for 30 s (task period). This trial was repeated five times for a total duration of 300 s while an electrocardiogram (ECG) was measured by radiofrequency ECG (Wireless Vital Sign Sensor, Micro Medical Device, Inc.). Heart rates were calculated from RR intervals in individual ECGs.

Fig. 5.20

A snapshot of the experiment being conducted

The period taken to present each picture was longer that than in previous experiments [23] in order to provide sufficient time to significantly change heart rates. The time interval between photograph selection and ECG measurement was limited to 10–15 min to ensure that the participant’s impressions of the photos remained unchanged.

6.3 Results

6.3.1 ECG Waveforms

A typical ECG waveform (Fig. 5.21) shows the amplitude [mV] as a function of time [s] within a trial. This waveform was extracted from a trial using an exciting kawaii photograph. Figures 5.22, 5.23, and 5.24 show typical heart rates obtained during each trial for a participant using exciting kawaii, relaxing kawaii, and uninteresting photographs, respectively. In Figs. 5.22, 5.23, and 5.24, numbers in legends are corresponding to the trials. Horizontal axes indicate the rest and task periods while vertical axes represent heart rates in a trial [bpm]. Heart rate distributions are shown for each trial.

Fig. 5.21

A typical waveform of the heartbeat

Fig. 5.22

Typical heart rates for each trial using an exciting kawaii photograph

Fig. 5.23

Typical heart rates for each trial using a relaxing kawaii photograph

Fig. 5.24

Typical heart rates for each trial using an uninteresting photograph

6.3.2 Average Heart Rate Difference

For each trial, a heartbeat modulation evoked by a photograph was defined as a change in heart rate average during the second half of the task period (15 s), which was assumed as a baseline, relative to the heart rate average during the first half of the task period (15 s). This change was called heart rate difference. For each participant, five heart rate differences were obtained and averaged for the five trials conducted using individual pictures. These averages were defined as average heart rate differences (Table 5.11).

Table 5.11 Average heart rate differences for each type of photograph and individual subject

6.3.3 Statistical Tests

Statistical tests were conducted using average heart rate differences (Table 5.11). To determine whether a type of photograph induced a heartbeat modulation, a t-test was performed for these average differences according to the stimulus. Exciting kawaii photographs produced a significant difference (p < 0.05), whereas relaxing kawaii and uninteresting photographs induced little changes.

ANOVA was conducted to evaluate whether heartbeat modulation depended on the type of photograph presented. A significant difference was observed between the three stimuli (p = 0.044). Table 5.12 shows post hoc multiple comparison results of the one-way ANOVA using the Bonferroni correction. A significant difference was obtained between exciting kawaii and uninteresting photographs (p < 0.05). No significant differences were detected between exciting and relaxing kawaii photographs or between relaxing kawaii and uninteresting photographs. Figure 5.25 shows one-way ANOVA and post hoc multiple comparison results. Average heart rate differences [bpm] are displayed for individual exciting kawaii, relaxing kawaii, and uninteresting photographs. Average heart rate difference distributions are also presented for the seven participants.

Table 5.12 Post hoc multiple comparison results of the one-way ANOVA

Fig. 5.25

A typical waveform of the heartbeat

6.4 Discussion

A significant t-test result was observed only for exciting kawaii photographs, indicating that these photographs increased heart rates. Moreover, significant differences were detected between exciting kawaii and uninteresting photographs, whereas relaxing kawaii and uninteresting photographs produced little changes. These results validate the hypothesis that the kawaii feeling evoked while watching photographs can be classified as excited and relaxed kawaii feelings. Exciting kawaii photographs produced a heartbeat modulation, unlike their relaxing kawaii counterparts.

The heart rate increase induced by the exciting kawaii photographs was consistent with previous observations for participants watching kawaii pictures [23]. Here, relaxing kawaii photographs did not boost heart rates to the same extent as their exciting kawaii counterparts, showing that the relaxing effects of healing photographs may suppress the increase in heart rate [28]. Furthermore, excited and relaxed kawaii feelings may correspond to two opposite directions on the arousal axis of Russell’s circumplex model of emotion.

6.5 Summary

This section verified the hypothesis that the kawaii feelings evoked while watching photographs can be classified as excited and relaxed kawaii feelings. Heartbeat modulations occurring in participants experiencing these feelings while watching photographs were measured. Exciting kawaii photographs significantly enhanced heart rates, whereas their relaxing kawaii counterparts induced little heart rate increases, validating our hypothesis. Overall, the heartbeat responses resulting from watching kawaii photographs exhibit different modulations depending on the two types of kawaii feeling, although only one word is used in our daily lives.

7 Detection of Relaxing Kawaii Using Stuffed Animals and ECG

7.1 Background

Our previous studies revealed that the physiological responses evoked by kawaii feelings can be measured by ECG (Sects. 5.3–5.5). We also verified that exciting kawaii photographs significantly enhanced heart rates, while their relaxing counterparts did not have this effect (Sect. 5.6). Exciting kawaii response is located in the first quadrant of Russell’s circumplex model, and that of relaxing kawaii is in its fourth quadrant. In this study, we employed stuffed animals as stimuli for relaxing kawaii and clarified the resulting effects using ECG measurements [5].

7.2 Experimental Method

The task was to look at six stuffed animals one by one. Pulse waves were measured as participants did this looking task. We used a pulse wave sensor that can measure heartbeat by simply attaching it to the left index finger, which lessens the psychological burden on participants. The sensor was connected to an amplifier (NEXUS-10 Mark II) to send the output to a PC. Consequently, heartbeat could be monitored in real time. Figure 5.26 shows our experimental system.

Fig. 5.26

Experimental system

The participant was a female in her forties. A two-stage procedure of a 30-s rest period, during which the participant does not look at anything, and a 30-s task period, during which she looks at a single stuffed animal, was repeated six times. Three out of the six stuffed animals were owned by the participant, but she had never seen the other three before. Figure 5.27 shows the stuffed animals used in the task and their presentation order.

Fig. 5.27

Stuffed animals in the task

7.3 Experimental Results

We calculated the participant’s heart rate at rest for 30s and during the 30s tasks based on the RR intervals of the pulse wave. Figure 5.28 shows an experimental scene. Figure 5.29 shows the heart rate calculated for each trial, where the vertical axis shows the heart rate and the horizontal axis shows the rest and task periods in each trial. As shown in the figure, the heart rate was lower in most trials while the participant looked at the stuffed animals than in the resting state. This is an important finding because the heart rate showed that the participant felt super-relaxing kawaii while looking at stuffed animals.

Fig. 5.28

Experiment scene

Fig. 5.29

Calculated heart rate for each trial

7.4 Summary

We performed an experiment using stuffed animals and ECG measurement. Calculated heart rates showed that the participant felt relaxing kawaii while looking at stuffed animals.

In the previous section (Sect. 5.6), the relaxing kawaii photograph gave no significant change from resting state. Therefore, the success of the detection of relaxing kawaii by heart rates decreases in this section is an important new finding.

8 Conclusions

This chapter described various researches on measurement of kawaii feeling using biological signals. In Sect. 5.1, we introduced the merits of using biological signals to measure positive dynamic emotions. In Sect. 5.2, we introduced the measurement of wakuwaku feeling. In Sects. 5.3–5.5, we introduced the measurement of exciting kawaii feelings by color, size, and visual stimulus using ECG. In Sects. 5.6 and 5.7, we introduced that kawaii feeling can be classified into excited kawaii and relaxed kawaii.