*2.6. Data Analysis*

*Physiological responses and facial expressions.* Facial expressions and heart rates were collected during the five-second intervals of the viewing, anticipation, and tasting phases. The duration of the anticipation phase was not exactly controlled because it ended when the participant was manually fed the test food and started chewing, and only the results of the viewing and tasting phases will be reported here. For each phase, averages per second were used for further analysis. PPG heart rate was not successfully collected for

five participants. To allow the proper comparison, RPPG heart rates of these participants were also omitted from further analysis. For the remaining participants, heart rates were missing in 4% of all cases, and facial expressions were missing in 1% of all cases. also omitted from further analysis. For the remaining participants, heart rates were missing in 4% of all cases, and facial expressions were missing in 1% of all cases. *EmojiGrid scores.* The positions of the scores on the EmojiGrid valence (X) and arousal (Y) axis were converted into numerical values (1–10).

the participant was manually fed the test food and started chewing, and only the results of the viewing and tasting phases will be reported here. For each phase, averages per second were used for further analysis. PPG heart rate was not successfully collected for five participants. To allow the proper comparison, RPPG heart rates of these participants were

*Foods* **2021**, *10*, x FOR PEER REVIEW 7 of 13

*EmojiGrid scores.* The positions of the scores on the EmojiGrid valence (X) and arousal (Y) axis were converted into numerical values (1–10). Results were analyzed with Mixed Model Anovas (IBM® SPSS® statistics, version 25, Armonk, New York, NY, USA) with the participant as a random factor, and with the fa-

Results were analyzed with Mixed Model Anovas (IBM® SPSS® statistics, version 25, Armonk, New York, NY, USA) with the participant as a random factor, and with the familiarity with soy sauce (familiar and unfamiliar users), soy sauce (3), duration (5 s), and condition (unbranded and branded) as fixed factors. The usage of soy sauce was a betweensubject factor, and the others were within-subject factors. LSD post-hoc tests were used for main effects. 95% Confidence intervals were used for post-hoc tests of interaction effects. miliarity with soy sauce (familiar and unfamiliar users), soy sauce (3), duration (5 s), and condition (unbranded and branded) as fixed factors. The usage of soy sauce was a between-subject factor, and the others were within-subject factors. LSD post-hoc tests were used for main effects. 95% Confidence intervals were used for post-hoc tests of interaction effects.

#### **3. Results 3. Results**

#### *3.1. General 3.1. General*

The results of facial expressions and RPPG/PPG heart rates will be presented for the viewing and tasting phases, together with Emoji valence and arousal scores, which were collected only in the tasting phase. The effects of each phase will be presented, as well as the effects of soy sauce, familiarity with soy sauce, and branding (no branding, matched branding). Post-hoc test results will only be reported when they are significant. The results of facial expressions and RPPG/PPG heart rates will be presented for the viewing and tasting phases, together with Emoji valence and arousal scores, which were collected only in the tasting phase. The effects of each phase will be presented, as well as the effects of soy sauce, familiarity with soy sauce, and branding (no branding, matched branding). Post-hoc test results will only be reported when they are significant.

#### *3.2. Viewing Phase: Facial Expressions 3.2. Viewing Phase: Facial Expressions*

*Facial expressions of valence*. Facial expressions to images of foods became more negatively valenced when branding information was added (F(1,1092) = 27.2, *p* < 0.001). The effects of branding varied with the type of soy sauce (interaction F(2,1092) = 3.9, *p* = 0.02). Facial expressions did not vary with time. *Facial expressions of valence*. Facial expressions to images of foods became more negatively valenced when branding information was added (F(1,1092) = 27.2, *p* < 0.001). The effects of branding varied with the type of soy sauce (interaction F(2,1092) = 3.9, *p* = 0.02). Facial expressions did not vary with time.

*Facial expressions of arousal*. Facial expressions related to arousal became intensified with branding information (F(1,1092) = 35.4, *p* < 0.001). Facial expressions varied overall between soy sauces (F(2, 1091) = 3.4, *p* = 0.03), and varied with combination of soy sauce and experience (interaction: F(2,1091) = 4.5, *p* = 0.01). Branding affected familiar and unfamiliar participants differently (interaction: F(1,1091) = 4.3, *p* = 0.04). Facial expressions did not vary with time. Results are shown in Figure 2. *Facial expressions of arousal*. Facial expressions related to arousal became intensified with branding information (F(1,1092) = 35.4, *p* < 0.001). Facial expressions varied overall between soy sauces (F(2, 1091) = 3.4, *p* = 0.03), and varied with combination of soy sauce and experience (interaction: F(2,1091) = 4.5, *p* = 0.01). Branding affected familiar and unfamiliar participants differently (interaction: F(1,1091) = 4.3, *p* = 0.04). Facial expressions did not vary with time. Results are shown in Figure 2.

**Figure 2.** Facial expressions related to valence (left) and arousal (right) during 5 s of viewing, 5 s of anticipation between viewing and tasting, and 6 s of tasting (6–11 s) of unbranded (solid line) and matched branded (dashed line) stimuli. **Figure 2.** Facial expressions related to valence (left) and arousal (right) during 5 s of viewing, 5 s of anticipation between viewing and tasting, and 6 s of tasting (6–11 s) of unbranded (solid line) and matched branded (dashed line) stimuli. Results were averaged across participants and stimuli. Significance levels of branded/unbranded differences are indicated by *p*-values. N.s. = not significant.

#### *3.3. Viewing Phase: RPPG and ECG Heart Rate* Results were averaged across participants and stimuli. Significance levels of branded/unbranded differences are indicated by *p*-values. N.s. = not significant.

*RPPG heart rate.* RPPG heart rates during viewing of the food lowered from 59.7 to 57.9 BPM when branding information was also presented (F(1,1080) = 24.0, *p* < 0.001). RPPG heart rates varied between soy sauces (F(2,1080) = 3.2, *p* = 0.04), and with combinations of soy sauce, familiarity, and branding (F(2,1080) = 5.5, *p* < 0.01). RPPG heart rate did not vary with time. *3.3. Viewing Phase: RPPG and ECG Heart Rate RPPG heart rate.* RPPG heart rates during viewing of the food lowered from 59.7 to 57.9 BPM when branding information was also presented (F(1,1080) = 24.0, *p* < 0.001). RPPG heart rates varied between soy sauces (F(2,1080) = 3.2, *p* = 0.04), and with combinations of soy sauce, familiarity, and branding (F(2,1080) = 5.5, *p* < 0.01). RPPG heart rate did

*Foods* **2021**, *10*, x FOR PEER REVIEW 8 of 13

*PPG heart rate.* PPG heart rates during viewing were 9–10 BPM higher than RPPG heart rates but showed a similar decrease with branding information (from 70.6 to 68.6 BPM, F(1,957) = 78.6, *p* < 0.001). The effect of branding varied with the participants' familiarity (F(1,957) = 10.2, *p* = 0.001). The lowering of PPG heart rate with branding information was larger for unfamiliar participants. PPG heart rate did not vary with time. Results are shown in Figure 3. not vary with time. *PPG heart rate.* PPG heart rates during viewing were 9–10 BPM higher than RPPG heart rates but showed a similar decrease with branding information (from 70.6 to 68.6 BPM, F(1,957) = 78.6, *p* < 0.001). The effect of branding varied with the participants' familiarity (F(1,957) = 10.2, *p* = 0.001). The lowering of PPG heart rate with branding information was larger for unfamiliar participants. PPG heart rate did not vary with time. Results are shown in Figure 3.

**Figure 3.** Heart rates measured by ECG and RPPG during 5 s of viewing, 5 s of anticipation between viewing and tasting, and 6 s of tasting (6–11 s) of unbranded (solid line) and matched branded (dashed line) stimuli. Significance levels of branded/unbranded differences are indicated by *p*-values. **Figure 3.** Heart rates measured by ECG and RPPG during 5 s of viewing, 5 s of anticipation between viewing and tasting, and 6 s of tasting (6–11 s) of unbranded (solid line) and matched branded (dashed line) stimuli. Significance levels of branded/unbranded differences are indicated by *p*-values.

#### *3.4. Tasting Phase: Facial Expressions 3.4. Tasting Phase: Facial Expressions*

*Facial expressions of valence*. Facial expressions became increasing more negatively valenced during tasting (effect of time: F(4,1099) = 7.4, *p* < 0.001). Brand information was associated with more negatively valenced facial expressions in low-frequency users, whereas branding did not affect facial expressions of high-frequency users (interaction: F(1,1099) = 5.8, *p* = 0.02). *Facial expressions of arousal*. Overall, arousal was more intense at the beginning of tast-*Facial expressions of valence*. Facial expressions became increasing more negatively valenced during tasting (effect of time: F(4,1099) = 7.4, *p* < 0.001). Brand information was associated with more negatively valenced facial expressions in low-frequency users, whereas branding did not affect facial expressions of high-frequency users (interaction: F(1,1099) = 5.8, *p* = 0.02).

ing and became less intense there after (F(4,1098) = 27.9, *p* < 0.001). Facial expressions during tasting became more aroused with branding information (F(1,1098) = 27.2, *p* < 0.001). The branding effect was strongest for inexperienced participants (F(1,1098) = 5.5, *p* = 0.02). Expressions of experienced participants were most aroused by Kikkoman, whereas those of inexperienced participants were most aroused by Inproba (interaction F(2,1098) = 3.3, *p*  = 0.04). Results are shown in Figure 2. *Facial expressions of arousal*. Overall, arousal was more intense at the beginning of tasting and became less intense there after (F(4,1098) = 27.9, *p* < 0.001). Facial expressions during tasting became more aroused with branding information (F(1,1098) = 27.2, *p* < 0.001). The branding effect was strongest for inexperienced participants (F(1,1098) = 5.5, *p* = 0.02). Expressions of experienced participants were most aroused by Kikkoman, whereas those of inexperienced participants were most aroused by Inproba (interaction F(2,1098) = 3.3, *p* = 0.04). Results are shown in Figure 2.
