Favored Blueberry Fruit Attributes by Japanese Consumers: Preferences and Sensory Evaluation of Blueberry Taste and Texture

Abstract

Fruit sensory characteristics are crucial in terms of decision-making for fruit production and breeding. However, due to the relatively short consumption history of blueberries in the Japanese market, their sensory characteristics have yet to be characterized by Japanese consumers. Over three years, Japanese consumer panels evaluated preferences and sensory attributes of 34 blueberry cultivars such as rabbiteye blueberry (Vaccinium virgatum Aiton) and highbush blueberry (V. corymbosum L.), including both northern highbush blueberry and southern highbush blueberry. The preferences and sensory attributes were rated using the general labeled magnitude scale (gLMS) method. The consumer panels recognized variations in sensory attributes and expressed their preferences for taste and texture. Overall liking was strongly correlated with texture liking (r = 0.66), sweetness (r = 0.58), and flavor intensity (r = 0.42), but weakly with firmness (r = 0.22) and not correlated with sourness or a rough feeling. Additionally, significant interaction effects between perceived and ideal taste intensities on overall liking were found, with sourness perception having a more pronounced effect. The results also suggested potential differences in blueberry preference between Japanese and U.S. consumer panelists. Furthermore, sensory attributes were well correlated with corresponding instrumentally measured taste and texture traits, supporting the usefulness of biochemical and mechanical measurements for evaluating blueberry quality characteristics, particularly texture. This study revealed that blueberry fruit preference was largely determined by sweetness intensity and texture preference, with sourness preference also having a significant influence. There is potential to satisfy Japanese preferences, especially regarding texture. The findings provide valuable insights into Japanese consumer preferences for blueberries and offer a roadmap for consumer-oriented cultivar selection and breeding.

Introduction

Sensory characteristics and preferences of high-value fruit are extremely important for breeders and producers because they determine the market value of their products. In recent years, with improvements in living standards and income, consumers have become more concerned about the quality of fresh produce and are willing to buy even expensive fruit. Sensory evaluation, where panelists (either consumers or specialists) interpret and rate their experiences of food consumption, is a commonly used method to characterize sensory attributes. However, sensory attributes are not equally linked to consumer preferences due to human genetic, ethnic, and empirical factors (Causse et al., 2010; Gilbert et al., 2014). Despite this, sensory evaluation has successfully analyzed sensory attributes and preferences of many fruit crops such as strawberries (Fragaria × ananassa), muscadine grape (Vitis rotundifolia), and blueberries (Vaccinium spp.) (Brown et al., 2016; Saftner et al., 2008; Schwieterman et al., 2014).

Blueberries (Vaccinium spp.) are important economic crops in temperate regions due to their excellent characteristics, such as a harmonious blend of sweet and sour tastes, crispy yet juicy texture, and numerous health benefits (Ma et al., 2018). In Japan, blueberry production has increased in the past few decades, covering an area of 1,052 ha and producing 2,268 t of fruit in 2020 (Ministry of Agriculture, Forestry and Fisheries, 2023). Currently, three blueberry species are commercially cultivated in Japan: northern highbush blueberry (V. corymbosum L.; NHB), southern highbush blueberry (V. corymbosum interspecific hybrid; SHB), and rabbiteye blueberry (V. virgatum Aiton; RE). SHB cultivars were developed from interspecific hybrids between the NHB germplasm and low-chill wild species such as V. darrowii Camp (Ban, 2014; Nishiyama et al., 2021).

As the popularity and consumption of blueberries increase, there is growing demand to develop cultivars that meet consumer preferences. Klee and Tieman (2018) suggested that consumers were dissatisfied with the flavor of many fruits, including blueberries, indicating that improving flavor quality can expand markets. Gilbert et al. (2014) found that U.S. consumers were most likely to purchase sweet berries with intense blueberry flavor. Gilbert et al. (2015) reported that overall liking of blueberries was significantly correlated with sweetness, texture liking, and flavor intensity based on sensory tests on U.S. consumers.

Fruit texture is a major trait influencing consumer preferences and postharvest quality. It has been suggested that the firmness is genotype-dependent in blueberries, so evaluating texture traits can be useful for cultivar selection and breeding (Finn and Luby, 1992). To evaluate berry texture and postharvest life, mechanical properties have been developed based on compression and penetration tests (Giongo et al., 2022; Rivera et al., 2021). Blaker et al. (2014) developed several texture descriptors, such as bursting energy, skin toughness, flesh firmness, mealiness, juiciness, and related them to mechanical measurements. However, few studies have explored the relationship between berry texture and consumer preferences.

The objectives of this study were as follows: 1) To characterize preferences and sensory attributes of flesh blueberries among Japanese consumers, with specific reference to a previous study conducted in the U.S. (Gilbert et al., 2015); and 2) To identify sensory properties and quality parameters that can estimate the preferences of Japanese consumers. The results obtained will provide valuable insights into consumer preferences for blueberries and offer a roadmap for consumer-oriented cultivar selection and breeding.

Materials and Methods

Ethics statement

This research was conducted with the approval of the Research Activity Promotion Committee of the Graduate School of Agriculture and Faculty of Agriculture, Kyoto University. Consumer panels were conducted in the Faculty of Agriculture, Graduate School of Kyoto University.

Plant materials

From 2020 to 2022, a total of 34 cultivars cultivated on a commercial farm located in Wakayama prefecture, Japan were used in this study (34°07'37"N, 135°12'21"E). These cultivars comprised 14 SHB, 10 NHB, and 10 RE blueberry cultivars. Among them, two interspecific hybrid cultivars were included: ‘Pearl River’ (60% V. virgatum and 40% northern highbush) in the SHB group, and ‘Pink Lemonade’ (50% V. virgatum, 17% V. darrowii, and 33% V. corymbosum) in the RE groups (Ehlenfeldt and Finn, 2007; Wang et al., 2012), respectively. Harvest dates of fruit samples are shown in Supplemental Table S1. On each harvest date, cultivars near their appropriate harvest time were selected, and fully purple-colored fruits, except for pink fruits of ‘Pink Lemonade’, were hand-harvested. After harvest, fruits were immediately placed on ice in a cool container and transported to Kyoto University in Kyoto prefecture, Japan (35°01'52"N, 135°47'04"E), where they were kept in the dark at 5 ± 1°C until the sensory tests. At least 200 fruits were harvested for each cultivar, and they were randomly separated into sets for sensory tests and quality measurements by instrumental tools. Berry samples for quality measurements were stored at −80°C until measurements.

Sensory evaluation

Over three years, sensory panels were recruited from students and staff in the Faculty of Agriculture, Graduate School of Kyoto University. Sixty-three blueberry samples were evaluated by 80 panelists, with an average of 26 panelists rating each sample (Table S1). The age range of panelists was from 21 to 38 years old. Prior to the sensory tests, panelists were given an explanation about sensory attributes and trained about the scaling method using the General Labeled Magnitude scale (gLMS) (Fig. 1). This scale allowed improved sample comparisons across panels from different seasons and years (Bartoshuk et al., 2004; Kalva et al., 2014). The panelists were instructed to rate texture and overall liking on a hedonic gLMS (Fig. 1a). The central point of this scale represented “neutral”. The anchor point on the right indicated a +100 rating, representing the ‘strongest imaginable liking’ of any kind. Conversely, the anchor point on the left indicated −100 rating, representing the ‘strongest imaginable disliking’ of any kind. Additionally, panelists evaluated the intensity of sweetness, sourness, flavor, firmness, and rough feeling using a sensory gLMS (Fig. 1b). On this scale, the anchor point on the left represented a rating of 0, indicating ‘no sensation’, while the anchor point on the right corresponded to a rating of 100, indicating the ‘strongest imaginable sensation’ of any kind. Considering the scale to be 100 units, the labels were placed as follows: weak (6 units), moderate (17 units), strong (34.7 units), and very strong (52.5 units) on both scales (Bartoshuk et al., 2004). During the training session, panelists were asked to write down the experiences associated with the endpoints of the two gLMS tests to personalize the scales (e.g., eating your favorite food, listening to your favorite music, time spent with someone you love). The sensory attributes were defined as follows: flavor (the combined perception of taste and aroma), firmness (the force required to rupture berry peels and flesh upon chewing), and rough feeling (texture of the fruit skin, stone cells, and seeds). Additionally, a total of 51 panelists were asked to rate their ideal taste intensities of blueberries on the gLMS upon completion of the sensory tests in 2021 and 2022.

Fig. 1

The hedonic and sensory gLMS used in the sensory evaluation. (a) Hedonic gLMS (−100 to +100; −100 = strongest imaginable disliking, +100 = strongest imaginable liking) and (b) sensory gLMS (0 to +100; 0 = no sensation, 100 = strongest imaginable sensation of any kind).

Sensory evaluations were conducted multiple times each year in June and July during the fruit harvest season. On the day of sensory tests, panelists were asked to refrain from eating any food or beverage except water one hour prior to the test. Three berries were put in a paper cup and served to the panelists, and four to seven samples from different cultivars were served in a random order. Panelists were instructed to eat all the blueberries in each sample in one sitting. They rinsed their mouths with water between samples. All berry samples were served at room temperature, around 25°C.

Determination of soluble solids content (SSC), acidity, fruit size, and mechanical properties

Soluble solids content (SSC) and acidity of berry samples were measured using a Pocket Brix‐Acidity Meter (PAL-BX | ACID F5 Master Kit; Atago, Tokyo, Japan). The berry flesh from three fruits was mixed and homogenized to form one replicate, and sample average scores were calculated from five replicates. In addition, the ratio of SSC to acidity (SSC/acidity) and BrimA (an alternative sweetness index of SSC/TA that measures the balance of Brix and acidity) were calculated (Jordan et al., 2001). BrimA was expressed as (SSC—k × acidity), where k is a constant determined by sensitivity to the acidity of fruits and was set as 6 in this study. This index allows smaller amounts of acid than SSC to make the same numerical change to BrimA.

For fruit size and mechanical measurements, 20 to 30 fruits were used in each sample. Each fruit was weighed, and the diameter and height were measured using digital vernier calipers. Mechanical properties were measured in penetration mode using a texture analyzer (EZ-SX; Shimadzu, Kyoto, Japan) with a 2 mm flat cylindrical probe. The probe punctured a berry at the equator and penetrated to a depth of 60% of deformation at a constant speed of 200 mm·min⁻¹. The force deformation curve was generated based on force (N) and distance (%, deformation) (Fig. 2). Mechanical parameters were defined according to previous studies (Giongo et al., 2022; Rivera et al., 2022). This penetration test analyzed skin toughness [maximum force to break the skin (N)], skin break distance (mm), skin stiffness (N·mm⁻¹), flesh firmness [average force measured during 50–80% of deformation (N)], and the number of peaks in the flesh (Np). Detailed descriptions are shown in Table 1.

Fig. 2

Graphical representation of the penetration test. A 2 mm flat cylindrical probe punctured the berry at the equator and penetrated to a depth of 60% deformation at a speed of 200 mm·min⁻¹.

Table 1

The list of mechanical parameters used for texture profiling in the penetration test.

Statistical analyses

The sensory data obtained from the panelists were subjected to one-way ANOVA, with the cultivars or panelists as variables. Pearson’s correlation analysis was conducted to assess the significance of correlations between sensory and instrumental measurements. Additionally, a two-way ANOVA was performed to test the interaction effects of ideal and perceived taste intensities on overall liking ratings. A Least Significant Difference (LSD) Test was used to test the significance of differences (P < 0.05) among the mean values of ideal and perceived tastes groups. These statistical analyses were carried out using R software (version 4.2.1).

Results

Evaluation of sensory properties and qualities of diverse blueberry samples

1.  Hedonic and sensory ratings by Japanese consumer panels

We observed broad variations in hedonic and sensory ratings across the samples from different cultivars and seasons (Figs. S1 and 2; Table S1). Overall liking varied from −2.4 to 28.3, with an overall mean of 14.9. Texture liking ranged from −2.9 to 26.2, with a mean value of 12.8. Sweetness and sourness ranged from 7.3 to 26.6 and 6.7 to 43.5, respectively. Flavor intensity, firmness, and rough feeling ranged between 15.0 to 37.5, 6.9 to 34.5, and 8.2 to 29.5, respectively. These results showed that the consumer panels could discern sensory differences among the berry samples. The results of ANOVA revealed significant differences in all hedonic and sensory attributes among samples (P < 0.01). There were also significant differences in sensory ratings among the panelists (P < 0.01), indicating that the panelist factor significantly impacted preference and sensory ratings. Additionally, we observed significant differences in sensory ratings between samples from different species. Notably, significant differences in texture attributes between species were observed (P < 0.01); RE samples showed significantly higher scores for firmness and rough feeling than NHB samples, while SHB samples showed intermediate traits (Fig. S2).

2.  Biochemical and mechanical properties

We also observed broad variations in biochemical and mechanical measurements across cultivars evaluated within each year (Table S1). SSC, acidity, SSC/acidity, and BrimA ranged between 7.9 to 15.9%, 0.22 to 1.66%, 7.1 to 58.0, and 1.9 to 12.8 respectively. Skin toughness and skin stiffness ranged between 0.32 to 1.07 N and 0.23 to 0.67 N·mm⁻¹, respectively. Skin break force, flesh firmness, and Nps ranged between 0.69 to 2.56 mm, 0.07 to 0.22 N, and 2.6 to 6.5, respectively. Fruit weight, diameter, and height ranged between 1.46 to 5.30 g, 13.6 to 22.9 mm, and 11.2 to 15.9 mm, respectively.

3.  A comparative study of blueberry preference between Japanese and American panels

To characterize Japanese consumer preference for fresh blueberries, we referred to the sample data and sensory ratings from a previous study reporting U.S. consumer preferences (Gilbert et al., 2015). The distributions of the scores and values in Japan and U.S. studies are shown in Figure 3. Both Japanese and U.S. consumers positively rated the berry samples, while the mean overall liking score (hedonic gLMS scale) given by Japanese consumers was 14.9 ± 6.2 SD, which was lower than the value of 22.6 ± 4.7 SD given by U.S. consumers (Fig. 3a). Moreover, the mean score of texture liking given by Japanese consumers was 12.8 ± 5.1 SD, which was much lower than the value of 25.1 ± 3.0 SD given by U.S. consumers (Fig. 3b). Although SSC values of the samples in both studies exhibited similar value distribution (Fig. 3c), sweetness score (sensory gLMS scale) by Japanese panels was 18.8 ± 4.4 SD, which was lower than the value of 23.0 ± 3.5 SD given by U.S. panels (Fig. 3d). Acidity of the berry samples was higher in this study than in the U.S. study (Fig. 3e), which could explain higher sourness scores by Japanese panels with a score of 20.4 ± 7.8 SD being higher than that of the 15.5 ± 5.7 SD by U.S. panels (Fig. 3f).

Fig. 3

Histograms of sensory ratings and quality parameters of blueberry samples used in the studies in Japan and the U.S. (Gilbert et al., 2015). Normalized distribution of (a) overall liking, (b) texture liking, (c) SSC, (d) sweetness, (e) acidity, and (f) sourness.

Relationship between sensory scores and quality parameters measured by instrumental tools

1.  Relationship between overall liking and sensory attributes

We first conducted a correlation analysis between hedonic and sensory data of individual ratings (Fig. 4). Overall liking showed significant positive correlations with texture liking (r = 0.66, P < 0.01; Fig. 4a) and sweetness (r = 0.58, P < 0.01; Fig. 4b). Overall liking also had a significant positive correlation with flavor intensity (r = 0.42, P < 0.01; Fig. 4c), although some samples were negatively rated despite strong flavor intensity, indicating undesirable flavor in some samples. Overall liking had a weak, but significant, correlation with firmness intensity (r = 0.22, P < 0.05; Fig. 4d). In contrast, we did not find any significant correlations between overall liking, sourness and rough feeling. Next, the panel means of each sample were used for correlation analysis (Fig. S3). Overall liking also had significant correlations with texture liking (r = 0.66, P < 0.01), sweetness (r = 0.72, P < 0.01), and flavor intensity (r = 0.41, P < 0.01).

Fig. 4

Relationship between overall liking and (a) texture liking, (b) sweetness, (c) flavor intensity, and (d) firmness. Pearson’s correlation coefficients (r) are shown within scatterplots. *, ** mean significant correlation at the 5%, 1% level, respectively. Dotted line represents linear regression.

2.  Relationship between sensory attributes and quality parameters measured by instrumental tools

Next, we examined the relationships between sensory ratings and biochemical indices using the mean values of each sample (Figs. 5 and S3). Sweetness was best correlated with BrimA (r = 0.65, P < 0.01; Fig. 5c), followed by SSC/acidity (r = 0.49, P < 0.01; Fig. 5b) and SSC (r = 0.48, P < 0.01; Fig. 5a). On the other hand, sweetness had a negative correlation with acidity (r = −0.45, P < 0.01; Fig. 5d). These results suggested that sweetness was more closely related to BrimA than solely SSC or SSC/acidity.

Fig. 5

Relationship between sweetness perception scores and (a) SSC, (b) SSC/acidity, (c) BrimA, and (d) acidity of RE (○), SHB (■) and NHB (▲). Pearson’s correlation coefficients (r) are shown within scatterplots. *, ** mean significant correlation at the 5%, 1% level, respectively. Mean scores of each sample were used for analysis. Dotted line represents linear regression.

Texture liking was also an important factor contributing to consumer preference for blueberries. Figure 6 shows a significant relationship between texture ratings and mechanical properties. Texture liking had significant positive correlations with skin toughness (r = 0.35, P < 0.01; Fig. 6a) and stiffness (r = 0.45, P < 0.01; Fig. 6b). Additionally, firmness had significant positive correlations with skin toughness (r = 0.46, P < 0.01; Fig. 6c). It could be observed that if the berry samples had similar skin toughness values, panelists experienced stronger firmness in RE samples than those of SHB and NHB. Furthermore, we observed a significant negative correlation between rough feeling and fruit diameter (r = −0.36, P < 0.01; Fig. 6d). This could be related to the ratio of the volume of peel and seeds to whole fruit; when the fruit size increases, the ratio of peel and seeds to whole fruit decreases, and rough feeling intensity also decreases.

Fig. 6

Relationship among texture liking and mechanical properties of RE (○), SHB (■) and NHB (▲) blueberry samples. Texture liking was significantly correlated with (a) skin toughness (SF) and (b) stiffness. (c) Firmness intensity was significantly correlated with SF and (d) Rough feeling was significantly correlated with diameter. Pearson’s correlation coefficients (r) are shown within scatterplots. *, ** indicate significant correlation at the 5%, 1% level, respectively. Dotted line represents linear regression. Mean scores of each sample were used for analysis.

Interaction effects of perceived and ideal taste intensities on overall liking

The ratings for ideal sweetness (Fig. 7a) and sourness (Fig. 7b) varied among the panelists, ranging from 7 to 71 and 0 to 51, respectively. The correlation between ideal sweetness and sourness was insignificant (r = 0.16, P = 0.27), implying that these two variables were independent. Subsequently, ideal and perceived taste intensities were classified into three classes based on the gLMS descriptors (weak: ≤10, medium: 11–30, strong: 31≤). Regarding ideal sweetness, most of the ratings were classified as strong (n = 42), followed by medium (n = 8) and weak (n = 1). As for ideal sourness, the ratings were classified as follows: weak (n = 10), medium (n = 32), and strong (n = 9).

Fig. 7

Effects of perceived and ideal taste intensities on overall liking. Histograms of ideal (a) sweetness and (b) sourness rated by the panelists. Interaction effects of (c) sourness × sweetness, (d) ideal sweetness × sweetness, (e) ideal sourness × sweetness, and (f) ideal sourness × sourness on overall liking (mean ± SE). The same letter indicates not significantly different by the Least Significant Difference (LSD) Test (α < 0.05). Panelists’ intensity ratings were classified as follows: weak: ≤10, medium: 11 to 30, strong: 31≤.

ANOVA revealed significant main effects of perceived sweetness and ideal sourness on the ratings of overall liking (Table 2). However, perceived sourness had no significant effect. Furthermore, we observed significant interaction effects between perceived sweetness × perceived sourness, ideal sweetness × perceived sweetness, ideal sourness × perceived sweetness, and ideal sourness × perceived sourness. Panelists generally made significantly higher ratings of overall liking for berries with strong sourness when perceiving medium or strong sweetness (Fig. 7c). Panelists with a high ideal sweetness preference gave significantly higher ratings for berries with strong sweetness, while they made lower ratings for berries with weak sweetness (Fig. 7d). Additionally, panelists with a high ideal sourness preference gave higher ratings for berries with weak and medium sweetness than other groups, but all panelist groups gave higher ratings for berries with strong sweetness (Fig. 7e). Furthermore, panelists giving a low ideal sourness rating gave higher ratings for berries with weak sourness, while panelists in the high ideal sourness group gave significantly higher ratings for berries with medium or strong sourness (Fig. 7f).

Table 2

Effects of perceived and ideal taste intensities on overall liking.

Discussion

Characterization of Japanese consumer preference for blueberry flesh: A comparison between Japanese and the U.S. panels

Based on the correlation analysis and rating of ideal taste intensities, Japanese panelists favored sweeter berries, and most of them preferred berries with moderate sourness (Figs. 4b, and 7a, b). This preference aligns with previous studies, where significant correlations were reported by Saftner et al. (2008) (r = 0.44, P < 0.01) and Gilbert et al. (2015) (R² = 0.55, P < 0.0001). These results indicated that Japanese consumers prefer sweeter berries, as do U.S. consumers. Conversely, the ideal sourness for Japanese panelists was rated as ‘moderate’ by 67% and ‘strong’ by 15%. Due to varying preferences for sourness, overall liking in this study did not significantly correlate with sourness, differing from the negative correlation found in U.S. consumer panels (R² = 0.55, P < 0.0001) (Gilbert et al., 2015). Additionally, significant interaction effects between ideal and perceived tastes indicated that both sample quality and personal preferences influenced overall liking. Our study highlighted crucial interaction effects between taste characteristics, specifically consumer preference for sourness.

Blueberries received generally lower ratings for overall liking by Japanese consumers compared to U.S. consumers (Fig. 3a), potentially due to multiple factors: the sample factor and the panelist factor. In this study, different blueberry species, including NHB, SHB, and RE cultivars, could have affected hedonic ratings, while only SHB cultivars were used in the U.S. study. SHB cultivars received higher texture liking ratings compared to NHB and RE cultivars (Fig. S4). In the SHB breeding program, fruit texture was a primary selection trait that enhanced consumer acceptance (Blaker et al., 2014). The exclusive use of SHB genotypes could explain the higher texture liking among U.S. consumers. Furthermore, Japanese panels generally rated sweetness lower and sourness higher. The higher acidity in this study’s samples, particularly in NHB cultivars, may have led to increased sourness perception and reduced sweetness due to the mixture suppression effect (Ehlenfeldt et al., 1994). The ideal sourness value given by Japanese consumers was 21 (sensory gLMS scale), which was higher than the value of 16 given by U.S. consumers. This suggests that a slightly stronger perception of sourness would not reduce the hedonic ratings. Regarding panelist factors, the U.S. panelists may have had more experience of consuming blueberries compared to the Japanese panelists, who had limited experience with fresh blueberries. In a U.S. consumer survey (total subjects = 300), 99% of the subjects answered that they had purchased blueberries (Gilbert et al., 2014). The U.S. is the largest consumer of blueberries globally, and the U.S. market consumes 36% of global blueberry production (Pienaar et al., 2022). In particular, the consumption of fresh blueberries has continuously increased while frozen blueberry consumption has remained stable (Trejo-Pech et al., 2024). Due to the lack of availability of fruit texture parameters measured by instrumental tools in the U.S. study, factors influencing the lower texture ratings remain unclear. Our investigation suggests that Japanese consumers initially exhibit a lower preference for blueberries than U.S. consumers, probably due to limited eating experience, but the results imply considerable potential for improving preference in terms of berry texture.

Sensory property and quality parameters to estimate Japanese consumer preferences

This section discusses how preferences and sensory attributes can be predicted using quality parameters measured by instrumental tools. Our findings suggest that sweetness is a crucial factor influencing consumer preference. In this study, sweetness correlated well with values measured by instrumental tools. BrimA showed a better correlation with sweetness than solely SSC or acidity, and SSC/acidity, likely because BrimA measures the balance of SSC and acidity, reflecting the mixture suppression effect of sweet and sour tastes (Frank et al., 1993). Besides SSC, Brix/acidity or BrimA as measures of sweetness, a sweetness index based on the proportion of individual sugar components is a common measure of acceptability of fresh horticultural produce (Beckles, 2012). Che et al. (2009) reported that blueberries mainly accumulate glucose and fructose in similar proportions across species, making it unlikely that sugar composition significantly affects sensory perception. Figure 4c indicates some berry samples had negative or unfavorable flavor. In this study, flavor was defined as a combination of taste and aroma. Over 100 distinct volatile organic compounds (VOCs) related to the blueberry aroma, such as fruity, floral, and minty notes, have been identified in Vaccinium species (Sater et al., 2020). These VOCs influence consumer preferences (Cheng et al., 2020; Ferrão et al., 2020; Ferrão et al., 2022; Gilbert et al., 2015). This study highlighted the importance of sweetness and BrimA, suggesting that more accurate sensory characterization can be achieved by considering blueberry aroma in future studies.

Texture properties were also related to the mechanical values measured instrumentally. Texture liking correlated best with peel properties, expressed as skin stiffness. Saftner et al. (2008) reported that eating quality correlated with sensory textural scores for juiciness, bursting energy, and texture during chewing. Blaker et al. (2014) reported that crispness, a desirable texture in the SHB germplasm, was related to bursting energy, flesh firmness, and skin toughness. Although crispness was not used as a sensory descriptor in this study, skin stiffness may relate to berry crispness because it reflects both skin and flesh properties as the ratio of skin toughness (N) to elasticity (mm). We found that preferred berries had a wide intensity range for firmness and rough feeling, while the cultivars that received the highest texture liking had strong firmness and rough feeling regions (Fig. S5a). Additionally, blueberries with various fruit sizes and skin stiffness levels were preferred, with some preferred berries having a skin stiffness close to 0.5 N·mm⁻¹ (Fig. S5b). These results suggest that berry skin properties, particularly skin stiffness, can predict consumer texture liking.

In conclusion, this study evaluated the preference and sensory attributes of 63 blueberry samples from 34 different cultivars and different seasons using Japanese consumer panels. Overall liking was strongly correlated with texture liking, sweetness, and flavor intensity, but not with sourness or rough feeling. Significant interaction effects were observed between perceived and ideal taste intensities, with sourness perception significantly impacting overall liking. Sensory attributes correlated well with quality parameters measured by instrumental tools, especially sweetness with BrimA and texture liking with skin stiffness. This study provides useful insights for future cultivar selection and breeding strategies to meet Japanese consumer preferences.

Acknowledgements

We deeply thank Mr. Eiji Ido, Japanese blueberry producer in Kainan city, Wakayama prefecture, for providing fruit materials. We are also grateful to the students and staff in the Faculty of Agriculture, Graduate school of Kyoto University for kindly participating in sensory tests.

Literature Cited

•  Ban,  T. 2014. Historical changes in cultivars and recent study trends of blueberries. Hort. Res. (Japan) 13: 185–191 (In Japanese with English abstract).
•  Bartoshuk,  L. M.,  V. B.  Duffy,  B. G.  Green,  H. J.  Hoffman,  C. W.  Ko,  L. A.  Lucchina,  L. E.  Marks,  D. J.  Snyder and  J. M.  Weiffenbach. 2004. Valid across-group comparisons with labeled scales: the gLMS versus magnitude matching. Physiol. Behav. 82: 109–114.
•  Beckles,  D. M. 2012. Factors affecting the postharvest soluble solids and sugar content of tomato (Solanum lycopersicum L.) fruit. Postharvest Biol. Technol. 63: 129–140.
•  Blaker,  K. M.,  A.  Plotto,  E. A.  Baldwin and  J. W.  Olmstead. 2014. Correlation between sensory and instrumental measurements of standard and crisp-texture southern highbush blueberries (Vaccinium corymbosum L. interspecific hybrids). J. Sci. Food Agric. 94: 2785–2793.
•  Brown,  K.,  C.  Sims,  A.  Odabasi,  L.  Bartoshuk,  P.  Conner and  D.  Gray. 2016. Consumer acceptability of fresh-market muscadine grapes. J. Food Sci. 81: S2808–S2816.
•  Causse,  M.,  C.  Friguet,  C.  Coiret,  M.  LéPicier,  B.  Navez,  M.  Lee,  N.  Holthuysen,  F.  Sinesio,  E.  Moneta and  S.  Grandillo. 2010. Consumer preferences for fresh tomato at the European scale: A common segmentation on taste and firmness. J. Food Sci. 75: S531–S541.
•  Che,  J.,  S.  Suzuki,  S.  Ishikawa,  H.  Koike and  I.  Ogiwara. 2009. Fruit ripening and quality profile of 64 cultivars in three species of blueberries grown in Tokyo. Hort. Res. (Japan) 8: 257–265 (In Japanese with English abstract).
•  Cheng,  K.,  B.  Peng and  F.  Yuan. 2020. Volatile composition of eight blueberry cultivars and their relationship with sensory attributes. Flavour Fragr. J. 35: 443–453.
•  Ehlenfeldt,  M. K. and  C. E.  Finn. 2007. G-435 and ARS 96-138, pink-fruited blueberry selections. HortScience 42: 172–173.
•  Ehlenfeldt,  M. K.,  F. I.  Meredith and  J. R.  Ballington. 1994. Unique organic acid profile of rabbiteye vs. highbush blueberries. HortScience 29: 321–323.
•  Ferrão,  L. F. V.,  T. S.  Johnson,  J.  Benevenuto,  P. P.  Edger,  T. A.  Colquhoun and  P. R.  Munoz. 2020. Genome-wide association of volatiles reveals candidate loci for blueberry flavor. New Phytol. 226: 1725–1737.
•  Ferrão,  L. F. V.,  H.  Sater,  P.  Lyrene,  R. R.  Amadeu,  C. A.  Sims,  D. M.  Tieman and  P. R.  Munoz. 2022. Terpene volatiles mediates the chemical basis of blueberry aroma and consumer acceptability. Food Res. Int. 158: 111468. DOI: 10.1016/j.foodres.2022.111468.
•  Finn,  C. E. and  J. J.  Luby. 1992. Inheritance of fruit quality traits in blueberry. J. Am. Soc. Hort. Sci. 117: 617–621.
•  Frank,  R. A.,  N. J.  van der Klaauw and  H. N. J.  Schifferstein. 1993. Both perceptual and conceptual factors influence taste-odor and taste-taste interactions. Percept Psychophys 54: 343–354.
•  Gilbert,  J. L.,  M. J.  Guthart,  S. A.  Gezan,  M. P.  De Carvalho,  M. L.  Schwieterman,  T. A.  Colquhoun,  L. M.  Bartoshuk,  C. A.  Sims,  D. G.  Clark and  J. W.  Olmstead. 2015. Identifying breeding priorities for blueberry flavor using biochemical, sensory, and genotype by environment analyses. PLoS ONE 10: e0138494. DOI: 10.1371/journal.pone.0138494.
•  Gilbert,  J. L.,  J. W.  Olmstead,  T. A.  Colquhoun,  L. A.  Levin,  D. G.  Clark and  H. R.  Moskowitz. 2014. Consumer-assisted selection of blueberry fruit quality traits. HortScience 49: 864–873.
•  Giongo,  L.,  M.  Ajelli,  M.  Pottorff,  P.  Perkins-Veazie and  M.  Iorizzo. 2022. Comparative multi-parameters approach to dissect texture subcomponents of highbush blueberry cultivars at harvest and postharvest. Postharvest Biol. Technol. 183: 111696. DOI: 10.1016/j.postharvbio.2021.111696.
•  Jordan,  R. B.,  R. J.  Seelye and  V. A.  McGlone. 2001. A sensory-based alternative to brix/acid ratio. Food Technol. 55: 36–44.
•  Kalva,  J. J.,  C. A.  Sims,  L. A.  Puentes,  D. J.  Snyder and  L. M.  Bartoshuk. 2014. Comparison of the hedonic general labeled magnitude scale with the hedonic 9-point scale. J. Food Sci. 79: S238–S245.
•  Klee,  H. J. and  D. M.  Tieman. 2018. The genetics of fruit flavour preferences. Nat. Rev. Genet. 19: 347–356.
•  Ma,  L.,  Z.  Sun,  Y.  Zeng,  M.  Luo and  J.  Yang. 2018. Molecular mechanism and health role of functional ingredients in blueberry for chronic disease in human beings. Int. J. Mol. Sci. 19: 2785. DOI: 10.3390/ijms19092785.
• Ministry of Agriculture, Forestry and Fisheries. 2023. Tokusan kaju seisan syukka zisseki chousa. <https://www.maff.go.jp/j/tokei/kouhyou/tokusan_kazyu/index.html> (In Japanese).
•  Nishiyama,  S.,  M.  Fujikawa,  H.  Yamane,  K.  Shirasawa,  E.  Babiker and  R.  Tao. 2021. Genomic insight into the developmental history of southern highbush blueberry populations. Heredity 126: 194–205.
• Pienaar, L., A. Smit, J. Hattingh and K. Cloete. 2022. Developing economic models for South African Blueberry industry: outlook 2022–2031. Bureau for Food and Agricultural Policy, Pretoria.
•  Rivera,  S.,  L.  Giongo,  F.  Cappai,  H.  Kerckhoffs,  S.  Sofkova-Bobcheva,  D.  Hutchins and  A.  East. 2022. Blueberry firmness—A review of the textural and mechanical properties used in quality evaluations. Postharvest Biol. Technol. 192: 112016. DOI: 10.1016/j.postharvbio.2022.112016.
•  Rivera,  S.,  H.  Kerckhoffs,  S.  Sofkova-Bobcheva,  D.  Hutchins and  A.  East. 2021. Influence of water loss on mechanical properties of stored blueberries. Postharvest Biol. Technol. 176: 111498. DOI: 10.1016/j.postharvbio.2021.111498.
•  Saftner,  R.,  J.  Polashock,  M.  Ehlenfeldt and  B.  Vinyard. 2008. Instrumental and sensory quality characteristics of blueberry fruit from twelve cultivars. Postharvest Biol. Technol. 49: 19–26.
•  Sater,  H. M.,  L. N.  Bizzio,  D. M.  Tieman and  P. D.  Muñoz. 2020. A review of the fruit volatiles found in blueberry and other Vaccinium species. J. Agr. Food Chem. 68: 5777–5786.
•  Schwieterman,  M. L.,  T. A.  Colquhoun,  E. A.  Jaworski,  L. M.  Bartoshuk,  J. L.  Gilbert,  D. M.  Tieman,  A. Z.  Odabasi,  H. R.  Moskowitz,  K. M.  Folta,  H. J.  Klee,  C. A.  Sims,  V. M.  Whitaker and  D. G.  Clark. 2014. Strawberry flavor: diverse chemical compositions, a seasonal influence, and effects on sensory perception. PLoS ONE 9: e88446. DOI: 10.1371/journal.pone.0088446.
•  Trejo-Pech,  C. O.,  A.  Rodríguez-Magaña,  H.  Briseño-Ramírez and  R.  Ahumada. 2024. A Monte Carlo simulation case study on blueberries from Mexico. Int. Food Agribus. Man. Rev. 27: 359–377.
•  Wang,  S. Y.,  H.  Chen,  M. J.  Camp and  M. K.  Ehlenfeldt. 2012. Flavonoid constituents and their contribution to antioxidant activity in cultivars and hybrids of rabbiteye blueberry (Vaccinium ashei Reade). Food Chem. 132: 855–864.