Current Research Related to Wine Sensory Perception Since 2010

: Due to socioeconomic changes, the demand for and interest in alcoholic beverages have recently been increasing. Among various alcoholic drinks, consumer preference for wine could be varied. It might be related with its complexity, as wine is known to have characteristics that are di ﬃ cult to understand. Several factors, such as intrinsic or extrinsic factors interact to inﬂuence wine complexity and consumer perception. Many studies have been conducted to understand the complexity of wine, allowing deeper insight into its perception by consumers. Studies over the last decade on wine sensory perception using descriptive analysis and consumer tests were reviewed. Additionally, chemical analysis studies were included because ﬂavor constituents information would contribute to a better understanding of sensory perception. Descriptive sensory analysis of wine was reviewed considering panels and ﬂavor characterization for better understanding wine-related research. Several consumer studies regarding liking / preference, emotion, context, and extrinsic factors to understand consumer preference or perception were reviewed. Research on chemical constituents may a ﬀ ect consumer perception and is thus presented in this review. However, most of the research was focused on a narrow range of wine, thus is missing overall wine category evaluations. Furthermore, evaluations were mainly in wine-producing countries, resulting in limited understanding of an emergent market. This review will help guide wine researchers and industry by providing information on factors inﬂuencing consumer wine perception.


Introduction
Wine is an alcoholic beverage made by fermenting wine grapes, particularly Vitis vinifera L. [1]. The top five wine producing countries are Italy, France, Spain, United States, and Argentina in 2013-2018 [2]. Recently, consumer preference for alcoholic beverages, especially wine, appears to be more affected by globalization, mixed cultures, or the open trade market rather than regional tradition [3][4][5]. In addition to globalization, economic change and the pursuit of pleasure may also have influenced regional drinking patterns [6,7]. These changes have occurred due to the blurring of the borders between nations and cultures [8]. In recent decades, the international wine trade has grown markedly [2]. Its growth was caused by increasing wine consumption in Northern Europe, North America (in the 1990s), and Asian countries [9,10].
Consumers have historically drunk wine for hedonic pleasure [11] and to experience the wide spectrum of sensations when drinking [12]. Perceived wine quality by consumers could be linked to their hedonic liking [13], while a preference for or liking of wine is quite different. In addition, the wine products range from economy to premium, ultra-premium, and artisan by economic value.
preference", "wine complexity", "wine sensory evaluation", "wine sensory volatile", and "wine sensory chemical". Among numerous studies, articles corresponding to each of subtopic were selected. This review will provide an overview of the sensory evaluation of wine, in addition to any related or useful academic insight into the wine industry and researchers.

Panels
Sensory analysis of wine should ideally be carried out by a trained panel to obtain accurate and reliable sensory data [34]. Sensory properties show the perceived differences amongst the samples and can be related to the hedonic responses of consumers [28]. Thus, a consumer-based trained panel has been advocated as best practice because they use consumer-relevant and non-technical-based language [35]. After a series of training and practice sessions including lexicon development, panelists evaluate and rate the attribute intensity of the samples [28] and are not asked to indicate liking or acceptability [36]. However, most wineries still rely on the judgement and opinion of wine experts [37] as opposed to general descriptive analysis conducted by a trained panel. It may be related to that novices find evaluating and detecting differences in wine quality difficult [38], but wine experts commonly screen wines and award a quality score to give a useful indication to consumers [19].
There are different perspectives on which panel would be more suitable for wine descriptive sensory analysis. Trained panels screened by their discriminative ability are specialized for descriptive analysis and typically participate in several training sessions to further develop their discrimination ability [39]. However, in some cases, wine experts were considered as having a better ability to describe and evaluate wine due to wine complexity [34,40,41]. In addition, experts enable to identify wine defects and evaluate how a wine typifies the variety, region, or style it represents [42]. Parr et al. [43] defined the term "experts" as including wine makers, wine-science researchers and teaching staff, professionals (e.g., wine master; wine judges; wine writers; wine retailers), graduate students who majored in viticulture and oenology, or individuals otherwise highly involved with wine. Maitre et al. [44] categorized panels by level of expertise, such as wine producers (working with the product), sensory assessors (tasting the products), a consumer group involving consumers (tasting for pleasure), and experienced consumers (looking for expertise).
Studies have been conducted to compare the discrimination and evaluation ability of trained panels. These aimed to determine their expertise regarding wine quality or characteristics, including complexity and astringency [19,34]. Hopfer and Heymann [19] compared the ability of experts and trained panelists to judge wine quality using 27 Cabernet Sauvignon wines based on the California Wine Competition awards. The evaluation process was different between the two judging groups. Descriptive analysis of the samples was conducted only by a trained panel, which developed terminology and evaluated the intensity of the characteristics. However, experts evaluated quality by dividing samples into five quality categories, ranging from the lowest to the highest quality. When comparing discriminative ability, the trained panel group showed a balanced profile with detectable differences between attributes, while the expert group had high inconsistency among wine judges. Vidal et al. [34] compared the assessment ability of trained assessors and experts based on astringency using six Tannat wines. In their study, both panels could discriminate samples, but trained assessors showed higher ability to discriminate sample astringency. The study mentioned that this may be linked to the training level of the panel and their familiarity with the evaluation process and attribute terms. Furthermore, when experts were asked to focus on only one characteristic, the halo-dumping effect may have occurred. The halo-dumping effect could occur when evaluating a sample that can evoke a mixture of sensations using only one intensity scale. Since the subject is forced to use only one question for two sensations, the second sensation is "dumped" onto the only available scale [45].
Unlike previous studies demonstrating that experts had greater ability to distinguish wine quality than novices [46][47][48][49], the aforementioned research showed that trained panels had better evaluation techniques. This may be explained by the fact that the studies did not just involve wine categorization, but also sensory evaluation involving developing sensory terms and rating intensities. Experts are known to evaluate wine based on their wine experience and knowledge [19,34,50], but this tendency may lead them to assess subjectively and inconsistently. To obtain reliable assessment by wine experts, it is crucial to advance their ability to objectively measure the sensory characteristics of wine [51]. To close the gap between the two types of panels, sensory training sessions will be needed to familiarize experts with descriptive analysis. Second, sufficient time and knowledge should be provided in the training session for descriptive panelists. A consumer-based trained panel has been recommended as the best option for enhancing validity, in addition to assessing relationships with consumer liking data [35].
Thus far, consensus has not been reached regarding whether trained panelists or wine experts are more appropriate for evaluating wine. A descriptive sensory study is conducted with a highly trained panel to objectively evaluate the attributes of the samples and their intensity. However, wine is considered as a very difficult beverage to evaluate due to its complex flavors, alcohol content, and the high knowledge requirements. Overall, the trained panel who were not wine experts had strengths in objectively evaluating samples, but lacked wine-related experience or knowledge. Conversely, experts appeared to have difficulty in objectively evaluating samples, but had abundant wine experience and knowledge. Although experts have some advantages and still participate in evaluation of the related field, the trained panel participating in traditional descriptive analysis would be more appropriate for evaluating wine. However, sufficient education and exposure to wine are required prior to evaluation in order to acquire knowledge and experience.

Flavor Characterization
Descriptive analysis is generally conducted for generating an attributes list of samples and then identifying the factors that most influence the sample characteristics, combined with consumer data. In most cases, panelists develop and evaluate overall characteristics. However, for wine, some specific attributes such as astringency, bitterness, or mouthfeel were evaluated because of their complexity. Different types of wines were evaluated for discrimination between samples, such as red wine [19], white wine [52,53], rosé wine [54], and sparkling white wine [15]. Other studies focusing on a specific attribute, such as astringency [34,55] or bitterness [56], were also conducted.
In those studies, different numbers of samples and ranges of attributes were evaluated. Except for sparkling white wine, most of those studies evaluated more aroma attributes than flavor or taste characteristics. This could potentially be because aroma is considered as an important factor of wine quality. In the study focusing on red wine [19], oak, sweet, and red/dark fruit aromas were positively correlated with quality, while brett, chemical, earthy, fresh/canned vegetable, and sulfur aromas showed a negative correlation with quality scores. In the white wine studies [52], different attributes were evaluated than those in the red wine studies. There were more fruity and floral attributes, and some different characteristics such as petrol and tropical fruit were included. Each sample was related to different attributes, and thus panelists could discriminate differences between samples despite their complexity. In a study comparing the mineral properties of white wine [53], mineral aroma was positively related with a chalky, reduced, and grassy aroma and bitter taste. However, it was negatively associated with ethanol taste and barrel, petrol, and fruit aromas. Rosé [54] and white sparkling wines [15] also had sensory differences in each sample. Interestingly, the characteristics of rosé wine were similar to those of both white and red wine, but white sparkling wine had more similarities to white wine attributes.
Astringency is a major contributor to the perceived quality and complexity of red wine and is regarded as one of the most complex mouthfeel sensations [30,34,57]. Astringency can be defined as a complex tactile sensation related to drying, puckering, and roughing of the mouth epithelia [58]. It is widely known that the phenolic compounds of red wine and their interaction with salivary proteins are closely related to astringent perception [59,60]. Moreover, other tastes such as bitterness, sourness, and even sweetness are able to modulate overall perception of astringency [61]. In a study made by Vidal et al. [34], wine astringency was evaluated by a time-intensity task, and check-all-that-apply (CATA) questions were assessed. Astringency was related to quality score and was highly associated with the complex and suede astringency subqualities. Rinaldi and Moio [55] analyzed the astringency intensities and subqualities of tannin-added wine. They mentioned that pucker, aggressive, adhesive, and dry attributes were negatively associated with astringency perception, whereas it was positively associated with the full-body, rich, and mouthcoat subqualities. Due to the use of different samples and the purpose of the studies differing, no consensus was achieved between the results of both studies. In further studies, it may be desirable to use the same terms for astringency subqualities.
Bitterness is also a major characteristic of wine, in addition to other properties such as sweetness or sourness [56]. Noticeable bitterness can be considered by consumers to be a negative property [56,62]. Sokolowsky and Fischer [56] conducted a study to compare methods for defining wine bitterness among descriptive analysis (DA), time-intensity (TI) analysis, and temporal dominance of sensation (TDS) analysis. Bitterness had a significant and positive correlation with the concentration of wine components, especially ethanol. DA was able to evaluate only the bitterness itself without being affected by other properties. TI could reflect the exact information of the properties related to time passing. However, as TI focused only on bitterness, a halo-dumping effect may have occurred. TDS demonstrated the average sequence of sensations, but as it does not measure intensity, it must be used in conjunction with other methods. It was recommended to combine all aforementioned methods and correlate with analytical measurements to facilitate a comprehensive understanding of bitterness which is a complex perception of wine. When this is not possible, a traditional descriptive sensory method, flavor profile, could be considered as it has "order of attributes perceived" and "aftertaste" aspects in addition to aroma and flavor-by-mouth [63].
Ethanol, the main component of alcohol in wine, is known to influence wine sensory profiles [64,65]. Ethanol concentration in wine is relevant to aroma, flavor, and mouthfeel [63][64][65][66][67]. In the study of Villamor et al. [66], an increased ethanol concentration in wine positively affected bitterness, burning sensation, and chemical/woody/spicy aroma and flavor. On the contrary, it negatively affected fruity, floral, and caramel aroma/flavor. Similarly, King and Heymann [68] showed a positive correlation between alcohol concentration and overall aroma intensity and hot mouthfeel. King et al. [65] also demonstrated the effect of alcohol concentration on the sensory perception using DA. In their study, alcohol had a positive relationship with bitterness, astringency, burning sensation, and mouthfeel sensation such as sharp and gritty, whereas a negative relation was shown between alcohol and fresh fruit and floral aromas. They also mentioned that tasting order by alcohol level may affect wine perception. Though the studies of King and Heymann [68] and King et al. [65] were about alcohol concentration, ethanol is expected to have a similar influence on other attributes of wine.
Studies carrying out descriptive analysis of wine were conducted focusing on specific attributes regarded as dominant aspects of wine, such as astringency, bitterness, or component (e.g., ethanol). The reason for this may be that these characteristics contribute to the complexity of wine and could affect consumer perception of wine as being difficult to understand. Though numerous descriptive sensory studies on some factors aforementioned in the text and mentioned in Table 1 were conducted in the past, more current DA studies paired with consumer acceptability would be helpful to understand present consumers because consumer preferences and wine market trends are dynamic. Different descriptive panels with different cultural backgrounds may use different terms to describe the characteristics of wine because of their languages, food culture, or experiences. In the cross-cultural DA studies of wine, there were differences between panels on verbalization by cultural background of French and Spanish [69], or perception of minerality by the different wine histories and production styles of France and New Zealand [70]. To elucidate general information regarding wine, more studies with considerations such as evaluating new products or panels from emerging markets in Eastern countries are required because the previous studies were focused on Western countries.

Consumer Studies
Consumers can be divided into subgroups by familiarity with wine, such as familiar and unfamiliar consumers [71,72]. The authors defined unfamiliar consumers as individuals "without any formal training or much exposure to product" and familiar consumers as those "without any formal training but with regular exposure to product" [72]. The common feature of these groups is that each individual has their own preference for consumer products, including wine. Consumer tests to measure the degree to which they like and dislike products are conducted with naïve consumers [39]. As diverse factors such as emotion, context, or extrinsic cues can affect consumer liking or perception, research was conducted considering the following factors: elicited emotion during consumption of wine, context effect of place, lighting, and music, and extrinsic factors such as price, label, brand, cultivar, and origin, among others. Furthermore, wine with different characteristics can influence both consumer liking and the emotion elicited [73,74]. The intention of this consumer studies part is to summarize and organize various consumer studies of wine from these aforementioned perspectives. The brief information of those studies is shown in Table 2.

Liking/Preference
Consumer liking or preference represents the interaction between humans and foods [86] and is closely linked with intent to purchase products. General consumer tests are conducted to analyze and compare the liking score between sample sets. For wine studies, consumer evaluations from various perspectives have also been conducted such as studies considering consumer subgroups, clusters, or demographic data, or time-related studies. Many factors must be considered due to the complexity of wine, with the involvement of sensory attributes including intrinsic/extrinsic factors and others such as target consumers and methods.
For understanding consumer opinions, cluster analysis was performed to identify consumer subgroups with different preferences [87][88][89]. Culbert et al. [15] investigated consumer acceptance of Australian sparkling white wine and determined subgroups showing different preferences. The younger consumer group preferred fruit-forward style wines, while the older group preferred those with more complexity. In a study by Hopfer and Heymann [19], consumers were divided by their tendency for liking Cabernet Sauvignon wine. Consumers had different preferences by subgroup but could not classify wines with similar qualities. Both studies obtained the demographic information of each subgroup to understand differences between clusters. However, the later study did not indicate liking and only showed consumer clusters in internal preference mapping. This may be because obtaining this liking data was not the main aim of their investigation.
Silva et al. [75] emphasized that because eating and drinking are dynamic processes, temporal evaluation is required. Indeed, evaluation at a single point provided only limited information of a continuous process. They measured liking at several points throughout the tasting period, namely temporal liking (TL), and the liking score was different per sip. The authors mentioned the importance of the multi-sip approach, as liking can be increased with an additional sip of wine. This is related to hedonic escalation, which is likely to occur in palatable food and beverages with complex flavor combinations, and typically occurs according to the number of sips [90]. It would be very critical to control the evaluation process to be consistent in measuring the acceptability of the same sample repeatedly such as when and how many times to ask liking. Repeated measure analysis of variance could be considered within the same sample evaluation and compared to the first sip of wine [91], but between-sample evaluation throughout the repeated liking measure would be difficult.
In general, little research has been conducted in recent years simply comparing consumer acceptance between samples. This has instead been used as an additional means to identify any factors affecting consumer perception, in addition to comparing or proving methods in the field of sensory science. To understand consumer perception of wine beyond liking and/or preference, simple and quick ways such as check-all-that-apply (CATA) and rate-all-that-apply (RATA) methodologies were used in recent sensory studies [92][93][94]. Comparison studies of wine between DA and CATA [92] or RATA [93] showed that consumer-based approaches had similar discriminability as well as DA. Perception of extrinsic factors [95], cultural differences [94,96], and wine defect [97] using CATA or RATA also showed significant results. Though increasing consumer preference research through various approaches may be considered positive, wine evaluation should be conducted with a careful and critical attitude and from various angles, even in the general consumer test. One of the most important reasons for this is that wine is an alcoholic beverage with complexity, and some sensory attributes such as bitterness or astringency may easily cause palate fatigue [98]. This may have a greater impact on the naïve consumer than on wine experts or trained panels. Further research on familiarity with wine utilizing different wine types, such as red, white, or sparkling, may be helpful for understanding how consumers perceive wine attributes differently by their degree of familiarity. Producers or marketers should also be able to refer to this to set marketing strategies in meeting consumer expectations.

Emotion
The sensory properties of wine can vary markedly and trigger different responses in consumers. When consuming food products, one of the aspects eliciting consumer responses is an emotional response [8]. Emotion research may provide deeper insight into consumer perception. This is different to consumer acceptance measured by the hedonic scale [99], as this could distinguish products of similar price, quality, and other factors in a competitive product space [100,101]. Emotion research was conducted alongside wine consumption or was related to other factors such as context or temporal dominance [8,73,75,76].
Danner et al. [73] pointed out that perception of quality by naïve consumers was highly associated with emotion. High-quality wines were more liked and elicited more positive emotions in consumers than low-quality wines. They also observed that the emotions evoked were highly correlated with willingness to pay. In another emotion study associated with mouthfeel [76], changes in emotions were observed. Adding a mouthfeel agent (e.g., grape seed extract) led to a decrease in liking score and positive emotions such as calm and relaxed, while negative emotions such as irritated increased. This indicates that liking is closely linked to emotions, wherein higher liking of wine is associated with positive emotions, and disliking is associated with negative emotions [73,74]. In a temporal emotion study [75], the three dominant emotions while consuming wine were pleased, comforted, and relaxed. When Mora et al. [8] analyzed the emotion evoked by commercial wine, different emotions were influenced according to the demographic information of consumers, such as sex and age. For joyful and hedonic, only female participants showed significant differences among samples. Overall, males gave higher ratings than females for emotion. Besides, older adults reported higher scores for active, enthusiastic, good natured, joyful, loving, mild, nostalgic, pleasant, satisfied, understanding, warm, and curious than the young and middle-aged groups. Furthermore, the attributes of white wine were associated with more positive emotions than those of aged wine, which were associated with negative emotions in their study.
All wine emotion studies generally adhere to the perspective that liking and emotion are positively associated. Additionally, drinking wine may be considered to generally arouse positive emotions. However, some factors related to disliking or low quality could induce negative emotions during consumption. Assessing emotion research regarding wine tasting has mostly been conducted in assorted contexts over the course of relevant studies. As context alone could influence the emotions of consumers, detailed attention is required to determine how the quality or characteristics of wine itself affect consumer perception. Unlike the research conducted by Mora et al. [8], Ashton et al. [102] showed that consumers generally feel tired and relaxed when consuming red wine. This was not the case for white wine, which did not evoke strong emotional responses. These results show that carrying out an emotion study of wine characteristics also requires careful preparation and attention to detail. Moreover, prudent interpretation is needed when analyzing results related to consumer demographic data such as age and gender to avoid generalization.

Context
In addition to emotion, context also affects consumer perception. Many researchers have considered that environmental context could affect how accurately one can predict real-life consumer behavior and observe interactions between product choice and evaluation [73,[103][104][105][106]. Food served in different dining settings, such as a restaurant versus a cafeteria, could affect consumer acceptance, perhaps due to consumer expectations [106]. Music or sound also changes consumer perception while consuming wine [107][108][109][110]. Though there are only a few contextual studies of wine evaluation, it is clear that context leads to differences in consumer perception.
Danner et al. [73] investigated the relationship between context and liking, emotions, and moods while consuming wine. In their study, the three contexts of a sensory laboratory, restaurant, and home environment were utilized. Presentation, social conditions, and environmental conditions were compared across the three contexts. Unlike previous context studies [105,106,111], there were no significant contextual effects on liking. However, context had a significant impact on several emotions and moods. Spence et al. [77] demonstrated the effect of light colors and background sound on the perception of wine by consumers. The liking and flavor intensity scores were significantly different with differing environmental contexts. Another study conducted by Spence et al. [78] revealed the effect of background music. Participants rated higher for enjoyment when they tasted wine while listening to music compared with consuming in silence. These studies compared how different contexts affect the perception of wine by consumers.
All of the aforementioned studies were well controlled for comparing context, but some limitations may be considered when conducting future research. Familiarity, expectations, and other factors could affect the association between context and acceptability. It is therefore believed that sufficient consideration is required for sample selection. The study conducted by Spence et al. [78] had very limited consumers (n = 26) participating. For a more general conclusion, it may be better to target a larger number of consumers. Though detailed results were only obtained for the wines used in the study, general differences between the three different contexts could be applied to other context studies.

Extrinsic Factors
As consumers often cannot taste wine at the moment of purchase, they make their selection based on product information such as the brand, price, and label [84,112,113]. This information is called extrinsic cues, which combine with intrinsic cues to constitute a product [114]. Compared with other commercial food products, wine has many more extrinsic factors that can affect consumer perception, such as information on origin, variety, wine ageing, awards, bottle shape or weight, and back label [85,115]. Factors such as origin, variety, or wine ageing influence intrinsic characteristics, however, information transforms them into extrinsic factors, as awards, bottle shape or weight, and back label information are. For example, consumers consider expert assessments of wine quality and attributes to be valuable information for their purchase decisions [34]. When regular wine consumers purchase wine, they frequently rely on recommendations from wine experts and other trusted sources [19,42].
Among the numerous studies on the extrinsic factors of wine, research conducted since 2010 includes the following: brand or price [79]; label/back label information [80,81]; label and brand [82]; grape variety/price/region/awards and others [83]; bottle weight [84]; and overall extrinsic cues [85]. Mueller et al. [79] determined that price was a strong positive driver for the liking and tasting experience of consumers, and strongly affected repurchase intent. Kelly et al. [80] mentioned that back label information, especially regarding food-wine pairing, appeals to consumers and may influence their purchase decisions. Similarly, Mueller et al. [81] stated that wine back label information discussing food pairing and taste descriptors was found to have a positive effect on consumer choice. In a study by Muller and Szolnoki [82], the degree of wine experience affected perceived extrinsic cues such as packaging, label, or brand. Chrea et al. [83] showed that price and region were the most important factors for consumers, followed by grape variety, awards, and vintage. Piqueras-Fiszman and Spence [84] mentioned that there was a positive correlation between wine price and quality and the weight of the wine bottle. However, they also stated that further research would be required to reveal the actual perception of consumers. A study carried out by Sáenz-Navajas et al. [85] determined that there was a correlation between quality perception and extrinsic cues such as origin, label design, bottling, awards, and vintage.
In these studies, extrinsic factors were associated with consumer perception and wine purchasing behavior. Extrinsic factors such as brand, label information, including origin and variety, and others may be associated with the wine knowledge of consumers or their familiarity with wine. However, these factors may be less influential to consumers in emerging markets and other factors may be more important such as product price compared to price willing to pay, and sensory properties communicated. More studies like choice experiments [116,117], conjoint analysis [80,95], fine wine instrument [118], or wine product involvement [119,120] paired with intrinsic evaluations will enable to investigate the relationship between extrinsic and intrinsic factors that influence consumer acceptability or selection behavior. Further, studying consumers from emergent markets, or finding consumer segmentation from existing markets, thus differing the influence of these factors, would be beneficial so that the wine industry could respond to meet consumers' expectations.

Chemical Analysis
The perception of wine flavor and aroma is the result of multi-interactions between the vast chemical components and sensory receptor of humans [52,121]. Wine aroma is necessary for identifying the factors involved in consumer preferences for all wine styles [54]. Additionally, other factors such as color, flavor, and mouthfeel could affect consumer perception. Astringency is regarded as one of the most important characteristics of wine. While mouthfeel is described as dryness and roughness rather than a taste [122,123], bitterness is a taste caused by a variety of molecules [124]. Sound knowledge is needed to understand the specific chemical compounds that impart the desirable sensory characteristics of wine [125]. Many researchers analyzed which components affect wine perception associated with aroma, mouthfeel, or other attributes [52,54,124,126,127].
Chemical analysis of wine was generally accompanied by sensory evaluation, such as descriptive analysis and/or consumer tests. This is likely because the purpose of chemical analysis is not just to determine constituents, but also to investigate which components interact with consumer perception of wine such as studies mentioned in Table 3. Gas chromatography (GC), liquid chromatography (LC), high-performance liquid chromatography (HPLC), mass spectrometry (MS), and other physicochemical assessments provided quantitative data on volatile and non-volatile composition [52,128]. Electronic nose (e-nose), electronic tongue (e-tongue), and gas chromatography-olfactometry (GC-O) are also used for characterization and classification [129,130]. Ebeler [131] already reviewed the analytical tools of wine about targeted/non-targeted analysis. However, when evaluated individually, the precise volatile compounds of the "sniffed aroma" of wine provide inadequate information regarding the flavor of the whole system as perceived by the human senses [52,132]. To understand the relationship between complex mixtures of compounds and sensory properties, multivariate statistical methods such as principal component analysis (PCA) and partial least square regression (PLS) are usually conducted [131]. To determine how chemical compounds affect the aroma and sensory properties of wine, volatile compound analysis was conducted. In a study by Álvarez et al. [52], 37 volatile compounds were identified and quantified, including terpenes, alcohols, acetates, ethyl esters, volatile phenols, volatile fatty acids, lactones, aldehydes, and sulfur compounds. Various compounds affect aroma perception, among which are ethyl ester/acetates, fatty acids, and terpenes. These markedly contribute to the flavor of Godello wine. Thiols are widely present in fruits and vegetables and were studied in Chardonnay wine by Capone et al. [126]. A total of 71 volatile aroma compounds were quantified, including thiols, esters, acids, acetate, C6 compounds, aliphatic γ-lactones, norisoprenoids, monoterpenes, ethyl cinnamate, and oxidation-related compounds. Highly potent thiols related to tropical fruit aroma and flavor, such as 3-mercaptohexan-1-ol (3-MH), could be major contributors to Chardonnay wine. In Australian rosé wines, a total of 51 chemical compounds were identified, including 47 volatile compounds (ethyl esters, acetate esters, other esters, alcohols, isoprenoids, acids, carbonyls, and other) and four thiols [54]. In this study, esters were a prominent aroma based on their odor activity value. Other chemical components such as β-damascenone, 3-methylbutyl acetate, ethyl hexanoate, and 3-mercaptohexyl acetate (3-MHA) were also regarded as important. Sauvignon blanc [133] and Malbec [134] wines were also investigated on the relationship between chemical and sensory properties depending on the different origin. Over and above, reviews of wine chemistry and flavor/aroma published before 2011 provide more information about wine perception-relevant flavor/aroma chemical components [121,135,136].
Chemical analysis can be conducted to understand taste and mouthfeel, such as bitterness and astringency, respectively. Chira et al. [124] investigated phenolic compounds, tannin composition, and other chemical components to compare Bordeaux wine (Cabernet Sauvignon and Merlot) with ageing. The correlation between astringency and chemical factors differed based on the variety. However, the study also suggested that chemical components such as tannin and phenolic compounds may regulate wine perception involved with astringency or bitterness.
When chemical analysis studies are accompanied with descriptive sensory analyses, relationships between chemical constituents and sensory perception by human senses can be analyzed. Gawel [30] reviewed the non-volatile chemical contents of wine associated with red wine astringency for a better understanding of the effect of the wine polyphenol-salivary protein interaction. To improve wine quality, Garrido and Borges [32] reviewed wine and grape polyphenols to obtain insight into wine characteristics and knowledge. More chemical and sensory studies considering wine with various cultivars, regions, winemaking techniques, processing methods, and other factors affecting wine quality might be helpful to researchers and the wine industry for developing high-quality wine.

Concluding Remarks
In this review, sensory studies on wine published over the past 10 years were broadly reviewed. These studies included descriptive sensory analysis, consumer evaluation, and chemical analysis. As aforementioned, wine is one of the most difficult beverages to evaluate because of its complex characteristics. The studies provided information regarding the attributes of wine and consumer behaviors for academics and the wine industry. Among the numerous sensory studies of wine, some limitations existed. Firstly, most wine sensory studies conducted in the last decade evaluated only a narrow range of wines as samples, such as the origin and cultivar, and aimed to compare wine-making methods, processing, or other treatments. However, the wine origin or grape cultivar are factors that directly affect the characteristics of wine, therefore, studies comparing different kinds of wine are required to better comprehend consumers' wine perception. Additionally, most wine studies were conducted in Western countries, which are generally where wine is produced. For example, European countries with a long tradition of winemaking such as France, Italy, Spain, and Germany, and more recently in other countries such as the United States, Argentina, Chile, Australia, New Zealand, and South Africa. This may be because of the availability of winemakers and academic training in viticulture and oenology, in addition to the high frequency of wine consumption in these countries. Wine sensory studies conducted in minor wine consumption regions where the wine market is currently expanding may provide a wider and deeper insight into the perception of wine by novice consumers. To obtain deeper insight into consumers, the sensory evaluation of wine has previously been reviewed [28,29]. This review also recommended that more studies should be conducted to understand the attitude of consumers towards wine. Evaluating this in natural situations (e.g., context, food pairing) would also be beneficial. This general review of studies related to wine perception would provide helpful information to the wine industry for understanding consumers and researchers, allowing further evaluation using various categories of wine. In addition, further evaluation to achieve insight into consumers and the relationship between wine quality and actual consumer perception is essential. The knowledge gained could be utilized to meet consumers' hedonic expectations and results satisfaction from wine consumption.
Author Contributions: J.Y. wrote this article, and J.L. conceived, guided, and revised this article. All authors have read and agreed to the published version of the manuscript.