Finding the Sweet Spot: Measurement, Modification, and Application of Sweet Hedonics in Humans

ABSTRACT

Sweetness is a sensation that contributes to the palatability of foods, which is the primary driver of food choice. Thus, understanding how to measure the appeal (hedonics) of sweetness and how to modify it are key to effecting dietary change for health. Sweet hedonics is multidimensional so can only be captured by multiple approaches including assessment of elements such as liking, preference, and consumption intent. There are both innate and learned components to the appeal of sweet foods and beverages. These are responsive to various behavioral and biological factors, suggesting the opportunity to modify intake. Given the high amount of added sugar intake in the United States and recommendations from many groups to reduce this, further exploration of current hypothesized approaches to moderate sugar intake (e.g., induced hedonic shift, use of low-calorie sweeteners) is warranted.

Introduction

Eating is an extremely complex behavior governed by a wide array of constantly changing environmental (e.g., food availability, health messaging), behavioral (e.g., food choice, portion size, temporal patterns), and physiological (e.g., sensory, endocrine, neural) determinants (1). However, consumers consistently report that sensory factors are the primary driver (2). The sensory properties of foods and beverages are not only determinants of purchasing decisions (2–4), but also the most important index of interest in new foods and beverages (5). In addition, they influence whether one self-reports as having “eaten well” (2, 6), outranking other factors such as nutritional quality, ethics/sustainability, or monetary considerations. Although extraordinary events like the COVID-19 pandemic and factors that lead to food insecurity highlight the importance of access to food as a critical element in food choice, the fact that Americans spend <10% of their disposable income on food (7) indicates most have the freedom to indulge their palates.

Sweetness is a sensation that contributes to the palatability of many foods. It is increasingly derived from low-calorie sweeteners (LCSs) (8) as well as nutritive sweeteners inherent in foods or as discretionary sources that are added to foods. The latter now directly contributes a mean of 12.7% of energy for the US population ≥1 y of age (9). There is also an undetermined amount of energy from the addition of nutritive tabletop sweeteners to foods and beverages by consumers, and this is not accounted for in added sugar intake measurements (9). Moreover, although there is definitely evidence to the contrary (10–13), sweetness and sweeteners also reportedly promote intake by contributing to sensory variety (14, 15), appetite (16, 17), digestion (18), endocrine responses (19, 20), and by signaling food safety (21). However, even these mechanisms work in varying degrees through an effect on palatability. Sweetness and sweeteners have been associated with increased (22–24) and decreased (25–27) body weight as well as improved (28, 29) and reduced (30) diet quality. To assess the true impact of sweetness on energy intake and diet quality, it will be essential to better define and measure the impact of sweetness on food and beverage palatability. This article focuses on the effects of the hedonics of sweetness on dietary intake and is complementary to a recently published article on scaling of sweetness and dietary intake (31). This narrative review will first introduce the problem of excess sugar intake by summarizing trends in sweetener use and recommendations to moderate nutritive sweetener intake. This will be followed by a consideration of measurement issues related to sweet taste, particularly hedonics, and selected factors that influence the affective sensation. Finally, strategies to harness sweet preference to meet health recommendations of reducing sugar intake are described.

Current Status of Knowledge

Trends in sweetener use

Added sugars

The US FDA defines added sugars as “sugars that are either added during the processing of foods or are packaged as such,” and these include brown sugar, high-fructose corn syrup, honey, lactose, maltose, molasses, and sucrose (32). The WHO uses a slightly more precise definition of “monosaccharides and disaccharides added to foods and beverages by the manufacturer, cook or consumer, and sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates” and refers to this as “free sugars” (33). Thus, added sugars are free sugars and the primary difference between the 2 terms is that free sugars include sugars that are naturally occurring in fruits and vegetables that have been mechanically disrupted. It must be emphasized that neither free nor added sugars necessarily reflect the overall sweetness level of the diet. Inherent sugars (such as those in fruit) also contribute to sweetness. In addition, LCSs add sweetness with intensities that vary between compounds by many orders of magnitude and in nonlinear ways with graded concentration (34, 35). Further, the sensory impact of added sugars may be diminished by nonsweet food components (e.g., ketchup contains added sugars but is not overtly sweet). Thus, measurement of total sweetness in the diet is not as straightforward as measuring the concentrations of compounds that impart a sweet sensation. Nevertheless, added sugars and LCS contributions can serve as a crude estimator. Trends described below depict data from nationally representative surveys of food intake in America (i.e., the NHANES).

Intake of added sugars increased between the end of the 1970s and the end of the 1990s, with the percentage of total energy from added sugars rising from 14% to 19% in American children and from 12% to 15% in American adults (36, 37). The next decade saw a 24% decline in absolute intake, as total added sugar consumption decreased from 401 kcal/d in 1999–2000 to 307 kcal/d in 2007–2008 (38). Two-thirds of this decrease was due to a reduction of sugars consumed in soft drinks (38).

In America during 2005–2008, boys between 2 and 19 y of age consumed a mean of 362 kcal added sugars/d (or 16.3% of total energy), whereas girls consumed significantly less at 282 kcal/d (or 15.5% of total energy) (39). Adolescent boys (12–19 y of age) consumed the most added sugar at nearly 450 kcal/d (or 17.5% of total energy) (39). In America from 2005 to 2010, the amount of energy consumed from added sugars was significantly higher in adult (>18 y of age) men (335 kcal/d) than in adult women (239 kcal/d), although the percentage of energy from added sugars was not different between the sexes (12.7% and 13.2% of total energy for men and women, respectively) (40). Among adults, individuals 20–39 y of age consumed the most added sugar, with a mean of 397 kcal/d in men (or 14.1% of total energy) and 275 kcal/d in women (or 14.5% of total energy) (40).

Although sugar-sweetened beverage (SSB) consumption has declined recently (41, 42), energy intake from SSBs remains high. In America, nearly two-thirds of children (43) and half of adults (44) consume ≥1 SSB on a given day. This results in SSBs contributing 155 kcal/d [or 41% of daily added sugar energy (39)] for children and 151 kcal/d [or 33% of daily added sugar energy (40)] for adults (42). In 2014, soft drinks provided ∼60% of the SSB calories per day, whereas juice and sports/energy drinks provided ∼15% and ∼25% of the SSB calories per day, respectively (45). On their own, soft drinks contribute 7.1% to total energy, making them the greatest individual contributor to energy intake (46).

LCSs

As added sugar consumption in America has decreased in recent years, LCS consumption has increased markedly (8). This is measured by proportions of consumers and use of products containing LCSs rather than actual intake in grams, because no data are available for the latter. In 1999–2000, only 8.7% of children and 26.9% of adults were LCS consumers, whereas in 2007–2008, 14.9% of children and 32.0% of adults were LCS consumers (47). However, the proportion of children and adults who are LCS consumers may be underestimated for a number of reasons, including the following: 1) consumers may not be aware of LCSs in food or beverage products; 2) producers and manufacturers are not required to provide information about LCS content on labels (48); and 3) current food databases might not accurately capture the rapid changes in the food supply (49). Low- and no-calorie beverages are the main vehicle for LCS consumption, but LCSs are also now found in tabletop sweeteners, grain products, dairy products, desserts, and condiments (8, 50). It was estimated in 2009–2010 that 19.5% of American adults consumed a beverage with LCSs on a given day, whereas only 11.4% and 4.6% consumed tabletop sweeteners or foods with LCSs, respectively (51). Beverages containing LCSs now account for up to one-third of beverages consumed in America (45, 47, 48).

Recommendations to Moderate Sweetener Intake

Added sugars

An upper limit of 10% of total energy was recommended in the 2015–2020 Dietary Guidelines for Americans (DGA), although it was noted that for most total energy amounts, the available energy after meeting nutrient requirements was, in fact, not sufficient to allow the addition of 10% of total energy intake from added sugars (52). The evidence was reviewed again by the 2020–2025 Dietary Guidelines Advisory Committee (DGAC). They concluded that there is strong evidence that eating patterns that limit added sugars are associated with reduced risk of cardiovascular disease and moderate evidence that these eating patterns are associated with a reduced risk of obesity, type 2 diabetes, and some types of cancers (52, 53). Their recommendation was that Americans limit added sugars to <6% energy/d (9), although this limit was not adopted and the recommendation to limit added sugar intake to <10% of total energy intake was retained in the 2020–2025 DGA (53). The basis for the 6% limit was modeling that indicated it would not be possible to meet all nutrient intake goals and remain within specified total energy intake bounds with an intake of added sugars >6% of total energy for diets ranging from 1000 kcal/d to 2800 kcal/d (9). Moreover, even this value is liberal because it was based on assumptions that food choices would be more nutrient-dense than customary choices and no alcohol consumption. It was noted that there can be a tradeoff between saturated fat and added sugar, but even if no saturated fat is ingested, the limit for added sugars increases only modestly. For a diet of 2000 kcal, shifting food choice so that the proportion of saturated fat to added sugar changes from 55%:45% (the ratio of current food choices) to 0%:100% results in an increased allowance of only 132 kcal from added sugars. Thus, the allowance for added sugar is very small.

Recommendations for added sugar, by both the DGAC and other food and health organizations, are largely focused on disease prevention rather than enhancing diet quality. Energy provided by added sugars is not any more obesogenic than energy provided by other foods; however, added sugars are often consumed in foods that are energy-dense and nutrient-poor. Nearly 70% of added sugars intake comes from 5 food categories: sweetened beverages, desserts and sweet snacks, coffee and tea, candy and sugars, and breakfast cereals and bars (54, 55). In a systematic review of added sugars and diet quality, 21 of 22 studies reported an association between higher added sugar intake and poorer diet quality (56). The intake of nutrients such as thiamin, riboflavin, niacin, folate, calcium, iron, zinc, and vitamins A, B-6, and B-12 is negatively associated with added sugar intake (56). In addition, children that consume ≤10% of their energy from added sugars have significantly higher intakes of total fruit, dairy, and whole grains than children that consume >10% of their energy from added sugars (55). There are several approaches for reducing added sugars intake and concurrently improving diet quality. The selection of items from the food categories providing the majority of added sugars may be reduced, especially those with the lowest nutrient density (i.e., items that are not enriched or fortified). Second, discretionary contributions (i.e., table sweeteners) may be moderated. Third, LCSs can be substituted for discretionary nutritive sweeteners. Fourth, foods that contain high added sugars content as well as desired nutrients can be replaced with versions where LCSs substitute for the added sugars.

The influence of total dietary sweetness on diet quality is less clear, however, because total sugar intake is positively associated with calcium and vitamin C intake (56). In addition, NHANES data from 1999–2008 demonstrated that LCS consumption was associated with higher scores on the Healthy Eating Index, which is a measure of diet quality used to assess how well a set of foods aligns with DGA recommendations (29). However, consuming a “diet” version of the same nutrient-poor food product could reduce energy intake to help with weight management, but would not necessarily increase nutrient intake. Sweeteners are not consumed on their own but are, instead, part of a food or beverage that contributes to the diet. Further research is needed to better understand how sweeteners and sweet products fit into a total diet.

Several international health organizations have developed similar recommendations to the 2020–2025 DGA for added/free sugars. In 2015, the WHO recommended that free sugar intake not exceed 10% of total energy intake, with an ultimate goal of reducing free sugar intake to 5% of total energy intake (57). These recommendations were based on evidence indicating an association between sugar and body weight and dental caries. Diabetes Canada, the Scientific Advisory Council on Nutrition in England, and Healthy People 2020 followed roughly the same reasoning as the WHO and recommended similar upper limits for added sugar intake (58–60). The American Heart Association (AHA) has recommended that women and men should consume no more than 100 and 150 kcal/d, respectively, from added sugars (61). This recommendation was made by the AHA in order to “achieve and maintain healthy weights and decrease cardiovascular disease risk while at the same time meeting essential nutrient needs” (61). However, the Institute of Medicine has recommended that added sugar intake not exceed 25% of total energy intake (62), because it reports no adverse effects related to added sugar consumption below this amount.

Overall, ≥60 countries have implemented guidelines or policies in an attempt to curb sugar consumption (63). The policies implemented fall under a number of different categories, including those intended to provide information (e.g., new Nutrition Facts label that includes an amount of added sugar), restrict or eliminate choice [e.g., fewer SSBs sent to schools (64)], guide choice through (dis)incentives (e.g., SSB tax), and enable or guide choice by changing the default [e.g., the food/drink industry reduces overall sugar in products (65)]. Other reviews (36, 65, 66) have conducted a more in-depth exploration of the variety of different policies enacted to reduce sugar intake.

Americans appear to implicitly understand the importance of these recommendations, because a recent survey found that 74% of consumers claimed that they were trying to avoid or limit sugar intake (2). In fact, low sugar is currently the top product claim consumers are looking for when purchasing groceries (67). Still, only ∼10% of the population consumes at or below the recommended amount for added sugars (9).

LCSs

Whereas there tends to be strong agreement that added sugar consumption should be limited, there is far less certainty surrounding recommendations for LCS consumption. The 2020–2025 DGA concluded “replacing added sugars with low- and no-calorie sweeteners may reduce calorie intake in the short-term and aid in weight management, yet questions remain about their effectiveness as a long-term weight management strategy” (53). Similarly, the AHA and American Diabetes Association published a joint statement in 2012 declaring that there are insufficient data to determine whether LCSs are useful for reducing added sugars or providing benefits to appetite, energy intake, weight management, or cardiometabolic risk factors (68). In 2018, the AHA concluded that LCS beverages may be useful for energy intake and weight control but advised against prolonged consumption of LCS beverages by children, because there is some observational evidence of potential adverse effects (e.g., risk of type 2 diabetes and stroke) (69). Public Health England states that replacing sugars with LCSs could be beneficial for energy intake and weight management (70). Alternatively, Canada's Food Guide states that LCSs are not needed for healthy eating and may actually make healthy eating more difficult because foods and drinks with LCSs may replace healthier foods or lead to a preference for sweet foods (71), although there is little direct evidence to support this claim (72). Importantly, various LCSs may have different effects on body weight, energy intake, and appetite, and it may not be appropriate to group them together, leading to uncertainties in recommendations (73).

Sweet Taste Measurement

The sense of taste can be measured by multiple attributes including threshold sensitivity, suprathreshold intensity, temporal patterns, quality, and hedonics. The first 4 dimensions contribute important information, but their impact on feeding patterns and energy intake largely results from how that information is interpreted hedonically (74–81) (Figure 1). Being able to detect sweetness motivates food approach or avoidance based on whether the sensation signals the item is palatable and/or safe. However, sweet hedonics is complex, with multiple measurement approaches and each with different implications for feeding.

FIGURE 1.

The measurement of sweet hedonics in relation to intake pattern and amount. Detection Threshold: the lowest concentration that can be detected relative to a given background at better than chance probability. Recognition Threshold: the lowest concentration to identify quality relative to a given background at better than chance probability. Scaling: intensity ratings. Time Intensity: changes in perceived intensity over time. Descriptive Analysis: assessment of the combination of sensory attributes that characterize selected dimensions (e.g., texture) or the totality of a food's properties.

Perception of sweetness

Sensitivity to sweetness and intensity judgments of sweetness are 2 distinct dimensions of sweetness measurement. Sensitivity (threshold) is a measure of the lowest limits of a stimulus concentration that can be detected (detection threshold), recognized (recognition threshold), or differentiated (discrimination threshold) under a given set of conditions (82, 83). The detection threshold is the lowest concentration of a sweetener in 1 stimulus (e.g., water, food) that permits the determination that it differs from the same vehicle without the sweetener. Thresholds are not an innate characteristic of an individual, rather they are an index of correct performance under a given set of conditions with some predetermined level of confidence. Values will vary not just under different testing conditions but also over time and with changes in a judge's attention, motivation, and physiological status. Recognition thresholds are the lowest concentrations at which a particular quality label (e.g., sweet) is reliably ascribed to the stimulus. Whereas thresholds reflect the lowest functional limits of a sensory system, intensity judgments (scaling) reveal how the magnitude of sensation from suprathreshold (above threshold) stimuli changes with stimulus concentration. Sweetness intensity sensation grows nonlinearly with increments of sweetener concentration. Typically, it increases at a rate that exceeds gradations of sweetener concentration (84). That is, except at very high concentrations, where discrimination is diminished, small changes in sweetener concentration result in disproportionately greater sweetness sensation (83). This holds important implications for determining how adjustments of sweetener concentration may affect hedonic ratings. Small additions or reductions at lower concentrations have a larger impact on sensation than similar changes at high concentrations.

Dimensions of hedonics and approaches to measure each

Hedonics is a multidimensional attribute defined in absolute and relative terms. Reported liking for an item is a judgment in absolute terms. In contrast, preferences are relative impressions that may be based on multiple properties such as ingredient concentrations that affect sensory quality, health beliefs [e.g., one sample is preferred over another because it is lower in sugar (regardless of taste)], brand [e.g., one source (food company, country of origin) is preferred over another], product category [e.g., one type of food (juice compared with solid food) is preferred over another], or social responsibility (e.g., one product is preferred because it has a lower carbon footprint). Preference for an item does not necessarily mean it is liked. It is possible to prefer one food/beverage over another and dislike both. Consequently, different approaches are required to measure liking and preference.

The simplest and most common method to measure liking is an “acceptability test” where the item of interest may be rated on a variety of response formats. Most commonly, liking ratings are made on categorical or visual analog scales (VASs), as shown in Figure 2 (76, 85). The categorical scale has a limited set of response options ranging from 2 to many categories. Larger numbers of categories offer finer sensitivity of responses, but the interpretation of small differences on scales with very large numbers of categories becomes less clear (e.g., a difference of 5 units on a 10-point scale is clear, but on a 100-point category scale, it is probably not perceptually meaningful). The optimal number of categories will be a function of the types of judges conducting the assessment (e.g., cognitive function, task familiarity), level of sensitivity desired by the experimenter, and practicality of analysis. The most common scale is a 9-point scale which allows respondents to provide reasonable discrimination at each level (86) (see Figure 2A). An odd number of categories allows the development of a bipolar scale with a neutral point in the center. The strength of category scales is that they are simple and familiar response scales that can be adapted to almost any audience. A weakness is that the liking difference between 2 sequential categories at the low end of the scale may not be perceptually equal to that between 2 sequential categories at the high end of the scale. Some scales use a quasi-logarithmic distribution of categories to more closely reflect human perceptional differences across graded sensation levels (87, 88). On the other hand, VASs provide a continuous response format without the need for number assignment. Respondents are asked to place a mark along a line anchored with semantic labels at each end that reflect their impression of the liking of the food item (see Figure 2B). Ratings are interpreted by measuring the distance from a set point (typically the low anchor) to the respondent's mark. VASs are slightly more cognitively demanding, may not be practical in some settings, and can pose some additional burden in analyses, although electronic means to administer and analyze such scales are now available.

FIGURE 2.

Hedonic scales. (A) 9-point categorical scale, (B) VAS. VAS, visual analog scale.

Food and beverage hedonics can also be assessed by indirect indexes such as intent to purchase, acute intake, and habitual intake (89). Purchase intent is also commonly assessed by category scales or VASs with scale descriptors such as “definitely would not buy,” “would maybe buy/would maybe not buy,” and “would definitely buy” (90). The assumption is that “intent” or “use” is an index of hedonics, but this is not necessarily the case because items may be consumed for a variety of reasons as long as they are not too unpalatable. In addition, purchase intent measures future intended use rather than actual usage and does not perfectly reflect liking of an item (90). Another indirect measure of liking is the acute intake of an item within a limited time, commonly measured by plate waste. Some evidence suggests this is a better predictor of liking in children (91, 92) than in adults (93, 94), for whom the contribution of nonhedonic factors in behavioral responses varies. In addition, liking may be inferred by measures of habitual intake as assessed by FFQs or dietary records (e.g., 24-h recall, 3-d food diary), but again this is an imperfect measure owing to contributions from many other factors such as cost, availability, convenience, or health beliefs. Whereas disliking is strongly related to low intake (95), the association between liking and intake is weaker (96).

Where a choice is an option, preferences can improve the predictive strength of a hedonic response, because it is possible to have a preference for one item over another despite the 2 being equally liked. In preference testing, respondents may be offered the option of indicating no preference. The advantages are that consumers with no real preference have a logical response option, and the researcher gains insight into the proportion of respondents falling into this category. When respondents are not required to make a decision, each respondent is free to adopt their own criteria as to how strong a feeling they must hold to be willing to choose one sample over the other. This characteristic introduces an uncontrolled amount of variability into the data and may mask subtle distinctions that would ultimately result in a choice of one item over another. It is also important to interpret ratings from measurements made in the laboratory cautiously because the fidelity of their translation to free-living situations where many additional, potentially confounding influences exist is uncertain.

Innate and Learned Influences on Hedonics

Innate influences

The common teleological explanation for the sensitivity to and liking of sweetness holds that these sensory responses developed and were conserved in the human genome during the Paleolithic period to permit hunters and gatherers to identify and ingest foods containing carbohydrate as an energy source. Human breast milk, the sole source of nutrition in early life, contains lactose, a weak sweetener, which may have contributed to the acceptance of breast milk, but the importance of this purported role is unknown. Infants with an intolerance to cow milk and other intact proteins will consume bitter casein hydrolysate formulas if this is the only option. Postweaning, overwhelmingly the availability of carbohydrates would have been in the form of starch, which humans may have some limited capacity to detect (97, 98), but is not sweet. The most effective sweet taste stimuli, such as glucose and sucrose, would have been extremely rare and, when encountered, would have been primarily in energy-dilute sources such as fruits (honey is the exception). Thus, the functionality of sweetness as an inherent determinant of food choice in early hominins is open to question (99).

Sensitivity to sweetness is present in utero (100) and its palatability is demonstrable in neonates (101). Interpreting the hedonic valence of stimuli in these types of trials is problematic. However, similar studies with stimuli rated as pleasant and unpleasant by adults indicate reactions are similar across the lifespan. When saccharin, a sweet substance, is injected into the amniotic fluid, there is an increase in fetal drinking (100). This is in contrast to a decrease in fetal drinking with exposure to lipiodol, a bitter substance (102). In addition, whereas newborns pucker and lick their lips and relax their faces after exposure to sweetness, sour and bitter tastes lead to a gaping response and salivation to dilute and expel the unpleasant stimulus (101). Sweetness is also rewarding to newborns. When exposed to a sweetened gelatin nipple, newborns suck harder and take fewer, shorter pauses than when they are provided an unsweetened latex nipple (103).

Learned influences

There is a well-documented innate basis for individual variability in perceived bitterness of selected compounds (104), but this is not the case for sweetness. Still, there is marked interindividual variability in the preferred level of sweetness of foods and preference for foods that are predominantly sweet compared with savory (105). This reflects a contribution of early and ongoing learning, which starts before birth. Flavors from the mother's diet are transmitted in the amniotic fluid and breast milk (106, 107). Therefore, what the mother eats during pregnancy and lactation can influence the child's flavor preferences (106). Whether the mother's diet influences liking and preference for sweetness is unknown.

There has been some evidence for associative learning about sweetness hedonics in infants. Repeated exposure to sweetness leads to a heightened preference for sweetness (108, 109). Infants who were repeatedly exposed to sugar water consumed more sugar water later in life than those who were not repeatedly exposed to sugar water (110). However, this was specific to the sugar water consumed as an infant, because they did not consume more of a fruit-flavored sugar water later in life. Thus, the associative learning was specific to the food source rather than total sweet food intake. Similar findings were observed in children whose mothers routinely added sugar to their foods (109). Another study reported children who were provided a sweet drink for 8 d developed a preference for sweeter foods generally (111), but long-term effects of sweetness exposure on hedonics are unclear (112). Repeated exposure to flavors in children may inform them which flavors are appropriate in a given food context (111). As opposed to learning through adding flavors into the diet, work with salt and fat demonstrates that the preferred level of these sensory stimuli in foods can be manipulated by reducing sensory exposure to the sensations they impart (113, 114). This is discussed in more detail in the “Approaches to modify sweet hedonics” section.

Individual Variability in Sweet Hedonics and Energy Intake

An innate liking for sweetness is well established (108, 115–117), but this is not a strong determinant of liking of sweet foods or beverages or preferred sweetness levels of specific foods. Sensitivity to sweetness varies between individuals (82, 118) and there is a substantive contribution of learning and social custom to liking (76, 77, 119). For example, within practical limits, liking of sweet beverages has been reported to be inversely associated with its sweet intensity (79, 119), whereas liking of cookies and ice cream has been reported to be positively correlated with sweet intensity (119). Furthermore, preferred sweetness intensity differs between ethnic groups generally and for specific foods (77). Differences may also reflect the various roles sweeteners play in foods and how these properties are valued. Sweeteners also contribute to flavor profiles through an impact on texture, aroma, and color (120). Indeed, sweetness and its role on palatability may be as much a function of other flavor-active compounds as of the sweetener itself (121).

The importance of sweetness liking on food choice also varies by food category (79, 81, 122). It may be more strongly associated with candy and snack intake than with sweet drink consumption, cereal/dairy/fruit products consumption, or added sugar consumption (122). This may reflect the variability in the primary motivation for ingesting different types of products, with sweetness playing a larger role for confections and snacks. Exposure frequency effects are also time-dependent, with recent exposure generally reducing the appeal of a sweet item (112).

Extremes of Hedonics

Cravings

Although there is currently no accepted definition of a craving, it is thought to be a special and intense case of liking that periodically motivates behavior toward obtaining the craved substance (123). In those that experience cravings, >50% of craved foods are predominantly sweet (117). There are several factors that influence sweet cravings (105, 124–131).

One of the greatest influencers of cravings is gender. Nearly all females state that they have experienced cravings, whereas only two-thirds to three-quarters of males state the same (105, 124, 125). Males are more likely to crave savory than sweet foods, whereas females are the opposite (105). Chocolate, traditionally viewed as sweet, is one of the most craved foods, especially among women in America (123, 124,132). Forty-eight percent of females identify chocolate as their most intensely craved food, compared with only 18% of males (126). Culture may help explain this discrepancy, because data from other countries reveal no difference in chocolate cravings between men and women (105,126). Another potential reason for these differences in cravings is that over half of females report cravings to be related to their menstrual cycle (123). In particular, females exhibit greater cravings for chocolate foods immediately before and during menses, although the exact mechanism underlying such cravings is currently unknown (133, 134). In addition, pregnant women are known to experience increased cravings (especially for sweet foods) (135).

Age (125) and alertness (127, 128) are positively associated with cravings, especially for energy-dense sweet foods. Although food cravings were previously theorized to be a result of a deficiency of energy or nutrients (136), a recent meta-analysis found that energy restriction is actually associated with reduced overall and sweet cravings (137). This is possibly due to energy restriction disrupting the association between a stimulus, environment, or occasion and the craving (137). That is, cravings can be specific to a time, place, and set of conditions. It is undetermined if there is a specific diet that works best to limit flavor cravings. Low-carbohydrate diets have been associated with both reduced (130, 138, 139) and increased (131) sweet cravings compared with habitual diets. A role of genetics is uncertain. One study reported those with ≥1 A allele of the fat mass and obesity associated gene (FTO) rs9939609 experienced greater food cravings than the TT homozygotes (129), whereas this was not observed in another study (140).

There is considerable debate as to whether an addiction can develop to sweetness. Evidence from animal studies suggests exposure to sugars can evoke neurochemical and behavioral responses similar to those observed with administration of drugs of abuse (141). However, these responses are more closely linked to palatability than a specific sweetener (including LCSs) (142). The evidence from human trials is less compelling (143, 144). Again, there is a lack of evidence implicating any specific sweetener and assessment of “food addiction” reveals a lack of association with sugar or carbohydrate intake (145).

Aversions

A food aversion is avoidance of a food after its ingestion has been temporally paired with illness (most commonly nausea or vomiting). Over one-third of the general population has formed a food aversion at some point in their life, whereas about one-quarter of the population currently has an aversion (146). However, sweet foods are not a common target of food aversions. In a study of aversive foods, only 18% of aversions involved sweet foods (117), a prevalence that is notably low given the high frequency of exposure to sweet foods. Aversions toward high-protein and high-fat foods are much more common (146, 147). This may be because protein and fat stimulate gastric acid secretion and an increased likelihood of gastric reflux, which may augment any other sensations of malaise (146). In addition, prior dislike of a food is known to facilitate both the acquisition and persistence of aversions, and sweet foods are generally well liked (148).

Monotony

Monotony stems from increased exposure to a food or foods with a common characteristic and may result in reduced appeal and intake of that food (149). The teleological explanation for monotony is that it promotes a greater variety in food choice, thereby increasing the probability of ingesting all needed nutrients and reducing nutrient imbalances and exposure to high amounts of toxins (150). The role of sweetness in monotony is unclear. Some evidence suggests sweet foods are relatively resistant to monotony effects (151, 152). However, the opposite has also been noted because sweet items with high initial pleasantness ratings (e.g., chocolate) may show greater reductions of both pleasantness and desire to eat over time (153, 154). The reason for this discrepancy may be related to consumption frequency. Selected highly liked foods, such as cakes and cookies, have a much lower desired consumption frequency than staple foods, such as bread and milk. Therefore, consumers may experience monotony for sweet foods much more quickly if they are instructed to eat these foods at a greater frequency than would be their habit (155).

Neophobia

Neophobia, the fear of novel foods, may have had survival value as a protection from food toxins for our human ancestors (156). However, neophobia is now relied upon less and may hold negative consequences for children's diet quality (156–158). Neophobia can be overcome by repeated exposure to an item such that preference for the once-feared food increases after it is tasted many times (159). Novel sweet (and salty) foods quickly overcome neophobia, as shown by consistent, rapid increases in pleasantness ratings with repeated exposure, whereas bitter and sour foods are more resistant to exposure effects (160). The reason for this is likely due to the fact that humans demonstrate innate preferences for sweet and salty tastes (161).

Select Influential Factors of Sweet Hedonics

Sex

Whereas some studies report men prefer higher concentrations of sucrose (162, 163) and give higher liking ratings of sweetness (164) than women, others find no differences in preference or liking ratings of sweetness between genders (165–167) and no differences between genders in sweet liking phenotypes (168, 169). Women's sweet preferences vary across the menstrual cycle; however, at which phase they are highest is inconsistent (170–174).

Age

There is a heightened preference for sweetness during childhood and early adolescence that decreases during young adulthood (175–178). Preference for sweetness may be heightened again in older age, because elderly adults preferred higher concentrations of sweetness than young adults (162, 179, 180). Why children and adolescents have a higher sweetness preference than adults is unknown. The effect of age on sweetness preferences is also seen in rats (181). A teleological argument holds that it is due to the energy requirement for physical growth in childhood (118, 176–178, 182). Evidence for this is weak and mixed (183–185) and, contrary to the hypothesis, some data show children have a lower preference for high-fat foods than adults (186). Heightened sweet preferences during childhood and early adolescence are also not related to sweet taste sensitivity (187), nor does a loss in taste sensitivity necessarily explain an increase for sweet preferences in the elderly (162).

Ethnicity and culture

The diverse cuisines around the world highlight the variety of preferred flavor principles of cultural groups (105) and how inherent predilections to favor (sweetness) or reject (bitterness) certain qualities can be overridden (188). That is, through personal experience and cultural influences, one learns where sweetness is appropriate, at what magnitude, and at what frequency. For instance, non-Hispanic black adults consume more added sugar than non-Hispanic white and Mexican-American adults (40), and black adults prefer higher concentrations of sweetness than whites (188) in America. In addition, westernized cultures (Americans, Europeans, Australians) prefer greater sweetness concentrations than do Asian cultures, with further differences between Asian cultures where South Asian cultures (Malaysian, Indian) have greater sweet preferences than Central Asian cultures (Chinese, Japanese, Korean) (189). This is in line with the recent DGA report that non-Hispanic Asians consume less sugar than do other racial and ethnic groups (9).

Genetics

Variations in the sweet taste receptor gene alter sweet taste sensitivity. However, few studies have documented gene variation effects on hedonics. Single-nucleotide polymorphisms located upstream of the TAS1R3 coding sequence strongly correlate with sweet taste thresholds and explain 16% of a population's variation in sweet sensitivity (190). Those with the CC allele have the ability to detect low concentrations of sucrose, whereas those with CT and TT alleles have lesser abilities to detect low concentrations of sucrose. Mothers who have TT alleles on the TAS1R3 sweet receptor gene rs35744813 variant prefer water with a higher concentration of sucrose than those with CC alleles (177, 186). This was not observed in children (177, 186). Other studies have assessed the relations of bitter receptor genotype and 6-n-propylthiouracil (PROP) taster status with sweet liking, but results are mixed (164, 188, 191, 192). It should also be noted that genetically based individual differences in sensitivity to other taste qualities, especially bitterness, can influence sweet exposure because a number of LCSs also activate bitter receptors (193). Thus, susceptibility to sweetness sensitivity determined by the TAS1R3 gene may influence sweetness preferences, but this requires further study.

Approaches to Modify Sweet Hedonics

Recommendations to moderate intake of added sugars to reduce energy intake and risk of chronic disease are widespread. However, reducing added sugars may reduce diet sweetness and palatability, resulting in poor acceptance. Replacing sources of added sugars with LCSs may reduce energy intake while maintaining palatability; however, with limited guidance on LCS consumption from health agencies and continued excess added sugar intake with LCSs widely available, more strategies are warranted. As opposed to maintaining palatability by replacing added sugar with LCSs, several approaches to modify the hedonic response to sweetness have been proposed to aid reduction of dietary sugar and energy intake.

Retrained palate

One approach to reduce added sugar intake is to reduce the preferred sweetness level of foods through a gradual reduction in oral exposure to sweetness. It is hypothesized that a gradual reduction of exposure will lead to a lower preferred sweetness level of foods (a hedonic shift) over time while maintaining palatability. This phenomenon has been observed with salt and fat reduction. In adults who adhered to a reduced-fat or -sodium diet for 12 wk, there was a downward hedonic shift such that pleasantness ratings for high-fat foods declined (114), the preferred fat content of foods declined (114), pleasantness ratings for lower concentrations of sodium increased (194), the acceptance of reduced-sodium foods increased (195), and there was lower sodium intake (195). The hedonic shift occurred and was reinforced by reduced oral sensory exposure to the taste quality rather than altered intake and/or metabolism of fat or salt. This interpretation is supported by the findings that those who reduced sensory exposure to salt or fat, with no change in total intake, experienced the hedonic shift, whereas those that reduced intake but maintained oral exposure did not experience a hedonic shift (114). The shift required ∼8–12 wk to occur. However, in contrast to this literature, a randomized controlled trial that tested this phenomenon with sugar found that a low-sweetness diet increased perceived sweet taste intensity, but differed from the studies with salt and fat in that it had no effect on rated pleasantness or preferred sweetness level in food and beverage samples (196). Whether the increased sweet taste intensity was due to reduced oral sensory exposure specifically was not tested. This change in sweetness intensity in the absence of a change in hedonics would likely not yield changes in dietary intake (80). Whether sweetness differs from tastes of salts and fats in susceptibility to exposure effects on hedonics is not known. Several clinical trials further exploring a hedonic shift in sweetness are underway.

A recent systematic review reported inconsistent evidence for the role of sweetness manipulation on generalized acceptance, preference, choice, and intake of sweet foods in the diet (112). This analysis indicated that in acute randomized controlled trials, higher sweetness exposure tended to reduce preference for sweetness, indicative of sensory-specific satiety. However, population cohort studies and longer-term randomized controlled trials reported equivocal evidence for sweetness exposure and sweetness hedonics and intake. Studies included in the review often only manipulated certain sweet foods or beverages within a diet without controlling sweetness of the entire diet. Future studies controlling for sweetness of the entire diet with no or minimal sweetness exposure compared with higher or present-day average sweetness exposure will better address whether retraining the palate through decreasing sensory exposure to sweetness in the entire diet is effective for reducing preference and liking of sweetness and reducing sugar intake. There is widespread recognition that reduced exposure to a single food can lead to altered preference for that food [e.g., low- compared with high-fat milk (197)]. Whether this would hold for sweet foods and whether targeting the primary sources of sugar intake would be a beneficial strategy for moderating sugar intake have not been tested.

Neonatal programming

Exposure to flavors from the mother's diet in amniotic fluid and breast milk can lead to acceptance of those flavors later in life (106, 107). This has been associated with short-term increased fruit and vegetable consumption. However, whether the mother's diet influences an infant's liking and preference for sweetness in general has not been measured.

Although not an approach to purposefully modify sweet hedonics, there is suggestive evidence of neonatal programming of food choice in those who were born preterm or were intrauterine growth restricted (IUGR). Adults who were IUGR or preterm have lower intake of fruits and vegetables (198) and increased energy and carbohydrate intake (199) compared with those who were born at term and are normal-weight. Studies also report that IUGR and preterm infants have a lower hedonic response to a sucrose solution than term and normal-weight infants. However, this was not the primary outcome of the work (200). Thus, sweet hedonics may be modified by IUGR or birth weight, but research specifically addressing this question is needed.

Conclusions

Humans are drawn to sweetness. Although sensitivity and intensity may play a role, the primary driver of consumption of sweet foods and beverages is their sensory appeal. To quantify sensory hedonics, multiple approaches are required including assessments of liking, preference, and consumption intent. There is an innate desire for sweetness but this is modified by dietary experience, culture, and biology. People learn where sweetness is appropriate and at what level of intensity as well as when such foods can/should be ingested. In the United States, the prevailing forces result in high amounts of sweet exposure through increasing use of LCS-containing products and intake of added sugars (currently 12.7% of daily energy). Although there has been some reduction of sugar intake over the past 2 decades, most recommendations are to reduce this further (e.g., <10% of daily energy according to the 2020–2025 DGA). Several strategies have been proposed to moderate the appeal of sweetness or replace nutritive sweeteners with LCSs to facilitate this change, but the efficacy of such approaches remains to be determined.

Notes

Supported by the Institute for the Advancement of Food and Nutrition Sciences (IAFNS) through an International Life Sciences Institute—North America Low-Calorie Sweeteners and Carbohydrates Committees grant (to RDM). IAFNS is a nonprofit science organization that pools funding from industry collaborators and advances science through the in-kind and financial contributions from the public and private sector participants.

Author disclosures: EC, EJR, SRH, and RDM were commissioned by the Institute for the Advancement of Food and Nutrition Sciences to write this review.

Abbreviations used: AHA, American Heart Association; DGA, Dietary Guidelines for Americans; DGAC, Dietary Guidelines Advisory Committee; IUGR, intrauterine growth restricted; LCS, low-calorie sweetener; SSB, sugar-sweetened beverage; VAS, visual analog scale.