Keywords

FormalPara Key Points

  • The types of foods that people choose to eat and the amounts of those foods consumed are influenced by many factors.

  • For humans, non-homeostatic or psychological factors may play a more important role in food selection and energy intake than do homeostatic or biological factors.

  • Energy density describes the number of calories per gram of food.

  • Foods that are high in energy density are often preferred and more reinforcing, but these foods are typically lower in nutrient density.

  • Motivation to get food is another major factor influencing food selection and intake.

  • Motivation can be measured empirically in the laboratory.

  • Hunger and food restriction can increase motivation to get food, while satiation and monotony can decrease motivation.

  • Finding ways to reduce motivation to eat unhealthier food while increasing motivation to eat healthier food may improve diet quality and reduce obesity.

Introduction

Eating occurs for many reasons. In some cases, eating is driven by physiological signals, such as low blood sugar or hunger pains. More often than not, eating is driven by other, non-homeostatic factors (reviewed by [1]). For example, school-aged children and many adults who have a fixed work schedule eat at a specific time of day, regardless of hunger or when and what was eaten previously. Another example of this is consumption of dessert, which typically occurs at the end of a meal, when hunger is diminished. In the current obesogenic environment, the list of non-homeostatic drivers of food intake is long and increasing, while the physiological mechanisms that regulate food intake remain stable and, may be becoming less relevant for humans [2]. This chapter will explore the factors that drive non-homeostatic feeding, why high-fat/high-sugar foods are more liked and more reinforcing, how factors that influence motivation to eat are measured, and how the current empirical research may be used to make positive changes in eating behavior and diet quality.

Energy Density and Food Choice

Have you ever wondered why we tend to like sweet and fat foods more than other foods? Perhaps it is because we have evolved to consume food that contains a large amount of energy in order to assure our survival through times of famine [3]. When we are born, we have an innate preference for sweet and fat foods [4]. In response to this preference (or perhaps as a cause of this preference), human milk has a high concentration of both lactose (48 % of energy) and fat (45 % of energy) [5]. As we are exposed to new foods throughout infancy and childhood, flavors that are predictive of energy content tend to be preferred over those that predict low energy [6]. Similarly, flavors that are predictive of food spoilage (sour) or poison (bitter) tend to be avoided [6]. This is often referred to as flavor-nutrient learning. Studies in both human and animal models consistently show pairing of novel flavors with greater energy results in greater conditioning and increases in liking for those flavors than when the flavor is paired with lower energy content [7]. This occurs even when an orally presented flavor is paired with an intragastric infusion of the nutrient, but not water, further supporting that the pairing of a flavor with energy increases its liking [8].

Although flavor-nutrient learning has been adaptive throughout evolution, in our current environment, this type of learning may contribute to higher energy intake, poor diet quality, and positive energy balance. For example, flavors that predict greater energy are found in foods that are high in fat (ex., potato chips) or high in sugar (ex., sweetened beverages) or high in both fat and sugar (ex., cookies, cakes, and ice cream). These foods are highly liked and widely consumed, which contributes to greater energy intake [9, 10]. This may also lead to poor diet quality if flavors that predict lower energy (ex., broccoli, carrots, or apples) are not preferred. Improving diet quality may, therefore, rely on fighting our innate dietary preferences in two ways: reducing liking of high-energy-density foods and increasing liking of low-energy-density foods. This is one of the major challenges in dietary interventions today.

One factor that influences the types of foods that we choose is energy density [11]. Energy density is defined as the amount of energy per gram of foods. For example, high-fat, high-sugar foods, such as potato chips, baked goods, and French fries, tend to have a high energy density (>4 kcal/g), whereas low fat, low sugar foods, such as fruits and vegetables, tend to have low energy density (<1 kcal/g). Energy density is often negatively associated with nutrient density (low-energy-density foods tend to be high in nutrient density and vice versa), but this is not always the case. For example, yogurt, cheese, and nuts can be high in energy density, but also contain many beneficial nutrients.

Motivation and Behavior

The choice to engage in a specific behavior (such as ingestive behavior) is driven by our motivations. Motivation can be thought of as a drive toward achieving goals. Motivation that comes from within an individual that is driven by enjoyment of the activity or task itself is referred to as intrinsic motivation. Examples of this are playing games, eating, or talking with friends. Extrinsic motivation is driven by factors that exist outside of an individual and may encourage an individual to engage in an activity that is not intrinsically motivating in order to reach a goal. Examples of extrinsic motivators are money, grades, or reduction in the probability of punishment. Intrinsic and extrinsic motivators may relate to one another. For example, if a student wants to get an A in an American Literature course and also finds reading American Literature enjoyable, both extrinsic and intrinsic motivations would support reading behavior. However, if a person in intrinsically motivated to eat donuts and watch television, but extrinsically motivated to lose weight, a decision must be made as to which motivations are going to prevail. These motivations are also tied into other factors, such as impulsivity, emotion, and memory and can represent very complex processes (reviewed in [12]). This chapter will focus on how these motivations relate to eating behavior specifically, but it should be noted that the interaction between intrinsic and extrinsic motivations plays a role in everyday decision making beyond just food selection.

It is important to study motivation as a way to understand factors that regulate food selection and diet quality. All animals have a primary motivation to reproduce and pass on genes to the next generation. However, in order to do that, energy intake must be sufficient not only for survival but for reproduction and assuring that the offspring survive to reproductive age. This requires sufficient energy intake. Thus, survival of the species depends on locating palatable food, being motivated to consume the food, and having a mechanism to store energy from that food to buffer times of food scarcity.

Once populations have stable access to food and are able to move beyond concerns about food scarcity, they can begin to address the topic of diet quality. Diet quality is also influenced by motivation, but instead of a general motivation to consume enough food, diet quality is influenced by trying to get food that is energy dense. For example, humans are, in general, intrinsically motivated to consume highly palatable foods that are typically of low quality [13]. However, for many people, there are extrinsic motivators which compete with the innate drives to reduce motivation to eat these foods. These motivators include desire to lose weight, social pressure, or knowledge of health risks [11]. It is the relationship between these different types of motivators that ultimately drives food selection and diet quality. In the next sections, we will discuss food-related motivation, how it is measured, and how it relates to diet quality and obesity.

Food Reinforcement: One Type of Motivation

A reinforcer is something that increases the probability of a behavior which it follows. For example, a child that is given a chocolate for cleaning his room may be more likely to clean his room than a child given no chocolate. Food reinforcement is the amount of behavior that a given amount or type of food will support. Food is a commonly used reinforcer for behaviors in both humans and animals. If you have ever been to a zoo or an aquarium and watched the trainers getting the animals to perform tricks, you have seen how food is given to shape and reinforce the desired behaviors. The use of food as a reinforcer throughout our life has also helped shape our behavior as well as our choices and preferences for food.

Measuring Food Reinforcement

There are a number of ways to assess food reinforcement. For example, in animals, one common technique used to measure food reinforcement is called lever pressing. In this case, an animal presses on a lever and, after a certain number of lever presses, a food pellet is given [14]. This is a classic technique used across species to examine, not only food reinforcement, but the reinforcing value of a variety of reinforcers, such as drugs and alcohol. In humans, there are different methods that are used as well. On method involves asking people to make a choice between a portion of a specified food and an alternative reinforcer, such as money. For example, a person may be asked “Would you rather have this piece of chocolate or $0.75?” In this case, the monetary value would keep increasing until the person chose the money over the chocolate. The point at which the person switches would be considered the reinforcing value. Another way to measure food reinforcement is to use a more objective testing paradigm. This can be accomplished using a task that has been modified from the animal literature (lever pressing) to be used in humans. Briefly, participants make responses on a computer mouse under conditions where individuals are working only for access to food or conditions where participants work for food or another alternative, or in some situations, two different types of food [15, 16]. Another important component of this task is that after a reinforcer is earned, it becomes harder to earn the next one. These schedules of reinforcement are referred to as progressive ratio schedules that can increase in different increments. For example, a person could be required to push the mouse button 50 times for the first reinforcer, and the number of button presses could increase by 10 button presses each time or could double each time. The reinforcing value of food is assessed by evaluating the number of responses made for food or alternatives on these progressive ratio schedules of reinforcement [15, 16].

The experimental environment includes two computer stations with a swivel chair in the middle (Fig. 8.1). At one station is the computer on which participants can earn food. The other station has a separate computer on which participants may work for a non-food alternative (such as reading magazines) or for a different type of food (ex., low-energy-density foods on one computer and high-energy-density foods on the other). Based on these relationships, we can analyze a number of relationships. First, the total number of responses can be used as an indication of food reinforcement, with more responses indicating a higher reinforcing value of food. Second, the pattern of responding can be assessed by analyzing the number of responses across the different schedules of reinforcement. Third, the relative reinforcing value of high-energy-density food to a low-energy-density food or non-food alternative can be measured by comparing responses for each type of reinforce within individuals within the same session.

Fig. 8.1
figure 00081

Photograph of the Nutrition and Health Research Laboratory room for testing reinforcing value of food. On the right are two computer stations where participants could work for food and non-food alternatives or for two different types of foods. The table at the top of the picture is where interviews, questionnaire completion, and eating take place. The chair can easily be moved around the room so the participants can engage in activities as they choose. This figure has not been previously published

Full size image

Advantages of the Computer-Based Reinforcement Task

One advantage of using the computer-based food reinforcement task is that it is less subjective than other measures that try and assess why people eat the way they do. For example, asking a person to report how much they like a food or how much they want to eat a food lets them know that the experimenter is measuring their feelings about food and may put an individual in a situation where they are self-conscious about their response. Using the computer-based task does not require people to report how much they want food. They are simply asked to make responses on the computer mouse for food, and when they no longer wish to earn more portions of food, they can stop. People often do not know exactly what is being measured, so they may not be as worried about the responses that they are making. This is not to say that there is no subjective influence over responding on this task. People certainly may be self-conscious about the amount of food that they have earned and may choose not to consume the food. However, we do believe that this task is more objective than other commonly used methods for assessing liking and motivation to eat.

Relationship Between Food Reinforcement and Energy Density

The influence of palatability on food reinforcement is well established. In general, high-energy-density foods that are high in sugar and/or fat are both highly liked and extremely reinforcing when compared to low sugar/low fat foods that may provide more nutrient value, but are rated as less palatable (reviewed in [17]). In addition, binge eating occurs for foods that are high in sugar, fat, or both (reviewed in [18]), but is not observed for foods that are low in fat, such as fruit, or low in sugar and fat, such as vegetables. This suggests that consumption of foods that are high in fat and/or sugar results in neural responses and perhaps neural adaptations that reinforce intake of those foods and, in some cases, lead to dysfunctional consumption patterns. Studies that have examined the reinforcing properties of food have been primarily restricted to foods that are high in fat (potato chips) or high in fat and sugar (cookies or candy bars). For example, Epstein and colleagues have used highly palatable snack foods in adults and children to demonstrate that weight status [19–22], dopamine receptor genotype [23, 24], and treatment with dopamine agonists [25] all influence the reinforcing value of highly palatable foods. In addition, our previous studies have demonstrated that daily intake of a high-energy-density snack food for 2 weeks significantly increases its reinforcing value in obese women while decreasing food reinforcement in non-obese women [19, 21]. All of these studies have been restricted to foods that are high in energy density, and our preliminary data suggest that these effects do not generalize to low-energy-density alternatives.

These findings support the work mentioned above demonstrating that flavor-nutrient learning is stronger for flavors that predict higher energy. These foods become liked and then they are preferred and become reinforcing. In addition to innate drives toward high-energy-density food consumption, these preferences are further reinforced behaviorally. For example, food is often used as a reward by parents for desirable behavior (“If you eat your broccoli, you can have dessert” or “If you use the potty, you can have some M&M’s”). Behind the explicit use of food as a reinforcer, there are implicit rewards in that high-energy-density foods are often associated with celebrations, such as birthday parties and holidays. This association is reinforced several times per year, where on happy occasions where we are socializing with family and friends, we are also eating high-energy-density foods. Finally, this pattern of explicitly and implicitly using food as a reinforcer also has the opposite effect on healthy food. For example, if broccoli eating is something that constantly needs to be rewarded, then broccoli eating is perceived as less desirable. Similarly, when healthy eating is restricted to times that are less celebratory, healthy eating is not perceived as reinforcing.

Relationship Between Food Reinforcement and Obesity

Obese individuals find high-energy-density food more reinforcing than non-obese individuals [16]. This has been shown in both adults [19, 21, 23] and in children [22]. Higher food reinforcement for high energy density is positively related to ad libitum energy intake in the laboratory [23, 24] and to self-reported energy intake outside of the laboratory [23, 24, 26]. These data suggest that obese individuals will work harder to get access to high-energy-density foods than will non-obese individuals. As high-energy-density food becomes more reinforcing for obese individuals, low-energy-density foods and non-food alternatives may become less reinforcing. This is problematic for a number of reasons. First, if low-energy-density foods become less reinforcing, then intake of these foods may be reduced in the diet, contributing to poor diet quality. Secondly, non-food alternatives, such as physical activity, may be important competing motivators with eating, but if these become less reinforcing, then individuals may choose to spend more time eating high-energy-density foods and less time engaged in non-eating activities.

These relationships have been demonstrated in cross-sectional studies, but, to date, it is unclear whether chronic overeating of high-energy-density foods increases their reinforcing value or if individuals who are at risk for obesity or who are already obese have higher levels of food reinforcement. Studies on food reinforcement in children have shown that 8–12-year-old children at risk for obesity find food more reinforcing than those who are normal weight [22]. This suggests that this cross-sectional relationship is established early, but still does not demonstrate the direction of the relationship, since the children with higher food reinforcement are already heavier. Other studies in adults have shown that food reinforcement can be modified by manipulating the intake of high-energy-density snack foods. Specifically, consumption of 300 kcal portions of high-energy-density snack foods for 2 weeks decreases food reinforcement in non-obese participants, but increases it in obese participants [19, 21, 27]. These findings lend support to the idea that the chronic intake of high-energy-density snack foods can shift food reinforcement, but because the increase was observed primarily in individuals who are already obese, we are still unable to determine the directionality of the relationship (Fig. 8.2).

Fig. 8.2
figure 00082

Schematic model of the bidirectional and unidirectional relationships among food reinforcement, intake of high-energy-density (HED) and low-energy-density (LED) foods, energy intake, and obesity. The bidirectional arrows indicate that these factors influence each other in a reciprocal manner, and the unidirectional arrows indicate that the relationship typically moves in only one direction. This figure has not been previously published

Full size image

The current obesogenic food environment may act as a moderator of the relationship between food reinforcement and obesity. For example, if the amount of work required to obtain, to prepare, and to consume food was equivalent between low-energy-density and high-energy-density foods, then we would have the ability to conduct a fair comparison of the relationship between food reinforcement and diet quality. However, our current food environment is such that obtaining, preparing, and consuming high-energy-density foods is often easier than for low-energy-density foods. Fast-food establishments and quick service restaurants are easy for most individuals in the US population to access [28]. The majority of packaged, easy-to-prepare meals are high energy density and are of lower quality than an equivalent meal prepared from scratch [29]. Meals consisting of low-energy-density foods may require more preparation and may take more time than a quick meal or a drive-through meal at a fast-food restaurant. Therefore, in individuals who find high-energy-density foods more reinforcing and low-energy-density foods less reinforcing, it is difficult to imagine a way to alter their diet quality when the work required to obtain, prepare, and consume low-energy-density foods is greater.

Can Food Reinforcement Be Altered?

Food reinforcement is a relatively stable trait. Foods that we find reinforcing and not reinforcing tend to be the same over time. When we test this empirically by measuring food reinforcement at separate time points, we find a strong correlation between tests [30]. These data would suggest that food reinforcement is an established trait that cannot be changed. However, other studies have shown that there are factors that can modify food reinforcement. For example, the reinforcing value of food is higher when individuals are hungry than when they are fed [31]. This suggests that people will consume foods that they might otherwise not be motivated to eat if they are presented when hungry. This strategy is often recommended to parents when introducing new foods to infants, but may also be worth exploring in adults and older children. For example, if you are hungry, eat fruits or vegetables first. This may increase liking of these foods and improve diet quality. Another factor that influences food reinforcement is food variety. People will work harder for different foods than they will for a single, highly liked food [32, 33]. Dietary variety has also been shown to increase energy intake and food liking [32, 34–36]. Individuals who are successful weight losers have diets that are more varied in high-energy-density foods and less varied in high-energy-density groups [37]. Weight loss programs have begun recommending increases in variety of low-energy-density foods as a way to promote lower energy intake and improve diet quality. Finally, recent experience with a food can reduce food reinforcement and food liking [31]. Studies examining sensory-specific satiety have shown that recently eaten food is rated liked less than an uneaten food [35, 38, 39].

One possible strategy for reducing reinforcing value of high energy density is to utilize the principle of monotony. By definition, monotony is when the hedonic properties of a food are reduced after repeated intake over days or weeks. This is an extension of sensory-specific satiety, which is a decrease in hedonic value of an eaten food after a single meal. Although liking is different from reinforcing value, often changes in liking are associated with changes in food reinforcement. Our laboratory sought to determine if monotony could reduce reinforcing value of highly liked snack foods. In order to do this, we first measured baseline food reinforcement. We then provided participants with 14, 300 kcal portions of a highly liked snack food to consume daily. After 2 weeks, participants had food reinforcement measured again. Our hypothesis was that the monotony would reduce food reinforcement. In our first study, in non-obese participants, we found that this treatment did, in fact, significantly reduce food reinforcement [27]. In our next study, we decided to examine two additional questions. First, would obese individuals show the same reductions in food reinforcement? Second, is this effect dependent on portion size? We used the same monotony paradigm described above, except we compared obese and non-obese individuals and examined 0, 100, and 300 kcal portion size conditions. We found that, on average, obese individuals showed an increase in food reinforcement after the 2-week monotony period and that the increase in obese individuals and the decrease in non-obese individuals were dependent on portion size, as it was not observed in the 0 and 100 kcal conditions (Fig. 8.3; [21]). Finally, in a third study, we examined whether we could increase the reinforcing value of healthier food in obese and non-obese individuals using the same paradigm. We found that this effect was specific to high-fat and/or high-sugar snack foods, as it was not observed for fruits and vegetables [19]. When taken together, this series of studies demonstrates that food reinforcement can be altered by recent and prolonged food exposure, but it is in the opposite direction from what would be desired in obese individuals. Future work in our laboratory and others is focusing on understanding the mechanisms that underlie changing food reinforcement in order to improve functional applications.

Fig. 8.3
figure 00083

Influence of snack food monotony on food reinforcement. These graphs represent data compiled across three studies in which participants had food reinforcement tested at baseline and again after 2 weeks of consumption of 300 kcal of high-energy-density snack food. Mean ± SEM number of button presses for high-energy-density snack food across different schedules of reinforcement in non-obese (a) and obese (b) participants at baseline (black circles) and after 2 weeks of daily snack food intake (monotony; white circles). The non-obese participants decrease food reinforcement after monotony, but the obese participants increase food reinforcement after this same manipulation. This figure has not been previously published

Full size image

Implications of Food Reinforcement Research

The goal of food reinforcement research is to understand motivation to eat food. Eating is motivated by many factors other than physiological hunger. For example, often people eat because they enjoy the taste of food or because food provides a source of comfort [1, 2]. Food reinforcement research allows for these factors to be taken into account when examining motivational aspects of food intake. In addition, because the food reinforcement paradigm used in our laboratory is objective, the data are less influenced by social and psychological factors that may motivate dishonesty, such as impression management. The food reinforcement task is simple enough to use in both adults and children and has been shown to be highly reliable in repeated testing (reviewed in [16]). The strength of the methodology allows us to design studies to examine methods to reduce or increase food reinforcement, which is relevant for improving diet quality and for reducing obesity.

One implication of the work we have done so far is that food reinforcement is a predictor of food choice [19], energy intake [26, 40], and obesity [20, 22, 23]. In addition, we have demonstrated repeatedly that food reinforcement can be altered by recent, prolonged exposure to food [19, 21, 27]. Our goal moving forward is to build upon our previous findings and determine ways in which the reinforcing value of high-energy-density foods can be reduced and the reinforcing value of low-energy-density foods can be increased. Some potential mechanisms for reduction of high-energy-density food reinforcement include a longer period of monotony, pairing high-energy-density foods with unpleasant foods or consuming high-energy-density foods only when full. None of these experiments have been conducted, so we do not have any idea whether they will work or not. In terms of increasing the reinforcing value of low-energy-density foods, we also have some potential methods, including increasing variety of fruits and vegetables, teaching people novel ways to prepare fruits and vegetables, and consuming fruits and vegetables when hungry. If we could both decrease the reinforcing value of high-energy-density food and increase the reinforcing value of low-energy-density foods, we could vastly improve diet quality and reduce obesity.

Barriers to Manipulating Food Reinforcement

Although food reinforcement can be measured and manipulated in laboratory studies, there are several barriers to achieving this in the real world. In order for the laboratory-based research to be translated into prevention and treatment strategies, these barriers need to be addressed. The first barrier is food availability. In order to shift food reinforcement away from high-energy-density foods and toward low-energy-density foods, people need to have access to a variety of low-energy-density foods and have limitations on access to high-energy-density foods. Right now, in populations with the poorest diet quality and the highest risk of obesity, access to fruits and vegetables is poor and access to highly palatable, high-energy-density foods is both convenient and relatively inexpensive [41]. This limits the ability of researchers, physicians, and public health practitioners to recommend increasing fruit and vegetable intake. A second barrier to shifting food reinforcement is motivation itself. Studies that have examined factors that contribute to food selection have shown that health-related concerns about diet quality are often less important than factors such as food taste and food cost in people of low socioeconomic status [42]. This suggests that in order to shift reinforcing value of food in a direction consistent with improved diet quality, fruits and vegetables would need to be relatively less expensive than high-energy-density foods and they will need to be perceived as tasting good. There have been many proposals to deal with food cost, but they have been met with resistance from the food industry as well as from consumers [43]. Finally, a third barrier to shifting food reinforcement is time. Often people report that they choose high-energy-density foods because they are convenient and require less time to prepare. In many families, both parents work and children are engaged in after-school activities. This leaves less time for food preparation and makes fast food a desirable option, even in families that place a greater value on health. To complicate things further, many people are dealing with multiple barriers at the same time, which makes shifting food reinforcement that much more difficult.

Conclusions

The reinforcing value of food is an important contributor to diet quality. There are many factors that promote increases in the reinforcing value of high-energy-density foods. In order to improve diet quality and reduce weight gain and obesity, we are going to find ways to shift food reinforcement away from high-energy-density foods and toward low-energy-density foods. Laboratory-based studies have begun to develop, test, and refine potential methods for shifting food reinforcement. Once we identify ways to achieve this, the laboratory-based research can be translated to prevention and treatment strategies to improve diet quality and reduce obesity rates.