Psychobiological mechanisms in food choice

4 Psychobiological mechanisms in food choice M. R. Yeomans, University of Sussex, UK 4.1 The importance of understanding psychobiological mechanisms in food choice Humans evolved an appetite control system that was designed to protect the body from nutrient shortages and to allow us to exploit food supplies which were scarce. The modern-day consumer no longer faces the everyday pressure of searching out the rare resource of nutritional food which occupied humans during our evolutionary history. However, the modern consumer retains a highly complex appetite control system which is predisposed to allow us to identify and consume safe and nutritious foods. The environment consumers now live in is very different from that which shaped our appetite control systems, and this mismatch has been suggested as a contributory factor to the worldwide increase in incidence of obesity and disordered eating (Zimmet and Thomas, 2003). A key factor in modern food development must be a recognition that our ability to control intake can be compromised by factors such as disguised energy content. This chapter reviews our current understanding of food choice and preference from a psychobiological perspective, highlighting the relationship between food selection and preference and the appetite control system. 4.2 Need-states and hedonic rewards in eating Two key concepts need to be understood. The first is the concept of need-state. This phrase originated in physiological models of controls of eating and drinking, and reflected a general approach to the physiology of motivation which was based on analogies between the way the body may regulate its internal 82 Consumer-led food product development environment (homeostasis) and the way physical control systems operated (e.g. Toates, 1986). This systems approach to motivation is still influential today, as evidenced by the continued use of phrases such as deprivation-induced eating. The basic notion of need-state is that in order for our bodies to function optimally, various biochemical parameters need to be maintained within a narrow physiological range. Examples include temperature, fluid levels and, in relation to food choice, the level of nutrients such as glucose. The basic control theory model of appetite suggests that the body detects perturbations from the ideal level of key physiological variables, and that our appetite is then driven by the need to restore homeostasis. These concepts are critical to food choice, since they make clear predictions about the relationship between the physical state of the body and the likely choices consumers will make. A second critical concept is that of hedonically driven eating: eating for the pleasure of the sensory experience. Could stimulation of appetite by sensory cues underlie the current obesity crisis by causing over-eating? Or are our hedonic responses to food in part a reflection of our need-states? For example, does food taste more pleasant when we are hungry? A full analysis of these questions is beyond the scope of this chapter, and has been attempted elsewhere (Yeomans et al., 2004). However, the current state of our understanding of these issues is summarised in the following discussions, and the need-state versus pleasure debate underlies each of the topics explored in this chapter. The chapter starts with the question of why we like foods, before exploring more explicit psychobiological influences on food choice. To understand these issues fully some knowledge of the mechanisms underlying appetite control is needed, and our current understanding of these complex mechanisms is summarised. The chapter ends by discussing the relevance of these ideas to food production, and how future findings may create new challenges for the food industry. 4.3 Psychobiological influences on acquisition and expression of food preferences The only unequivocal innate flavour preference is for sweet taste, alongside an innate aversion to bitterness (Steiner et al., 2001). However, we have the potential to develop limitless flavour preferences, in the process reversing some of our instinctive reactions such as our aversion to bitterness. Although the complex learning processes that underlie the development of flavour preference are far from fully understood, experimental studies of flavour preference development in humans and other animals has helped identify many of the key processes. The two most important theories are highlighted here, while a more detailed analysis is provided by several reviews (Brunstrom, 2004; Yeomans, 2006a; Zellner, 1991). Psychobiological mechanisms in food choice 83 4.3.1 Flavour±consequence learning (FCL) If we eat a novel food and subsequently become ill, we develop a strong aversion to the flavour of the food even if we know this was not the cause of the illness. Originally described in rats as conditioned taste aversion (CTA: Garcia and Koelling, 1966), this form of learning is now known to be based on similar principles to the conditioning process originally described by Pavlov in his classic work on salivation in dogs. Most consumers are likely to have developed some flavour aversions in this way. However, the principles underlying development of CTA have been used to offer an explanation for how we acquire food preferences as well as aversions. Essentially, one way in which we conceive flavour preferences and aversions is through learned associations between the flavour of the food we are consuming and the effects the food then has on our body once ingested. CTA represents learning where the post-ingestive effect is aversive. However, if the effects are positive (for example, increased energy, reduced hunger or a specific pharmacological benefit such as the effects of alcohol or caffeine), then a flavour preference should develop. This broadening of CTA to a general psychobiological model of preference development through flavour± consequence learning (FCL) remains the most widely cited theory of how our appetite system is coupled with learning to generate food preferences. The ideas behind FCL are heavily influenced by broader concepts in associative learning, based on the original learning principles set out in Pavlov's Nobel Prize winning work on appetitive responses to cues predicting the arrival of food to dogs. Pavlov identified a general principle of learning where responses to a biologically important stimulus (e.g. food, drink: defined as the unconditioned stimulus or UCS) could transfer to a second, neutral stimulus (the conditioned stimulus or CS). Accordingly, in FCL the primary association is thought to be between the perceived sensory characteristics of the ingested food or drink (acting as CS) and the post-ingestive effects of the food or drink (UCS). One way of conceptualising FCL is shown in Fig. 4.1a. Thus, as with other forms of learning, it is predicted that in most situations, changes in preferences generated by FCL will proceed progressively, with repeated experiences of flavour and consequences strengthening the change in preference. But what features of foods are most likely to support FCL? 4.3.2 FCL based on nutrients It is clear that our appetite control system predisposes us to develop a preference for nutritional food while learning to avoid harmful substances and ignore those that have no benefit, such as grass. Most important are flavours that signal sources of energy, since the primary function of eating is to maintain the supply of energy to our bodies. There is a clear relationship between the energy density of foods and self-reported flavour preference (Drewnowski, 1998). Similar data are seen when considering preferences for fruit and vegetables by children, with children preferring those fruit that have the highest energy density (Gibson and 84 Consumer-led food product development Fig. 4.1 A Pavlovian model of the nature of the associations in (a) flavour±consequence learning and (b) flavour±flavour learning. Wardle, 2003). These observations are consistent with the predictions of FCL: past associations between flavours and the ability of foods to reduce hunger (which will depend on the amount of energy we consume) shape our preferences. The strongest evidence for the importance of nutrients in generating FCL comes from studies in animals, where neutral flavours are selectively paired with the post-ingestive delivery of energy (Capaldi, 1992; Sclafani, 1999). A large number of energy-containing nutrients have been shown to induce FCL in rats: sucrose (Capaldi et al., 1994; Fedorchak and Bolles, 1987; Harris et al., 2000; Sclafani, 2002), glucose (Myers and Sclafani, 2001a,b), starch (Elizalde and Sclafani, 1988; Ramirez, 1994; Sclafani and Nissenbaum, 1988), fats (Lucas and Sclafani, 1989) and alcohol (Ackroff and Sclafani, 2001, 2002, 2003). The strength of the acquired preference and consistency of findings both point to FCL as a key mechanism in the development of flavour preferences by animals, consistent with the idea that FCL evolved as a solution to the `omnivore's paradox': how to select nutritious and safe food from the huge variety of potential foods encountered in nature (Rozin and Vollmecke, 1986). FCL has been discussed as an important component of the development of liking for foods with high energy density in humans (Stubbs and Whybrow, 2004). The most convincing laboratory-based studies have been with children, with the reported increase in preference for high-energy (carbohydrate) containing drinks relative to low-energy drinks (Birch et al., 1990) and for high-fat relative to low-fat yoghurt (Johnson et al., 1991; Kern et al., 1993). In adults, consumption of flavours associated with high or low levels of protein at breakfast resulted in a preference for the high-protein version, (Gibson et al., 1995), while preference for the flavour of a soup with added starch increased Psychobiological mechanisms in food choice 85 relative to a soup with no starch (Booth et al., 1982). Thus nutrient-based FCL appears to be a component of flavour preference development in humans. 4.3.3 FCL based on pharmacological components of foods and drinks Some of the most consistent acquired flavour preferences are for drinks that contain substances with psychoactive consequences, such as alcohol and caffeine. Journey to another part of the world and the cuisine may be unfamiliar, but coffee or tea and familiar soft drinks typically dominate the drinks being consumed. Indeed, in these instances liking seems to be contrary to our natural instinct to dislike bitter tastes. FCL provides an obvious framework through which to explain this acquired liking: the specific flavour of the drink becomes a reliable and contingent predictor of the positive post-ingestive effect of the drink. The clearest examples in the laboratory use caffeine as the consequence, and have repeatedly shown acquired preferences for the flavours of drinks paired with caffeine (Richardson et al., 1996; Rogers et al., 1995; Tinley et al., 2003; Yeomans et al., 1998). More recently, the same approach has demonstrated caffeine-based FCL under more naturalistic conditions (Mobini et al., 2005). Here, consumers evaluated their liking for two novel-flavoured iced tea drinks, and then they consumed these at home either at breakfast, at night or whenever they wished. Unbeknown to the participants, one drink contained caffeine and one did not. As can be seen (Fig. 4.2), regular caffeine consumers who drank Fig. 4.2 Changes in rated pleasantness of drink flavours associated with either caffeine (n) or the absence of caffeine (ú) after consuming these drinks at home for four weeks either as a breakfast drink, a night-time drink, consumed anytime or rated but not consumed (control). Adapted from Mobini et al. (2005) with permission. (From Food Quality and Preference, 16, Mobini, Elliman and Yeomans, Conditioned liking for flavours paired with caffeine: a naturalistic study, 659±666, Copyright (2005), with permission from Elsevier.) 86 Consumer-led food product development these drinks at breakfast developed a clear preference for the caffeine-paired flavour relative to the caffeine-free flavour. 4.3.4 Flavour±flavour models of evaluative conditioning (FFL) It is common practice to add something we already like to a new food in order to increase overall acceptability. Thus we may start drinking coffee by adding milk and sugar, or try versions of alcohol where the bitter taste is disguised by a sweet mixer. Or a mother may encourage a toddler to try vegetables by smothering them in cheese sauce. These examples could all be explained by a second form of conditioned flavour association called evaluative conditioning (EC), which involves a change in evaluation of one stimulus by association with a second stimulus that is already liked or disliked (De Houwer et al., 2001; Field and Davey, 1999). Flavour-based EC, also known as flavour±flavour learning (FFL), involves pairing a neutral flavour CS with a liked or disliked flavour UCS. As with FCL, changes in liking are usually interpreted within an associative learning framework based on the principles of Pavlovian conditioning (Fig. 4.1b). For example, sweetness is innately liked, and the addition of sweetness to a wide variety of foods and drinks increases their immediate acceptability. FFL predicts that the simultaneous experience of a new flavour and sweetness results in increased liking for the sweet-associated flavour on its own. This is consistent with the development of liking for flavours as diverse as coffee, tea, beer and wine, yoghurt, etc., all of which are initially consumed in a sweetened form. This idea is supported by laboratory-based studies showing increased liking for sweet-paired food-odours and flavours (Yeomans et al., 2006; Zellner et al., 1983). Conversely, concurrent experience of a new flavour with a disliked flavour such as tween (Baeyens et al., 1990, 1995, 1996) or for a food-related odour with a bitter taste (Yeomans et al., 2006) reliably causes a flavour aversion to develop. Overall, development of a dislike for flavour components consistently paired with an aversive flavour UCS is robust, and alongside CTA explain how human flavour aversions develop. Flavour preferences through FFL with sweet taste UCS may also be influential, particularly in consumers with strong sweet preferences. As with FCL, liking change with FFL requires relatively few pairings of CS and UCS flavours. FFL may therefore be important in human flavour preference development, although as with FCL more research is needed to determine the full scope and importance of these associations. So what is the relevance of all these learning studies for developers of new food products? These studies have important considerations when considering the likely acceptability of new products. For example: · That the reaction of consumers to products will change over time as they acquire associations with consequences through FCL. Thus product development that concentrates solely on first reactions may be a poor predictors of the likely success of a product. Psychobiological mechanisms in food choice 87 · That initial responses to flavours can be modified by knowledge of how existing preferences can be modified through FFL. For example, if a product is developed that has a less than optimal flavour but that has the potential to become liked over time because of its post-ingestive effects, initial flavour acceptability could be enhanced by the addition of sweetness, which will act both to increase initial liking and promote liking for other elements of the overall product flavour through FFL. · That product development must consider closely the expected effects consumers are looking for. A product to be used at mealtimes must be satiating, and this has to be true even of reduced-energy foods. In contrast, a snack aimed at boosting short-term energy needs to be formulated appropriately. In both cases, learning technology could be usefully applied to product development to ensure that the predicted effects deliver what the consumer wants in a way which enhances product acceptability over time. 4.4 Motivational influences on food preferences As we have developed clear and well-supported models of flavour-preference development in the form of FFL and FCL, so it has also become clear that the nutritional needs of the consumer greatly affect both the ability to learn new flavour preferences and how and when these acquired preferences are subsequently expressed. 4.4.1 Motivational states and expression of flavour±consequence learning The idea that the current motivational needs of a consumer influence their hedonic evaluation of a product makes intuitive sense. For instance, an acquired liking for a flavour that predicts that a food has a high energy content would be appropriate when the consumer was hungry, but responding to that same acquired liking when sated could lead to over-eating. Thus, it may be that the body has evolved a mechanism for not only acquiring food preferences, but also determining whether expression of these acquired preferences is an appropriate response. There is empirical support for this idea. In the study of FCL in children with drinks containing energy in the form of carbohydrate (Birch et al., 1990), the acquired preference was much less when the children evaluated the drink when sated than when hungry. If this worked perfectly, and all flavour preferences were based on FCL, then it appears the appetite system underlying these types of preference has evolved in such a way as to guard against overconsumption. The idea that preferences acquired through FCL are sensitive to acute need state are strongly supported by studies with caffeine as UCS (Yeomans et al., 2000a,b). In this instance, state-dependency was found in two ways. Firstly, consumers had to be both caffeine-dependent (Rogers et al., 1995; Tinley et al., 2003, 2004) and acutely deprived of caffeine (Yeomans et al., 1998) in order to 88 Consumer-led food product development develop liking for a novel flavour predictive of caffeine. Thus, the consumer had to be in an appropriate motivational state (both chronically dependent and acutely deprived of caffeine) in order for caffeine to be an effective reinforcer of FCL. Secondly, once caffeine-consumers had acquired a liking for a novel caffeinepaired drink, they expressed that acquired liking only when acutely in need of caffeine (Yeomans et al., 2000a,b). Thus the body seems not only sensitive to the state needed to support the acquisition (learning) of FCL but also acutely sensitive to the relevance of these acquired preferences to the current need, in this instance for the effects of caffeine. Our recent study of acquired liking for caffeine-paired flavours in home consumers adds support to this conclusion. When consumers were allowed to choose when to consume these drinks themselves, those who consistently consumed the caffeinated drink after a period without caffeine developed a preference for the drink, whereas those who consumed their drinks within a couple of hours of consuming tea or coffee did not (Mobini et al., 2005). 4.4.2 Motivational state and expression of flavour±flavour learning While there is clear evidence that expression of flavour preferences developed through FCL are sensitive to the current needs of the consumer, whether this is so with FFL is much less clear. Studies in rats suggest that preferences generated by FFL are unaffected by hunger, whereas preferences based on FCL do. For example, rats were trained with simultaneous pairings of either an odour and sucrose or an odour with saccharin when hungry (Fedorchak and Bolles, 1987) and subsequently tested the expression of these preferences when food-deprived or sated. Hunger modified expression of the sucrose-, but not saccharin-, based preferences, implying that the flavour-energy component of the flavour±sucrose association was sensitive to current energetic needs. Likewise, preferences for odours associated with sucrose, but not saccharin, were greater if rats were trained hungry rather than sated (Capaldi et al., 1994). However, a recent study in humans suggests that FFL is dependent on hunger-state at testing (Yeomans and Mobini, 2006). Here, liking for food-related odours acquired by pairing with a sweet taste, but not a bitter taste or water, was expressed when hungry, but not when sated (Fig. 4.3). This too implies that liking for flavours developed through FFL in humans should be regulated by needs. If both FFL and FCL-based preferences are sensitive to hunger, how then do we explain the over-eating that is partly responsible for the current increases in obesity? The answer probably lies in the accuracy of our appetite control system. As discussed earlier, it makes no sense for an animal to eat just sufficient of a nutritious food if the same food will have disappeared the next time the animal comes to find it. So a system which allows some degree of over-eating makes evolutionary sense. So again, what do these findings mean for product development? The full implications are discussed later, but in brief the key implications are the following: Psychobiological mechanisms in food choice 89 Fig. 4.3 The effects of high- and low-energy soup or water preloads on expression of acquired hedonic orthonasal evaluations of odours following repeated retronasal experience of the same odours paired with either 10% sucrose, 0.01% quinine hydrochloride or water: (a) sucrose-paired odours; (b) quinine-paired odour; (c) control odour. Adapted from Yeomans and Mobini (2006). (From Food Quality and Preference, 16, Mobini, Elliman and Yeomans, Conditioned liking for flavours paired with caffeine: a naturalistic study, 659±666, Copyright (2005), with permission from Elsevier.) · A development process that tests potential new products without taking into account the relevant motivational state of the assessor is likely to fail. For example, if assessors are hungry at the point of testing, but consumers use the product in a low-hunger situation, this mismatch may lead to products failing. Conversely, developing a product with a panel tested at different times of day may be uninformative about acceptability for a product consumers will use in a consistent way: potential snack products need to be assessed at suitable snack times, and meal products at mealtimes. · The development process has to think about likely changes in product acceptability with repeated consumption, consider what components of the product may facilitate these changes, and consider when consumers are likely to use the product. Thus liking may increase with repeated consumption for a product where the flavour incorporates novel aspects alongside a known liked quality (e.g. a sweet taste). But a product developed to enhance satiety, where liking change may be through FCL, needs to ensure the product is consumed primarily when hungry. This adds a role for how products are marketed to the development process. 4.5 Motivational influences on food choice The previous section concentrated on how our motivational state altered our liking for foods, with clear evidence that flavours are liked more when hungry than when sated. But how does motivation relate to our selection of foods given a number of alternative options (food choice)? Psychobiological studies in this area are much more limited in scope because of the inherent complexity in what is being asked. 90 Consumer-led food product development As we have seen, liking for the flavour of foods may increase by association with other food qualities (either in terms of flavour or nutrients). It is also known that liking is a very important determinant of food choice: people rarely consume food they do not like (De Castro, 2000), instead choosing foods for which they have acquired a liking. Thus the primary psychobiological influence on food choice is sensory hedonics, eating for pleasure. We also know a great deal about the biological basis of the pleasure experience from eating (Berridge, 2003). The main neural system underlying hedonic experience in general involves activation of opioid peptides (Kelley et al., 2002), our own brain opiates. When opioid receptors are blocked, food tastes less pleasant (Yeomans and Gray, 2002), and both humans (Yeomans et al., 1990) and animals (Cooper et al., 1985) consequently are less likely to choose foods normally considered palatable over less palatable options. These data together show that hedonic experience is critical to food choice. How then does manipulation of motivational state alter food choice? The answer here is less clear. One important study, however, showed clearly that hunger-state was an important factor in food choice independent of sensory hedonics (Tuorila et al., 2001). In that study, women rated their liking for fatfree or regular energy hot fudge to be consumed on ice-cream which also contained either no fat or regular fat. They were then informed of their evaluations so that their choice was made with the explicit knowledge of which flavour they had rated as the more pleasant. They were then given a choice test between the two samples. Importantly, when tested, less hungry participants were more likely to chose the fat-free versions even though many of these participants had rated this as the less pleasant taste. The implication is that people may use information about energy content to guide food choice, contrary to their hedonic preference, but that hunger state can reverse this effect. A further important influence on food choice relates to personal goals of the consumer. Thus people who have adopted a restrained style of eating either in order to reduce the risk of weight gain or in an attempt to lose weight may actively seek out foods that are less preferred over higher-energy, more preferred options. Although well intentioned, it is also clear that factors such as attempts to diet, especially where they have failed repeatedly (yo-yo dieting), may be counter-productive by disrupting more appropriate food choice behaviours (Mela, 2001). Indeed, the very act of restricting access to favourite foods may have the counter-productive effect of making these overvalued. For example, when women were required to carry a bag of chocolates with them for a day without eating them, they subsequently consumed twice as much when given free access to the same food relative to controls who had not had to deal with the temptation of free availability of a preferred food (Stirling and Yeomans, 2004). This effect was exacerbated in women with a history of restrained eating (Fig. 4.4), clearly showing that repeated attempts to avoid certain foods may lead to a reduced ability to resist these foods. In summary, this section highlights the importance of anticipated pleasure in food choice, and how this may be modified by hunger state at the time when Psychobiological mechanisms in food choice 91 Fig. 4.4 The effects of self-denial on intake: chocolate consumed by high- (ú) and lowrestraint (ú) women who had been required to carry but not consume chocolates for 24 h beforehand (temptation) and a control group with free access to chocolate. Adapted from Stirling and Yeomans (2004). food is chosen. Longer-term self-restriction appears to heighten sensitivity to highly palatable foods, suggesting that, although we are able to direct our food choices based on what we consider healthy, over-rigid self-restriction is counterproductive. However, these conclusions have to be seen as tentative since the literature on motivational effects on food choice remains sketchy relative to the equivalent literature on acquired food preferences and on our understanding of the control of food intake. In order that future products can be developed that maximise our understanding of psychobiological influences on consumer behaviour, further research is urgently needed to help clarify these issues. 4.6 Motivational influences on food intake Historically, the study of motivational influences on control of food intake has been the main focus of the psychobiological investigation of appetite, and a full discussion of factors involved in food intake control is beyond the scope of this section. Instead, the section attempts to summarise the key components of control of food intake, highlighting in particular the rapid advances made in recent years. The section also focuses on those aspects of control of food intake that may be most relevant to food product development, particularly with a view of considering what features of foods may contribute most strongly to prevention of over-consumption. 92 Consumer-led food product development An important first question is whether meal-size is driven by hunger (i.e. energy and nutrient deficits generated by metabolism and nutrient use since the last meal) or satiety (the more we eat, the longer it is until we eat again). A classic way of discriminating these two options has been to explore the relationship between the size of meals and the period of time before and after each meal (De Castro, 1996). Based on extensive diary records, the size of voluntary meals shows a stronger hunger-related (preprandial) correlation than satiety-based (postprandial) relationship (De Castro and Elmore, 1988). This contrasts with rats that ate more frequently and showed a satiety-driven relationship, before switching to a hunger-driven relationship when the opportunity to eat was restricted to fewer meals per day (De Castro, 1988). Interestingly, when people were placed in an environment with no external cues (daylight, time, etc.) to help dictate mealtimes, their eating switched to a satiety-driven pattern (Bernstein et al., 1981). The overall implication is that our meal pattern is controlled mainly by habits: we associate certain times of day with eating, and these time cues act as critical stimuli for meal initiation. 4.6.1 Control of meal-size How then is meal-size controlled? Intuitively, we might construct a simple model where the use of energy and nutrients by the body leads progressively to a state of hunger, while consumption reverses these deficits (satiety). However, meal duration is far too short for the body to assimilate the nutrients and use the signals generated by the nutrients to reverse hunger during each meal. Thus the processes that promote the initiation of eating (hunger) and those that determine when meals terminate have to be different, and here the main factors involved in both sets of processes are briefly summarised. 4.6.2 Psychobiological factors in meal initiation The physiological basis of meal initiation has been long researched, but is still poorly understood. In recent years two different physiological cues have received particular attention. The first relates to a classic theory of hunger, the glucostatic model (Mayer, 1953), which was based on the general idea that feelings of hunger were generated by changes in the availability of blood glucose. Given that glucose is the primary energy source for the brain, ensuring a constant supply of blood glucose is clearly important. Although the simplistic idea that overall blood glucose levels equate with hunger does not fit with our current understanding of control of blood glucose levels (Campfield and Smith, 1990), an important observation was that spontaneous meal-taking by rats was preceded by a transient decline and then increase in blood glucose, with feeding starting during the ascending phase. Similar findings have since been reported in humans, with the spontaneous request for food by time-blinded participants again preceded by transient changes in glucose (Melanson et al., 1999). Thus transient changes in blood glucose levels may lead to an experience of hunger Psychobiological mechanisms in food choice 93 and so meal initiation. What remains unclear is whether the glucose changes are the key signal, or represent metabolic adjustments in anticipation of eating. We know that the rapid influx of nutrients generated by eating is a major challenge to homeostatic control (Woods, 1991), and so determining whether glucose changes are really a cause or consequence of the decision to eat remains uncertain. For many years scientists have hypothesised that the body may have a specific hormone, the actions of which underlie our experience of hunger. However, until 1999 all appetite-related hormones were known to reduce appetite (i.e. were satiety signals). The discovery of a new hormone, called ghrelin, which stimulates food intake in animals (Nakazato et al., 2001) and humans (Cummings et al., 2001, 2004; Wren et al., 2001), have led to the suggestion that it may act as a hunger hormone, a notion supported by the clear relationship between overall body-size and ghrelin levels, with abnormally high levels in patients with anorexia (Tanaka et al., 2003), and low levels in obese patients (Kanumakala et al., 2005; Shiiya et al., 2002; Tschop et al., 2001). The latter finding is also strong evidence that ghrelin is not a prerequisite for eating, since obese people clearly over-consume relative to their energy needs despite having low levels of ghrelin. Thus ghrelin appears to be a hormonal signal which works alongside leptin to signal the status of long-term energy stores (Hellstrom et al., 2004). 4.6.3 Control of meal-size: palatability±satiety interactions How then is meal-size controlled? It is now recognised that meal-size reflects the interaction of two different sorts of feedback systems (Smith, 2000). The first set of signals (satiation) relate to the processes underlying the development of satiety, and involve a complex sequence of cues incorporating learned, orosensory, gastric and post-gastric cues (the satiety cascade: Blundell and Tremblay, 1995). These processes act through negative feedback to reduce desire to eat and so lead to meal termination. In contrast, the second factor is seen as a feed-forward control, and relates to the ability of sensory characteristics of foods to stimulate short-term appetite (often referred to as palatability effects). Clear evidence for the importance of the latter processes is that the experience of hunger can increase in the early stages of a meal if the food is perceived as pleasant tasting (palatable) relative to bland or unpleasant tasting (Yeomans, 1996; Fig. 4.5). However, as eating progresses so hunger decreases and fullness increases until the meal ends. Once the meal has been ingested, post-absorptive satiety cues inhibit further feeding. While a full discussion of the processes involved in satiation and satiety is beyond the scope of this chapter, a brief summary is warranted in order to evaluate the importance of the nature of the food consumed in determining the rate at which satiety develops. A discussion of the neural controls of appetite is not given here; the reader is referred to one of several recent reviews for summaries of the complex brain systems that underlie hunger and satiety (Berthoud, 2003; Schwartz et al., 2000). 94 Consumer-led food product development Fig. 4.5 Changes in rated hunger for normal weight men eating a palatable, bland or overly strong flavoured test meal. Redrawn from Yeomans (1996) with permission. 4.6.4 The role of volume and energy in determining meal-size In the context of the current chapter, the key question is the extent to which the nutrient content of the ingested food controls intake. To what extent is meal-size a function of nutrient intake? The answer appears to be that actual nutrient content has little impact on intake until the consumer has been able to learn about the nutritional consequences of ingestion after the meal. The strongest evidence that this is so comes from studies looking at the short-term effects of nutrient supplementation or dilution. For example, on first exposure volunteers ate the same amount of porridge when this was served in low-energy or highenergy dense versions (Fig. 4.6), but ate more of the low-energy version once they had had a chance to learn the relationship between the porridge flavour and the post-ingestive consequence (Yeomans et al., 2005). Similar conclusions have come from studies of energy density in relation to portion-size (Bell and Rolls, 2001; Kral and Rolls, 2004), where manipulated portion sizes have consistent effects on voluntary short-term intake (the more that was served, the more people consume), while manipulation of energy density had minimal effects. In contrast to the lack of evidence for effects of energy density or nutrients, there is strong evidence that people regulate the volume which they Psychobiological mechanisms in food choice 95 Fig. 4.6 Effects of high-energy (360 kJ) and low-energy (60 kJ) soup preloads on mass (left hand panel) and energy (right hand panel) of a lunch presented in bland or palatable forms. Data modified from Yeomans et al. (2001b), and reprinted with permission. ingest (Rolls and Roe, 2002; Rolls et al., 1998, 2000). Thus in the short term, how much food is served combined with orosensory and gastric metering of the volume ingested seem to be the strongest determinants of meal-size. 4.6.5 Gastric cues What role does the stomach play in satiety? Is the progressive increase in feelings of fullness as a meal progresses a reflection of specific sensors measuring gastric filling? Current research identifies gastric stretch receptors as an element in satiety (Read et al., 1994). However, although cues reflecting gastric distension play a role in controlling food intake, the effects of gastric stretch are a short-lived cue. It is also possible for people to discriminate the feelings of fullness generated by gastric stretch alone from the feelings of fullness generated by eating (French and Cecil, 2001). Thus the stomach may play an important role in control of meal-size only where people eat meals which approach gastric capacity. 4.6.6 Post-gastric cues: the satiety hormones Food entering the small intestine is known to stimulate a number of hormones which are implicated in satiety. The best known of these is cholecystokinin (CCK), which appears to fit all the characteristics needed to be defined as a satiety hormone: (a) it is released in response to food cues, (b) CCK injection inhibits eating and can reduce feelings of hunger (Kissileff et al., 1981; Muurahainen et al., 1991) and (c) levels of circulating CCK correlate with feelings of fullness (French et al., 1993). The CCK signalling system is now well 96 Consumer-led food product development characterised, and is clearly an important physiological component of satiety (Strader and Woods, 2005). Two other hormones, polypeptide-Y 3-36 (PYY36) and glucagon-like peptide 1 (GLP1) also appear to have satiety-like properties. In humans, most, but not all (Long et al., 1999), studies report decreased food intake after administration of GLP-1, including studies in normal weight, diabetic and obese participants (Flint et al., 1998, 2000; Gutzwiller et al., 1999a,b; NaÈslund et al., 1998, 1999b; ToftNielsen et al., 1999). Reduced food intake is consistent with ratings of reduced hunger, and increased fullness following GLP-1 infusion (Gutzwiller et al., 1999b; NaÈslund et al., 1999a). PYY36 is also known to reduce food intake in animals, is released postprandially from the gastro-intestinal tract in proportion to the energy content of the ingested meal and induces satiety in humans (Batterham et al., 2002). Infusions of PYY36 have been shown to reduce food intake in obese and lean subjects (Batterham et al., 2003). Thus food arriving in the gut generates a sequence of signals which act to suppress appetite. 4.6.7 Palatability and the control of meal-size As discussed earlier, hedonic motivation (eating for pleasure) is an important influence on short-term ingestion. In laboratory studies there is a linear relationship between the rated pleasantness of a food and subsequent voluntary food intake (Yeomans, 2006b). Likewise, in diary-based studies of behaviour under naturalistic conditions, the overall palatability of a meal predicts voluntary mealsize independently of hunger state (De Castro et al., 2000a,b). Indeed, in laboratory-based studies, enhancing the palatability of a test meal counter-acts the ability of energy consumed beforehand to suppress intake (Yeomans et al., 2001b), suggesting that palatability overrides satiety and can therefore lead to over-consumption. Palatability effects are also clear in animal studies (Davis, 1989), and as discussed briefly when evaluating influences on food choice, a great deal is now known about the neural basis of hedonic stimulation of appetite, with a clear role for opioid peptides in both animal (Bodnar, 2004) and human studies (Yeomans and Gray, 2002). This has led to increased interest in the hedonic component of eating as a possible contribution to the increased incidence of obesity, as well as parallels between the pleasures associated with eating and drug-based reward (Grigson, 2002; Kelley, 2004; Pelchat, 2002; Wang et al., 2004). 4.6.8 Learning, satiation and satiety This brief discussion of motivational controls of food intake ends by returning to considering further the role of learning in control of food intake. Despite the clear evidence for clear physiological influences on meal-size, these controls should be seen as influences rather than absolute controls. Learning to eat in response to cues in our environment can have dramatic impacts on food intake, and must override many of the physiological satiety signals discussed earlier. A Psychobiological mechanisms in food choice 97 classic example came from a study in rats which were restricted to a single meal a day, signalled by a noise in their cage (Weingarten, 1983). The rats were then allowed free access to food. Despite their lack of food deprivation, hearing the same noise they had experienced as a meal cue when food deprived, reliably induced eating, with rats consuming as much as 20% of their daily intake in the period immediately after the noise. It is hard to construct similar tests in people, but another classic study did demonstrate that believing it was now their usual dinner-time was a sufficient cue for obese volunteers to eat more in a laboratory test (Schachter and Gross, 1968). A great deal of our daily eating may by similarly stimulated by such cues, and recent evidence in animals that the pattern of release of the appetite hormone ghrelin across the day itself depends on how many meals are habitually consumed (Sugino et al., 2002) suggests that habitual meal patterns may entrain the physiological controls of meal initiation. 4.7 Understanding psychobiological mechanisms in food choice for food product development How might the large body of information of psychobiological influences on food preference, choice and appetite control help in the development of new food products? In truth, there are a myriad potential uses and influences of our enhanced understanding of the psychobiology of appetite that can be of value. This discussion focuses on three areas that may be of particular value. 4.7.1 Sensory impact and appetite A simple response to the observation that enhanced palatability results in increased intake might be to suggest that future food products need to be less palatable. The widespread belief (with surprisingly little evidence) that the palatability of the foods available to consumers is higher today than in the past has been cited as an explanation for the increase in obesity. It is certainly well established that voluntary intake increases as a function of palatability (Yeomans, 2006b). However, it is self-apparent that new food products will not be successful if consumers do not like the flavour of these products! How then might foods be designed to utilise our knowledge of the factors involved in meal-size control to try to militate against over-eating as a consequence of palatability? The answer will lie in the ability to build into products elements that maximise the ability of the food to satiate so that the appetite-stimulating effects of flavour are counter-acted by enhanced satiety so leading to no net increase in intake. There are many ways this could be achieved. For example, macronutrients differ in the extent to which they lead to satiety, with fat recognised as the least satiating (Blundell and Burley, 1992) and protein the most satiating (Blundell et al., 1996; De Graaf et al., 1992; Vandewater and Vickers, 1996). Thus altering the macronutrient balance in products should alter the level of post-ingestive satiety. 98 Consumer-led food product development 4.7.2 The importance of learning One obvious conclusion from the present review is that the relationship between a consumer and a food product is not a fixed one. The most important changes will occur over the first few times a product is consumed, as the consumer associates the flavour with other features of the product. It is important to remember that a great deal of this learning is not something consumers are aware of. For example, consumers develop a preference for the flavour of caffeinated drinks over drinks without caffeine without any awareness that caffeine was present (Rogers et al., 1995). However, the observation that caffeinated drinks remain the most popular beverages worldwide is proof of the power of flavour±caffeine associations. Thus developing an appreciation that consumers should learn to associate flavours and after-effects will be an important element of successful product development, particularly since this learning will alter consumers' appreciation of food flavour. Just as learned associations will alter consumers liking for flavours, they may also promote learning of sensible consumption patterns by learning not to overconsume foods that make them feel more full (Booth, 1991). The important consideration here will be ensuring that products have a flavour that promotes learning: the product must stand out as different for this to happen. The prediction from this is that inconsistency between the relationship between flavours and consequences should lead to dysregulation of appetite, and so may lead to overeating (Stubbs and Whybrow, 2004). For example, if you consume a sugar-based soda one day and a diet version with the same (or very similar) flavour another, the body cannot detect a reliable pattern between flavour and consequence and so you cannot learn to moderate your intake of the energy-rich version. This implies that successful products need to have an optimal balance of macronutrients and carefully constructed flavours that facilitate flavour-nutrient learning. 4.7.3 Expectations and appetite control The final issue is the extent to which information about a product may generate an explicit knowledge which itself may interact with the experience of a product once ingested and so direct consumer behaviour. The most obvious examples of this in the psychobiological literature have examined whether providing labels about nutritive content modify the ability of foods to satiate a consumer. Most evidence suggests that food labels have minimal effects on actual appetite regulation. Thus soup labels implying high fat or low energy had no impact on subsequent food intake, whereas actual energy in the soup modified lunch intake (Yeomans et al., 2001a). Likewise, the rate at which pleasantness of an eaten food declines (sensory-specific satiety: SSS) is similarly insensitive to labelled nutrient information. Thus the labelled fat content of potato chips did not modify the rate at which SSS developed even after multiple-exposures to the snack (Miller et al., 2000). Do these results imply that providing clear nutritional advice on a label is of Psychobiological mechanisms in food choice 99 no value? The answer is of course no, since the label serves to help educate consumers. However, it would be wrong to assume that food labelling itself can help consumers control intake. The important factor remains the actual nutritional content of the food consumed regardless of the awareness of the consumer of that content. 4.8 Future trends Unlike the trend for increased body size of consumers, which shows no sign of slowing, predicting how our psychobiological understanding of the relationship between consumer and food will develop is unclear. Much research effort at present is directed at the development of drug treatments to reduce body size in obese patients (Halford, 2006), but this increased understanding of these appetite suppressants will have knock-on effects for our broader understanding of appetite control. Thus a few years ago we knew that cannabis was associated with cravings for carbohydrate-rich foods (the munchies) but had no knowledge of the mechanism underlying this. Today we know a great deal about the role of our endogenous cannabinoids (Kirkham, 2005), leading to new treatments for obesity. The pace at which our understanding of the neural controls of feeding has developed has accelerated markedly, driven by pressures to help understand and treat obesity. Areas where current understanding is incomplete include the role of learning in appetite control, the inability to respond in the short term to energy density and the realisation that not all consumers are the same. All of these areas will impact heavily on future product development. A critical issue will be finding ways to reduce energy density without impacting negatively on consumers' perception of flavour and satisfaction with the product. An ideal product would be an enjoyable eating experience which satisfies short-term appetite without promoting over-consumption, and only an appreciation of how the consumer experiences both the hedonic and sensory qualities of foods will allow development of products that meet this aim. The final lesson from this brief review of psychobiological influences on food choice and appetite is that development of products has to be conducted in a way that is realistic to the situation where the product will be consumed. Most important is the data showing sensitivity of liking ratings to current needs. Developing the sensory qualities of products to taste good in a test situation that is different from the way in which consumers may use that product is unlikely to succeed. For example, developing a product using panels that are not hungry at the time of sampling may not be helpful in developing a product used by consumers as a snack to reduce hunger between meals. Likewise, product development has to include measures of intake, and the impact of repeated intake on product acceptability. 100 Consumer-led food product development 4.9 Sources of further information and advice The reference list in Section 4.10 details all the papers cited in this review. However, there are many useful reviews that can be used to expand on the material from this chapter, and these are broken up by topic here. Useful books include: (2004) The Psychology of Eating and Drinking, 3rd edition. Hove: BrunnerRoutledge. HETHERINGTON, M.M. (2001) Food Cravings and Addiction. Leatherhead: Leatherhead Food RA. MELA, D.J. and ROGERS, P.J. (1998) Food, Eating and Obesity: The Psychobiological Basis of Appetite and Weight Control. London: Chapman & Hall. LOGUE, A.W. For information on how flavour preference develops, the following are accessible reviews: (1991). How foods get to be liked: some general mechanisms and some special cases. In R. C. Bolles (Ed.), The Hedonics of Taste (pp. 199±217). Hillsdale, NJ: Lawrence Erlbaum Associates. YEOMANS, M. R. (2006). The role of learning in development of food preferences. In R. Shepherd and M. Raats (Eds.), Psychology of Food Choice (pp. 93±112). Wallingford: CABI. ZELLNER, D. A. The control of appetite is another area that has been extensively reviewed, with the following recommended: (2003). Neural systems controlling food intake and energy balance in the modern world. Current Opinion in Clinical Nutrition and Metabolic Care, 6(6), 615±620. SMALL, C. J. and BLOOM, S. R. (2004). Gut hormones and the control of appetite. Trends in Endocrinology and Metabolism, 15(6), 259±263. BERTHOUD, H. R. 4.10 References and SCLAFANI, A. (2001). Flavor preferences conditioned by intragastric infusion of ethanol in rats. Pharmacology Biochemistry and Behavior, 68, 327± 338. ACKROFF, K. and SCLAFANI, A. (2002). Flavor quality and ethanol concentration affect ethanol-conditioned flavor preferences. Pharmacology Biochemistry and Behavior, 74(1), 229±240. ACKROFF, K. and SCLAFANI, A. (2003). Flavor preferences conditioned by intragastric ethanol with limited access training. Pharmacology Biochemistry and Behavior, 75, 223±233. BAEYENS, F., EELEN, P., VAN DEN BURGH, O. and CROMBEZ, G. (1990). Flavor-flavor and color±flavor conditioning in humans. Learning and Motivation, 21, 434±455. BAEYENS, F., CROMBEZ, G., HENDRICKX, H. and EELEN, P. (1995). Parameters of human evaluative flavour±flavor conditioning. Learning and Motivation, 26, 141±160. BAEYENS, F., CROMBEZ, G., DE HOUWER, J. and EELEN, P. (1996). No evidence for modulation ACKROFF, K. Psychobiological mechanisms in food choice 101 of evaluative flavour±flavor associations in humans. Learning and Motivation, 27, 200±241. BATTERHAM, R. L., COHEN, M. A., ELLIS, S. M., LE ROUX, C. W., WITHERS, D. J., FROST, G. S., et al. (2003). Inhibition of food intake in obese subjects by peptide yy3-36. New England Journal of Medicine, 349(10), 941±948. BATTERHAM, R. L., COWLEY, M. A., SMALL, C. J., HERZOG, H., COHEN, M. A., DAKIN, C. L., et al. (2002). Gut hormone pyy(3-36) physiologically inhibits food intake. Nature, 418, 650±654. BELL, E. A. and ROLLS, B. J. (2001). Energy density of foods affects energy intake across multiple levels of fat content in lean and obese women. American Journal of Clinical Nutrition, 73, 1010±1018. BERNSTEIN, I. L., ZIMMERMAN, J. C., CZEISLER, C. A. and WEITZMAN, E. D. (1981). Meal patterns in `free-running' humans. Physiology and Behavior, 27(4), 621±623. BERRIDGE, K. C. (2003). Pleasures of the brain. Brain and Cognition, 52(1), 106±128. BERTHOUD, H. R. (2003). Neural systems controlling food intake and energy balance in the modern world. Current Opinion in Clinical Nutrition and Metabolic Care, 6(6), 615±620. BIRCH, L. L., MCPHEE, L., STEINBERG, L. and SULLIVAN, S. (1990). Conditioned flavor preferences in young children. Physiology and Behavior, 47, 501±505. BLUNDELL, J. E. and BURLEY, V. (1992). Evaluation of the satiating power of dietary fat in man. In Y. Oomura, S. Baba and T. Shimazu (Eds.), Progress in Obesity Research (pp. 543±457). London: Libbey. BLUNDELL, J. E. and TREMBLAY, A. (1995). Appetite control and energy (fuel) balance. Nutrition Research Reviews, 8, 225±242. BLUNDELL, J. E., LAWTON, C. L., COTTON, J. R. and MACDIARMID, J. I. (1996). Control of human appetite ± implications for the intake of dietary-fat. Annual Review of Nutrition, 16, 285±319. BODNAR, R. J. (2004). Endogenous opioids and feeding behavior: a 30-year historical perspective. Peptides, 25(4), 697±725. BOOTH, D. A. (1991). Learned ingestive motivation and the pleasures of the palate. In R. C. Bolles (Ed.), The Hedonics of Taste. Hillsdale, NJ: Lawrence Erlbaum Associates. BOOTH, D. A., MATHER, P. and FULLER, J. (1982). Starch content of ordinary foods associatively conditions human appetite and satiation, indexed by intake and pleasantness of starch-paired flavours. Appetite, 3, 163±184. BRUNSTROM, J. M. (2004). Does dietary learning occur outside awareness? Consciousness and Cognition, 13(3), 453±470. CAMPFIELD, L. A. and SMITH, F. J. (1990). Systemic factors in the control of food intake. In E. M. Stricker (Ed.), Handbook of Neurobiology: Volume 10, Neurobiology of Food and Fluid Intake (pp. 183±206). New York: Plenum Press. CAPALDI, E. D. (1992). Conditioned food preferences. Psychology of Learning and Motivation, 28, 1±33. CAPALDI, E. D., OWENS, J. and PALMER, K. A. (1994). Effects of food deprivation on learning and expression of flavor preferences conditioned by saccharin or sucrose. Animal Learning and Behavior, 22, 173±180. COOPER, S. J., JACKSON, A., MORGAN, R. and CARTER, R. (1985). Evidence for opiate receptor involvement in the consumption of a high palatability diet in nondeprived rats. Neuropeptides, 5, 345±348. CUMMINGS, D. E., PURNELL, J. Q., FRAYO, R. S., SCHMIDOVA, K., WISSE, B. E. and WEIGLE, D. S. (2001). A preprandial rise in plasma ghrelin levels suggests a role in meal initiation 102 Consumer-led food product development in humans. Diabetes, 50(8), 1714±1719. and CHAPELOT, D. (2004). Plasma ghrelin levels and hunger scores in humans initiating meals voluntarily without time- and food-related cues. American Journal of Physiology: Endocrinology and Metabolism, 287(2), E297±304. DAVIS, J. D. (1989). The microstructure of ingestive behavior. Annals of the New York Academy of Sciences, 575, 106±121. DE CASTRO, J. M. (1988). The meal pattern of rats shifts from postprandial regulation to preprandial regulation when only five meals per day are scheduled. Physiology and Behavior, 43, 739±746. DE CASTRO, J. M. (1996). How can eating behavior be regulated in the complex environments of free-living humans? Neuroscience and Biobehavioral Reviews, 20(1), 119±131. DE CASTRO, J. M. (2000). Eating behaviour: lessons from the real world of humans. Nutrition, 16, 800±813. DE CASTRO, J. M. and ELMORE, D. K. (1988). Subjective hunger relationships with meal patterns in the spontaneous feeding behaviour of humans: evidence for a causal connection. Physiology and Behavior, 43(2), 159±165. DE CASTRO, J. M., BELLISLE, F. and DALIX, A.-M. (2000a). Palatability and intake relationships in free-living humans: measurement and characterisation in the french. Physiology and Behavior, 68, 271±277. DE CASTRO, J. M., BELLISLE, F., DALIX, A.-M. and PEARCEY, S. M. (2000b). Palatability and intake relationships in free-living humans: characterization and independence of influence in north americans. Physiology and Behavior, 70, 343±350. DE GRAAF, C., HULSHOF, T., WESTSTRATE, J. A. and JAS, P. (1992). Short-term effects of different amounts of protein, fats, and carbohydrates on satiety. American Journal of Clinical Nutrition, 55, 33±38. DE HOUWER, J., THOMAS, S. and BAEYENS, F. (2001). Associative learning of likes and dislikes: a review of 25 years of research on human evaluative conditioning. Psychological Bulletin, 127(6), 853±869. DREWNOWSKI, A. (1998). Energy density, palatability and satiety: implications for weight control. Nutrition Reviews, 56(12), 347±353. ELIZALDE, G. and SCLAFANI, A. (1988). Starch-based conditioned flavor preferences in rats: influence of taste, calories and Cs±Us delay. Appetite, 11, 179±200. FEDORCHAK, P. M. and BOLLES, R. C. (1987). Hunger enhances the expression of calorie- but not taste-mediated conditioned flavour preferences. Animal Behavior Processes, 13, 73±79. FIELD, A. P. and DAVEY, G. C. L. (1999). Re-evaluating evaluative conditioning: an exemplar comparison model of evaluative conditioning effects. Journal of Experimental Psychology: Animal Behavior Processes, 25, 211±224. FLINT, A., RABEN, A., ASTRUP, A. and HOLST, J. J. (1998). Glucagon-like peptide-1 promotes satiety and suppresses energy intake in humans. Journal of Clinical Investigation, 101, 515±520. FLINT, A., RABEN, A., REHFELD, J. F., HOLST, J. J. and ASTRUP, A. (2000). The effect of glucagon-like peptide-1 on energy expenditure and substrate metabolism in humans. International Journal of Obesity, 24, 288±298. FRENCH, S. J. and CECIL, J. E. (2001). Oral, gastric and intestinal influences on human feeding. Physiology and Behavior, 74, 729±734. FRENCH, S. J., MURRAY, B., RUMSEY, R. D., SEPPLE, C. P. and READ, N. W. (1993). Is CUMMINGS, D. E., FRAYO, R. S., MARMONIER, C., AUBERT, R. Psychobiological mechanisms in food choice 103 cholecystokinin a satiety hormone? Correlations of plasma cholecystokinin with hunger, satiety and gastric emptying in normal volunteers. Appetite, 21(2), 95±104. GARCIA, J. and KOELLING, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4, 123±124. GIBSON, E. L. and WARDLE, J. (2003). Energy density predicts preferences for fruit and vegetables in 4-year-old children. Appetite, 41(1), 97±98. GIBSON, E. L., WAINWRIGHT, C. J. and BOOTH, D. A. (1995). Disguised protein in lunch after low-protein breakfast conditions food-flavor preferences dependent on recent lack of protein intake. Physiology and Behavior, 58, 363±371. GRIGSON, P. S. (2002). Like drugs for chocolate: separate rewards modulated by common mechanisms. Physiology and Behavior, 76, 389±395. GUTZWILLER, J. P., DREWE, J., GOKE, B., SCHMIDT, H., ROHRER, B., LAREIDA, J., et al. (1999a). Glucagon-like peptide-1 promotes satiety and reduces food intake in patients with diabetes mellitus type 2. American Journal of Physiology, 276(5 Pt 2), R1541± 1544. GUTZWILLER, J. P., GOKE, B., DREWE, J., HILDEBRAND, P., KETTERER, S., HANDSCHIN, D., et al. (1999b). Glucagon-like peptide-1: a potent regulator of food intake in humans. Gut, 44(1), 81±86. HALFORD, J. C. (2006). Pharmacotherapy for obesity. Appetite, 46(1), 6±10. HARRIS, J. A., GORISSEN, M. C., BAILEY, G. K. and WESTBROOK, R. F. (2000). Motivational state regulates the content of learned flavor preferences. Journal of Experimental Psychology: Animal Behavior Processes, 26(1), 15±30. HELLSTROM, P. M., GELIEBTER, A., NASLUND, E., SCHMIDT, P. T., GLUCK, M., YAHAV, E., et al. (2004). Peripheral and central signals in control of eating in normal, obese and binge-eating human subjects. British Journal of Nutrition, 92, S47±S57. JOHNSON, S. L., MCPHEE, L. and BIRCH, L. L. (1991). Conditioned preferences: young children prefer flavors associated with high dietary fat. Physiology and Behavior, 50(6), 1245±1251. KANUMAKALA, S., GREAVES, R., PEDREIRA, C. C., DONATH, S., WARNE, G. L., ZACHARIN, M. R., et al. (2005). Fasting ghrelin levels are not elevated in children with hypothalamic obesity. Journal of Clinical Endocrinology and Metabolism, 90(5), 2691±2695. KELLEY, A. E. (2004). Ventral striatal control of appetitive motivation: role in ingestive behavior and reward-related learning. Neuroscience and Biobehavioral Reviews, 27(8), 765±776. KELLEY, A. E., BAKSHI, V. P., HABER, S. N., STEININGER, T. L., WILL, M. J. and ZHANG, M. (2002). Opioid modulation of taste hedonics within the ventral striatum. Physiology and Behavior, 76, 365±377. KERN, D. L., MCPHEE, L., FISHER, J., JOHNSON, S. and BIRCH, L. L. (1993). The postingestive consequences of fat condition preferences for flavors associated with high dietary fat. Physiology and Behavior, 54, 71±76. KIRKHAM, T. C. (2005). Endocannabinoids in the regulation of appetite and body weight. Behavioral Pharmacology, 16(5±6), 297±313. KISSILEFF, H. R., PI-SUNYER, F. X., THORNTON, J. and SMITH, G. P. (1981). C-terminal octapeptide of cholecystokinin decreases food intake in man. American Journal of Clinical Nutrition, 34, 154±160. KRAL, T. V. and ROLLS, B. J. (2004). Energy density and portion size: their independent and combined effects on energy intake. Physiology and Behavior, 82(1), 131±138. LONG, S. J., SUTTON, J. A., AMAEE, W. B., GIOUVANOUDI, A., SPYROU, N. M., ROGERS, P. J., et al. (1999). No effect of glucagon-like peptide-1 on short-term satiety and energy 104 Consumer-led food product development intake in man. British Journal of Nutrition, 81(4), 273±279. and SCLAFANI, A. (1989). Flavor preferences conditioned by intragastric fat infusions in rats. Physiology and Behavior, 46, 403±412. MAYER, J. (1953). Genetic, traumatic and environmental factors in the etiology of obesity. Physiological Reviews, 33, 472±508. MELA, D. J. (2001). Determinants of food choice: relationships with obesity and weight control. Obesity Research, 9(Supplement 4), 249S±255S. MELANSON, K. J., WESTERTERP-PLANTENGA, M. S., SARIS, W. H. M., SMITH, F. J. and CAMPFIELD, L. A. (1999). Blood glucose patterns and appetite in time-blinded humans: carbohydrate versus fat. American Journal of Physiology, 277, R337±R345. MILLER, D. L., BELL, E. A., PELKMAN, C. L., PETERS, J. C. and ROLLS, B. J. (2000). Effects of dietary fat, nutrition labels, and repeated consumption on sensory specific satiety. Physiology and Behavior, 71, 153±158. MOBINI, S., ELLIMAN, T. D. and YEOMANS, M. R. (2005). Changes in the pleasantness of caffeine-associated flavours consumed at home. Food Quality and Preference, 16, 659±666. MUURAHAINEN, N. E., KISSILEFF, H. R., LACHAUSSEE, J. and PI-SUNYER, F. X. (1991). Effect of a soup preload on reduction of food intake by cholecystokinin in humans. American Journal of Physiology, 260, R672±R680. MYERS, K. P. and SCLAFANI, A. (2001a). Conditioned enhancement of flavor evaluation reinforced by intragastric glucose. Physiology and Behavior, 74, 495±505. MYERS, K. P. and SCLAFANI, A. (2001b). Conditioned enhancement of flavor evaluation reinforced by intragastric glucose. I. Intake acceptance and preference analysis. Physiology and Behavior, 74, 481±493. NAKAZATO, M., MURAKAMI, N., DATE, Y., KOJIMA, M., MATSUO, H., KANGAWA, K., et al. (2001). A role for ghrelin in the central regulation of feeding. Nature, 409(6817), 194±198. È SLUND, E., GUTNIAK, M., SKOGAR, S., ROSSNER, S. and HELLSTROM, P. M. (1998). GlucagonNA like peptide 1 increases the period of postprandial satiety and slows gastric emptying in obese men. American Journal of Clinical Nutrition, 68(3), 525±530. È SLUND, E., BARKELING, B., KING, N., GUTNIAK, M., BLUNDELL, J. E., HOLST, J. J., et al. NA (1999a). Energy intake and appetite are suppressed by glucagon-like peptide-1 (glp-1) in obese men. International Journal of Obesity, 23(3), 304±311. È SLUND, E., BOGEFORS, J., SKOGAR, S., GRYBACK, P., JACOBSSON, H., HOLST, J. J., et al. NA (1999b). Glp-1 slows solid gastric emptying and inhibits insulin, glucagon, and pyy release in humans. American Journal of Physiology, 277(3 Pt 2), R910±R916. PELCHAT, M. L. (2002). Of human bondage: food craving, obsession, compulsion, and addiction. Physiology and Behavior, 76, 347±352. RAMIREZ, I. (1994). Flavor preferences conditioned with starch in rats. Animal Learning and Behavior, 22, 181±187. READ, N., FRENCH, S. A. and CUNNINGHAM, K. (1994). The role of the gut in regulating food intake in man. Nutrition Reviews, 52, 1±10. RICHARDSON, N. J., ROGERS, P. J. and ELLIMAN, N. A. (1996). Conditioned flavour preferences reinforced by caffeine consumed after lunch. Physiology and Behavior, 60, 257± 263. ROGERS, P. J., RICHARDSON, N. J. and ELLIMAN, N. A. (1995). Overnight caffeine abstinence and negative reinforcement of preference for caffeine-containing drinks. Psychopharmacology, 120, 457±462. ROLLS, B. J. and ROE, L. S. (2002). Effect of the volume of liquid food infused intragastrically on satiety in women. Physiology and Behavior, 76, 623±631. LUCAS, F. Psychobiological mechanisms in food choice 105 and PELKMAN, C. L. (1998). Volume of food consumed affects satiety in man. American Journal of Clinical Nutrition, 67, 1170±1177. ROLLS, B. J., BELL, E. A. and WAUGH, B. A. (2000). Increasing the volume of a food by incorporating air affects satiety in man. American Journal of Clinical Nutrition, 72, 361±368. ROZIN, P. and VOLLMECKE, T. A. (1986). Food likes and dislikes. Annual Review of Nutrition, 6, 433±456. SCHACHTER, S. and GROSS, L. P. (1968). Manipulated time and eating behavior. Journal of Personality and Social Psychology, 10, 98±106. SCHWARTZ, M. W., WOODS, S. C., PORTE, D., JR., SEELEY, R. J. and BASKIN, D. G. (2000). Central nervous system control of food intake. Nature, 404(6778), 661±671. SCLAFANI, A. (1999). Macronutrient-conditioned flavor preferences. In H.-R. Berthoud and R. J. Seeley (Eds.), Neural Control of Macronutrient Selection (pp. 93±106). Boca Raton, FL: CRC Press. SCLAFANI, A. (2002). Flavor preferences conditioned by sucrose depend upon training and testing methods: two-bottle tests revisited. Physiology and Behavior, 76, 633±644. SCLAFANI, A. and NISSENBAUM, J. W. (1988). Robust conditioned flavor preference produced by intragastric starch infusions in rats. American Journal of Physiology, 255, R672±R675. SHIIYA, T., NAKAZATO, M., MIZUTA, M., DATE, Y., MONDAL, M. S., TANAKA, M., et al. (2002). Plasma ghrelin levels in lean and obese humans and the effect of glucose on ghrelin secretion. Journal of Clinical Endocrinology and Metabolism, 87(1), 240± 244. SMITH, G. P. (2000). The controls of eating: a shift from nutritional homeostasis to behavioral neuroscience. Nutrition, 16, 814±820. STEINER, J. E., GLASER, D., HAWILO, M. E. and BERRIDGE, K. C. (2001). Comparative expression of hedonic impact: affective reactions to taste by human infants and other primates. Neuroscience and Biobehavioral Reviews, 25, 53±74. STIRLING, L. J. and YEOMANS, M. R. (2004). Effects of exposure to a forbidden food on eating in restrained and unrestrained women. International Journal of Eating Disorders, 35, 59±68. STRADER, A. D. and WOODS, S. C. (2005). Gastrointestinal hormones and food intake. Gastroenterology, 128(1), 175±191. STUBBS, R. J. and WHYBROW, S. (2004). Energy density, diet composition and palatability: influences on overall food energy intake in humans. Physiology and Behavior, 81, 755±764. SUGINO, T., YAMAURA, J., YAMAGISHI, M., OGURA, A., HAYASHI, R., KUROSE, Y., et al. (2002). A transient surge of ghrelin secretion before feeding is modified by different feeding regimens in sheep. Biochemical and Biophysical Research Communications, 298(5), 785±788. TANAKA, M., NARUO, T., YASUHARA, D., TATEBE, Y., NAGAI, N., SHIIYA, T., et al. (2003). Fasting plasma ghrelin levels in subtypes of anorexia nervosa. Psychoneuroendocrinology, 28(7), 829±835. TINLEY, E. M., YEOMANS, M. R. and DURLACH, P. J. (2003). Caffeine does not reinforce conditioned flavour liking in fully abstinent caffeine consumers. Psychopharmacology, 166, 416±423. TINLEY, E. M., DURLACH, P. J. and YEOMANS, M. R. (2004). How habitual caffeine consumption and dose influence flavour preference conditioning with caffeine. Physiology ROLLS, B. J., CASTELLANOS, V. H., HALFORD, J. C., KILARA, A., PANYAM, D. 106 Consumer-led food product development and Behavior, 82(2/3), 317±324. (1986). Motivational Systems. Cambridge: Cambridge University Press. TOFT-NIELSEN, M. B., MADSBAD, S. and HOLST, J. J. (1999). Continuous subcutaneous infusion of glucagon-like peptide-1 lowers plasma glucose and reduces appetite in type 2 diabetic patients. Diabetes Care, 22, 1137±1143. TSCHOP, M., WEYER, C., TATARANNI, P. A., DEVANARAYAN, V., RAVUSSIN, E. and HEIMAN, M. L. (2001). Circulating ghrelin levels are decreased in human obesity. Diabetes, 50(4), 707±709. TUORILA, H., KRAMER, F. M. and ENGELL, D. (2001). The choice of fat-free vs. regular-fat fudge: the effects of liking for the alternative and the restraint status. Appetite, 37(1), 27±32. VANDEWATER, K. and VICKERS, Z. (1996). Higher-protein foods produce greater sensoryspecific satiety. Physiology and Behavior, 59, 579±583. WANG, G. J., VOLKOW, N. D., THANOS, P. K. and FOWLER, J. S. (2004). Similarity between obesity and drug addiction as assessed by neurofunctional imaging: a concept review. Journal of Addictive Diseases, 23(3), 39±53. WEINGARTEN, H. P. (1983). Conditioned cues elicit feeding in sated rats: a role of learning in meal initiation. Science, 220, 431±433. WOODS, S. C. (1991). The eating paradox: how we tolerate food. Psychological Review, 98(4), 488±505. WREN, A. M., SEAL, L. J., COHEN, M. A., BRYNES, A. E., FROST, G. S., MURPHY, K. G., et al. (2001). Ghrelin enhances appetite and increases food intake in humans. Journal of Clinical Endocrinology and Metabolism, 86(12), 5992. YEOMANS, M. R. (1996). Palatability and the microstructure of eating in humans: the appetiser effect. Appetite, 27(2), 119±133. YEOMANS, M. R. (2006a). The role of learning in development of food preferences. In R. Shepherd and M. Raats (Eds.), Psychology of food choice (pp. 93±112). Wallingford, Oxford, UK: CABI. YEOMANS, M. R. (2006b). The role of palatability in control of food intake: Implications for understanding and treating obesity. In S. J. Cooper and T. C. Kirkham (Eds.), Progress in Brain Research: Appetite and Body Weight: Integrative Systems and the Development of Anti-obesity Drugs (in press). Elsevier. YEOMANS, M. R. and GRAY, R. W. (2002). Opioids and human ingestive behaviour. Neuroscience and Biobehavioral Reviews, 26, 713±728. YEOMANS, M. R. and MOBINI, S. (2006). Hunger alters the expression of acquired hedonic but not sensory qualities of food-paired odors in humans. Journal of Experimental Psychology: Animal Behavior Processes, 32, 460±466. YEOMANS, M. R., WRIGHT, P., MACLEOD, H. A. and CRITCHLEY, J. A. J. H. (1990). Effects of nalmefene on feeding in humans: dissociation of hunger and palatability. Psychopharmacology, 100(3), 426±432. YEOMANS, M. R., SPETCH, H. and ROGERS, P. J. (1998). Conditioned flavour preference negatively reinforced by caffeine in human volunteers. Psychopharmacology, 137(4), 401±409. YEOMANS, M. R., JACKSON, A., LEE, M. D., NESIC, J. S. and DURLACH, P. J. (2000a). Expression of flavour preferences conditioned by caffeine is dependent on caffeine deprivation state. Psychopharmacology, 150(2), 208±215. YEOMANS, M. R., JACKSON, A., LEE, M. D., STEER, B., TINLEY, E. M., DURLACH, P. J., et al. (2000b). Acquisition and extinction of flavour preferences conditioned by caffeine in humans. Appetite, 35, 131±141. TOATES, F. M. Psychobiological mechanisms in food choice 107 and GRAY, R. W. (2001a). The actual, but not labelled, fat content of a soup preload alters short-term appetite in healthy men. Physiology and Behavior, 73(4), 533±540. YEOMANS, M. R., LEE, M. D., GRAY, R. W. and FRENCH, S. J. (2001b). Effects of test-meal palatability on compensatory eating following disguised fat and carbohydrate preloads. International Journal of Obesity, 25, 1215±1224. YEOMANS, M. R., BLUNDELL, J. E. and LESHAM, M. (2004). Palatability: response to nutritional need or need-free stimulation of appetite? British Journal of Nutrition, 92, S3±S14. YEOMANS, M. R., WEINBERG, L. and JAMES, S. (2005). Effects of palatability and learned satiety on energy density influences on breakfast intake in humans. Physiology and Behavior, 86, 487±499. YEOMANS, M. R., MOBINI, S., ELLIMAN, T. D., WALKER, H. C. and STEVENSON, R. J. (2006). Hedonic and sensory characteristics of odors conditioned by pairing with tastants in humans. Journal of Experimental Psychology: Animal Behavior Processes, 32, 215±228. ZELLNER, D. A. (1991). How foods get to be liked: some general mechanisms and some special cases. In R. C. Bolles (Ed.), The Hedonics of Taste (pp. 199±217). Hillsdale, NJ: Lawrence Erlbaum Associates. ZELLNER, D. A., ROZIN, P., ARON, M. and KULISH, C. (1983). Conditioned enhancement of human's liking for flavor by pairing with sweetness. Learning and Motivation, 14, 338±350. ZIMMET, P. and THOMAS, C. R. (2003). Genotype, obesity and cardiovascular disease ± has technical and social advancement outstripped evolution? Journal of Internal Medicine, 254(2), 114±125. YEOMANS, M. R., LARTAMO, S., PROCTER, E. L., LEE, M. D.