food reinforcement and eating: a multilevel analysisabcwest/pmwiki/cafe/2007 - epstein et al -...

Food Reinforcement and Eating: A Multilevel Analysis

Leonard H. Epstein, John J. Leddy, andJennifer L. Temple

University at Buffalo School of Medicine and BiomedicalSciences

Myles S. FaithUniversity of Pennsylvania School of Medicine

Eating represents a choice among many alternative behaviors. The purpose of this review is to providean overview of how food reinforcement and behavioral choice theory are related to eating and to showhow this theoretical approach may help organize research on eating from molecular genetics throughtreatment and prevention of obesity. Special emphasis is placed on how food reinforcement andbehavioral choice theory are relevant to understanding excess energy intake and obesity and how theyprovide a framework for examining factors that may influence eating and are outside of those that mayregulate energy homeostasis. Methods to measure food reinforcement are reviewed, along with factorsthat influence the reinforcing value of eating. Contributions of neuroscience and genetics to the study offood reinforcement are illustrated by using the example of dopamine. Implications of food reinforcementfor obesity and positive energy balance are explored, with suggestions for novel approaches to obesitytreatment based on the synthesis of behavioral and pharmacological approaches to food reinforcement.

Keywords: food reinforcement, food intake, choice, dopamine, behavioral genetics

Choice is ubiquitous. There are many choices that people makethat influence their eating and body weight, beginning with howearly to get up, to exercise before breakfast or to sleep in, what tohave for breakfast, and so on. These choices are made in thecontext of alternatives, many of which are pleasurable and exertconsiderable influence to do these behaviors rather than others.The morning is a good time for many people to choose to exercise,but a warm bed and a little extra sleep time represent a powerfulalternative that even the most enthusiastic exerciser confronts andoften succumbs to. Similarly, eating a good breakfast starts the dayoff right, but a pastry and coffee may be particularly invitingalternative foods that taste very good and may tempt many awayfrom healthy cereal and fruit. This article provides a theoreticalframework for understanding motivation to eat, how people makechoices about eating versus engaging in other behaviors, and howthese theoretical approaches can provide insights into obesity,mechanisms that regulate food choice, and implications of choicetheory for interventions. One of the unique aspects of food rein-

forcement and behavioral theories of choice is the way in whichthey can be used to understand factors that influence eating atmultiple levels. Table 1 provides examples of the multiple levels ofanalysis, using the framework of micro to environmental levels ofanalysis. This framework will be used to organize the literaturethat is reviewed.

The Reinforcing Value of Food

A reinforcer is a stimulus that increases the rate of a behaviorthat it follows. For example, when a dog learns a trick to get a dogtreat, the treat is the reinforcer. The reinforcing value, or reinforcerefficacy (Bickel, Marsch, & Carroll, 2000), of a stimulus refers tohow much behavior the stimulus will support, or the response-strengthening function of reinforcement. Strong reinforcers canmotivate a lot of behavior, whereas weaker reinforcers do notsupport very much behavior. It is not surprising to consider thatmany drugs of abuse are strong reinforcers as drug addicts willengage in a substantial amount of behavior to gain access to thesestrong reinforcers. Food is a strong reinforcer and, in some con-texts, may be a more powerful reinforcer than are drugs of abuse(Hursh & Bauman, 1987).

Consumption of some reinforcers may be regulated, as thepatterns of responding adjust to match the consumption of aregulated reinforcer. For example, when the dose of alcohol isvaried the number of alcohol self-administrations is inverselyrelated to dose, so that alcohol consumption can be regulated(Karoly, Winger, Ikomi, & Woods, 1978). An analogy to foodregulation might be that if subjects are working for a slice of pizzaas the reinforcer and the portion size is reduced in half, thensubjects would double the amount of work they are willing to doto maintain the same amount of pizza. On the other hand, respond-ing may also be a function of the concentration of the substance.Rats will increase responding in a dose–response manner as the

Leonard H. Epstein and Jennifer L. Temple, Department of Pediatrics,University at Buffalo School of Medicine and Biomedical Sciences; JohnJ. Leddy, Department of Orthopedics, University at Buffalo School ofMedicine and Biomedical Sciences, and Myles S. Faith, Department ofPsychiatry, University of Pennsylvania School of Medicine.

The development of this article was funded in part by National Instituteof Child Health and Human Development Grants HD 39792 and HD 44725to Leonard H. Epstein. We thank Tony Caggiula and Ken Perkins, whohelped develop the initial research program on food reinforcement at theUniversity of Pittsburgh.

Correspondence concerning this article should be addressed to LeonardH. Epstein, Department of Pediatrics, School of Medicine and BiomedicalSciences, State University of New York at Buffalo, Farber Hall, RoomG56, 3435 Main Street, Building #26, Buffalo, New York 14214-3000.E-mail: [email protected]

Psychological Bulletin Copyright 2007 by the American Psychological Association2007, Vol. 133, No. 5, 884–906 0033-2909/07/$12.00 DOI: 10.1037/0033-2909.133.5.884

884

concentration of sucrose increases (Sclafani & Ackroff, 2003). Ifthe animals were working to gain a regulated amount of sucrose,then they would respond less, rather than more, as the concentra-tion increased.

The process of presenting a reinforcer contingent on behaviorand observing increases in the rate of the behavior is calledpositive reinforcement. Positive reinforcement is one example ofthe law of effect, which is defined by the more general rule that therate of instrumental behavior depends on its effect on the environ-ment (Bouton, 2007). The process of a behavior producing apositive consequence is also called reward learning, and the pos-itive events can also be called rewards. In many cases, reward andreinforcer have similar meanings, but we have tried to use theterms reinforcement or reward consistently with the author’s useof the terms. The behavior that is associated with presentation ofthe contingent event can be described as instrumental or operantbehavior. In situations in which we refer to the differences in theamount of work that subjects engage in to get access to reinforcers,we refer to these behaviors as motivated behaviors.

Behavioral Theories of Choice: Relative ReinforcingValue and Choice Among Multiple Alternatives

Reinforcer efficacy, or reinforcer value, can be defined in termsof how many responses are made to obtain food (Epstein &Saelens, 2000). Schedules of reinforcement establish how muchwork is needed to gain access to the reinforcer. The most commonmethod for evaluating the efficacy of a single reinforcer, which isreferred to as absolute (rather than relative) reinforcing value, is toarrange for a subject to gain access to a reinforcer by respondingon progressive ratio schedules of reinforcement. In progressiveratio schedules, the schedule requirements are progressively in-creased after a reinforcer is obtained. For example, a subject mayhave to make 10 responses to obtain the first reinforcer, 100 to getthe second, 1,000 to get the third, and so on. Reinforcer efficacy ismost commonly defined in terms of the breakpoint—the point atwhich subjects stop responding. A reinforcer that maintains a

higher breakpoint is considered to be more reinforcing than is acommodity for which subjects terminate responding earlier (Rich-ardson & Roberts, 1996). The amount of behavior that a specificreinforcer will support can be compared across substances just asthe abuse liability of drugs can be compared in animals (Ator &Griffiths, 2003; Balster & Bigelow, 2003) and humans (Griffiths,Bigelow, & Ator, 2003).

Outside of the laboratory, people usually are not presented withonly one option but rather have to choose how to allocate time,behavior, and resources among several alternatives. The paradigmfor evaluating reinforcer efficacy can be extended to two or morechoices, and the theoretical base for considering choice amongmultiple reinforcers is behavioral choice theory (Vuchinich &Tucker, 1988). The core paradigm for studying choice is to provideaccess to two or more alternatives and to vary the amount of work(schedules of reinforcement) required to gain access to each of thealternatives. This allows for the determination of the relativereinforcing value of the alternatives. In the concurrent scheduleparadigm, the subject needs to choose which alternative to workfor. If the schedules of reinforcement are the same for bothalternatives, then the pattern of responses will most likely reflectthe reinforcing value of each alternative. Choice responding,which is often referred to as preference, depends on the reinforcingvalue of the alternatives and constraints on the alternatives. Forexample, if someone has equal access to two types of food in thehome, and both are ready to eat and require the same behavioralcost to prepare, then they are likely to choose the meal that theyfind most reinforcing. However, if the behavioral cost of thepreferred meal is much higher than that of the less-preferred meal,and it requires a lot of preparation and perhaps leaving the houseto obtain new ingredients, the person may shift preference from themore-preferred to less-preferred meal. If the behavioral cost of analternative is increased, then someone will shift choice from pre-ferred to less-preferred foods (Lappalainen & Epstein, 1990).

One way to define how reinforcing an alternative is in relation-ship to a second alternative is to establish baseline responding for

Table 1Levels of Analysis of Behavioral Economics of Food Reinforcement

Level of analysis Research area

GeneticMolecular Identify genes associated with excess energy intakeBehavioral Use co-twin design to control for the influence of genetics and test the influence of the

environment on food reinforcementPhysiological

Neurobiological Identify brain mechanisms associated with food as a reinforcer and substitution of nonfood forfood reinforcers

Metabolic Determine the role of insulin and glucose in food reinforcementBehavioral

Nutritional Compare food reinforcement value for carbohydrates, fat, and proteinModerators of food reinforcement Identify role of deprivation and satiation on food reinforcement value

ClinicalPharmacological Test agents to reduce reinforcing value of foodBehavioral Identify alternatives to food as a reinforcer

PolicySchools Increase variety for healthy foods and reduce variety for less healthy foodsPricing Increase prices (taxes) on high energy dense, low nutrient dense foods, and/or reduce prices

(subsidies) on low energy dense, high nutrient dense foods

885FOOD REINFORCEMENT

both alternatives and then place constraints on access to the morereinforcing alternative while the behavioral cost of responding forthe least-preferred alternative stays the same. In this paradigm,there will be a greater rate of responding for the more-preferredalternative when the behavioral costs are equal, but as the behav-ioral cost for the more-preferred alternative increases relative tothe cost for the less-preferred alternative, the responses shouldindicate a shift from responding for the more- to the less-preferredalternative. The point at which the work shifts from the most-preferred to least-preferred alternative defines the crossover orswitch point, which provides an estimate of the relative reinforcingvalue of that commodity. Another approach is to increase thebehavioral cost for each alternative separately as that alternative isearned. In this paradigm, the behavioral costs for each alternativeincrease independently on the basis of responding for each alter-native. As described by Bickel (Bickel et al., 2000), increasing thebehavioral cost can shift the pattern of responding. For example,there may be no differences in choice between two alternativeswhen the behavioral costs to obtain them are the same, but definitepreferences for one of the alternatives may be observed when thebehavioral cost increases. Alternatively, someone may have arelative preference for one alternative when the cost to obtain thealternatives is low, but as the behavioral cost increases, the pref-erences may shift to the alternative that was initially least preferred(Bickel et al., 2000).

Comparing the reinforcing value of alternative reinforcers withstudy-relative reinforcing value is a time-consuming process. Analternative method for determining relative reinforcing value maybe to use a questionnaire. Griffiths and colleagues (Griffiths, Rush,& Puhala, 1996) developed a questionnaire to assess the relativereinforcing value of nicotine, which has been adapted for food byGoldfield and colleagues (Goldfield, Epstein, Davidson, & Saad,2005). Questionnaires have been used in previous research todetermine a food reinforcement phenotype that interacted withdopamine genotypes to predict food intake (Epstein et al., 2004b).The use of a questionnaire to get an estimate of relative reinforcingvalue provides the opportunity to test a wider number of alterna-tives that can then be further studied with more precise behavioralmeasures in laboratory settings.

The relationship between responding and response requirementsas defined by the schedule of reinforcement can be considered byconstructing demand curves. Demand curves provide an indicationof the proportional change in responding as a function of propor-tional change in behavioral requirements to gain access to food(Figure 1). Figure 1, adapted from a theoretical analysis of demandcurves and reinforcer efficacy in behavioral pharmacology (Bickelet al., 2000), presents rates of responding and energy consumed asa function of the amount of work required to earn a reinforcer,which is specified at fixed ratio (FR) schedules of FR 1, FR 10,through FR 10,000. An FR 10,000 schedule requires 10,000 re-sponses to gain access to a reinforcer, which is a higher behavioralcost than an FR 1,000 schedule, which requires 1,000 responses togain access to the reinforcer, and so on. As the figure shows, asbehavioral cost increases from FR 1,000 to FR 10,000, responsesto obtain food decrease, and so does energy consumption. Thisrelationship can be considered as an indication of demand elastic-ity. Elasticity is a function of behavioral cost such that from an FRof 1 to an FR of 1,000 the curve is relatively inelastic, that is,subjects increase their responding linearly to maintain a relatively

constant energy intake. However, the relationship between re-sponding and food is elastic for FRs greater than 1,000. Theseshifts in the pattern of responding as a function of changes in thebehavioral cost to obtain the reinforcer provide a strong rationalethat reinforcer efficacy cannot be established on the basis ofcomparisons at one response requirement. Because preferencedepends in part on the behavioral cost, evaluation of choice onreinforcer efficacy at one behavioral cost provides only a verylimited view of absolute or relative reinforcing values of behav-iors. Just as the effects of a drug cannot be determined by studyingone dose, reinforcer effectiveness needs to be tested across mul-tiple response requirements.

One of the most relevant aspects of the demand curve forapplication to eating is the cause of the shift from elastic toinelastic consumption, or a shift from increasing responding forfood to reducing responding for food. The change in the shape ofthe curve is likely due to a combination of satiation of the rein-forcer and demands placed on the subject by the increased workrequired by the schedule. The use of the term satiation in rein-forcement theory refers to the point at which subjects have con-sumed enough of the reinforcer so that they will not engage in anymore behavior to obtain it. We return to other uses of the termsatiation in a later section.

Absolute Versus Relative Reinforcing Value

Food reinforcement can be measured with either absolute orrelative reinforcing value, or whether only one or multiple optionsare available. The patterns of responding in single versus concur-

Cal

orie

s C

onsu

med

Hypothetical Demand CurveOutput Function

10 100

1,00

0

10,0

001

10

1

100

Res

pons

es P

er S

essi

on

100

1,000

10,0000

10,000

Omax

ElasticInelastic

Pmax

Fixed Ratio

Figure 1. Hypothetical demand curve for food. The Omax represents themaximal amount of responding for food, and the Pmax the behavioral priceat which the greatest amount of responding is observed. As the behavioralcost to obtain food increases, responding increases up to an FR 1,000, andthe relationship between behavioral cost and consumption is inelastic.After behavioral costs of FR 1,000 or greater, response rate decreases, andthe relationship is elastic. The reduction in responding may be due to acombination of reinforcer satiation and the behavioral cost of obtaining thefood. Adapted from a figure describing demand curves for drug reinforcersby Bickel and colleagues (Bickel et al., 2000).

886 EPSTEIN, LEDDY, TEMPLE, AND FAITH

rent schedules may be similar, and a behavior that is very rein-forcing in one paradigm may be determined to be very reinforcingin the other paradigm (Bickel et al., 2000). Whereas this is true formany comparisons, there may be important exceptions that warrantuse of choice paradigms. One of the basic assumptions of choiceparadigms is that choice depends in part on the alternative that isavailable (Bickel, Madden, & Petry, 1998). Thus, if the alternativehas low reinforcing value, then there may be few differencesbetween absolute and relative reinforcing value. However, if thealternative is very reinforcing, then the absolute and relative rein-forcing values may be quite different. Take for example assess-ment of the absolute reinforcing value of an apple or the relativereinforcing value of the apple versus broccoli. Because for mostpeople broccoli is not a very reinforcing alternative, the determi-nation of absolute and relative reinforcing value for the appleshould be similar. However, consider if chocolate is the alternativeto the apple. In this case, the relative reinforcing value of the appleversus chocolate would likely be lower in comparison with theabsolute reinforcing value of the apple when studied alone.

In addition, a choice paradigm provides information on relativereinforcer efficacy that is not available when only one reinforcer isstudied. There is ecological validity to a choice paradigm becausein the real world eating is almost always a choice, and studyingfood intake when only one behavioral option is available may notprovide a realistic appraisal of the reinforcing value of food. In aninteresting study on measuring liking of food in laboratory andfield settings, investigators showed that the correspondence be-tween ratings was highest when the laboratory setting involved achoice among foods, rather than presenting individual foods, asproviding a choice may simulate the natural environment (de Graafet al., 2005).

Integrating General Reinforcement Theory With FoodReinforcement

Food reinforcement can be considered to be a specialized ex-ample of more general reinforcement theory. There are generaltheories of reinforcement that describe the conditions under whichbehaviors function as reinforcers and provide a priori determina-tion of what types of activities or stimuli can be used to motivatepeople to change (Timberlake & Farmer-Dougan, 1991). A prioridefinition of what types of behaviors or activities can function asreinforcers reduces the circularity in definitions of reinforcers thatare based only in terms of their function, which is an importantaddition to reinforcement theory. A theoretical approach to under-standing how something becomes reinforcing is disequilibriumtheory, which defines reinforcers in terms of constraints on accessto usual responding (Timberlake & Farmer-Dougan, 1991). This isan approach that extends the Premack principle (Premack, 1959),which states that the unconstrained rate of a behavior is an indi-cation of the reinforcing value of that behavior, as high probabilityresponses can be used to reinforce lower probability responses. Insimple terms, this means that most people engage in behaviors theyfind reinforcing and that you can motivate people to perform a lowrate, nonreinforcing behavior by scheduling a higher rate, morereinforcing behavior contingent on completing the lower rate be-havior. Disequilibrium theory also uses response rates as the basisfor reinforcer effectiveness, but disequilibrium theory argues thatbase rate alone is not sufficient to assess reinforcer efficacy; rather,

constraint on access to behavior is critical for something to bereinforcing and for that to motivate behavior. In this account, anybehavior can be reinforcing if that behavior is constrained to bereduced below baseline levels. Although reinforcing value is usu-ally conceptualized in relationship to high preference behaviorsthat people are very motivated to engage in, the concepts ofresponse deprivation and disequilibrium theory bring these con-structs to bear on everyday behaviors, such as reading, sewing,painting, and candle making (Bernstein & Ebbesen, 1978).

Response deprivation is a condition that can increase the rein-forcing value of a behavior, just as food deprivation is a conditionthat can increase the reinforcing value of food (H. A. Raynor &Epstein, 2003), which will be discussed later in the review. Sim-ilarly, response satiation can reduce the reinforcing value of abehavior, just as food satiation reduces the motivation to consumemore food. These analogies argue that at least some of the effectsof food deprivation or satiation are examples of more generalprinciples of behavior. Research is needed to determine the com-monalities and differences between food-specific effects on moti-vation because of food deprivation and the more general effectsthat may be due to response deprivation.

Factors That Influence the Reinforcing Value of Food

Relationships Between Alternatives in a ChoiceParadigm: Substitutes and Complements

In a choice situation the reinforcing value of food depends onthe reinforcing value of the alternatives as well as the constraintson access to food. For example, adults and children chose to workfor the more-preferred food when given a choice between pre-ferred versus less-preferred foods, but as constraints on the pre-ferred food increased, subjects chose to work for the less-preferredfood (Lappalainen & Epstein, 1990; J. A. Smith & Epstein, 1991).Likewise, when provided a choice between preferred snack foodsversus fruits and vegetables or pleasurable sedentary activities,subjects chose snacks but shifted their choice to the alternatives asthe work needed to gain access to snacks increased (Goldfield &Epstein, 2002). Thus, increasing the behavioral cost to obtain thepreferred commodity reduces responding for that commodity (in-creased behavioral cost for A reduces consumption of A). Inaddition, increasing the behavioral cost to obtain the preferredcommodity can increase responding for the less-preferred com-modity (increased behavioral cost for A increases consumption ofB). The alternative commodity is called a substitute. The substitutemay not be as reinforcing but may be chosen when the behavioralcost to obtain the preferred commodity becomes too high. Forexample, assume that someone regularly drinks coffee as a stim-ulant to wake up in the morning. The person is out of coffee thatmorning but has access to a supply of breakfast tea. The choice isto get dressed and go out to get coffee, which requires more work,or to drink breakfast tea, which also has caffeine, but is not aspreferred, but is more accessible. The difference in the amount ofwork needed to switch from the usual choice to a new choiceprovides a way to index how substitutable are different commod-ities.

Figure 2 provides graphs of hypothetical youths who substituteor do not substitute healthy food for less healthy food (e.g., potatochips) when the behavioral cost of obtaining the less healthy food


is increased and the behavioral cost of obtaining the healthier food(apples, for example) stays the same. The left graph demonstratessubstitutions. Consumption of less healthy foods decreases as thebehavioral cost of less healthy foods increases, whereas consump-tion of healthy foods concomitantly increases. In the right graph,the youths also reduce consumption of less healthy foods asbehavioral cost increases but do not increase consumption ofhealthier alternatives. They do not substitute healthier eating as thebehavioral cost of less healthy eating increases.

Responding for alternatives in a choice methodology can berelated in another way. When responding to gain access to abehavior decreases, responding for a different, related behaviormay also decrease (increased cost for A reduces consumption ofB). This is evidence for a complementary relationship. For exam-ple, food intake is often associated with television watching, andreducing television watching can reduce food intake (Epstein,Roemmich, Paluch, & Raynor, 2005a). Thus, food intake can bereduced by increasing the behavioral cost to obtain the food,providing reinforcing alternatives to eating, or reducing access tovariables that are reliably associated with eating.

Food Deprivation or Restriction

Recent experience with a reinforcer will alter reinforcing effec-tiveness. Deprivation can increase reinforcer effectiveness and themotivation to consume the reinforcer (H. A. Raynor & Epstein,2003). This occurs after as little as 4 hours, when the person isexpecting the next meal, and prefeeding prior to a meal reduces thereinforcing value of that meal (H. A. Raynor & Epstein, 2003).Reinforcer sensitivity theory (Reiss & Havercamp, 1996) arguesthat individual differences in reinforcement should not be tested

under deprivation conditions because deprivation increases rein-forcer effectiveness for everyone. Reinforcer effectiveness shouldrather be tested when the person is not deprived of the reinforcerbut remains motivated to obtain it. This suggests that in studies ofindividual differences in food reinforcer effectiveness, deprivationshould be limited. One way to do that is to preload, or provide foodprior to beginning the experimental trials, to remove respondingdue to hunger to provide a better indication of the influence ofreinforcement effects on responding. This is the strategy that wasused in research on the influence of interaction of food reinforce-ment and dopamine genotypes as markers of dopaminergic activityin smokers (Epstein et al., 2004b).

In addition to the acute effects of reinforcer deprivation, theremay also be long-term effects of food deprivation on reinforcereffectiveness. For example, Carr and associates have shown thatfood deprivation can increase bar pressing of rats to gain access toelectrical stimulation in the lateral hypothalamus (Carr, 1996).This increased responding has been described as sensitization toreward effects. Sensitization is a term often used to define in-creases in drug reinforcement for the same drug dose over time. Ifrepeated or chronic food deprivation can also sensitize food rein-forcer effects, then repeated deprivation that occurs with veryrestrictive diets may increase the reinforcing value of food overtime.

A construct that may be related to deprivation is dietary re-straint. People who evidence dietary restraint inhibit eating insituations in which they may want to eat (Lowe, 1993). There is animportant conceptual difference between deprivation and restraintin that many people with dietary restraint are not in fact reducingtheir energy intake (Lowe, 1993). It is also possible that dietary

Example of substitution of healthyfor less healthy foods

Behavioral cost

Ene

rgy

Inta

ke

100

120

140

160

180

200

220

240

260

Less healthy foodsHealthy foods

Behavioral costE

nerg

y In

take

100

120

140

160

180

200

220

240

260

Example of no substitution of healthyfor less healthy foods

Less healthy 10 20 30 40 50 Healthy 10 10 10 10 10

Less healthy 10 20 30 40 50 Healthy 10 10 10 10 10

Figure 2. Hypothetical examples of youths who substitute healthy foods for less healthy foods (left graph) ordo not substitute healthy foods for less healthy foods (right graph) when the behavioral cost of obtaining the lesshealthy food is increased.


restraint, although not always associated with negative energybalance and weight loss, does involve intermittent periods of fooddeprivation followed by periods of usual eating or overeating thatresult in energy homeostasis and weight maintenance or gain(Lowe, 1993). Several studies have now shown that girls andadolescents who self-reported dieting are heavier later (Field et al.,2003; Stice, Cameron, Killen, Hayward, & Taylor, 1999; Stice,Presnell, Shaw, & Rohde, 2005). If dietary restraint does involveintermittent periods of brief food deprivation, or deprivation ofspecific foods, these girls may paradoxically be increasing thereinforcing value of these foods. There are of course a number ofother potential explanations for this effect. For example, the girlswho restrict energy intake may be those girls who recognize thatthey are gaining weight and restrict energy intake as a protectivestrategy, so the weight problem may be the cause rather than theresult of the restriction. Research is needed to assess whetherdietary restraint, characterized by not consuming as much food asone may want, involves brief periods of energy restriction, whichmay increase the reinforcing value of food.

The effects of dieting on craving have been studied. Craving isrelated to the conditioned incentive value of a reward (Robinson &Berridge, 1993). A series of studies on dieting subjects, who wouldbe expected to be deprived of both total energy intake as well asspecific foods, showed a reduction in food cravings (Harvey,Wing, & Mullen, 1993; Martin, O’Neil, & Pawlow, 2006), withgreater reduction for those who eliminated more of these foodsfrom their diet (Martin et al., 2006). One possible explanation fora reduction in cravings during dieting is based on the idea thatcraving is in part a conditioned response, and removing the con-ditioned stimuli could reduce craving. Programs that remove theopportunity for presentation of cues related to restricted foodswould reduce cravings for these foods. Craving and incentivesalience are theoretically distinct constructs from reinforcing value(Berridge & Robinson, 2003), and it is important to understandconditions that are associated with similar versus distinct patternsof change in these constructs.

Reinforcer Satiation

Recent experience with a reinforcer can reduce the motivation toobtain that reinforcer, which is called reinforcer satiation. Thisgeneral principle applies to food as well because preloading mayreduce energy intake in comparison with food deprivation (Shide,Caballerro, Reidelberger, & Rolls, 1995; Spiegel, Shrager, & Stel-lar, 1989). On the basis of analysis of demand curves (Bickel et al.,2000), responding to obtain a reinforcer might show an increase inresponding as the session is initiated, with a reduction in rate ofresponding over time until reinforcer satiation, when there isreduced responding to obtain food. The observation of shifts inmotivation to eat is mirrored by data on eating rate, which show agreater eating rate in the beginning of a meal followed by areduced eating rate over time (Bellisle, Lucas, Amrani, & LeMag-nen, 1984; Spiegel et al., 1989).

The shifts in response rate depend on the types and variety offood that are available, not only on cessation of responding fromreduction of energy depletion. A consistent body of research hasshown that animals (McSweeney, Hinson, & Cannon, 1996; Mc-Sweeney & Swindell, 1999) and humans (Epstein, Saad, et al.,2003; Temple, Kent, et al., 2006) will reduce responding if the

same food is presented over time but will reinstate instrumentalresponding for food when a new food is presented. This suggeststhat reinforcer satiation does not depend only on shifts in energystores.

It is relevant to consider that ingestive behavior researchers alsouse the term satiation. In eating behavior studies, satiation usuallyrefers to the termination of eating, and the amount of food con-sumed is used as an objective measure of satiation. There may besituations in which the findings in experiments used to studysatiation are related to changes in reinforcer effectiveness. Forexample, in studies in which preference for two foods is studiedand the amount consumed is used as the measure of preference,shifts in responding for the two foods may involve reinforcementmechanisms. However, there are also many variables in ingestivebehavior research that are not likely to be related to reinforcement.For example, shifts in energy intake as a function of energy densityand food volume may be examples of dietary factors that alter foodintake through mechanisms other than food reinforcement (B. J.Rolls et al., 1998; B. J. Rolls, Roe, & Meengs, 2006).

Macronutrient Composition

Although the construct of food reinforcement is often used todescribe the motivation to eat, it probably is more appropriate toconsider the reinforcing value of particular types of food ratherthan that of food in general. Someone who finds apple pie rein-forcing may not work hard for broccoli. It is possible for someoneto work hard for several different types of foods, and it is likelythat there are particular classes of foods that the person would notbe very motivated to obtain. There has been extensive research onhow macronutrient composition of food may influence food intake(B. J. Rolls, 1995), and some of these effects may be related tofood reinforcement. People usually select foods because they tastegood, not on the basis of whether they are primarily constituted ofcarbohydrates, proteins, or fats (Levine, Kotz, & Gosnell, 2003).Foods that taste good are often high in sugars and/or fat (e.g., icecream; Drewnowski & Greenwood, 1983). Obesity has been as-sociated with a craving for foods containing mixtures of fats andsugars (Drewnowski, Krahn, Demitrack, Nairn, & Gosnell, 1992;Drewnowski, Kurth, Holden-Wiltse, & Saari, 1992).

Sugar preference may be established early in development(Desor, Maller, & Turner, 1973), as sugar can calm stressed infants(Blass & Watt, 1999; B. A. Smith & Blass, 1996), and suckling ismotivated in part in infant rats by sugar (Blass, 1990). Sugarreinforcement may be primary because animals prefer sugar towater in their initial exposure (Ackerman, Albert, Shindledecker,Gayle, & Smith, 1992). Research suggests that the opioid systemis in part responsible for the pleasantness of sweet taste (Ackermanet al., 1992; Bertino, Beauchamp, & Engelman, 1991) and in partfor food reward in general (Kelley et al., 2002). Animals aremotivated to work for sugar reward (Sclafani & Ackroff, 2003),and sugar has effects similar to some drugs of abuse, as intermit-tent bingeing on sucrose and drugs of abuse repeatedly increasesextracellular dopamine in the nucleus accumbens shell (Rada,Avena, & Hoebel, 2005).

Dietary fats may also influence the reinforcing value of food.Infant rats respond for dietary fat (Ackerman et al., 1992), and ratsprefer petroleum or mineral oil, which have no calories, as well aslard or Crisco-blended lab chow, suggesting that at least some of


the effects of fat as a reinforcer may not depend on energy(Ackroff, Vigorito, & Sclafani, 1990; Hamilton, 1964). Freed andGreen (1998) explored the relative reinforcing value of corn oilversus mineral oil and of sucrose or saccharin solutions versusplain water in rats. The demand and consumption for corn oildepended on the changes in response requirements (reinforcementschedule) for corn oil as well as availability of alternatives. Whenthe behavioral cost to obtain mineral oil, sucrose, or saccharinalternatives was lower than the behavioral cost of corn oil, then therats responded for the alternatives. However, plain water did notserve as a substitute for corn oil. These results do not suggest thatthere is a basic need for dietary fat that is defended as thebehavioral cost increases but that dietary fat is a reinforcer that canbe substituted for by other substances differing in energy and intaste.

To our knowledge, there is no research on the specific reinforc-ing value of dietary protein. This research may be useful inshedding light on the way in which protein influences dietaryintake because dietary protein has a major influence in satiety(Vandewater & Vickers, 1996).

Food Variety

A consistent body of research has shown that people consumemore energy when given a variety of food than when given thesame food (H. A. Raynor & Epstein, 2001; B. J. Rolls et al., 1981),even when variety involves small differences in food characteris-tics, such as differently shaped pasta or different flavors of yogurt(B. J. Rolls, Rowe, & Rolls, 1982). Several studies have demon-strated that meal variety increases energy intake (H. A. Raynor &Epstein, 2001). For example, subjects receiving a four-course mealconsisting of different foods in each course consumed 60% morecalories than did subjects receiving the same food in each course(B. J. Rolls, van Duijvenvoorde, & Rolls, 1984). Even when thevariation in the food was more subtle, such as different-flavoredcream cheese in sandwiches, subjects consumed more in the vari-ety group than in the group receiving the same-flavored sandwich(B. J. Rolls, Rolls, & Rowe, 1982). Studies have also investigatedthe influence of variety when all foods are served in a single courseand the same phenomenon is observed, even when the total energydensity presented is held constant (Pliner, Polivy, Herman, &Zakalusny, 1980; Spiegel & Stellar, 1990).

A propensity to seek out and respond to food variety may ensurea balanced nutrient intake (H. A. Raynor & Epstein, 2001). How-ever, in the current obesiogenic environment, dietary variety maybe contributing to the growing obesity epidemic (McCrory et al.,1999; H. A. Raynor & Epstein, 2001). Consumption of a variety ofhigh energy density foods (energy density � kj/g), coupled withlow intake of fruits and vegetables and a lack of physical activity,leads to maintenance of positive energy balance and, possibly, toobesity (McCrory et al., 1999; H. A. Raynor & Epstein, 2001).McCrory and colleagues (McCrory et al., 1999) reported that therewas a positive relationship between body mass index (BMI) andvariety of sweets, snacks, carbohydrates, entrees, and condimentsconsumed as well as a negative relationship between BMI andconsumption of a variety of vegetables (excluding potatoes) inweight-stable adults who accurately reported food intake for 6months. Likewise, those who lost the most weight and maintainedthe greatest degree of weight loss in an obesity treatment program

consumed the smallest variety of high-fat foods, oils, and sweetsand the largest variety of low-fat breads and vegetables (H. A.Raynor, Jeffery, Tate, & Wing, 2004).

Variety of foods also increases responding for food in compar-ison with presentation of one food in adults (Myers & Epstein,2002) and children (Temple, Giacomelli, Roemmich, & Epstein, inpress). Similarly, introducing a new food after responding for thatfood has decreased is associated with an increase in responding toobtain the new food (Epstein, Saad, et al., 2003). Introducing avariety of foods, or new foods after reinforcer satiation, maymaintain the motivation to eat.

Food variety can be conceptualized more broadly than the foodsthat are available within a meal in terms of foods available incafeterias, stores, or in the food supply. The increase in foodvariety is greater for higher energy dense snack foods than forhealthier alternatives, such as fruits and vegetables (McCrory etal., 1999). A potentially important public health and policy direc-tion might be to limit the varieties of less healthy alternatives inschools to promote healthier eating and perhaps to encouragedevelopment of new varieties of forms of healthier foods, whichmay provide more alternatives to people when they are making thechoice of what types of foods to eat as well as potentially increas-ing intake of healthier alternatives.

Stimulus Control and Conditioned Reinforcing Value ofFood

When a behavior is consistently reinforced in the presence of aunique stimulus, that stimulus begins to influence the rate of thebehavior. This is called stimulus control. There are many examplesof how environmental stimuli can influence eating. For example,when a unique environmental stimulus is reliably paired witheating, then presentation of that stimulus will initiate eating insated rats (Weingarten, 1983). Similarly, time of day can serve toinitiate eating, and humans eat at regular intervals that are definedby time rather than by nutrient depletion (Schachter, 1968).

In addition to the association of neutral external stimuli onmotivated responding for food, stimulus characteristics of food,such as smell or sight of food, may alter motivated responding(E. T. Rolls, 1997). Small portions of food may prime an increasein motivated responding for food similar to low doses of drugspriming increases in drug self-administration (de Wit, 1996). Theeffect of priming is another example of how changes in foodreinforcement cannot be explained by shifts in energy depletion,because an energy depletion model would argue that providing apriming amount of food would reduce, rather than increase, sub-sequent food intake.

There may be individual differences in response to stimuluscontrol of eating that are important in the development of obesity.For example, obese children are more responsive to olfactory cuesthan are leaner youths (Jansen et al., 2003). Responsivity to foodcues could lead to an increased motivation to eat and, thus, topositive energy balance and obesity. In addition, the construct ofeating in the absence of hunger may be related to cue control ofeating. In this paradigm, youths are allowed to eat to fullness andthen presented with additional food. Individual differences ineating in the absence of hunger have been related to pediatricobesity (Fisher & Birch, 2002).


Because food is a primary reinforcer, activities associated withfood may derive secondary reinforcing value. For example, ifyouths regularly eat in association with television watching, andeating is reinforcing for these youths, then television watching canbecome a conditioned reinforcer. This represents a powerful wayin which a response class of behaviors associated with eating canbecome reinforcing and can motivate further behavior. The greaterthe association between food and activities associated with eating,the stronger the conditioned reinforcing value of these behaviors(Mazur, 1995). When youths are in situations in which they cannoteat and have to choose alternative activities, conditioned reinforc-ing activities that have been associated with eating may takeprecedence over activities that are incompatible with, and have notbeen paired with, eating. This may stimulate further eating whenfood is available. Pairing of environments with eating may repre-sent an analogue to conditioned place preference in animal studies.Animals reliably prefer to allocate time to environments that havebeen associated with drug delivery (Bardo & Bevins, 2000), andconditioned place preference is used to assess motivational valueof drugs (Meririnne, Kankaanpaa, & Seppala, 2001). Similarly,people who find food reinforcing may prefer to spend time inenvironments associated with food, which may then stimulatefurther eating when food is available.

Depression and the Reinforcing Value of Food

There are many psychological factors that may influence eating(Greeno & Wing, 1994; Laitinen, Ek, & Sovio, 2002; Pecoraro,Reyes, Gomez, Bhargava, & Dallman, 2004), and some of thesemay in part be mediated by changes in reinforcing value of food.It is interesting to note that basic research in both animals andhumans (Willner et al., 1998) has shown that subjects find sweetfoods to be more reinforcing after depressive mood induction.There are wide individual differences in the influence of depres-sion on body weight, with some people gaining weight whiledepressed and others losing weight (Barefoot et al., 1998). Onereliable individual difference between who may eat more or less isa history of weight fluctuation or obesity. Individuals with highlevels of depressive symptoms with a higher baseline BMI showgreater weight gain over time than do those with lower BMI levels(Barefoot et al., 1998). Likewise, heavier individuals show lessreduction in appetite with major depression than do those whoweighed less (Berlin & Lavergne, 2003). Obesity has been asso-ciated with depressive moods (Lee, Kim, Beck, Lee, & Oh, 2005).Additional research is needed to test whether depressive moodinfluences a variety of foods, in addition to the reinforcing value ofsweet foods and whether the reinforcing value of food is increasedunder other psychological states that increase food intake, such asstress (Greeno & Wing, 1994; Roemmich, Wright, & Epstein,2002).

Validity and Reliability of Food Reinforcement

Reinforcement has been an important area of research, with along history of relevance to basic and applied behavioral science aswell as an intense area of study in neuroscience (Salamone &Correa, 2002). It is interesting that the same basic paradigms canbe used to study reinforcer efficacy in animals and humans (Ator& Griffiths, 2003; Griffiths et al., 2003). The majority of research

on reinforcement is laboratory research to study reinforcementprocesses and variables that influence and mediate reinforcement.Outside of the laboratory in the natural environment, people do notrespond on an operant task to gain access to food, though there aremany aspects of food procurement, preparation, and consumptionvarying in their behavioral costs that may be related to foodreinforcement. For example, foods may be stored in your homeand easy to access or may have to be purchased outside of thehome and are thus less accessible. Foods may be prepared or easyto eat versus in their raw form and much less accessible. A goodexample of this is snack food. High energy dense snack foodsgenerally come ready to eat, whereas healthier, lower energy densesnack foods often require some preparation. Some fruits and veg-etables require preparation to clean, cut, and make them intoappropriate portions sizes, though some important advances inconvenience are fruit and vegetable snacks such as baby carrots orprepared celery sticks. Given the differences between the labora-tory and extralaboratory contexts, it is worthwhile to consider thevalidity of laboratory measures of food reinforcement.

There are three ways in which the validity of food reinforcementhas been studied. First, the relationship between food earned andfood consumed within a food reinforcement session has beenstudied. In laboratory measures of food reinforcement, subjectscan consume the food they earned, and thus it would be expectedthat there would be a relationship between the amount and types offood earned and the amount of food consumed in that session.Research has shown that when the reinforcing value of food isincreased, the amount of food consumed and energy intake alsoincreases (Temple, Giacomelli, et al., in press). Second, it wouldbe expected that subjects high in food reinforcement would con-sume more food and energy in a separate ad-lib eating task, whichhas been demonstrated (Epstein et al., in press; Epstein et al.,2004a). Third, if elevated food reinforcement is associated withincreased energy intake, then it would be expected that obesepersons, who are in positive energy balance because of higherenergy intake than that of leaner peers, would respond more forfood in laboratory reinforcement tasks than would leaner subjects.This relationship has been observed (Johnson, 1974; Saelens &Epstein, 1996).

An extension of reinforcement theory is to consider reinforcersas commodities and behavioral cost as equivalent to price by usingeconomic principles (Staddon & Ettenger, 1989). A basic principleof economics is that increasing the price of a commodity reducespurchases, just as increasing the behavioral cost of obtaining areinforcer reduces consumption (Bickel et al., 2000). Consistentwith this notion, we have hypothesized that willingness to paymore for a commodity is related to reinforcing value of thatcommodity (Epstein, Dearing, Paluch, Roemmich, & Cho, inpress). Additional implications of an economic analysis of rein-forcement are discussed later in this review.

In many animal laboratory studies, animals are studied overrepeated sessions until a steady state of responding is observed,which ensures that the effects observed in each condition arereproducible. In many human laboratory studies, particularly thosein which individual differences are studied, only one measure ofreinforcing value is obtained, and it is important to assess in thesestudies whether patterns of responding are reliable over days. Toprovide initial data on the test–retest reliability of reinforcing valueof food, keeping the conditions of measurement constant, we


retested 20 subjects who participated in a study on the interactionof food reinforcement and dopamine genotypes on eating andfound a correlation of .80 between the switch point in the twosessions (Epstein et al., in press).

Food Reinforcement and Obesity

Obesity is a disorder of positive energy balance, in which energyintake exceeds energy expenditure. It is possible that obese personsare more motivated to eat than are nonobese persons, and researchshows that obese persons find food more reinforcing than dononobese persons (Johnson, 1974; Saelens & Epstein, 1996). Inaddition, when pleasurable activities are considered for obese andnonobese persons, eating is at the top of the list for the obese(Doell & Hawkins, 1982; Jacobs & Wagner, 1984; Westenhoefer& Pudel, 1993). Not only is the reinforcing value of food greaterfor obese than nonobese subjects, but craving for food may begreater, which may promote greater food seeking (Gendall, Joyce,Sullivan, & Bulik, 1998).

In addition to the influence of individual differences in foodreinforcement in the development of obesity, food reinforcementmay play a role in challenges associated with weight loss. One ofthe major problems in obesity research is the difficulty in long-term maintenance of weight loss (Jeffery et al., 2000). This may bedue in part to the fact that reducing energy to produce negativeenergy balance and weight loss necessarily involves some degreeof food deprivation, which may increase the reinforcing value offood (H. A. Raynor & Epstein, 2003) and energy intake (Telch &Agras, 1996). This creates a tension for obese persons, because thelonger they are in negative energy balance and the more weightthey lose, the more food reinforcement may increase and the moremotivated they may be to eat.

The energy depletion model of eating suggests that people eatbecause they are hungry, and eating should reduce hunger and thusthe amount of food subsequently consumed. Preloads should re-duce eating in a meal, and this pattern is often shown in nonobese,nondieting subjects (Spiegel et al., 1989). Obese subjects, how-ever, often increase eating and energy intake after a preload or afirst-course appetizer (Jansen et al., 2003; Spiegel et al., 1989).This effect may be due in part to the fact that obese individualsoften restrict energy intake, and it is the food restriction thatenhances the effect of the preload (Lowe, Foster, Kerzhnerman,Swain, & Wadden, 2001; McCann, Perri, Nezu, & Lowe, 1992).The observation that consumption of a preload or a small amountof food increases energy intake in a person who is food deprivedmay be a similar phenomenon to priming, in which presenting asmall amount of a drug increases drug self-administration (de Wit,1996). Priming effects do not necessarily depend on food depri-vation or restriction, as priming subjects with a palatable food canincrease eating even in satiated subjects (Cornell, Rodin, &Weingarten, 1989).

Dopamine and the Neurobiology of Food Reinforcement

Given the importance of food for survival, it is not surprisingthat there are multiple pathways and physiological systems work-ing to ensure that adequate energy intake is achieved. The regu-lation of food intake or appetite has both central and peripheralinfluences (Ahima & Osei, 2001; Schwartz, Woods, Porte, Seeley,

& Baskin, 2000) and is responsive to sensory information relatedto the sight and smell of food, previous feeding experiences,satiety signals elicited by ingestion, and hormonal signals relatedto energy balance (Ahima & Osei, 2001; Schwartz et al., 2000).Imaging studies show a comprehensive neuronal network control-ling human feeding behavior that includes parts of the cortex,hypothalamus, thalamus, and the limbic system (Gautier et al.,2000; Tataranni et al., 1999). This network connects the regions ofthe brain that control the perception of sensations and flavors (thecortex); that influence the reinforcing value of food (the limbicsystem); and that influence appetite, body weight, and energybalance (the thalamus and hypothalamus). The goal of this sectionis not to provide a complete review of the neurobiology of foodreinforcement but to focus on one factor that may be related tofood reinforcement, dopaminergic activity, and illustrate how do-paminergic activity may be related to factors that influence foodreinforcement.

Dopamine is generally considered to be a primary neurotrans-mitter involved in food reinforcement (Wise, 2006), though themechanisms for this effect are debated (Salamone & Correa,2002). Dopaminergic neurons in the substantia nigra and ventraltegmental area project to the caudate putamen and nucleus accum-bens, where they modulate movement and reward (Koob, 1999;Schultz, 1998). Other projections from the nucleus accumbens tothe lateral hypothalamus in part regulate feeding (Maldonado-Irizarry, Swanson, & Kelley, 1995). The process whereby food isor becomes reinforcing depends, in part, on the release, transport(reuptake), and receptor binding of synaptic dopamine (Wise,2006). The levels of extracellular dopamine in the brain depend ontwo processes: tonic dopamine release (Schultz, 1998; Stricker &Zigmond, 1984) and dopamine release in response to a stimulus(phasic dopamine release; Grace, 2000).

There are a number of ways in which understanding the neuro-biology of eating may elucidate mechanisms for some of thebehavioral phenomena that have been discussed. Eating elicitsdopamine release, which may be related to appetitive aspects offood seeking (Berridge, 1996). Food deprivation reduces messen-ger RNA for the dopamine transporter and reduces extracellulardopamine (Patterson et al., 1998; Pothos, Hernandez, & Hoebel,1995), which may stimulate the increase in reinforcing value offood after food deprivation. The dopamine release observed withconsuming food and then with stimuli associated with food may bein part the mechanism for two behavioral constructs discussed inthe previous section: stimulus control of eating and conditionedreinforcement value for stimuli associated with eating. It is possi-ble that conditioned increases in dopaminergic activity that areelicited by stimuli associated with eating play a role in craving forfood (Berridge, 1996). Initially, dopamine is released when food isconsumed (Hoebel, Hernandez, Schwartz, Mark, & Hunter, 1989),and its activity in the nucleus accumbens may relate not only toconsumption but also to rewarding properties of food (Hernandez& Hoebel, 1988). Over repeated pairings of an environmentalstimulus with eating, dopamine is released after presentation of acue paired with food in anticipation of eating. Dopamine thenbinds to dopamine D2 receptors in the nucleus accumbens, perhapsinitiating food-seeking behaviors (Kiyatkin & Gratton, 1994).Positron emission tomography studies in humans have shown thatbrain dopamine increases both in anticipation of and during theingestion of food (Small, Jones-Gotman, & Dagher, 2003; Volkow


et al., 2003). The influence of food variety on increased motivationto eat may also be regulated in part by dopamine. Researchsuggests that brain dopamine activity increases in response tonovel stimuli (Salamone, Correa, Mingote, & Weber, 2005), whichcould elicit the motivation to consume new foods and increaseenergy intake when a variety of foods rather than just one food isavailable.

An interesting aspect of research on behavioral choice and foodreinforcement is that dopamine may be involved in the demandelasticity for food (Salamone & Correa, 2002). Salamone (Salam-one & Correa, 2002) has used a behavioral choice paradigm inwhich rats can choose to work for very palatable food or have freeaccess to chow. The rats will normally choose to engage in workto gain access to favorite foods, but dopamine depletion shiftschoice to the freely available chow. In many conditions, depletionswill have a greater effect on higher ratio schedules. This has ledSalamone to conclude that dopamine is involved in the amount ofinstrumental behavior that is performed to gain access to foodrewards (Salamone & Correa, 2002). Thus, relative reinforcingvalue and behavioral choice paradigms may provide unique infor-mation about how dopamine influences behavior that cannot bedetermined by studying reinforcing value with single schedules ofreinforcement in the absence of choice.

Obese Versus Lean: Altered Dopaminergic Activity

Differences in neuronal activity in response to food and satiationin several brain regions of humans have been identified in obesefemale subjects compared with lean female subjects (Karhunen,Lappalainen, Vanninen, Kuikka, & Uusitupa, 1997) and malesubjects (Gautier et al., 2000). For example, obese subjects havesignificantly higher metabolic activity in the parietal somatosen-sory cortex, where sensation to the mouth, lips, and tongue islocated (Wang et al., 2002). Enhanced activity in the corticalregions involved with sensory processing of food in the obesecould make them more sensitive to the reinforcing properties offood related to palatability and could be one of the variablescontributing to their excess food consumption.

There is an extensive animal literature that supports the rela-tionship between obesity and altered dopamine metabolism. Obeserats release more dopamine during eating than do their lean coun-terparts (Yang & Meguid, 1995). Dopamine may also inhibitfeeding through its influence on other central mediators of appe-tite. In the hyperphagic obese leptin-deficient mouse model, in-creasing central dopaminergic tone with dopamine agonists elim-inates the hyperphagia via inhibition of Neuropeptide Y, one of themost potent orexigenic (appetite stimulating) agents known (Bina& Cincotta, 2000). Central dopamine activity may also be involvedin the appetite suppression associated with Interleukin-1 alpha-associated anorexia (Yang et al., 1999) and with the central effectsof cholecystokinin (Chung et al., 1994).

Deficiencies in dopaminergic receptors may also play a role inthe development of obesity. For example, obese Zucker rats havefewer dopamine D2 receptors and reduced hypothalamic dopamineactivity when fasting but release more dopamine when eating anddo not stop eating in response to the peripheral administration ofinsulin and glucose when compared with the same in lean Zuckerrats (Orosco, Rouch, & Nicolaidis, 1996). Obese male Sprague-Dawley rats have reduced dopamine turnover in the hypothalamus

compared with that of the diet-resistant strain before they becomeobese (Levin & Dunn-Meynell, 1997), and develop the obesephenotype only when given access to a palatable high energy diet(Levin & Dunn-Meynell, 2002). Dopamine D2 receptor blockadecauses obese but not lean rats to overeat (Fetissov, Meguid, Sato,& Zhang, 2002; Orosco, Gerozissis, Rouch, Meile, & Nicolaidis,1995). Because a chronic change in receptor expression is aprerequisite for behavioral sensitization (Kuczenski & Segal,1999), these data suggest that the blockage of an already low D2receptor level may sensitize obese rats to food.

Common Neurobiological Substrates for Obesity andDrug Addiction

Recent research suggests that common neurobiological sub-strates underlie obesity and drug addiction (Volkow & Wise,2005). Similar to drug use, ingestive behavior may be unrelated tohomeostatic need, that is, driven not by the need to maintainenergy balance but by cognitive and environmental factors that arenot compensated for by the body’s internal feedback system tomaintain stable body weight (Berthoud, 2004), and individualdifferences in the responsiveness of brain dopamine to this “non-regulatory ingestive behavior” have been identified in animals(Mittleman, Castaneda, Robinson, & Valenstein, 1986) and hu-mans (B. J. Rolls, 2003). There are other parallels between hy-perphagia and drug addiction. One interesting phenomenon thatmay suggest common pathways for food consumption and druguse is that chronic food restriction increases food reinforcement(H. A. Raynor & Epstein, 2003) and also increases the self-administration and motor-activating effects of abused drugs (Carr,2002; Carroll & Meisch, 1984; Pothos, Creese, & Hoebel, 1995),along with improving dopamine receptor function (Wilson, Nomi-kos, Collu, & Fibiger, 1995). In addition, an animal’s level ofsucrose preference can predict its desire to self-administer cocaine(Levine et al., 2003), and sweets will reduce cocaine’s reinforcingvalue (Comer, Lac, Wyvell, & Carroll, 1996), suggesting a relationbetween sweet taste and drug reward. Both obese rats and chronicdrug users have low basal dopamine levels (Hamdi, Porter, &Prasad, 1992), experience periodic exaggerated dopamine releaseassociated with either food (Fetissov et al., 2002) or drug intake(Worsley et al., 2000), and have reduced dopamine D2 receptorand increased D1 receptor expression (Fetissov et al., 2002). Anumber of addictive behaviors (alcoholism; cocaine, heroin, mar-ijuana, and nicotine use; and glucose bingeing) have been associ-ated with low expression or dysfunction of D2 receptors (Comings& Blum, 2000). Thus, reduced expression of D2 receptors mayrepresent a common pathway in obesity and drug addiction thatelicits behavior to release dopamine into areas of the brain that“crave” dopamine.

There are many similarities between food and drugs; there alsomay be important differences (Bassareo et al., 2003; Kelley &Berridge, 2002; Parkinson, Fudge, Hurd, Pennartz, & Peoples,2000; Salamone & Correa, 2002). For example, whereas food anddrugs of abuse both elicit dopamine release in the shell of thenucleus accumbens, habituation (reduced response with repeatedexposure) occurs with food (specific to the food ingested) that isresponsive to the motivational state of the animal (i.e., food re-stricted or not; Bassareo & Di Chiara, 1999). Such habituation isnot always seen with drug use (Bassareo et al., 2003). Research is


needed to examine the similarities and differences between eatingand drug self-administration to understand how best to applyresearch from drug addiction to eating and obesity.

This brief overview of dopamine and food reinforcement fo-cuses on dopamine as an example of how a more basic under-standing of the neurobiology of reinforcement may provide abetter understanding of mechanisms that influence food reinforce-ment. Although understanding the role of dopamine in reinforce-ment is a major issue in contemporary neuroscience (Salamone &Correa, 2002), new research may provide a better understanding ofthe role of dopamine in food reinforcement (Salamone & Correa,2002) and may identify other neurobiological variables that influ-ence food reinforcement in addition to, or interacting with, dopa-mine.

Behavioral Genetics of Food Reinforcement

If dopamine is involved in food reinforcement, then individualdifferences in dopaminergic activity that are related to dopaminegenotypes may predict factors related to dopaminergic activity,food reinforcement, and obesity. Most of the focus on dopamineand eating involves the D2 receptor and the DRD2 gene. Both theD1 and D2 dopamine receptors act synergistically to inhibit feed-ing (Terry, Gilbert, & Cooper, 1995). It is the DRD2 gene, how-ever, that appears to function primarily as a reinforcement orreward gene (Noble, 2003) having been associated with feedingand addictive behaviors (Baptista, 1999).

The DRD2 gene has three allelic variants (A1/A1, A1/A2, andA2/A2). Individuals with genotypes containing either one or twocopies of the A1 allele have fewer D2 receptors than have thoselacking an A1 allele and so are associated with reduced braindopamine signaling (Jonsson et al., 1999; Pohjalainen et al., 1998;Ritchie & Noble, 2003). In genetically homogeneous populations,such as the Pima Indians, a missense substitution in the DRD2gene has been associated with reduced resting energy expenditure(Tataranni et al., 2001), greater BMI, and type II diabetes (Jen-kinson et al., 2000). Epidemiological studies in genetically heter-ogeneous human populations have demonstrated a higher preva-lence of the A1 DRD2 allele in obese individuals (�50%)compared with lean individuals (�30%; Blum et al., 1996; Com-ings et al., 1993; Comings, Gade, MacMurray, Muhleman, &Peters, 1996; Noble et al., 1994). Low D2 receptor availability inthe presence of the DRD2 A1 allele likely makes subjects lesssensitive to stimulation of dopamine-regulated reward circuits(Blum et al., 2000), and so they may seek powerful reinforcers toovercome their “dopamine deficit,” increasing the risk not only forobesity but also for other related addictive behaviors (Blum et al.,1996). The DRD2 A1 allele may reduce the efficiency of thedopaminergic system, rewarding behaviors that increase braindopamine levels (Volkow, Fowler, & Wang, 2002), such as foodingestion. Consistent with this hypothesis, Wang et al. (2001) haveshown that D2 receptors are reduced in the striatum of the brainsof very obese (BMI � 40) humans in proportion to their BMI.They argued that persons with fewer D2 receptors seek powerfulreinforcers to overcome their “dopamine deficit,” increasing therisk of those with fewer receptors for obesity as well as for drugabuse (Wang et al., 2001). They and others (Baumann et al., 2000)have suggested that raising brain dopamine levels may be a treat-ment for obesity.

The SLC6A3 gene codes for the dopamine transporter (DAT-1),a protein that controls the levels of dopamine in the synapse byrapidly carrying the neurotransmitter back into nerve terminalsafter its release, limiting the level and duration of dopaminereceptor activation (Bannon et al., 1992). Synaptic dopamine in-creases when, for example, cocaine inhibits the reuptake of dopa-mine by binding to the DAT-1 or when the dopamine transportershave been eliminated by knockout of the SLC6A3 gene (Leshner,1996). The majority of studies examining polymorphisms of theSLC6A3 gene have focused on the variable number of tandemrepeats at the 3�UTR of the gene with reported evidence forassociation with the 10R allele (Need, Ahmadi, Spector, & Gold-stein, 2006). This allele may be associated with increased dopa-mine transporter density (Kirley et al., 2002) and transport (Swan-son et al., 2000) and therefore to lower levels of postsynapticdopamine (Heinz et al., 2000) when compared with the 9R allele,although this is not consistent across all studies (Jacobsen et al.,2000; Martinez et al., 2001). Recent data suggest that DAT-1 mayhave a role in the development or perpetuation of obesity. Elevatedlevels of DAT-1 have been found in obese rats (Figlewicz et al.,1998), and food deprivation (Patterson et al., 1998), food restric-tion (Bello, Sweigart, Lakoski, Norgren, & Hajnal, 2003), and theadministration of insulin (Figlewicz, Szot, Chavez, Woods, &Veith, 1994) to the brains of rats increases DAT-1 messenger RNAand alters DAT-1 function (Baskin et al., 1999; Figlewicz, 2003a,2003b). The up-regulation of DAT-1 during food restriction in ratsis similar to the neuroadaptation in the mesolimbic dopaminecircuitry resulting from drugs of abuse and is associated withincreased intake of sucrose, a palatable food (Bello et al., 2003).This mechanism may have relevance to pathological human feed-ing behaviors such as restrictive eating disorders, bingeing, andhyperphagia.

One application of the knowledge about genotypes related todopaminergic activity would be to use the genotype as a marker ofneurotransmitter activity. For example, if the 10R allele of theSLC6A3 gene is associated with higher density of DAT-1 protein,then individuals with this allele would presumably have less do-pamine available in the synapse. Similarly, if the A1 allele of thedopamine D2 receptor gene is related to fewer D2 receptors, thenindividuals with this allele would have fewer D2 receptors andlower dopaminergic activity. Therefore, these genotypes may be amarker of dopamine neurotransmission. This provides for theopportunity to indirectly assess relationships between dopaminer-gic activity and eating. Currently, we are not able to assess dy-namic changes in dopaminergic activity in human brains. Someways to study the neurobiology of eating include using positronemission tomography scans to examine dopamine receptor binding(Elsinga, Hatano, & Ishiwata, 2006), functional magnetic reso-nance imaging studies to examine regional blood flow and activa-tion of specific brain regions (Holsen et al., 2005), and postmortembrain analyses to examine dopamine receptor density (Piggott etal., 1999). The use of genotypes as markers of dopaminergicactivity may provide a useful tool to better understand how dopa-minergic activity is related to eating in behaviorally active con-texts.

As an example of this strategy, we used the DRD2 and SLC6A3genotypes as markers of dopaminergic activity in combinationwith a food reinforcement phenotype that has been examined insmokers. The food reinforcement phenotype was defined by using


a relative reinforcing value questionnaire that asked how hardsubjects would work for food or for money. Food reinforcementhad a main effect on energy intake as those high in food reinforce-ment consumed more food than did those low in food reinforce-ment. High food reinforcement interacted with both the SLC6A310R/10R and the DRD2 Taq 1 A1 genotypes (A1/A1 � A1/A2) toincrease energy intake beyond high food reinforcement with othergenotypes (Epstein et al., 2004b). This study suggests there is aninteraction of dopamine genotypes with food reinforcement thatinfluences energy intake. The genotypes provide a framework forhow two aspects of dopaminergic activity, the number of dopa-mine D2 receptors and the density of dopamine transporters, mayrelate to food reinforcement, energy intake, and obesity.

A behavioral genetic approach to eating may provide insightsinto individual differences that affect the substitution of alterna-tives to food. There may be genetic differences that facilitatechoice of nonfood alternatives when a person has the choice to eator engage in alternative behaviors. Likewise, there may be indi-vidual differences in how motivating are different types of rein-forcers, ranging from food to pharmacological reinforcers to phys-ical activity. The expression of a behavioral phenotype, such ashigh food reinforcement or response to food deprivation, mayresult from differences in gene expression based on the interactionof genes with biological, nutritional, or behavioral factors (Martin-Gronert & Ozanne, 2005). For example, in both animals (deFourmestraux et al., 2004) and humans (Pietilainen et al., 2004)with identical genes there are differences in phenotypes that resultfrom interactions between the genes and the environment. Epige-netic effects on dopamine gene expression may provide insightinto how dopamine genotypes interact with experiences that shapefood reinforcement to influence excess energy intake and obesity.

There are many ways in which genetic predispositions mayinteract with behavioral and environmental factors to influencefood reinforcement. The above discussion is just a beginning in theattempt to understand how genes may influence food reinforce-ment. Studying different genotypes and developing new ways tostudy how genetics can influence behavior may yield importantinsights into food reinforcement.

Behavioral and Pharmacological Methods to Modify FoodReinforcement and Obesity

Behavioral Methods

There are several approaches that may be relevant for improvingobesity treatment that are based on factors that influence foodreinforcement. Obesity treatment generally involves decreasingfood, which obese persons find very reinforcing (Saelens & Ep-stein, 1996), and increasing physical activity, which is not veryreinforcing for obese persons (Epstein, Smith, Vara, & Rodefer,1991). One option would be to identify nonfood alternatives thatmaintain the overall level of perceived positive reinforcement butshift the source of reinforcers from food to nonfood alternatives.This approach has proven to be very successful for drug abuse(Higgins, Alessi, & Dantona, 2002) and may be generalized toobesity. In drug abuse treatment, former addicts are reinforced forabstinence by using nondrug reinforcers. Laboratory research hasshown that nonfood alternatives can substitute for highly reinforc-ing foods (Goldfield & Epstein, 2002), but clinical and field

research has not been successful in developing programs thatenhance eating change or weight loss by using nonfood alterna-tives (Epstein, Roemmich, Stein, Paluch, & Kilanowski, 2005).

Variety increases energy intake and the reinforcing value offood. It may be advisable for people who are attempting to loseweight to reduce the variety of high energy dense, low nutrientdense foods they store in their homes and eat (H. A. Raynor,Jeffery, Phelan, Hill, & Wing, 2005; H. A. Raynor et al., 2004).The ensuing reduction in the reinforcing value of food could resultin a decrease in the motivation to eat and lower energy intake.Food craving and stimulus control of eating become conditioned tocues associated with eating (Jansen et al., 2003; Weingarten,1983). People who are reducing energy intake should limit theplaces in which they eat, to narrow stimulus control to the diningroom for example, and should not engage in alternative behaviorswhen eating, to reduce the possibility that other behaviors set theoccasion for eating.

Pharmacological Treatments to Increase Brain DopamineLevels in the Treatment of Obesity

If the desire to consume more food and high levels of foodreinforcement is based in part on low levels of brain dopamine,then drugs that increase brain dopamine may reduce food rein-forcement and reduce energy intake. One way to increase braindopamine is to inhibit dopamine reuptake. Methylphenidate, aDAT inhibitor used in the treatment of childhood and adult atten-tion-deficit/hyperactivity disorder (Wilens, Biederman, Spencer,& Prince, 1995), increases brain synaptic dopamine (Volkow et al.,2001). A commonly reported side effect of methylphenidate isanorexia (Spencer et al., 1995). Chronic administration of meth-ylphenidate reduces the responsiveness to drugs of abuse (Carle-zon, Mague, & Andersen, 2003) and natural rewards such assucrose (Comings & Blum, 2000). Weight loss has been reportedin children given methylphenidate with largest reductions for theheaviest children (Schertz, Adesman, Alfieri, & Bienkowski,1996). In adults, methylphenidate significantly reduced energyintake over one meal by 33% in obese male subjects comparedwith placebo (Leddy et al., 2004). In combination with diet andexercise counseling, buproprion, another DAT inhibitor, was ef-fective in reducing weight in obese men and women over 26 weeks(Jain et al., 2002) to 48 weeks (Anderson et al., 2002) whencompared with placebo.

Another way to enhance central dopamine function is to use adopamine receptor agonist. In animal experiments, bromocriptine,a dopamine D2 receptor agonist, reduced hyperinsulinemia andbody fat stores (Cincotta, MacEachern, & Meier, 1993), purport-edly by resetting hypothalamic circadian rhythms under dopami-nergic control (that are altered in obesity; Cincotta et al., 1993),and other dopamine agonists such as mesulergine will reduce foodintake in rats (Giannakopoulos, Galanopoulou, Daifotis, & Cou-varis, 1998). In obese humans, bromocriptine improved fasting andpostprandial glucose, free fatty acid, and triglyceride levels (Ka-math et al., 1997); reduced body fat stores (Meier, Cincotta, &Lovell, 1992); and improved glycemic control and glucose toler-ance in type II diabetics (Meier et al., 1992). In overweight andobese patients with prolactinomas, increasing dopaminergic tonewith bromocriptine significantly reduced serum leptin levels and


BMI over 6–24 months, without a diet or exercise prescription(Doknic et al., 2002).

Longer term studies will have to evaluate whether more sus-tained changes in brain dopamine activity are effective and safebecause in animal studies the chronic administration of amphet-amines (which raise brain dopamine levels) induces a brief periodof anorexia followed by hyperphagia and chronic obesity, accom-panied by long-term depletion of dopamine in the striatum and ofnorepinephrine in the hypothalamus (Hernandez, Parada, &Hoebel, 1983). An additional consideration in using pharmacolog-ical agents that modify brain dopamine levels is the abuse potentialfor some of these drugs. The reinforcing and subjective effects ofmethylphenidate are similar to amphetamine in non-drug-abusinghumans (Rush, Essman, Simpson, & Baker, 2001). In addition,there is evidence that some adolescents use methylphenidate forthe psychoactive effects (Musser et al., 1998). Other dopamineagonists such as bromocriptine, a specific D2 receptor agonist,may have less abuse potential, but also may not be as effective inreducing eating (Inoue et al., 1997). If there are common pathwaysthat influence eating and other behaviors (Kelley & Berridge,2002), an alternative to the use of pharmacological strategies is toidentify nonpharmacological agents that increase brain dopaminelevels, such as exercise (Ingram, 2000) or other natural rewards(Kelley & Berridge, 2002).

Combining Behavioral Interventions PlusPharmacological Manipulations of Dopamine

There are innovative studies that have combined behavioral andpharmacological interventions (Wadden, Berkowitz, Sarwer, Prus-Wisniewski, & Steinberg, 2001; Wadden et al., 2005), but theseapproaches have not focused the behavioral and pharmacologicalinterventions to alter the reinforcing value of food or the motiva-tion to eat. Novel treatments for obesity could target increasing themotivation for low energy dense, high nutrient dense foods or foralternative reinforcers to energy dense foods by pairing thesestimuli with dopamine agonists to increase their motivationalsalience. Such an approach might motivate people to sustain be-havior change to maintain reduced weight over the long term.Similarly, behavioral interventions could be designed to focus onenhancing the changes in motivation to eat specific foods thatcould result from pharmacological interventions. For example, itmay be easier to build a repertoire of alternatives to food if thereinforcing value of food is reduced by pharmacological interven-tion.

Fulton, Woodside, and Shizgal (2000) provided experimentalevidence that it may be possible simultaneously to reduce thereinforcing value of food and increase the reinforcing value ofnonfood alternatives. By using an animal model, Fulton and col-leagues showed that sensitivity to electrical stimulation in thelateral hypothalamus was reduced by leptin, as indicated by areduction in the rate of responding for lateral hypothalamic stim-ulation, whereas leptin increased sensitivity for responding toobtain electrical stimulation in adjacent, and potentially non-food-reinforcement-related, areas. As noted by Fulton, the ideal treat-ment would be one that reduced the motivation to obtain foodwhile increasing the motivation for non-food-related behaviors.

The effectiveness of behaviorally or pharmacologically alteringbrain dopamine activity may depend on individual dopaminergic

tone during specific contexts (Volkow, Wang, et al., 2002). Meth-ylphenidate’s effects are context dependent: It amplifies stimuli-induced dopamine increases when administered with a salientstimulus (for example, visual display of food in food-deprivedsubjects) when compared with a neutral stimulus (Volkow, Wanget al., 2002) and enhances the motivational saliency of a previouslyunmotivating or unexciting task (Volkow et al., 2004). A veryexciting opportunity would be to associate behaviors with dopa-mine increases with the goal of increasing the motivation toengage in these behaviors.

Energy Balance and Activity Reinforcement

Obesity is a result of a positive energy balance, with energyintake exceeding energy expenditure. A complete analysis of howreinforcement processes may relate to obesity requires consideringthe role of energy expenditure and physical activity in the energybalance equation. There is a considerable body of research onreinforcing value of physical activity (reviewed in Epstein &Saelens, 2000). For example, given the choice of sedentary oractive alternatives, obese youths find physical activity less rein-forcing than do leaner youths (Epstein et al., 1991). For youths,increasing the behavioral cost of gaining access to sedentarybehaviors results in a reduction in sedentary behaviors and aswitch to being more active (Epstein, Saelens, Myers, & Vito,1997; Epstein, Saelens, & O’Brien, 1995; Epstein et al., 1991) andfor adults (D. A. Raynor, Coleman, & Epstein, 1998; Vara &Epstein, 1993). For example, if youths have a choice betweenequally accessible sedentary or active alternatives, and being sed-entary is more reinforcing, they will choose to be sedentary.However if access to specific sedentary behaviors is reduced, thenthe youth has to choose how to reallocate the increased availabilityof leisure time.

Substitution of physical activity for sedentary behaviors is ob-served when reduction in access to one sedentary behavior resultsin increases in physically active behaviors. For example, moving acounsel video game system to an unheated, unfinished basementmay reduce video game playing while increasing physical activityin the sunlight. In this case, physical activity is a substitute forplaying video games. Similarly, inclement weather may reducerunning or bicycling outside yet increase treadmill running orusing a cycle ergometer or bicycle trainer indoors. Being activecan substitute for sedentary behaviors in the laboratory (Epstein etal., 1995, 1991; D. A. Raynor et al., 1998), and even very smalldifferences in accessibility of sedentary behaviors can havemarked effects on physical activity (D. A. Raynor et al., 1998). Notall youths substitute being active for being sedentary (Epstein,Roemmich, Paluch, & Raynor, 2005b), so understanding themechanisms for substitution has great theoretical importance fordetermining individual differences in substitution of active behav-iors for sedentary behaviors. Substitution of physical activity forsedentary behaviors depends in part on individual differences suchas child gender and child weight status (Epstein, Roemmich,Paluch, & Raynor, 2005b). In addition, the macro environmentinfluences the substitution of physically active behaviors for sed-entary behaviors (Epstein, Roemmich, Paluch, & Raynor, 2005b).

Dopamine is also involved in physical activity (Ingram, 2000).Dopamine agonists increase (Scislowski et al., 1999), whereasdopamine antagonists reduce (Fujiwara, 1992) physical activity.


Disorders that involve destruction of dopamine neurons are asso-ciated with substantial reductions in physical activity in humans(Comings et al., 1991), whereas exercise can increase dopaminelevels (de Castro & Duncan, 1985). Dopamine genotypes may berelated to differences in activity levels (Perusse, Tremblay, Le-blanc, & Bouchard, 1989). The absence of DRD2 produces pro-found bradykinesia and hypothermia in mice (Baik et al., 1995). Inhumans, a mutation of the DRD2 gene that impairs D2 receptorfunction is associated with lower total and 24-hour resting energyexpenditure and greater BMI in Pima Indians (Tataranni et al.,2001). Polymorphisms of the DRD2 gene have been associatedwith reported yearly physical activity levels in white female sub-jects (Simonen et al., 2003). Given that patterns of dopaminerelease can change in association with familiarity or repeatedexposure (e.g., training) as well as with individual experience(Schultz, 2002; Wise, 2002), increasing physical activity in theobese may improve central dopamine function and serve to rein-force liking of regular physical activity (Meeusen et al., 1997). Aswith the study of eating, using dopamine genotypes as markers fordopaminergic activity may provide insight into how dopamine isrelated to physical activity.

Alternative Theoretical Approaches to FoodReinforcement

There are a wide variety of alternative explanations beyond foodreinforcement as to why people eat. Taste, culture, class, andexperience can influence eating, but a thorough review of thesefactors is not central to understanding the role of food reinforce-ment and is beyond the scope of this article. There are twoalternative theoretical models, however, that are also designed tounderstand how food motivates food-seeking behavior and eatingand should be considered in a discussion of food reinforcement.These models will be briefly described.

Incentive Salience Models of Ingestive Behavior

Incentive salience models of motivated behavior such as eatingor drug use separate the process of reward into two components:wanting and liking (Berridge & Robinson, 1998). Liking refers tothe psychological processes that produce pleasure, hedonics, orliking of food. Liking thus refers to subjective characteristicsassociated with food. Wanting refers to the psychological processthat instigates goal-directed behavior, attraction to an incentivestimulus, and consumption of the goal object. Wanting is charac-terized by the process often referred to as craving, or the motiva-tion to obtain food even when it is not available (Berridge, 1996).The separation of wanting and liking is important to distinguishfactors that motivate people to eat.

One important aspect of incentive salience theory is developinga better understanding of the neurobiological substrates for likingand wanting. These processes can be distinguished behaviorallyand may be influenced by different neurobiological systems, withthe opiate system being more important for liking and the dopa-minergic system more important for wanting (Robinson & Ber-ridge, 2000). Conditioned, or cue-triggered, wanting is measuredby using conditioned incentive paradigms (Wyvell & Berridge,2000). In an interesting discussion of ways to conceptualize re-ward, Berridge and Robinson (2003) differentiated incentive sa-

lience models from instrumental conditioning models. They ar-gued that reward includes learning, affect, or emotion, as well asmotivation components. In their model, instrumental conditioningis an example of learning, whereas incentive salience is an exam-ple of motivation. They recognize that dopamine can changeinstrumental performance but does not influence liking (Berridge& Robinson, 2003). They also indicated that incentive saliencedoes not mediate the influence of the reinforcement contingency inreinforcement. Berridge (2004) has followed that article with aperspective on motivation concepts in neuroscience that placesincentive salience within the history of motivation. Salamone andCorrea (2002) have argued that wanting can be divided into the“appetite to consume” and the “working to obtain,” which servesas the basis for his contention that dopamine depletions, and byanalogy dopaminergic activity, primarily influence instrumentalperformance to obtain reinforcers. As noted previously, Salamonehas used choice experiments to elucidate the influence of dopa-mine depletions on instrumental performance, which may lead toa greater appreciation of behavioral choice approaches for study-ing motivation and reinforcement (Salamone & Correa, 2002).Given the contemporary importance of incentive salience theory,research may sharpen the ways in which wanting and reinforcingvalue are related, extending the theorizing of Salamone (Salamone& Correa, 2002).

The distinction between wanting and liking also may be relevantfor differences between liking and the reinforcing value of food.The motivation to obtain food may be a more powerful determi-nant of eating than is liking or hedonic value of food. Not surpris-ingly, the reinforcing value and liking value of food may beindependent processes, and increasing the reinforcing value offood in humans via food deprivation may not alter the liking ofselected foods (Epstein, Truesdale, Wojcik, Paluch, & Raynor,2003). In addition, the reinforcing value of food is a better pre-dictor of food intake than are food hedonics in adult smokers(Epstein et al., 2004a).

Habituation Theory and Food Reinforcement

Habituation refers to a decrease in responding after repeatedpresentations of a stimulus. The application of habituation theoryto ingestive behavior has been consistently supported by researchshowing that a variety of responses related to ingestive behaviors,and eating itself, are reduced with repeated presentation of visual,olfactory, or gustatory cues (Epstein, Rodefer, Wisniewski, &Caggiula, 1992; Epstein, Saad, Giacomelli, & Roemmich, 2005;Epstein, Saad, et al., 2003; Wisniewski, Epstein, & Caggiula,1992; Wisniewski, Epstein, Marcus, & Kaye, 1997). Paradigmsthat test habituation assess whether the reductions are stimulusspecific, whether response strength can be reinstated by presentinga new stimulus, and whether the reductions can be prevented byaltering stimuli during presentations of repeated food cues (Epsteinet al., 1992; Epstein, Saad, et al., 2005; Epstein, Saad, et al., 2003;Wisniewski et al., 1992, 1997). Increased intake with food varietymay be due in part to dishabituation or the failure to habituatewhen a variety of foods is presented (H. A. Raynor & Epstein,2001).

Theoretical advances by McSweeney and colleagues (Mc-Sweeney et al., 1996; McSweeney & Roll, 1998; McSweeney &Swindell, 1999) have shown that instrumental behavior can also


habituate. For example, responses engaged in to obtain food usu-ally are high in the beginning of an experimental session butdecrease within the session. Providing a new food reinstates re-sponding, consistent with a habituation model of instrumentalbehavior (Myers & Epstein, 2002) and inconsistent with energydepletion or homeostatic models of eating. The usual mechanismthought to inhibit further eating is satiation, which McSweeneyargued cannot provide a complete explanation of why people stopeating (McSweeney & Roll, 1998).

Research Gaps and Ideas for Research

There are many gaps in research on food reinforcement. Ideasfor research have been presented throughout this review, butseveral areas for future research on food reinforcement are high-lighted.

Naturalistic Analogues to Laboratory Measures of FoodReinforcement

Laboratory methods for assessing food reinforcement provide aflexible, objective way to assess food reinforcement. The constructof reinforcement may not be limited to laboratory measures butmay be readily generalized to a variety of naturalistic measures.There are natural differences in access to foods based on foodpurchases and storage, convenience, and amount of preparationrequired. The reinforcing value of a processed, portion-controlled,and prepared high energy dense snack food may be more reinforc-ing than that of a package of raw vegetables that requires washing,cutting up, and creating servings (consider a bag of potato chipsversus a bag of whole carrots), but the situation may be verydifferent if you compare a bag of potato chips versus a portion-controlled package of baby carrots served with a nonfat ranchdressing. Similarly, a full-fat hamburger may be preferred whencompared with a turkey burger, but what about the comparison ofa cooked turkey burger on a bun with appropriate condimentsversus a frozen hamburger that has to be cooked?

Reinforcer efficacy in the natural environment can be quantifiedby using token economies, or systems in which points or tokens areearned and then exchanged for reinforcers. It would be possible tovary the amount of work required to earn tokens, and strongerreinforcers should support more work. The amount of work can beconceptualized in different ways. For example, if the basic unit ofwork was conceptualized as an office task, such as folding letters,or piece work, such as sewing a garment, then work could beconceptualized as the number of units of work completed. It is alsopossible to conceptualize the amount of work in relationship to thedifficulty of the task, with the hypothesis that more reinforcingalternatives would support effort for more challenging tasks. Workcan also be defined in terms of the amount of physical workcompleted, such that more reinforcing alternatives would supportmore work as defined by watts on a calibrated bicycle ergometer.

An analogue for response effort may be price, and behavioralcost to obtain a reinforcer may influence response rates similarly tohow price influences purchases of a commodity. For example, asprice increases, purchases of a product decrease, similar to therelationship between increases in behavioral cost and decreases inresponse rate and consumption. Likewise, increasing purchases ofa product when the price of that product stays the same but the

price of an alternative product increases is evidence of substitution(or cross–price elasticity), similar to the increases in respondingfor access to a behavior when the behavioral cost of that behavioris stable but the behavioral cost of the alternative is increased(Goldfield & Epstein, 2002). The theoretical analysis of the rela-tionship between behavioral cost and response rate in terms ofdemand functions presented in the earlier section on reinforcingvalue is based on economic principles. Bickel and Madden (1999)suggest choice of responding when presented with concurrentalternatives provides similar information to demand curves ob-served when single alternatives are studied. A consistent body ofresearch has shown that manipulating price is a reliable method formodifying food purchases (French et al., 2001; French, Jeffery,Story, Hannan, & Snyder, 1997; French, Story, et al., 1997; Hor-gen & Brownell, 2002). Changes in price could be effective inreducing energy consumption even if obese people found foodmore reinforcing and were more motivated to purchase unhealthyfoods. It would be important to complement changes in prices ofless healthy alternatives with incentives to purchase healthierfoods in combination with educational and media programs edu-cating consumers about what types of foods to purchase.

Although there are similarities between price and behavioralcost, to our knowledge there have been no direct comparisons ofhow changes in behavioral cost might be related to changes inprice. Research should compare if changes in behavioral cost andchanges in price are associated with similar changes in behavior byusing laboratory analogues of price changes (Epstein, Dearing,Handley, Roemmich, & Paluch, 2006; Epstein, Handley, et al.,2006) as well as naturalistic experiments that manipulate price andobserve purchases and consumption (French et al., 2001; French,Jeffery, et al., 1997; French, Story, et al., 1997; Horgen &Brownell, 2002). Because changes in price, either by reducing therelative price of healthier alternatives or increasing the price of lesshealthy alternatives, could provide policy implications for improv-ing eating and health, these studies could represent an importantbridge between basic laboratory research and public policy. Bickeland colleagues have discussed the use of these behavioral para-digms to examine public policy implications for drug abuse(Bickel & DeGrandpre, 1995, 1996).

Behavioral Choice and Eating

One common goal in eating and obesity research is to shiftpreference from less healthy, highly reinforcing food to healthier,but perhaps less reinforcing, food. Many people attempt to shiftconsumption from less healthy to healthier foods to lose weight,reduce cholesterol, and so on. One of the biggest challenges forthese behavior changes is that there is substantial relapse in thesebehavior changes (Jeffery et al., 2000), which suggests that regularconsumption of these healthier foods does not result in these foodsbecoming very reinforcing in the context of less healthy alterna-tives. Research is needed to study how to increase the reinforcingvalue of these healthier alternatives. For example, it would be idealif procedures could be developed to increase the motivation ofpeople to consume fruits and vegetables. If an increase in thereinforcing value of these healthier alternatives cannot beachieved, research is needed to evaluate how much constraint onaccess to the less healthy, more-preferred alternatives is requiredbefore people shift choice on a long-term basis to healthier foods.


An attractive alternative to healthier foods as an option forbehavior change is to identify or develop behavioral, nonfoodalternatives or substitutes to food reinforcement. The developmentof nonfood alternatives may directly apply to situations in whichexcess energy intake is consumed in snacks and a behavioral goalis snack reduction. An alternative goal may be to reduce the timespent in meal consumption by providing attractive alternatives thatreduce the motivation to consume excess energy. Many peoplehave the experiences of being engaged in interesting or motivatedactivities that cause missing meals or of only allocating sufficienttime to the meal to consume needed energy but not allocatingenough time to eating to result in excess energy consumption.Although initial attempts to develop alternatives to eating have notresulted in reductions in energy intake or body weight (Epstein,Roemmich, Stein, Paluch, & Kilanowski, 2005), more research isneeded to assess other types of alternatives, or new ways to framethe choice situations.

There may be individual differences in who substitutes healthierfoods for less healthy foods or who substitutes alternative activitiesfor less healthy foods. These could be behavioral differences basedon experience or the history of eating or could be genetic differ-ences that are related to the reinforcing value of food. Identifyingindividual differences in food reinforcement could help identifywho may benefit most from specific interventions that focus onsubstitution of healthier foods, such as fruits and vegetables, forless healthy alternatives (Goldfield & Epstein, 2002).

Parameters of Food Deprivation/Satiation andReinforcing Value

One of the most powerful ways to increase the reinforcing valueof food is food deprivation. This is important because manyapproaches to changing eating behavior require or are associatedwith reductions in energy intake. Research is needed to identify therelationship between food deprivation and changes in food rein-forcement.

One hypothesis is that food reinforcement increases when en-ergy depletion occurs, and there is a dose–response relationshipbetween the degree of deprivation and the increase in reinforcingvalue. Parametric work is needed to assess how food deprivation isrelated to changes in reinforcing value and if there a window inwhich energy can be restricted without increasing reinforcingvalue. For example, is an energy restriction of 50 calories, 100calories, or 250 calories enough to increase the motivation to eat?Although a reduction of 250 calories per day may represent smallbehavior changes, this could lead to a weight loss of one-halfpound per week, or 26 pounds for the year, which would representsubstantial weight change. Research does suggest that more mod-erate changes that are maintained can result in better long-termweight loss than can traditional dietary approaches (Sbrocco,Nedegaard, Stone, & Lewis, 1999).

It is also possible that it is not the energy deficit that causes theincrease in reinforcing value but rather the deprivation of specificfoods. It would be interesting to assess changes in reinforcingvalue over time in conditions that reduced energy intake in part byremoving unhealthy often-preferred alternatives versus a conditionthat reduced energy intake but ensured that the person had accessto usual preferred foods, though the portion sizes of these foodswould be controlled.

Satiation is the opposite of deprivation, and the reinforcingvalue decreases as food is consumed until reinforcer satiation isreached. It would be interesting to assess the dose–response effectfor satiation to determine whether regular consumption of smallportions of a favorite food could result in a reduction in reinforcingvalue of that food, which may make it easier to reduce energyintake without increasing food deprivation. It may be possible toconsume small portions of a favorite food within an energy-restricted diet if consuming that favorite food reduces the motiva-tion to overeat and be in positive energy balance.

Determinants of Food Reinforcement

Research is needed to identify the determinants of food rein-forcement. This includes the characteristics of foods as well asdifferences in experience and learning history. Is reinforcing valuerelated to macronutrients of food, or to sensory characteristics offood, such as taste or smell? There is a very large research base onhow these factors relate to eating energy/macronutrient intake butlimited research that generalizes this to the mechanism of foodreinforcement. It would be important to know whether there areindividual differences in food reinforcement, as some people mayjust be motivated to eat, whereas others are motivated to eat onlyspecific foods.

Research is needed on experience or learning history factors thatcould influence food reinforcement. Food may become more re-inforcing on the basis of pairing eating and specific foods withvery pleasurable activities. For example, it is common to eat insocial situations (de Castro, 1994), which may facilitate eating, andthe pairing of eating with a pleasurable situation may increase thereinforcing value of food in that situation. The use of food as areward is common within families (Baughcum, Burklow, Deeks,Powers, & Whitaker, 1998; Sherry et al., 2004), and there may besituations in which parents or other adults use food as rewards ina positive context (Birch, Zimmerman, & Hind, 1980) that canincrease the reinforcing value of food. If reinforcing value iscentral to the motivation to eat and is a factor in obesity, thendeveloping a better understanding of these factors may be impor-tant for the prevention or treatment of obesity.

Summary

This article was designed to provide an overview of how foodreinforcement and behavioral theories of choice can be used tounderstand eating and energy intake. One of the unique aspects ofbehavioral theories of choice is the way in which they can be usedto understand factors that influence eating at multiple levels. Theselevels range from molecular genetics to the neurobiology of foodreinforcement through how to treat and prevent obesity.

An important distinction between food reinforcement and moretraditional approaches to eating is the potential for understandingnonregulatory ingestive behavior as opposed to eating that ismaintained by nutrient or energy depletion or homeostatic need. Itis likely that both of these areas of research are important inunderstanding factors that regulate intake. A focus on factors thatinfluence eating outside of homeostatic need provides the oppor-tunity to examine how nonfood reinforcers or behaviors influenceeating as well as how factors such as accessibility and behavioralcost can shift consumption.


In a field that is potentially so broad, there is the need forresearch at each level of analysis, including research that focuseson behavioral conditions that influence food reinforcement as wellas the biological basis for food reinforcement. There is much to belearned about how food reinforcement develops, what maintainsfood reinforcement, and how behavioral or pharmacological inter-ventions modify food reinforcement. Perhaps the most informativeresearch will be research that extends to multiple levels, that is,that combines the best behavioral and neurobiological approachesto better understand variables that influence the reinforcing valueof food and the motivation to eat. Likewise, developing a betterunderstanding of behavioral factors that influence food reinforce-ment and food intake may be important to improving the effec-tiveness of public health efforts to reduce the number of peoplewith obesity and comorbid medical conditions associated withobesity.

References

Ackerman, S. H., Albert, M., Shindledecker, R. D., Gayle, C., & Smith,G. P. (1992). Intake of different concentrations of sucrose and corn oilin preweanling rats. American Journal of Physiology, 262, R624–627.

Ackroff, K., Vigorito, M., & Sclafani, A. (1990). Fat appetite in rats: Theresponse of infant and adult rats to nutritive and non-nutritive oilemulsions. Appetite, 15, 171–188.

Ahima, R. S., & Osei, S. Y. (2001). Molecular regulation of eatingbehavior: New insights and prospects for therapeutic strategies. Trendsin Molecular Medicine, 7, 205–213.

Anderson, J. W., Greenway, F. L., Fujioka, K., Gadde, K. M., McKenney,J., & O’Neil, P. M. (2002). Bupropion SR enhances weight loss: A48-week double-blind, placebo-controlled trial. Obesity Research, 10,633–641.

Ator, N. A., & Griffiths, R. R. (2003). Principles of drug abuse liabilityassessment in laboratory animals. Drug and Alcohol Dependence, 70,S55–72.

Baik, J. H., Picetti, R., Saiardi, A., Thiriet, G., Dierich, A., Depaulis, A., etal. (1995, October 5). Parkinsonian-like locomotor impairment in micelacking dopamine D2 receptors. Nature, 377, 424–428.

Balster, R. L., & Bigelow, G. E. (2003). Guidelines and methodologicalreviews concerning drug abuse liability assessment. Drug and AlcoholDependence, 70, S13–40.

Bannon, M. J., Poosch, M. S., Xia, Y., Goebel, D. J., Cassin, B., &Kapatos, G. (1992). Dopamine transporter mRNA content in humansubstantia nigra decreases precipitously with age. Proceedings of theNational Academy of Sciences of the United States of America, 89,7095–7099.

Baptista, T. (1999). Body weight gain induced by antipsychotic drugs:Mechanisms and management. Acta Psychiatrica Scandinavica, 100,3–16.

Bardo, M. T., & Bevins, R. A. (2000). Conditioned place preference: Whatdoes it add to our preclinical understanding of drug reward? Psychop-harmacology (Berlin), 153, 31–43.

Barefoot, J. C., Heitmann, B. L., Helms, M. J., Williams, R. B., Surwit,R. S., & Siegler, I. C. (1998). Symptoms of depression and changes inbody weight from adolescence to mid-life. International Journal ofObesity and Related Metabolic Disorders, 22, 688–694.

Baskin, D. G., Figlewicz Lattemann, D., Seeley, R. J., Woods, S. C., Porte,D., Jr., & Schwartz, M. W. (1999). Insulin and leptin: Dual adipositysignals to the brain for the regulation of food intake and body weight.Brain Research, 848, 114–123.

Bassareo, V., De Luca, M. A., Aresu, M., Aste, A., Ariu, T., & Di Chiara,G. (2003). Differential adaptive properties of accumbens shell dopamine

responses to ethanol as a drug and as a motivational stimulus. EuropeanJournal of Neuroscience, 17, 1465–1472.

Bassareo, V., & Di Chiara, G. (1999). Differential responsiveness ofdopamine transmission to food-stimuli in nucleus accumbens shell/corecompartments. Neuroscience, 89, 637–641.

Baughcum, A. E., Burklow, K. A., Deeks, C. M., Powers, S. W., &Whitaker, R. C. (1998). Maternal feeding practices and childhood obe-sity: A focus group study of low-income mothers. Archives of Pediatricsand Adolescent Medicine, 152, 1010–1014.

Baumann, M. H., Ayestas, M. A., Dersch, C. M., Brockington, A., Rice,K. C., & Rothman, R. B. (2000). Effects of phentermine and fenflura-mine on extracellular dopamine and serotonin in rat nucleus accumbens:Therapeutic implications. Synapse, 36, 102–113.

Bellisle, F., Lucas, F., Amrani, R., & LeMagnen, J. (1984). Deprivation,palatability and the micro-structure of meals in human subjects. Appe-tite, 5, 85–94.

Bello, N. T., Sweigart, K. L., Lakoski, J. M., Norgren, R., & Hajnal, A.(2003). Restricted feeding with scheduled sucrose access results in anupregulation of the rat dopamine transporter. American Journal of Phys-iology: Regulatory, Integrative and Comparative Physiology, 284,R1260–1268.

Berlin, I., & Lavergne, F. (2003). Relationship between body-mass indexand depressive symptoms in patients with major depression. EuropeanPsychiatry, 18, 85–88.

Bernstein, D. J., & Ebbesen, E. B. (1978). Reinforcement and substitutionin humans: A multiple-response analysis. Journal of the ExperimentalAnalysis of Behavior, 30, 243–253.

Berridge, K. C. (1996). Food reward: Brain substrates of wanting andliking. Neuroscience and Biobehavioral Reviews, 20, 1–25.

Berridge, K. C. (2004). Motivation concepts in behavioral neuroscience.Physiology & Behavior, 81, 179–209.

Berridge, K. C., & Robinson, T. E. (1998). What is the role of dopaminein reward: Hedonic impact, reward learning, or incentive salience? BrainResearch/Brain Research Reviews, 28, 309–369.

Berridge, K. C., & Robinson, T. E. (2003). Parsing reward. Trends inNeurosciences, 26, 507–513.

Berthoud, H. R. (2004). Mind versus metabolism in the control of foodintake and energy balance. Physiology & Behavior, 81, 781–793.

Bertino, M., Beauchamp, G. K., & Engelman, K. (1991). Naltrexone, anopioid blocker, alters taste perception and nutrient intake in humans.American Journal of Physiology, 261, R59–63.

Bickel, W. K., & DeGrandpre, R. J. (1995). Modeling drug abuse policy inthe behavioral economics laboratory. In L. Green & J. H. Kagel (Eds.),Advances in behavioral economics (Vol. 3, pp. 69–95). New York:Ablex.

Bickel, W. K., & DeGrandpre, R. J. (1996). Basic psychological sciencespeaks to drug policy: Drug cost and competing reinforcement. InW. B. R. DeGrandpre (Ed.), Psychological perspectives on the control,prevention and treatment of illicit drug use (pp. 31–52). New York:Plenum.

Bickel, W. K., & Madden, G. J. (1999). A comparison of measures ofrelative reinforcing efficacy and behavioral economics: Cigarettes andmoney in smokers. Behavioral Pharmacology, 10, 627–637.

Bickel, W. K., Madden, G. J., & Petry, N. M. (1998). The price of change:The behavioral economics of drug dependence. Behavior Therapy, 29,545–565.

Bickel, W. K., Marsch, L. A., & Carroll, M. E. (2000). Deconstructingrelative reinforcing efficacy and situating the measures of pharmacolog-ical reinforcement with behavioral economics: A theoretical proposal.Psychopharmacology, 153, 44–56.

Bina, K. G., & Cincotta, A. H. (2000). Dopaminergic agonists normalizeelevated hypothalamic neuropeptide Y and corticotropin-releasing hor-mone, body weight gain, and hyperglycemia in ob/ob mice. Neuroen-docrinology, 71, 68–78.


Birch, L. L., Zimmerman, S. I., & Hind, H. (1980). The influence of socialaffective context on the formation of children’s food preferences. ChildDevelopment, 51, 856–861.

Blass, E. M. (1990). Suckling: Determinants, changes, mechanisms, andlasting impressions. Developmental Psychology, 26, 520–533.

Blass, E. M., & Watt, L. B. (1999). Suckling- and sucrose-induced anal-gesia in human newborns. Pain, 83, 611–623.

Blum, K., Braverman, E. R., Holder, J. M., Lubar, J. F., Monastra, V. J.,Miller, D., et al. (2000). Reward deficiency syndrome: A biogeneticmodel for the diagnosis and treatment of impulsive, addictive, andcompulsive behaviors. Journal of Psychoactive Drugs, 32(Suppl. i–iv),1–112.

Blum, K., Braverman, E. R., Wood, R. C., Gill, J., Li, C., Chen, T. J., etal. (1996). Increased prevalence of the Taq I A1 allelle of the dopaminereceptor gene (DRD2) in obesity with comorbid substance use disorder:A preliminary report. Pharmacogenetics, 6, 297–305.

Bouton, M. E. (2007). Learning and behavior: A contemporary synthesis.Sunderland, MA: Sinauer Associates, Inc.

Carlezon, W. A., Jr., Mague, S. D., & Andersen, S. L. (2003). Enduringbehavioral effects of early exposure to methylphenidate in rats. Biolog-ical Psychiatry, 54, 1330–1337.

Carr, K. D. (1996). Feeding, drug abuse, and the sensitization of reward bymetabolic need. Neurochemical Research, 21, 1455–1467.

Carr, K. D. (2002). Augmentation of drug reward by chronic food restric-tion: Behavioral evidence and underlying mechanisms. Physiology &Behavior, 76, 353–364.

Carroll, M. E., & Meisch, R. A. (1984). Increased drug-reinforced behaviordue to food deprivation. Advances in Behavioral Pharmacology, 4,47–88.

Chung, J., Kim, Y., Yim, S., Park, S., Ko, K., & Jung, J. (1994). Immu-nohistochemical and biochemical studies on dopamine content in ratbrain during cholecystokinin-induced suppression of feeding. KoreanJournal of Pharmacology, 30, 39–48.

Cincotta, A. H., MacEachern, T. A., & Meier, A. H. (1993). Bromocriptineredirects metabolism and prevents seasonal onset of obese hyperinsu-linemic state in Syrian hamsters. American Journal of Physiology, 264,E285–293.

Comer, S. D., Lac, S. T., Wyvell, C. L., & Carroll, M. E. (1996). Combinedeffects of buprenorphine and a nondrug alternative reinforcer on I.V.cocaine self-administration in rats maintained under FR schedules. Psy-chopharmacology (Berlin), 125, 355–360.

Comings, D. E., & Blum, K. (2000). Reward deficiency syndrome: Geneticaspects of behavioral disorders. Progress in Brain Research, 126, 325–341.

Comings, D. E., Comings, B. G., Muhleman, D., Dietz, G., Shahbahrami,B., Tast, D., et al. (1991). The dopamine D2 receptor locus as amodifying gene in neuropsychiatric disorders. Journal of the AmericanMedical Association, 266, 1793–1800.

Comings, D. E., Flanagan, S. D., Dietz, G., Muhleman, D., Knell, E., &Gysin, R. (1993). The dopamine D2 receptor (DRD2) as a major gene inobesity and height. Biochemical Medicine and Metabolic Biology, 50,176–185.

Comings, D. E., Gade, R., MacMurray, J. P., Muhleman, D., & Peters,W. R. (1996). Genetic variants of the human obesity (OB) gene: Asso-ciation with body mass index in young women, psychiatric symptoms,and interaction with the dopamine D2 receptor (DRD2) gene. MolecularPsychiatry, 1, 325–335.

Cornell, C. E., Rodin, J., & Weingarten, H. P. (1989). Stimulus-inducedeating when satiated. Physiology & Behavior, 45, 695–704.

de Castro, J. M. (1994). Family and friends produce greater social facili-tation of food intake than other companions. Physiology & Behavior, 56,445–455.

de Castro, J. M., & Duncan, G. (1985). Operantly conditioned running:

Effects on brain catecholamine concentrations and receptor densities inthe rat. Pharmacology, Biochemistry and Behavior, 23, 495–500.

de Fourmestraux, V., Neubauer, H., Poussin, C., Farmer, P., Falquet, L.,Burcelin, R., et al. (2004). Transcript profiling suggests that differentialmetabolic adaptation of mice to a high fat diet is associated with changesin liver to muscle lipid fluxes. Journal of Biological Chemistry, 279,50743–50753.

de Graaf, C., Cardello, A. V., Matthew Kramer, F., Lesher, L. L., Meisel-man, H. L., & Schutz, H. G. (2005). A comparison between likingratings obtained under laboratory and field conditions: The role ofchoice. Appetite, 44, 15–22.

Desor, J. A., Maller, O., & Turner, R. E. (1973). Taste and acceptance ofsugars by human infants. Journal of Comparative and PhysiologicalPsychology, 84, 496–501.

de Wit, H. (1996). Priming effects with drugs and other reinforcers.Experimental and Clinical Psychopharmacology, 4, 5–10.

Doell, S. R., & Hawkins, R. C. (1982). Pleasures and pounds: An explor-atory study. Addictive Behaviors, 7, 65–69.

Doknic, M., Pekic, S., Zarkovic, M., Medic-Stojanoska, M., Dieguez, C.,Casanueva, F., et al. (2002). Dopaminergic tone and obesity: An insightfrom prolactinomas treated with bromocriptine. European Journal ofEndocrinology, 147, 77–84.

Drewnowski, A., & Greenwood, M. R. C. (1983). Cream and sugar:Human preferences for high-fat foods. Physiology & Behavior, 30,629–633.

Drewnowski, A., Krahn, D. D., Demitrack, M. A., Nairn, K., & Gosnell,B. A. (1992). Taste responses and preferences for sweet high-fat foods:Evidence for opioid involvement. Physiology & Behavior, 51, 371–379.

Drewnowski, A., Kurth, C., Holden-Wiltse, J., & Saari, J. (1992). Foodpreferences in human obesity: Carbohydrates versus fats. Appetite, 18,207–221.

Elsinga, P. H., Hatano, K., & Ishiwata, K. (2006). PET tracers for imagingof the dopaminergic system. Current Medicinal Chemistry, 13, 2139–2153.

Epstein, L. H., Dearing, K. K., Handley, E. A., Roemmich, J. N., & Paluch,R. A. (2006). Relationship of mother and child food purchases as afunction of price: A pilot study. Appetite, 47, 115–118.

Epstein, L. H., Dearing, K. K., Paluch, R. A., Roemmich, J. N., & Cho, D.(in press). Price and maternal obesity influence purchases of low andhigh energy dense foods. American Journal of Clinical Nutrition.

Epstein, L. H., Handley, E. A., Dearing, K. K., Cho, D. D., Roemmich,J. N., Paluch, R. A., et al. (2006). Purchases of food in youth: Influenceof price and income. Psychological Science, 17, 82–89.

Epstein, L. H., Rodefer, J. S., Wisniewski, L., & Caggiula, A. R. (1992).Habituation and dishabituation of human salivary response. Physiology& Behavior, 51, 945–950.

Epstein, L. H., Roemmich, J. N., Paluch, R. A., & Raynor, H. A. (2005a).The influence of changes in sedentary behavior on energy and macro-nutrient intake in youth. American Journal of Clinical Nutrition, 81,361–366.

Epstein, L. H., Roemmich, J. N., Paluch, R. A., & Raynor, H. A. (2005b).Physical activity as a substitute for sedentary behavior in youth. Annalsof Behavioral Medicine, 29, 200–209.

Epstein, L. H., Roemmich, J. N., Stein, R. I., Paluch, R. A., & Kilanowski,C. K. (2005). The challenge of identifying behavioral alternatives tofood: Clinic and field studies. Annals of Behavioral Medicine, 30,201–209.

Epstein, L. H., Saad, F. G., Giacomelli, A. M., & Roemmich, J. N. (2005).Effects of allocation of attention on habituation to olfactory and visualfood stimuli in children. Physiology & Behavior, 84, 313–319.

Epstein, L. H., Saad, F. G., Handley, E. A., Roemmich, J. N., Hawk, L. W.,& McSweeney, F. K. (2003). Habituation of salivation and motivatedresponding for food in children. Appetite, 41, 283–289.

Epstein, L. H., & Saelens, B. E. (2000). Behavioral economics of obesity:


Food intake and energy expenditure. In W. K. Bickel & R. E. Vuchinich(Eds.), Reframing health behavior change with behavioral economics(pp. 293–311). Mahwah, NJ: Erlbaum.

Epstein, L. H., Saelens, B. E., Myers, M. D., & Vito, D. (1997). Effects ofdecreasing sedentary behaviors on activity choice in obese children.Health Psychology, 16, 107–113.

Epstein, L. H., Saelens, B. E., & O’Brien, J. G. (1995). Effects of rein-forcing increases in active behavior versus decreases in sedentary be-havior for obese children. International Journal of Behavioral Medicine,2, 41–50.

Epstein, L. H., Smith, J. A., Vara, L. S., & Rodefer, J. S. (1991). Behav-ioral economic analysis of activity choice in obese children. HealthPsychology, 10, 311–316.

Epstein, L. H., Temple, J. L., Neaderhiser, B. J., Salis, R. J., Erbe, R. W.,& Leddy, J. J. (in press). Food reinforcement, the dopamine D2 receptorgenotype and energy intake in obese and non-obese humans. BehavioralNeuroscience.

Epstein, L. H., Truesdale, R., Wojcik, A., Paluch, R. A., & Raynor, H. A.(2003). Effects of deprivation on hedonics and reinforcing value of food.Physiology & Behavior, 78, 221–227.

Epstein, L. H., Wright, S. M., Paluch, R. A., Leddy, J., Hawk, L. W.,Jaroni, J. L., et al. (2004a). Food hedonics and reinforcement as deter-minants of laboratory food intake in smokers. Physiology & Behavior,81, 511–517.

Epstein, L. H., Wright, S. M., Paluch, R. A., Leddy, J. J., Hawk, L. W.,Jaroni, J. L., et al. (2004b). The relationship between food reinforcementand dopamine genotypes on food intake in smokers. American Journalof Clinical Nutrition, 80, 82–88.

Fetissov, S. O., Meguid, M. M., Sato, T., & Zhang, L. H. (2002). Expres-sion of dopaminergic receptors in the hypothalamus of lean and obeseZucker rats and food intake. American Journal of Physiology—Regulatory, Integrative and Comparative Physiology, 283, R905–910.

Field, A. E., Austin, S. B., Taylor, C. B., Malspeis, S., Rosner, B., Rockett,H. R., et al. (2003). Relation between dieting and weight change amongpreadolescents and adolescents. Pediatrics, 112, 900–906.

Figlewicz, D. P. (2003a). Adiposity signals and food reward: Expandingthe CNS roles of insulin and leptin. American Journal of Physiology:Regulatory, Integrative and Comparative Physiology, 284, R882–892.

Figlewicz, D. P. (2003b). Insulin, food intake, and reward. Seminars inClinical Neuropsychiatry, 8, 82–93.

Figlewicz, D. P., Patterson, T. A., Johnson, L. B., Zavosh, A., Israel, P. A.,& Szot, P. (1998). Dopamine transporter mRNA is increased in the CNSof Zucker fatty (fa/fa) rats. Brain Research Bulletin, 46, 199–202.

Figlewicz, D. P., Szot, P., Chavez, M., Woods, S. C., & Veith, R. C.(1994). Intraventricular insulin increases dopamine transporter mRNA inrat VTA/substantia nigra. Brain Research, 644, 331–334.

Fisher, J. O., & Birch, L. L. (2002). Eating in the absence of hunger andoverweight in girls from 5 to 7 y of age. American Journal of ClinicalNutrition, 76, 226–231.

Freed, D. E., & Green, L. (1998). A behavioral economic analysis of fatappetite in rats. Appetite, 31, 333–349.

French, S. A., Jeffery, R. W., Story, M., Breitlow, K. K., Baxter, J. S.,Hannan, P., et al. (2001). Pricing and promotion effects on low-fatvending snack purchases: The CHIPS Study. American Journal of Pub-lic Health, 91, 112–117.

French, S. A., Jeffery, R. W., Story, M., Hannan, P., & Snyder, M. P.(1997). A pricing strategy to promote low-fat snack choices throughvending machines. American Journal of Public Health, 87, 849–851.

French, S. A., Story, M., Jeffery, R. W., Snyder, P., Eisenberg, M.,Sidebottom, A., et al. (1997). Pricing strategy to promote fruit andvegetable purchase in high school cafeterias. Journal of the AmericanDietetic Association, 97, 1008–1010.

Fujiwara, H. (1992). Comparative studies of sulpiride and classical neuro-leptics on induction of catalepsy, locomotor activity, and brain dopamine

metabolism in mice. Pharmacology Biochemistry and Behavior, 41,301–308.

Fulton, S., Woodside, B., & Shizgal, P. (2000, January 7). Modulation ofbrain reward circuitry by leptin. Science, 287, 125–128.

Gautier, J. F., Chen, K., Salbe, A. D., Bandy, D., Pratley, R. E., Heiman,M., et al. (2000). Differential brain responses to satiation in obese andlean men. Diabetes, 49, 838–846.

Gendall, K. A., Joyce, P. R., Sullivan, P. F., & Bulik, C. M. (1998). Foodcravers: Characteristics of those who binge. International Journal ofEating Disorders, 23, 353–360.

Giannakopoulos, G., Galanopoulou, P., Daifotis, Z., & Couvaris, C.(1998). Effects of mesulergine treatment on diet selection, brain sero-tonin (5-HT) and dopamine (DA) turnover in free feeding rats. Progressin Neuro-Psychopharmacology & Biological Psychiatry, 22, 803–813.

Goldfield, G. S., & Epstein, L. H. (2002). Can fruits and vegetables andactivities substitute for snack foods? Health Psychology, 21, 299–303.

Goldfield, G. S., Epstein, L. H., Davidson, M., & Saad, F. G. (2005).Validation of a multiple choice questionnaire measure of the relativereinforcing value of food. Eating Behaviors, 6, 283–292.

Grace, A. A. (2000). The tonic/phasic model of dopamine system regula-tion and its implications for understanding alcohol and psychostimulantcraving. Addiction, 95(Suppl. 2), S119–128.

Greeno, C. G., & Wing, R. R. (1994). Stress-induced eating. PsychologicalBulletin, 115, 444–464.

Griffiths, R. R., Bigelow, G. E., & Ator, N. A. (2003). Principles of initialexperimental drug abuse liability assessment in humans. Drug andAlcohol Dependence, 70, S41–54.

Griffiths, R. R., Rush, C. R., & Puhala, K. A. (1996). Validation of themultiple-choice procedure for investigating drug reinforcement in hu-mans. Experimental and Clinical Psychopharmacology, 4, 97–106.

Hamdi, A., Porter, J., & Prasad, C. (1992). Decreased striatal D2 dopaminereceptors in obese Zucker rats: Changes during aging. Brain Research,589, 338–340.

Hamilton, C. L. (1964). Rats’ preference for high fat diets. Journal ofComparative and Physiological Psychology, 58, 459–460.

Harvey, J., Wing, R. R., & Mullen, M. (1993). Effects on food cravings ofa very low calorie diet or a balanced, low calorie diet. Appetite, 21,105–115.

Heinz, A., Goldman, D., Jones, D. W., Palmour, R., Hommer, D., Gorey,J. G., et al. (2000). Genotype influences in vivo dopamine transporteravailability in human striatum. Neuropsychopharmacology, 22, 133–139.

Hernandez, L., & Hoebel, B. G. (1988). Food reward and cocaine increaseextracellular dopamine in the nucleus accumbens as measured by mi-crodialysis. Life Sciences, 42, 1705–1712.

Hernandez, L., Parada, M., & Hoebel, B. G. (1983). Amphetamine-inducedhyperphagia and obesity caused by intraventricular or lateral hypotha-lamic injections in rats. Journal of Pharmacology & ExperimentalTherapeutics, 227, 524–530.

Higgins, S. T., Alessi, S. M., & Dantona, R. L. (2002). Voucher-basedincentives: A substance abuse treatment innovation. Addictive Behav-iors, 27, 887–910.

Hoebel, B. G., Hernandez, L., Schwartz, D. H., Mark, G. P., & Hunter,G. A. (1989). Microdialysis studies of brain norepinephrine, serotonin,and dopamine release during ingestive behavior: Theoretical and clinicalimplications. Annals of the New York Academy of Sciences, 575, 171–193.

Holsen, L. M., Zarcone, J. R., Thompson, T. I., Brooks, W. M., Anderson,M. F., Ahluwalia, J. S., et al. (2005). Neural mechanisms underlyingfood motivation in children and adolescents. NeuroImage, 27, 669–676.

Horgen, K. B., & Brownell, K. D. (2002). Comparison of price change andhealth message interventions in promoting healthy food choices. HealthPsychology, 21, 505–512.

Hursh, S. R., & Bauman, R. A. (1987). The behavioral analysis of demand.


In L. Green & J. H. Kagel (Eds.), Advances in behavioral economics(Vol. 1, pp. 117–165). Norwood, NJ: Ablex.

Ingram, D. K. (2000). Age-related decline in physical activity: Generali-zation to nonhumans. Medicine & Science in Sports & Exercise, 32,1623–1629.

Inoue, K., Kiriike, N., Kurioka, M., Fujisaki, Y., Iwasaki, S., & Yamagami,S. (1997). Bromocriptine enhances feeding behavior without changingdopamine metabolism. Pharmacology, Biochemistry and Behavior, 58,183–188.

Jacobs, S. B., & Wagner, M. K. (1984). Obese and nonobese individuals:Behavioral and personality characteristics. Addictive Behaviors, 9, 223–226.

Jacobsen, L. K., Staley, J. K., Zoghbi, S. S., Seibyl, J. P., Kosten, T. R.,Innis, R. B., et al. (2000). Prediction of dopamine transporter bindingavailability by genotype: A preliminary report. American Journal ofPsychiatry, 157, 1700–1703.

Jain, A. K., Kaplan, R. A., Gadde, K. M., Wadden, T. A., Allison, D. B.,Brewer, E. R., et al. (2002). Bupropion SR vs. placebo for weight lossin obese patients with depressive symptoms. Obesity Research, 10,1049–1056.

Jansen, A., Theunissen, N., Slechten, K., Nederkoorn, C., Boob, B., Mulk-ens, S., et al. (2003). Overweight children overeat after exposure to foodcues. Eating Behaviors, 4, 197–209.

Jeffery, R. W., Drewnowski, A., Epstein, L. H., Stunkard, A. J., Wilson,G. T., Wing, R. R., et al. (2000). Long-term maintenance of weight loss:Current status. Health Psychology, 19(Suppl.), 5–16.

Jenkinson, C. P., Hanson, R., Cray, K., Wiedrich, C., Knowler, W. C.,Bogardus, C., et al. (2000). Association of dopamine D2 receptor poly-morphisms Ser311Cys and TaqIA with obesity or type 2 diabetes mel-litus in Pima Indians. International Journal of Obesity and RelatedMetabolic Disorders, 24, 1233–1238.

Johnson, W. G. (1974). Effect of cue prominence and subject weight onhuman food-directed performance. Journal of Personality and SocialPsychology, 29, 843–848.

Jonsson, E. G., Nothen, M. M., Grunhage, F., Farde, L., Nakashima, Y.,Propping, P., et al. (1999). Polymorphisms in the dopamine D2 receptorgene and their relationships to striatal dopamine receptor density ofhealthy volunteers. Molecular Psychiatry, 4, 290–296.

Kamath, V., Jones, C. N., Yip, J. C., Varasteh, B. B., Cincotta, A. H.,Reaven, G. M., et al. (1997). Effects of a quick-release form of bro-mocriptine (Ergoset) on fasting and postprandial plasma glucose, insu-lin, lipid, and lipoprotein concentrations in obese nondiabetic hyperin-sulinemic women. Diabetes Care, 20, 1697–1701.

Karhunen, L. J., Lappalainen, R. I., Vanninen, E. J., Kuikka, J. T., &Uusitupa, M. I. (1997). Regional cerebral blood flow during food expo-sure in obese and normal-weight women. Brain, 120, 1675–1684.

Karoly, A. J., Winger, G., Ikomi, F., & Woods, J. H. (1978). The rein-forcing property of ethanol in the rhesus monkey: II. Some variablesrelated to the maintenance of intravenous ethanol-reinforced responding.Psychopharmacology (Berlin), 58, 19–25.

Kelley, A. E., Bakshi, V. P., Haber, S. N., Steininger, T. L., Will, M. J., &Zhang, M. (2002). Opioid modulation of taste hedonics within theventral striatum. Physiology & Behavior, 76, 365–377.

Kelley, A. E., & Berridge, K. C. (2002). The neuroscience of naturalrewards: Relevance to addictive drugs. Journal of Neuroscience, 22,3306–3311.

Kirley, A., Hawi, Z., Daly, G., McCarron, M., Mullins, C., Millar, N., et al.(2002). Dopaminergic system genes in ADHD: Toward a biologicalhypothesis. Neuropsychopharmacology, 27, 607–619.

Kiyatkin, E. A., & Gratton, A. (1994). Electrochemical monitoring ofextracellular dopamine in nucleus accumbens of rats lever-pressing forfood. Brain Research, 652, 225–234.

Koob, G. F. (1999). The role of the striatopallidal and extended amygdala

systems in drug addiction. Annals of the New York Academy of Sciences,877, 445–460.

Kuczenski, R., & Segal, D. S. (1999). Sensitization of amphetamine-induced stereotyped behaviors during the acute response. Journal ofPharmacology & Experimental Therapeutics, 288, 699–709.

Laitinen, J., Ek, E., & Sovio, U. (2002). Stress-related eating and drinkingbehavior and body mass index and predictors of this behavior. Preven-tive Medicine, 34, 29–39.

Lappalainen, R., & Epstein, L. H. (1990). A behavioral economics analysisof food choice in humans. Appetite, 14, 81–93.

Leddy, J. J., Epstein, L. H., Jaroni, J. L., Roemmich, J. N., Paluch, R. A.,Goldfield, G. S., et al. (2004). The influence of methylphenidate oneating in obese men. Obesity Research, 12, 224–232.

Lee, E. S., Kim, Y. H., Beck, S. H., Lee, S., & Oh, S. W. (2005).Depressive mood and abdominal fat distribution in overweight premeno-pausal women. Obesity Research, 13, 320–325.

Leshner, A. I. (1996). Molecular mechanisms of cocaine addiction. NewEngland Journal of Medicine, 335, 128–129.

Levin, B. E., & Dunn-Meynell, A. A. (1997). Dysregulation of arcuatenucleus preproneuropeptide Y mRNA in diet-induced obese rats. Amer-ican Journal of Physiology—Regulatory, Integrative and ComparativePhysiology, 272, R1365–1370.

Levin, B. E., & Dunn-Meynell, A. A. (2002). Defense of body weightdepends on dietary composition and palatability in rats with diet-inducedobesity. American Journal of Physiology—Regulatory, Integrative andComparative Physiology, 282, R46–54.

Levine, A. S., Kotz, C. M., & Gosnell, B. A. (2003). Sugars and fats: Theneurobiology of preference. Journal of Nutrition, 133, 831S–834S.

Lowe, M. R. (1993). The effects of dieting on eating behavior: A three-factor model. Psychological Bulletin, 114, 100–121.

Lowe, M. R., Foster, G. D., Kerzhnerman, I., Swain, R. M., & Wadden,T. A. (2001). Restrictive dieting vs. “undieting” effects on eating regu-lation in obese clinic attenders. Addictive Behaviors, 26, 253–266.

Maldonado-Irizarry, C. S., Swanson, C. J., & Kelley, A. E. (1995). Glu-tamate receptors in the nucleus accumbens shell control feeding behav-ior via the lateral hypothalamus. Journal of Neuroscience, 15, 6779–6788.

Martin, C. K., O’Neil, P. M., & Pawlow, L. (2006). Changes in foodcravings during low-calorie and very-low-calorie diets. Obesity (SilverSpring), 14, 115–121.

Martinez, D., Gelernter, J., Abi-Dargham, A., van Dyck, C. H., Kegeles,L., Innis, R. B., et al. (2001). The variable number of tandem repeatspolymorphism of the dopamine transporter gene is not associated withsignificant change in dopamine transporter phenotype in humans. Neu-ropsychopharmacology, 24, 553–560.

Martin-Gronert, M. S., & Ozanne, S. E. (2005). Programming of appetiteand type 2 diabetes. Early Human Development, 81, 981–988.

Mazur, J. E. (1995). Conditioned reinforcement and choice with delayedand uncertain primary reinforcers. Journal of the Experimental Analysisof Behavior, 63, 139–150.

McCann, K. L., Perri, M. G., Nezu, A. M., & Lowe, M. R. (1992). Aninvestigation of counterregulatory eating in obese clinic attenders. In-ternational Journal of Eating Disorders, 12, 161–169.

McCrory, M. A., Fuss, P. J., McCallum, J. E., Yao, M., Vinken, A. G.,Hays, N. P., et al. (1999). Dietary variety within food groups: Associ-ation with energy intake and body fatness in men and women. AmericanJournal of Clinical Nutrition, 69, 440–447.

McSweeney, F. K., Hinson, J. M., & Cannon, C. B. (1996). Sensitization–habituation may occur during operant conditioning. Psychological Bul-letin, 120, 256–271.

McSweeney, F. K., & Roll, J. M. (1998). Do animals satiate or habituateto repeatedly presented reinforcers? Psychonomic Bulletin & Review, 53,428–442.

McSweeney, F. K., & Swindell, S. (1999). General-process theories of


motivation revisited: The role of habituation. Psychological Bulletin,125, 437–457.

Meeusen, R., Smolders, I., Sarre, S., de Meirleir, K., Keizer, H., Serneels,M., et al. (1997). Endurance training effects on neurotransmitter releasein rat striatum: An in vivo microdialysis study. Acta PhysiologicaScandinavica, 159, 335–341.

Meier, A. H., Cincotta, A. H., & Lovell, W. C. (1992). Timed bromocrip-tine administration reduces body fat stores in obese subjects and hyper-glycemia in type II diabetics. Experientia, 48, 248–253.

Meririnne, E., Kankaanpaa, A., & Seppala, T. (2001). Rewarding proper-ties of methylphenidate: Sensitization by prior exposure to the drug andeffects of dopamine D1- and D2-receptor antagonists. Journal of Phar-macology & Experimental Therapeutics, 298, 539–550.

Mittleman, G., Castaneda, E., Robinson, T. E., & Valenstein, E. S. (1986).The propensity for nonregulatory ingestive behavior is related to differ-ences in dopamine systems: Behavioral and biochemical evidence. Be-havioral Neuroscience, 100, 213–220.

Musser, C. J., Ahmann, P. A., Theye, F. W., Mundt, P., Broste, S. K., &Mueller-Rizner, N. (1998). Stimulant use and the potential for abuse inWisconsin as reported by school administrators and longitudinally fol-lowed children. Journal of Developmental and Behavioral Pediatrics,19, 187–192.

Myers, M. D., & Epstein, L. H. (2002). Habituation of responding for foodin humans. Appetite, 38, 224–234.

Need, A. C., Ahmadi, K. R., Spector, T. D., & Goldstein, D. B. (2006).Obesity is associated with genetic variants that alter dopamine availabil-ity. Annals of Human Genetics, 70, 293–303.

Noble, E. P. (2003). D2 dopamine receptor gene in psychiatric and neu-rologic disorders and its phenotypes. American Journal of MedicalGenetics, 116B, 103–125.

Noble, E. P., Noble, R. E., Ritchie, T., Syndulko, K., Bohlman, M. C.,Noble, L. A., et al. (1994). D2 dopamine receptor gene and obesity.International Journal of Eating Disorders, 15, 205–217.

Orosco, M., Gerozissis, K., Rouch, C., Meile, M. J., & Nicolaidis, S.(1995). Hypothalamic monoamines and insulin in relation to feeding inthe genetically obese Zucker rat as revealed by microdialysis. ObesityResearch, 3, S655–665.

Orosco, M., Rouch, C., & Nicolaidis, S. (1996). Rostromedial hypotha-lamic monoamine changes in response to intravenous infusions of insu-lin and glucose in freely feeding obese Zucker rats: A microdialysisstudy. Appetite, 26, 1–20.

Parkinson, J., Fudge, J., Hurd, Y., Pennartz, C., & Peoples, L. (2000).Finding motivation at Seabrook Island: The ventral striatum, learningand plasticity. Trends in Neurosciences, 23, 383–384.

Patterson, T. A., Brot, M. D., Zavosh, A., Schenk, J. O., Szot, P., &Figlewicz, D. P. (1998). Food deprivation decreases mRNA and activityof the rat dopamine transporter. Neuroendocrinology, 68, 11–20.

Pecoraro, N., Reyes, F., Gomez, F., Bhargava, A., & Dallman, M. F.(2004). Chronic stress promotes palatable feeding, which reduces signsof stress: Feedforward and feedback effects of chronic stress. Endocri-nology, 145, 3754–3762.

Perusse, L., Tremblay, A., Leblanc, C., & Bouchard, C. (1989). Geneticand environmental influences on level of habitual physical activity andexercise participation. American Journal of Epidemiology, 129, 1012–1022.

Pietilainen, K. H., Rissanen, A., Laamanen, M., Lindholm, A. K.,Markkula, H., Yki-Jarvinen, H., et al. (2004). Growth patterns in youngadult monozygotic twin pairs discordant and concordant for obesity.Twin Research, 7, 421–429.

Piggott, M. A., Marshall, E. F., Thomas, N., Lloyd, S., Court, J. A., Jaros,E., et al. (1999). Dopaminergic activities in the human striatum: Ros-trocaudal gradients of uptake sites and of D1 and D2 but not of D3receptor binding or dopamine. Neuroscience, 90, 433–445.

Pliner, P., Polivy, J., Herman, C. P., & Zakalusny, I. (1980). Short-term

intake of overweight individuals and normal weight dieters and non-dieters with and without choice among a variety of foods. Appetite, 1,203–213.

Pohjalainen, T., Rinne, J. O., Nagren, K., Lehikoinen, P., Anttila, K.,Syvalahti, E. K., et al. (1998). The A1 allele of the human D2 dopaminereceptor gene predicts low D2 receptor availability in healthy volunteers.Molecular Psychiatry, 3, 256–260.

Pothos, E. N., Creese, I., & Hoebel, B. G. (1995). Restricted eating withweight loss selectively decreases extracellular dopamine in the nucleusaccumbens and alters dopamine response to amphetamine, morphine,and food intake. Journal of Neuroscience, 15, 6640–6650.

Pothos, E. N., Hernandez, L., & Hoebel, B. G. (1995). Chronic fooddeprivation decreases extracellular dopamine in the nucleus accumbens:Implications for a possible neurochemical link between weight loss anddrug abuse. Obesity Research, 3, 525S–529S.

Premack, D. (1959). Toward empirical behavioral laws: I. Positive rein-forcement. Psychological Review, 66, 219–233.

Rada, P., Avena, N. M., & Hoebel, B. G. (2005). Daily bingeing on sugarrepeatedly releases dopamine in the accumbens shell. Neuroscience,134, 737–744.

Raynor, D. A., Coleman, K. J., & Epstein, L. H. (1998). Effects ofproximity on the choice to be physically active or sedentary. ResearchQuarterly for Exercise and Sport, 69, 99–103.

Raynor, H. A., & Epstein, L. H. (2001). Dietary variety, energy regulationand obesity. Psychological Bulletin, 127, 325–341.

Raynor, H. A., & Epstein, L. H. (2003). The relative-reinforcing value offood under differing levels of food deprivation and restriction. Appetite,40, 15–24.

Raynor, H. A., Jeffery, R. W., Phelan, S., Hill, J. O., & Wing, R. R. (2005).Amount of food group variety consumed in the diet and long-termweight loss maintenance. Obesity Research, 13, 883–890.

Raynor, H. A., Jeffery, R. W., Tate, D. F., & Wing, R. R. (2004).Relationship between changes in food group variety, dietary intake, andweight during obesity treatment. International Journal of Obesity andRelated Metabolic Disorders, 28, 813–820.

Reiss, S., & Havercamp, S. (1996). The sensitivity theory of motivation:Implications for psychopathology. Behaviour Research and Therapy,34, 621–632.

Richardson, N. R., & Roberts, D. C. S. (1996). Progressive ratio schedulesin drug self-administration studies in rats: A method to evaluate rein-forcing efficacy. Journal of Neuroscience Methods, 66, 1–11.

Ritchie, T., & Noble, E. P. (2003). Association of seven polymorphisms ofthe D2 dopamine receptor gene with brain receptor-binding character-istics. Neurochemistry Research, 28, 73–82.

Robinson, T. E., & Berridge, K. C. (1993). The neural basis of drugcraving: An incentive-sensitization theory of addiction. Brain ResearchReviews, 18, 247–291.

Robinson, T. E., & Berridge, K. C. (2000). The psychology and neurobi-ology of addiction: An incentive-sensitization view. Addiction, 95, S91–117.

Roemmich, J. N., Wright, S. M., & Epstein, L. H. (2002). Dietary restraintand stress-induced snacking in youth. Obesity Research, 10, 1120–1126.

Rolls, B. J. (1995). Carbohydrates, fats, and satiety. American Journal ofClinical Nutrition, 61, 960S–967S.

Rolls, B. J. (2003). The supersizing of America: Portion size and theobesity epidemic. Nutrition Today, 38, 42–53.

Rolls, B. J., Castellanos, V. H., Halford, J. C., Kilara, A., Panyam, D.,Pelkman, C. L., et al. (1998). Volume of food consumed affects satietyin men. American Journal of Clinical Nutrition, 67, 1170–1177.

Rolls, B. J., Roe, L. S., & Meengs, J. S. (2006). Reductions in portion sizeand energy density of foods are additive and lead to sustained decreasesin energy intake. American Journal of Clinical Nutrition, 83, 11–17.

Rolls, B. J., Rolls, E. T., & Rowe, E. A. (1982). The influence of varietyon human food selection and intake. In L. M. Barker (Ed.), The psycho-


biology of human food selection (pp. 101–122). Westport, CT: AVIPublishing.

Rolls, B. J., Rowe, E. A., & Rolls, E. T. (1982). How sensory properties offoods affect human feeding behavior. Physiology & Behavior, 29, 409–417.

Rolls, B. J., Rowe, E. A., Rolls, E. T., Kingston, B., Megson, A., &Gunary, R. (1981). Variety in a meal enhances food intake in man.Physiology & Behavior, 26, 215–221.

Rolls, B. J., van Duijvenvoorde, P. M., & Rolls, E. T. (1984). Pleasantnesschanges and food intake in a varied four course meal. Appetite, 5,337–348.

Rolls, E. T. (1997). Taste and olfactory processing in the brain and itsrelation to the control of eating. Critical Reviews in Neurobiology, 11,263–287.

Rush, C. R., Essman, W. D., Simpson, C. A., & Baker, R. W. (2001).Reinforcing and subject-rated effects of methylphenidate andd-amphetamine in non-drug-abusing humans. Journal of Clinical Psy-chopharmacology, 21, 273–286.

Saelens, B. E., & Epstein, L. H. (1996). The reinforcing value of food inobese and non-obese women. Appetite, 27, 41–50.

Salamone, J. D., & Correa, M. (2002). Motivational views of reinforce-ment: Implications for understanding the behavioral functions of nucleusaccumbens dopamine. Behavioural Brain Research, 137, 3–25.

Salamone, J. D., Correa, M., Mingote, S. M., & Weber, S. M. (2005).Beyond the reward hypothesis: Alternative functions of nucleus accum-bens dopamine. Current Opinion in Pharmacology, 5, 34–41.

Sbrocco, T., Nedegaard, R. C., Stone, J. M., & Lewis, E. L. (1999).Behavioral choice treatment promotes continuing weight loss: Prelimi-nary results of a cognitive-behavioral decision-based treatment for obe-sity. Journal of Consulting and Clinical Psychology, 67, 260–266.

Schachter, S. (1968, August 23). Obesity and eating: Internal and externalcues differentially affect the eating behavior of obese and normal sub-jects. Science, 161, 751–756.

Schertz, M., Adesman, A. R., Alfieri, N. E., & Bienkowski, R. S. (1996).Predictors of weight loss in children with attention deficit hyperactivitydisorder treated with stimulant medication. Pediatrics, 98, 763–769.

Schultz, W. (1998). Predictive reward signal of dopamine neurons. Journalof Neurophysiology, 80, 1–27.

Schultz, W. (2002). Getting formal with dopamine and reward. Neuron, 36,241–263.

Schwartz, M. W., Woods, S. C., Porte, D., Jr., Seeley, R. J., & Baskin,D. G. (2000, April 6). Central nervous system control of food intake.Nature, 404, 661–671.

Scislowski, P. W., Tozzo, E., Zhang, Y., Phaneuf, S., Prevelige, R., &Cincotta, A. H. (1999). Biochemical mechanisms responsible for theattenuation of diabetic and obese conditions in ob/ob mice treated withdopaminergic agonists. International Journal of Obesity, 23, 425–431.

Sclafani, A., & Ackroff, K. (2003). Reinforcement value of sucrose mea-sured by progressive ratio operant licking in the rat. Physiology &Behavior, 79, 663–670.

Sherry, B., McDivitt, J., Birch, L. L., Cook, F. H., Sanders, S., Prish, J. L.,et al. (2004). Attitudes, practices, and concerns about child feeding andchild weight status among socioeconomically diverse white, Hispanic,and African-American mothers. Journal of the American Dietetic Asso-ciation, 104, 215–221.

Shide, D. J., Caballerro, B., Reidelberger, R., & Rolls, B. J. (1995).Accurate energy compensation for intragastric and oral nutrients in leanmales. American Journal of Clinical Nutrition, 61, 754–764.

Simonen, R. L., Rankinen, T., Perusse, L., Leon, A. S., Skinner, J. S.,Wilmore, J. H., et al. (2003). A dopamine D2 receptor gene polymor-phism and physical activity in two family studies. Physiology & Behav-ior, 78, 751–757.

Small, D. M., Jones-Gotman, M., & Dagher, A. (2003). Feeding-induced

dopamine release in dorsal striatum correlates with meal pleasantnessratings in healthy human volunteers. NeuroImage, 19, 1709–1715.

Smith, B. A., & Blass, E. M. (1996). Taste-mediated calming in premature,preterm, and full-term human infants. Developmental Psychology, 32,1084–1089.

Smith, J. A., & Epstein, L. H. (1991). Behavioral economic analysis offood choice in obese children. Appetite, 17, 91–95.

Spencer, T., Wilens, T., Biederman, J., Faraone, S. V., Ablon, J. S., &Lapey, K. (1995). A double-blind, crossover comparison of methyl-phenidate and placebo in adults with childhood-onset attention-deficithyperactivity disorder. Archives of General Psychiatry, 52, 434–443.

Spiegel, T. A., Shrager, E. E., & Stellar, E. (1989). Responses of lean andobese subjects to preloads, deprivation, and palatability. Appetite, 13,45–69.

Spiegel, T. A., & Stellar, E. (1990). Effects of variety on food intake ofunderweight, normal-weight and overweight women. Appetite, 15, 47–61.

Staddon, J. E. R., & Ettenger, R. H. (1989). Learning: An introduction tothe principles of adaptive behavior. New York: Harcourt Brace andJovanovitch.

Stice, E., Cameron, R. P., Killen, J. D., Hayward, C., & Taylor, C. B.(1999). Naturalistic weight-reduction efforts prospectively predictgrowth in relative weight and onset of obesity among female adoles-cents. Journal of Consulting and Clinical Psychology, 67, 967–974.

Stice, E., Presnell, K., Shaw, H., & Rohde, P. (2005). Psychological andbehavioral risk factors for obesity onset in adolescent girls: A prospec-tive study. Journal of Consulting and Clinical Psychology, 73, 195–202.

Stricker, E. M., & Zigmond, M. J. (1984). Brain catecholamines and thecentral control of food intake. International Journal of Obesity, 8,S39–50.

Swanson, J. M., Flodman, P., Kennedy, J., Spence, M. A., Moyzis, R.,Schuck, S., et al. (2000). Dopamine genes and ADHD. Neuroscienceand Biobehavioral Reviews, 24, 21–25.

Tataranni, P. A., Baier, L., Jenkinson, C., Harper, I., Del Parigi, A., &Bogardus, C. (2001). A Ser311Cys mutation in the human dopaminereceptor D2 gene is associated with reduced energy expenditure. Dia-betes, 50, 901–904.

Tataranni, P. A., Gautier, J. F., Chen, K., Uecker, A., Bandy, D., Salbe,A. D., et al. (1999). Neuroanatomical correlates of hunger and satiationin humans using positron emission tomography. Proceedings of theNational Academy of Sciences of the United States of America, 96,4569–4574.

Telch, C. F., & Agras, W. S. (1996). The effects of short-term fooddeprivation on caloric intake in eating-disordered subjects. Appetite, 26,221–233.

Temple, J. L., Giacomelli, A. M., Roemmich, J. N., & Epstein, L. H. (inpress). Dietary variety impairs habituation in youth. Health Psychology.

Temple, J. L., Kent, K. M., Giacomelli, A. M., Paluch, R. A., Roemmich,J. N., & Epstein, L. H. (2006). Habituation and recovery of salivationand motivated responding for food in children. Appetite, 46, 280–284.

Terry, P., Gilbert, D. B., & Cooper, S. J. (1995). Dopamine receptorsubtype agonists and feeding behavior. Obesity Research, 3, 515S–523S.

Timberlake, W., & Farmer-Dougan, V. A. (1991). Reinforcement in ap-plied settings: Figuring out ahead of time what will work. PsychologicalBulletin, 110, 379–391.

Vandewater, K., & Vickers, Z. (1996). Higher-protein foods producegreater sensory-specific satiety. Physiology & Behavior, 59, 579–583.

Vara, L. S., & Epstein, L. H. (1993). Laboratory assessment of choicebetween exercise or sedentary behaviors. Research Quarterly for Exer-cise and Sport, 64, 356–360.

Volkow, N. D., Fowler, J. S., & Wang, G. J. (2002). Role of dopamine indrug reinforcement and addiction in humans: Results from imagingstudies. Behavioral Pharmacology, 13, 355–366.

Volkow, N. D., Wang, G., Fowler, J. S., Logan, J., Gerasimov, M.,


Maynard, L., et al. (2001). Therapeutic doses of oral methylphenidatesignificantly increase extracellular dopamine in the human brain. Jour-nal of Neuroscience, 21, RC121.

Volkow, N. D., Wang, G. J., Fowler, J. S., Logan, J., Jayne, M., Franceschi,D., et al. (2002). “Nonhedonic” food motivation in humans involvesdopamine in the dorsal striatum and methylphenidate amplifies thiseffect. Synapse, 44, 175–180.

Volkow, N. D., Wang, G. J., Fowler, J. S., Telang, F., Maynard, L., Logan,J., et al. (2004). Evidence that methylphenidate enhances the saliency ofa mathematical task by increasing dopamine in the human brain. Amer-ican Journal of Psychiatry, 161, 1173–1180.

Volkow, N. D., Wang, G. J., Maynard, L., Jayne, M., Fowler, J. S., Zhu,W., et al. (2003). Brain dopamine is associated with eating behaviors inhumans. International Journal of Eating Disorders, 33, 136–142.

Volkow, N. D., & Wise, R. A. (2005). How can drug addiction help usunderstand obesity? Nature Neuroscience, 8, 555–560.

Vuchinich, R. E., & Tucker, J. A. (1988). Contributions from behavioraltheories of choice to an analysis of alcohol abuse. Journal of AbnormalPsychology, 97, 181–195.

Wadden, T. A., Berkowitz, R. I., Sarwer, D. B., Prus-Wisniewski, R., &Steinberg, C. (2001). Benefits of lifestyle modification in the pharma-cologic treatment of obesity: A randomized trial. Archives of InternalMedicine, 161, 218–227.

Wadden, T. A., Berkowitz, R. I., Womble, L. G., Sarwer, D. B., Phelan, S.,Cato, R. K., et al. (2005). Randomized trial of lifestyle modification andpharmacotherapy for obesity. New England Journal of Medicine, 353,2111–2120.

Wang, G. J., Volkow, N. D., Felder, C., Fowler, J. S., Levy, A. V., Pappas,N. R., et al. (2002). Enhanced resting activity of the oral somatosensorycortex in obese subjects. NeuroReport, 13, 1151–1155.

Wang, G. J., Volkow, N. D., Logan, J., Pappas, N. R., Wong, C. T., Zhu,W., et al. (2001). Brain dopamine and obesity. Lancet, 357, 354–357.

Weingarten, H. P. (1983, April 22). Conditioned cues elicit feeding in satedrats: A role for learning in meal initiation. Science, 220, 431–433.

Westenhoefer, J., & Pudel, V. (1993). Pleasure from food: Importance forfood choice and consequences of deliberate restriction. Appetite, 20,246–249.

Wilens, T. E., Biederman, J., Spencer, T. J., & Prince, J. (1995). Pharma-

cotherapy of adult attention deficit/hyperactivity disorder: A review.Journal of Clinical Psychopharmacology, 15, 270–279.

Willner, P., Benton, D., Brown, E., Cheets, S., Davies, G., Morgan, J., etal. (1998). “Depression” increases “craving” for sweet rewards in animaland human models of depression and craving. Psychopharmacology,136, 272–283.

Wilson, C., Nomikos, G. G., Collu, M., & Fibiger, H. C. (1995). Dopa-minergic correlates of motivated behavior: Importance of drive. Journalof Neuroscience, 15, 5169–5178.

Wise, R. A. (2002). Brain reward circuitry: Insights from unsensed incen-tives. Neuron, 36, 229–240.

Wise, R. A. (2006). Role of brain dopamine in food reward and reinforce-ment. Philosophical Transactions of the Royal Society, Series B: Bio-logical Sciences, 361, 1149–1158.

Wisniewski, L., Epstein, L. H., & Caggiula, A. R. (1992). Effect of foodchange on consumption, hedonics, and salivation. Physiology & Behav-ior, 52, 21–26.

Wisniewski, L., Epstein, L. H., Marcus, M. D., & Kaye, W. (1997).Differences in salivary habituation to palatable foods in bulimia nervosapatients and controls. Psychosomatic Medicine, 59, 427–433.

Worsley, J. N., Moszczynska, A., Falardeau, P., Kalasinsky, K. S.,Schmunk, G., Guttman, M., et al. (2000). Dopamine D1 receptor proteinis elevated in nucleus accumbens of human, chronic methamphetamineusers. Molecular Psychiatry, 5, 664–672.

Wyvell, C. L., & Berridge, K. C. (2000). Intra-accumbens amphetamineincreases the conditioned incentive salience of sucrose reward: Enhance-ment of reward “wanting” without enhanced ’liking’ or response rein-forcement. Journal of Neuroscience, 20, 8122–8130.

Yang, Z. J., Blaha, V., Meguid, M. M., Laviano, A., Oler, A., & Zadak, Z.(1999). Interleukin-1 alpha injection into ventromedial hypothalamicnucleus of normal rats depresses food intake and increases release ofdopamine and serotonin. Pharmacology Biochemistry and Behavior, 62,61–65.

Yang, Z. J., & Meguid, M. M. (1995). LHA dopaminergic activity in obeseand lean Zucker rats. NeuroReport, 6, 1191–1194.

Received June 10, 2006Revision received March 26, 2007

Accepted March 30, 2007 �


food reinforcement and eating: a multilevel analysisabcwest/pmwiki/cafe/2007 - epstein et al -...

Documents