Elsevier

Appetite

Volume 48, Issue 1, January 2007, Pages 12-19
Appetite

Research report
From motivation to behaviour: A model of reward sensitivity, overeating, and food preferences in the risk profile for obesity

https://doi.org/10.1016/j.appet.2006.05.016Get rights and content

Abstract

The reinforcing effects of addictive drugs and palatable foods are regulated, at least in part, by a common biological mechanism. The reactivity or sensitivity of these brain reward regions have been found to correlate significantly with the risk for a variety of drug addictions. Sensitivity to Reward (STR) is conceptualised as a psycho-biological personality trait rooted firmly in the availability of dopamine in the mesocorticolimbic (‘common reward’) pathways, and as such is a good candidate for studying motivational factors and eating behaviours. The purpose of the present study was to examine whether STR was related to behaviours that contribute to excess body weight. Structural equation modelling procedures were used with a sample of healthy adult women (n=151). We hypothesised that STR would positively predict overeating and a preference for foods high in fat and sugar; and that these two behaviour would, in turn, predict a higher Body Mass Index. Results provided an excellent fit of the model to our data confirming our view that a personality trait like STR can only influence a physical condition like body weight indirectly by the way it co-varies with behaviours that contribute directly to variation in the outcome variable.

Introduction

The reinforcing effects of addictive drugs and palatable foods are regulated, at least in part, by a common brain mechanism depending crucially on the level of dopamine (DA) activation in mesocorticolimbic regions (e.g. Di Chiara et al., 2004; Risinger, Freeman, Rubinstein, Low, & Grandy, 2000). Indeed, the sub-cortical brain does not seem to differentiate among rewards provoked by natural reinforcers like food, illicit drugs like cocaine, or behaviours like gambling (Kelley, Schiltz, & Landry, 2005). For example, two recent studies found that obese women had lower rates of alcohol (Kleiner et al., 2004) and marijuana use (Warren, Frost-Pineda, & Gold, 2005) than their normal weight, age-matched counterparts, and that, in both cases, an inverse relationship existed between body mass index (BMI) and drug use. The authors concluded that overeating competes with pharmacologic agents for brain reward sites, and thereby may serve as a buffer for the use and abuse of other addictive behaviours. Complementary to this viewpoint is evidence that appetite suppression is a major pharmacological effect of chronic drug use (see Cochrane, Malcolm, & Brewerton, 1998).

Clearly many factors influence the kinds of pleasure we pursue in our daily lives. The context—through classical and operant conditioning processes—is particularly important in determining the choices we make and the frequency of their use (Corwin & Hajnal, 2005). In most Western societies, the convenience, the availability, and the relatively low cost, make tasty foods a highly salient reward for many people. Aggressive marketing by the fast-food industry further enhances the temptation to overindulge.

In view of the many commonalities between food and drug reward, some have argued that chronic overeating can be modelled as an addictive behaviour similar to other substance-dependent disorders (e.g. Corwin, 2006; Davis, Strachan, & Berkson, 2004; James, Gold, & Liu, 2004; Wang, Volkow, Thanos, & Fowler, 2004). Consequently, there is a growing interest in examining the role of DA neurotransmission in the risk profile for obesity—especially how its variation in the population affects individual differences in vulnerability to overeating.

Undoubtedly there are many routes to obesity so it is reasonable to assume that weight-gaining individuals possess a variable set of biological and behavioural susceptibility factors (Blundell et al., 2005). On the one hand, there is evidence that hypo-dopaminergic functioning—what has been called a Reward Deficiency Syndrome (RDS)1—underlies a range of addictions including alcoholism, cocaine abuse, and pathological gambling (see Bowirrat & Oscar-Berman, 2005). Some have recently argued that RDS is also a factor in the development of obesity (e.g. Wang et al., 2004). On the other hand, enhanced DA functioning, which is characterized by a heightened hedonic capacity and greater behavioural activation (e.g. Cohen, Young, Baek, Kessler, & Ranganath, 2005), fosters strong appetitive responses to the natural pleasures in life. For instance, amplification of the DA signal in human participants via a small dose of oral methylphenidate2 increased their desire to eat in response to a palatable food cue (Volkow et al., 2002). There is also evidence that obese individuals have enhanced sensitivity in brain areas associated with the sensory (e.g. lips, tongue, mouth) processing of food (Wang et al., 2002).

Regrettably, our current obesigenic environment can exploit those with a high sensitivity to reward (STR) by promoting consumption beyond caloric need. Non-homeostatic eating can take several forms including eating that is driven by emotional states (‘comfort’ eating) or by environmental cues such as the sight and smell of food stimuli. It is also characterised by frequent snacking and episodes of binge eating. In addition, hedonic processes—commonly regarded as the pleasure that is associated with food—are of central relevance to other aspects of eating. A primary target of such influence is an enhanced preference for foods that are fat and sweet since they typically provide us with a greater reinforcement value than bland food (Epstein & Leddy, 2006). In a recent study of high and low-fat phenotypes—those habitually consuming a diet containing >43% or <32% fat respectively—the former group comprised a significantly greater number of obese individuals (Blundell et al., 2005). However, among the high-fat group there was considerable variability in BMI, and results indicated that those who were prone to weight gain reported higher hedonic responsiveness to eating, and a greater intensity of pleasurable sensations from the taste of food. We predicted that those who are highly sensitive to reward would therefore also be more responsive (than their anhedonic counterparts) to the perceived palatability of sweet and fatty foods.

To date, only a handful of studies has investigated the relationship between reward sensitivity, eating behaviours, and body weight, and are all supportive of positive links (Davis et al., 2004; Franken & Muris, 2005; Loxton & Dawe, 2001). For example, Franken and Muris (2005) found that young women who were more sensitive to reward reported stronger food cravings and had a higher BMI. Likewise, Davis et al. (2004) found that anhedonic women were less likely to overeat, after controlling for depression in their regression model. However, as a body of work these studies are limited by the general use of small samples of primarily young, normal-weight women, by self-report instead of objective measures of body weight, and by single markers of overeating.

The present study expands on this research by using a more comprehensive set of measures to reflect the constructs of interest, and by testing a large sample of women who are more representative of the adult population in terms of BMI and (pre-menopausal) age. We used structural equation modelling (SEM) to test the prediction that sensitivity to reward is a phenotypic positive influence on weight gain—especially in an environment that proliferates with tempting and available foods—via its influence on overeating and food preferences (see Fig. 1). The first three constructs in the path diagram were modelled as multi-factorial latent variables. Sensitivity to Reward had two commonly used measures of this personality trait. The measured observations for Overeating comprised binge eating, emotionally-driven eating, and eating prompted by external stimuli rather than hunger. All these aspects of overeating have been implicated in the risk profile for obesity (e.g. Delahanty, Meigs, Hayden, Williamson, & Nathan, 2002; Lluch, Herbeth, Mejean, & Siest, 2000; Wansink, Painter, & North, 2005). Finally, Food Preferences had two observed measures: a preference for high fat food and a preference for sweet food.

Section snippets

Subjects

One hundred and fifty-one healthy pre-menopausal women between the ages of 25 and 50 years (mean=33.5 years; SD=7.1) took part in the study. They were solicited from posters placed at two university campuses, and at various hospital and community centres in the urban core of a large Canadian city, asking for volunteers to participate in a “health psychology study”. Participants were screened initially during a structured telephone interview and excluded if they had any serious medical

Results

The proposed model (see Fig. 1) was tested using SEM and Amos 6.0 software. SEM is a useful statistical procedure for researchers who want to test a theory involving causal processes, and therefore is well suited to the management of cross-sectional data for inferential purposes (Byrne, 2001). The kurtosis and skew of the eight variables in the model were between −1.04 and 0.50 and −0.53 and 0.96, respectively—values which are well within the acceptable range to proceed with SEM according to

Discussion

In a large sample of healthy adult women, we tested the theory that STR would predict (i) the tendency to eat beyond caloric need and in the absence of hunger and, (ii) a heightened preference for sweet and fatty foods—two factors which consistently co-vary with body weight in the positive direction. SEM procedures are advantageous in behavioural sciences research because they allow us to study constructs that are difficult to observe directly or which have complex manifestations—like

Conclusions

Personality can only influence a physical condition like body weight indirectly by the way it co-varies with behaviours that contribute directly to its variation in the population. STR is a biologically-based trait rooted firmly in the reactivity of the mesocorticolimbic (‘common reward’) DA pathways and as such is a good candidate for study in the area of ingestive behaviours. As predicted, a high STR was related to overeating as well as an heightened preference for sweet and fatty foods; and

References (42)

  • I.H.A. Franken et al.

    Individual differences in reward sensitivity are related to food craving and relative body weight in healthy women

    Appetite

    (2005)
  • P.J. Geiselman et al.

    Reliability and validity of the macronutrient self-selection paradigm and a Food Preference Questionnaire

    Physiology & Behavior

    (1998)
  • R. Torrubia et al.

    The Sensitivity to Punishment and Sensitivity to Reward Questionnaire (SPSRQ) as a measure of Gray's anxiety and impulsivity dimensions

    Personality and Individual Differences

    (2001)
  • G.-J. Wang et al.

    Brain DA and obesity

    The Lancet

    (2001)
  • B. Wansink et al.

    Exploring comfort food preferences across gender and age

    Physiology and Behavior

    (2003)
  • K. Blum et al.

    “Reward deficiency syndrome”: An emerging concept

    Molecular Psychiatry

    (2001)
  • K. Blum et al.

    Increased prevalence of Taq 1 A(1) allele of the DA receptor gene (DRD(2)) in obesisty with comorbid substance use disorder: A preliminary report

    Pharmacogenetics

    (1996)
  • A. Boomsma

    Reporting analyses of covariance structures

    Structural Equation Modeling

    (2000)
  • A. Bowirrat et al.

    Relationship between dopaminergic neurotransmission, alcoholism, and reward deficiency syndrome

    American Journal of Medical Genetics Part B-Neuropsychiatric Genetics

    (2005)
  • B.M. Byrne

    Structural equation modelling with AMOS

    (2001)
  • C.S. Carver et al.

    Behavioral inhibition, behavioral activation, and affective responese to impending reward and punishment: The BIS/BAS Scales

    Journal of Personality and Social Psychology

    (1994)
  • Cited by (290)

    View all citing articles on Scopus
    View full text