Obesity is a chronic disease that is credited with over 111,000 deaths annually in the US, which largely result from atherosclerotic cerebrovascular disease, coronary heart disease, colorectal cancer, hyperlipidemia, hypertension, gallbladder disease, and diabetes mellitus (Flegal, Graubard, Williamson, & Gail, 2005
). Regrettably, the treatment of choice for obesity only results in transitory weight loss (Jeffery et al., 2000
) and most obesity prevention programs do not reduce risk for future weight gain (Stice, Shaw, & Marti, 2006
). These interventions may have limited efficacy because our understanding of the etiologic processes is still incomplete. Although it has been established that obesity is the result of a positive energy balance, it is unclear why some individuals have such a difficult time balancing caloric intake with expenditure.
One possible explanation is that some individuals have abnormalities in subjective reward from food intake or anticipated intake that increase risk for obesity. Some scholars hypothesize that obese individuals experience greater activation of the meso-limbic reward system in response to food intake (consummatory food reward), which may increase risk for overeating (Davis, Strachan, & Berkson, 2004
; Dawe & Loxton, 2004
). This is akin to the reinforcement sensitivity model of substance abuse, which posits that certain people show greater reactivity of reward circuitry to psychoactive drugs (Dawe & Loxton, 2004
). In contrast, others hypothesize that obese individuals experience less activation of the meso-limbic reward system in response to food intake, which leads them to overeat to compensate for this deficiency (Comings & Blum, 2000
; Wang, Volkow, & Fowler, 2002
). This is similar to the reward deficiency syndrome thesis, which suggests that people turn to alcohol and drug use to stimulate sluggish reward circuitry (Comings & Blum, 2000
). A third hypothesis is that greater anticipated reward from food intake (anticipatory food reward) increases risk for overeating (Pelchat, Johnson, Chan, Valdez, & Ragland, 2004
; Roefs, Herman, MacLeod, Smulders, & Jansen, 2005
Two lines of evidence imply it may be useful to conceptually distinguish between consummatory food reward and anticipatory food reward. First, animal studies suggest that the reward value of food shifts from the consumption of food to the anticipated consumption of food after conditioning, wherein cues associated with food consumption begin to elicit anticipatory food reward. Naive monkeys that had not experienced rewards in a setting showed activation of mesotelencephalic dopamine neurons only in response to food taste; however, after conditioning, dopaminergic activity began to precede reward delivery and eventually maximal activity was elicited by the conditioned stimuli that predicted the impending reward rather than by actual food receipt (Schultz, Apicella, & Ljungberg, 1993
; Schultz, & Romo, 1990
). Kiyatkin and Gratton (1994)
found that the greatest dopaminergic activation occurred in an anticipatory fashion as rats approached and pressed the bar that produced food reward and activation actually decreased as the rat received and ate the food. Indeed, Blackburn, Phillips, Jakubovic, and Fibiger (1989)
found that dopamine activity was greater in the nucleus accumbens of rats after presentation of a conditioned stimulus that usually signaled food receipt than after delivery of an unexpected meal. Second, how hard participants work to earn snack food in an operant task (which they are later permitted to consume) is a stronger predictor of ad lib
caloric intake than are pleasantness ratings of tastes of the snack foods (Epstein, Temple et al., 2007
; Epstein et al., 2004a
). These data also seem to imply that anticipated reward from food intake is a stronger determinant of caloric intake than the reward experienced when the food is actually consumed. Collectively, these data imply that it may be useful to distinguish between consummatory food reward and anticipatory food reward when examining potential risk factors for obesity.
Brain imaging studies have identified regions that appear to encode consummatory food reward in normal weight individuals. Consumption of palatable foods, relative to consumption of unpalatable foods or tasteless foods, results in greater activation of the orbitofrontal cortex (OFC) and frontal operculum/insula, as well as greater release of dopamine in the dorsal striatum (O’Doherty, Deichmann, Critchley, & Dolan, 2002
; Small, Jones-Gotman, & Dagher, 2003
; Volkow et al., 2003
). Other brain imaging studies have identified regions that appear to encode anticipatory food reward in normal weight humans. Anticipated receipt of a palatable food, versus anticipated receipt of unpalatable food or a tasteless food, results in greater activation in the OFC, amygdala, cingulate gyrus, striatum (caudate nucleus and putamen), ventral tegmental area, midbrain, parahippocampal gyrus, and fusiform gyrus (O’Doherty et al., 2002
; Pelchat et al., 2004
). These studies suggest that somewhat distinct brain regions are implicated in anticipatory and consummatory food reward, but that there is some overlap (OFC and striatum). To date only two studies have directly compared activation in response to anticipatory and consummatory food reward to isolate regions that show greater activation in response to one phase of food reward versus the other. Anticipation of a pleasant taste, versus actual taste, resulted in greater activation in the dopaminergic midbrain, nucleus accumbens, and the posterior right amygdala (O’Doherty et al., 2002
). Another study found that anticipation of a pleasant drink resulted in greater activation in the amygdala and mediodorsal thalamus, whereas the receipt of the drink resulted in greater activation in the left insula/operculum (Small et al, 2008). These two studies suggest that the amygdala, midbrain, nucleus accumbens, and mediodorsal thalamus are more responsive to anticipated consumption versus consumption of food, whereas the frontal operculum/insula is more responsive to consumption versus anticipated consumption of food. Thus, available evidence seems to suggest that distinct brain regions have been implicated in encoding anticipatory and consummatory food reward, although more research will be necessary before firm conclusions are possible.
Certain findings appear to be consistent with the thesis that obese individuals experience greater food reward, though it is not clear whether findings are reflective of disturbances in consummatory versus anticipatory food reward. Obese relative to lean individuals recall that high-fat and high-sugar foods are more pleasant tasting and report that eating is more reinforcing (Rissanen et al., 2002
; Saelens & Epstein, 1996
; Westenhoefer & Pudel, 1993
). Children at risk for obesity by virtue of parental obesity rate tastes of high-fat food as more pleasant and show a more avid feeding style than children of lean parents (Stunkard, Berkowitz, Stallings, & Schoeller, 1999
; Wardle, Guthrie, Sanderson, Birch, & Plomin, 2001
). Obese children are more likely to eat in the absence of hunger (Fisher & Birch, 2002
) and work harder for food than lean children (Temple et al., in press
). Self-reported food cravings correlated positively with body mass and objectively measured caloric intake (Delahanty, Meigs, Hayden, Williamson, & Nathan, 2002
; Forman et al., 2007
; Franken & Muris, 2005
; Nederkoorn, Smulders, & Jansen, 2000
). Obese adults report stronger craving of high-fat, high-sugar foods (Drewnowski, Kurth, Holden-Wiltse, & Saari, 1992
; White, Whisenhunt, Williamson, Greenway, & Netemeyer, 2002
) and work for more food than lean adults (Epstein et al., 2007
; Saelens & Epstein, 1996
). Morbidly obese relative to lean individuals showed greater resting metabolic activity in the oral somatosensory cortex, a region associated with sensation in the mouth, lips, and tongue (Wang, Volkow, Felder et al., 2002
), which may render the former more sensitive to the rewarding properties of food intake and increase risk for overeating.
To date, few brain imaging studies have compared the brain activation in response to presentation of pictured food or actual food among obese verse lean individuals. One study found increased activation in the right parietal and temporal cortices after exposure to pictured food in obese but not lean women and that this activation correlated positively with hunger ratings (Karhunen, Lappalainen, Vanninen, Kuikka, & Uusitupa, 1997
). Rothemund and associates (2007)
found greater dorsal striatum response to pictures of high-calorie foods in obese verse lean adults and that body mass correlated positively with response in insula, claustrum, cingulate, somatosensory cortex, and lateral OFC. Stoeckel and associates (2008)
found greater activation in the medial and lateral OFC, amygdala, ventral striatum, medial prefrontal cortex, insula, anterior cingulate cortex, ventral pallidum, caudate, and hippocampus response to pictures of high-calorie foods (versus low-calorie foods) for obese relative to lean individuals. However, activation of the OFC and cingulate in response to viewing pictures of palatable foods correlated negatively with BMI among normal weight women (Killgore & Yargelun-Todd, 2005). Del Parigi et al (2004)
found that the dorsal insula and posterior hippocampus remain abnormally responsive to consumption of food in previously obese compared to lean individuals, leading to the conclusion that these abnormal responses may increase risk for obesity.
Other findings are more consistent with the notion that obese individuals may experience less food reward. Wang et al. (2001)
found that D2 receptors are reduced in the striatum in morbidly obese individuals in proportion to their body mass, suggesting that they exhibit decreased dopamine receptor binding in the meso-limbic system. Although it has yet to be determined whether obese individuals show reduced D2 receptor density relative to lean individuals, obese rats have lower basal dopamine levels and reduced D2 receptor expression than lean rats (Fetissov, Meguid, Sato, & Zhang, 2002
; Hamdi, Porter, & Prasad, 1992
; Orosco, Rouch, & Nicolaidis, 1996
), yet obese rats show more phasic release of dopamine during feeding than lean rats (Yang & Meguid (1995
). Furthermore, lean and obese adults with the TaqI A1 allele, which is associated with reduced D2 receptors and weaker dopamine signaling, work more to earn food in operant paradigms (Epstein et al., 2004b
). These results echo evidence that addictive behaviors such as alcohol, nicotine, marijuana, cocaine, and heroin abuse are associated with reduced D2 receptor density and blunted sensitivity of mesolimbic circuitry to reward (Comings & Blum, 2000
; Martinez et al., 2005
). Wang, Volkow, and Fowler (2002)
posit that deficits in D2 receptors may predispose individuals to use psychoactive drugs or overeat to boost a sluggish dopamine reward system. However, it is possible that overeating high-fat and high-sugar food leads to down-regulation of D2 receptors (Davis et al., 2004
), paralleling neural response to chronic use of psychoactive drugs (Volkow, Fowler, & Wang, 2002
). Indeed, animal studies suggest that repeated intake of sweet and fatty foods results in down-regulation of D2 receptors and decreased D2 sensitivity (Bello, Lucas, & Hajnal, 2002
; Kelley, Will, Steininger, Xhang, & Haber, 2003); changes that occurs in response to substance abuse.
In sum, there is emerging evidence that obese individuals may show general abnormalities in food reward relative to lean individuals. Specifically, obese relative to lean individuals report greater craving for high-fat/high-sugar foods, find eating more reinforcing, show greater resting activation of the somatosensory cortex, and show greater reactivity of the gustatory cortex to food intake and presentation of food or pictured food. Yet, there is also evidence that obese individuals show a hypofunctioning striatum, which may prompt them to overeat to boost a sluggish reward network or may be a result of receptor down-regulation. One factor that might have contributed to the mixed findings is that many studies used self-report measures, which could be misleading because those who struggle with overeating may assume that food is more rewarding to them, which influences how they complete the scales. Furthermore, self-report scales likely tap anticipated reward from food intake, or the memory of reward from food intake, rather than reward experienced during food consumption, as the studies did not measure perceived reward during food intake. In addition, findings from self-report and behavioral measures are vulnerable to social desirability biases. In addition, few studies have actually involved food intake or exposure to real food, which may limit the ecological validity of the findings. Perhaps most importantly, previous studies have not used paradigms that were specifically designed to assess individual differences in consummatory and anticipatory food reward when comparing obese to lean individuals. Thus, we think it may be useful to use objective brain imaging paradigms that directly measure activation of reward circuitry in response to food intake and anticipated food intake. To our knowledge, studies have not used brain-imaging to test whether obese individuals show differential activation of food reward circuitry during food consumption or anticipated consumption relative to lean individuals.
The present study sought to more fully characterize the nature of individual differences in neural response to food using objective brain imaging methodology, with the hope that an improved understanding of neurological substrates that increase risk for obesity will advance etiologic models and the design of more effective preventive and treatment interventions. We extended previous findings by examining activation in response to receipt of chocolate milkshake versus tasteless solution (consummatory food reward) and in response to cues signaling impending delivery of chocolate milkshake versus tasteless solution (anticipatory food reward) among obese and lean individuals. We hypothesized that obese relative to lean individuals would show greater activation in the gustatory cortex and the somatosensory cortex, and less activation in the striatum, in response to the anticipation and consumption of milkshake. We also hypothesized that the body mass of participants would show linear relations to activation in these brain regions. We studied adolescents because we wanted to reduce the risk that a long history of obesity might result in receptor down-regulation secondary to a chronically rich diet. We studied females because the primary goal of this study was to test whether food reward abnormalities correlate with bulimic pathology, which is rare in males.