In Study 1, the duration of elective acute caloric deprivation correlated positively with activation in the OFC in response to pictures of appetizing food. Activation in the OFC has been associated with the subjective evaluation of (food) reward (Kringelbach, 2004
) and craving (Wang et al., 1999
). This finding extends previous evidence that activation in the medial and lateral OFC in response to pictures of high-calorie over low-calorie foods is greater in a fasted state compared to a sated state (Goldstone et al., 2009
). Duration of acute caloric deprivation also correlated with elevated activation in the precentral gyrus. Activation in this region is associated with motor responses and is thought to be related to planning to acquire or consume food (Geliebter et al. 2006
). Activation in this region in response to palatable food pictures extends findings from previous fMRI studies in healthy people who report a craving for food after experimentally manipulated caloric deprivation (Goldstone et al., 2009
, Siep et al., 2009
). Increased motor responses to food images may therefore reflect an anticipated desire for food consumption.
In Study 2, participants reporting a longer duration of elective acute caloric deprivation exhibited a pattern of neural activation in regions implicated in (food) reward, motivation, and attention in response to receipt and anticipated receipt of palatable food, such as the thalamus, parahippocampal gyrus, dlPFC, dorsal ACC, MFG, putamen, and mid insula (Haase et al., 2009
; Koob & Volkow, 2010
; Pessoa et al., 2002
; Stice, Spoor, Bohon, Veldhuizen, & Small, 2008b
; Uher et al., 2006
). The dlPFC has also found to be associated with planning and goal-directed behavior (Heller, 2004
). The positive association between duration of acute caloric deprivation and activation in the dlPFC extends the finding of Uher et al. (2006)
who found that dlPFC activation during food pictures was stronger in the fasted compared to sated state. In response to anticipated food receipt, duration of acute caloric deprivation was also positively correlated with activation in regions associated with visual processing, such as the IOG, lingual gyrus, MTG (Pessoa et al., 2002
, Hahn et al., 2006
) and motor responses, such as the precentral gyrus (Geliebter et al., 2006
). Thus, a longer duration of acute caloric deprivation may prompt greater attention to food cues and increased motivation to obtain the food. Duration of caloric deprivation was also positively related to activation in the cerebellum in response to anticipated food receipt. Cerebellum activation has been associated with gustatory and olfactory stimulation (Sobel et al., 1998
), images of high-calorie versus low-calorie foods (Killgore et al., 2003
) and oral glucose intake (Liu et al., 2000
). The cerebellum finding dovetails the findings of past positron emission tomography (PET) studies that have shown that increased regional cerebral blood flow in the cerebellum is associated with hunger and appetite (Tataranni et al., 1999
), whereas decreased blood flow in this region is linked with satiation (Gautier et al., 2001
). Overall, the findings from Study 1 and Study 2 suggest that duration of acute calorie deprivation is related to greater reward valuation of and attention to palatable food pictures and food cues and potentially with greater motivation to obtain or consume the food.
Participants in a negative energy balance state during the scan in Study 2 compared to those in a positive or stable energy balance showed greater activation in the precuneus, mPFC, ventral ACC, cuneus, dlPFC, and PCC in response to anticipated receipt of milkshake. Activations in the precuneus and cuneus are related to visual attention and memory retrieval (Cavanna & Trimble, 2006
) and have found to be associated with experimental caloric deprivation (Uher et al., 2006
) and cue-induced craving (Due et al., 2002
). The mPFC and ventral ACC have been linked with motivation, emotional decision-making, and reward processing (Dolcos, LaBar, & Caneza, 2004
; Koob & Volkow, 2010
; Haase et al., 2009
). The PCC has found to be associated with processing of emotional salient stimuli (Maddock, 1999
). For example, the PCC showed activation in states of both high motivation and aversion of eating, while activation was decreased in a neutral condition (Small et al., 2001
). This region has also been linked to spatial attention (e.g., Small et al., 2003
) and visual imagery (Hassabis et al., 2007
). Overall, results suggest that a longer-term negative energy balance state is associated with elevated activation in regions implicated in reward, attention, and motivation during anticipated palatable food receipt.
Participants in a negative energy balance relative to those in a stable or positive energy balance state also showed greater activation in the caudate, precentral gyrus, hippocampus, MOG, and fusiform gyrus in response to actual intake of palatable food. The caudate is involved in incentive motivation and is thought to encode consummatory food reward (Small et al., 2001
; Stice et al., 2008b
). The hippocampus modulates saliency of stimuli through regulation of ventral striatum dopamine (DA) release (Berridge & Robinson, 1998
) and has been implicated in the development of memories of food (Van Vugt, 2010
), food craving (Pelchat et al., 2004
), physiological hunger (Haase et al., 2009
), and processing of food tastes (Gautier et al., 1999
). Both the MOG and fusiform gyrus have found to be involved in visual processing particularly in response to food images (Hahn et al., 2006
). The elevated activation in the fusiform gyrus dovetails with evidence that experimentally manipulated caloric deprivation results in elevated activation in this region in response to food stimuli (LaBar et al., 2001
; Uher et al., 2006
). Results may suggest that individuals in a longer-term negative energy balance state are more sensitive to the hedonic sensations associated with the palatable food intake.
It was noteworthy that caloric deprivation was related to altered neural response in a much wider array of reward, attention, and motivation regions in response to receipt and anticipated receipt of palatable foods versus images of palatable foods herein. Past fMRI studies have not investigated the effects of caloric deprivation on neural response to all three of these events. Yet, Uher et al. (2006)
did find that experimentally manipulated caloric deprivation increased responsivity of reward and sensory/hedonic regions to receipt of chocolate milk and chicken soup, but not images of palatable foods versus non-deprived participants. Collectively, results imply that caloric deprivation increases responsivity to real food much more than it does to images of food, probably because pictures of food hold no caloric value for hungry people. However, it is also important to consider alternative explanations for this pattern of findings. First, participant may have shown habituation to the palatable and unpalatable food images because they viewed and rated palatability two days before completing the scan on average. However, an earlier paper that examined the relation of BMI to BOLD response to the images of palatable foods versus unpalatable foods and water glasses using data from Study 1 (Stice et al., 2010
) found 16 significant peaks, which is within the range of peaks identified in the other studies that have examined this question with subjects viewing the food images for the first time (range 3 – 26 peaks; Bruce et al., 2010; Martin et al., 2010; Rothemund et al., 2007; Stoeckel et al., 2008). Second, it is possible that the fact that Study 2 included both sexes, whereas Study 1 included only females, contributed to the differential findings. However, post hoc
analyses confirmed that all effects remained significant when we controlled for sex in the Study 2 analyses. Third, it is also possible that the fact that Study 2 involved a much narrower range of BMI values contributed to the differential effects. Again, post hoc
analyses indicated that only 3 of the 34 peaks became non-significant when we controlled for BMI in analyses.
The evidence that elective caloric restriction in the present studies was associated with greater responsivity of attention and reward regions to palatable food images, anticipated receipt of palatable food, and receipt of palatable food more broadly extend findings from caloric deprivation experiments. In those experiments, manipulated caloric restriction was similarly associated with elevated response in attention (ACC), reward valuation (OFC, amygdala), reward (ventral striatum, insula), and memory (hippocampus) regions in response to images of palatable foods (e.g., Fuhrer et al., 2008
; Goldstone et al., 2009
; Leidy et al, 2011
). The similarity in the findings from studies that investigated elective caloric deprivation and experimentally manipulated caloric deprivation suggest that these effects are very robust. Experimental research with rats has revealed that acute food deprivation and chronic caloric deprivation results in increased DA release at feeding and increased D2 receptor binding (Thanos et al., 2008
), which suggests that caloric deprivation results in greater signaling capacity of DA-based reward circuitry in response to food. Interesting, animal experiments show that caloric deprivation selectively increases preferences for high-fat foods (Lucas & Sclafani, 1992
One potential mechanism for the effects observed herein comes from an experiment with rats involving discontinuation of cocaine use (Cameron & Carelli, 2012
). Abstinence from cocaine after a period of regular use resulted in a 17% increase in cells that selectively fire in response to cocaine directed behavior, suggesting that caloric deprivation may likewise result in an increase in DA-neurons that fire in response to food cues. Another potential mechanism for the effects observed herein comes from animal experiment involving caloric deprivation. In vivo
dialysis experiments show higher food-induced DA release in fasted versus satiated rats (Wilson et al., 1995
). Intracerebroventricular injection of the DR2 agonist quinpirole produced a more pronounced striatal neuronal activation in caloric restricted rats versus ad lib
fed controls (Carr et al., 2003
). Administration of the D1R agonist SFK-82958 produced enhanced striatal activation and elevated D2Rs showed enhanced effector striatal coupling (Carr et al., 2003
). Accumbens DA levels have been shown to increase in response to caloric intake more following longer versus shorter caloric deprivation periods (Yoshida et al., 1992
) and in response to caloric deprivation weight loss diets versus baseline (Avena et al., 2008
). These data collectively suggest that caloric deprivation increases D2R receptor sensitization, which could explain the greater reward value of food after caloric deprivation.
Thus, results collectively suggest that caloric deprivation increases the reward value of food, particularly high-calorie palatable food and cues predicting food receipt. One implication of these data is that weight loss diets characterized by caloric deprivation may be bound to fail because they increase the reward value of food with every passing hour of deprivation. Ironically, our findings imply that the more successful people are at caloric-restriction dieting, the greater likelihood that it will not last. Indeed, this phenomenon may explain why most weight loss diets are ineffective in producing lasting weight loss. Future studies should investigate whether absolute caloric deprivation increases the reward value of food more than reducing overall caloric intake by replacing high-fat/high-sugar foods with low-fat/low-sugar foods. However, trials have not found that weight loss interventions involving fewer meals produce significantly greater weight loss than those that involve more frequent meals (Groesz & Stice, 2007
; Schlundt, Hill, Sbrocco, Pope-Cordle, & Sharp, 1992
). Future fMRI studies should also investigate whether negative energy balance induced by reducing caloric intake has similar effects on neural responsivity as an energy deficit induced by increasing physical activity.
The present findings may also explain why people who attempt to diet do not typically succeed per objective measures of caloric intake (e.g., Hetherington et al., 2000
; Stice et al., 2004
; Sysko et al., 2007
). Our results imply that if people attempt to lose weight by abstaining from food intake for long durations of time, it will have the effect of increasing the reward value of food, which may lead to poor food choices when the individual eventually does eat. Thus, results imply that dieting that is characterized by meal skipping and fasting would be less successful than weight loss efforts characterized by intake of low energy dense foods. Hours since last caloric intake and negative energy balance status showed no correlation with scores on the Dutch Restrained Eating Scale, implying that restraint scale scores identify individuals who are attempting to reduce caloric intake, rather than those who typically in a negative energy balanced state. These findings imply that future studies should measure duration of fasting efforts or whether participants are actually in a negative energy balance state, rather than continue to use dietary restraint scales. Interestingly, hours of caloric deprivation and negative energy balance status showed markedly stronger relations to intake, anticipated intake, and images of palatable foods than to the Dutch Restrained Eating Scale in the present data sets: dietary restraint scores correlated positively with activation in the OFC and dlPFC in response to milkshake receipt, but did not correlate with activation in response to anticipated receipt of milkshake or pictures of palatable food (Burger & Stice, 2011
Another implication of the present results is that it will be important for future fMRI studies that investigate neural response to palatable food images, receipt, and anticipated receipt to more carefully standardize caloric intake before the scans. It may be optimal to invite participants to the lab several hours before the scan so that they can be fed a standardized meal.
The limitations of the present studies should be considered. First, although the repeated measures of weight at the beginning and end of the 2-week period during which the fMRI scans occurred allowed us to classify participants into those who were in an energy deficit on average versus energy balance or a positive energy state in Study 2, there was no way to confirm that on the day of the actual scan that they consumed fewer caloric than needed to maintain their weight. Yet, the fact that they showed increased responsivity in attention and reward regions in response to food that were similar to regions that showed elevated responsivity in response to food in those who reported a longer period of caloric deprivation and participants in previous studies who were assigned to caloric deprivation conditions increases confidence in the negative energy balance findings. Second, we investigated neural response to receipt and anticipated receipt of only one palatable food and it was a beverage. Results should be generalized with caution to other palatable foods, particularly solid foods. Third, because participants in these two studies were adolescents, results may not generalize to adults. Fourth, although the samples used in these studies are larger than those typically used in brain imaging studies, they were still only moderate in magnitude, which may have limited our statistical power and the generalizability of our findings.
The results from the current elective caloric deprivation studies taken in conjunction with findings from the earlier caloric deprivation experiments suggest that acute and prolonged caloric deprivation increases the reward value of food, particularly energy dense palatable foods. These data may explain for why weight loss diets typically do not produce lasting weight loss. Most critically, findings imply that weight loss diets that involve the replacement of unhealthy energy dense foods with healthy low energy density foods should be more effective than diets that involve long periods of caloric deprivation.