Before diving in, I'd like to address the critique that the food reward concept is a tautology or relies on circular reasoning (or is not testable/falsifiable). This critique has no logical basis. The reward and palatability value of a food is not defined by its effect on energy intake or body fatness. In the research setting, food reward is measured by the ability of food or food-related stimuli to reinforce or motivate behavior (e.g., 1). In humans, palatability is measured by having a person taste a food and rate its pleasantness in a standardized, quantifiable manner, or sometimes by looking at brain activity by fMRI or related techniques (2). In rodents, it is measured by observing stereotyped facial responses to palatable and unpalatable foods, which are similar to those seen in human infants. It is not a tautology or circular reasoning to say that the reinforcing value or pleasantness of food influences food intake and body fatness. These are quantifiable concepts and as I will explain, their relationship with food intake and body fatness can be, and already has been, tested in a controlled manner.
1. Increasing the reward/palatability value of the diet should cause fat gain in animals and humans
The most highly rewarding and palatable rodent diet I know of is the "cafeteria diet", composed of human junk food. This diet is unmatched among solid-food diets in its ability to cause persistent overeating and rapid obesity in rodents, easily surpassing high-fat and high-sugar diets, although these also cause fat gain to a lesser degree (3). Rodents will voluntarily endure extreme cold or foot shocks to obtain this food, even when regular chow is freely available (4, 5).
Other diets that cause marked obesity in rodents, such as chocolate Ensure and certain purified high-fat diets, are highly preferred to normal chow. In my own experience, rats really really like the purified high-fat diet we feed them to make them obese. When you put a pellet of it in their cage, they will drop everything and devour it immediately, even if they just finished a meal of regular chow and even if it's in the middle of the night for them (the light cycle; they're nocturnal). If you drop a pellet of regular chow into their cage, they will generally ignore it unless they've been food deprived.
As I reported in a previous post, exposing humans to a similar "cafeteria diet" causes overeating and rapid fat gain as well (6).
Some people may object that these experiments are not properly controlled to determine if food reward/palatability are the critical factors, i.e., there are multiple differences between diet conditions. For the most part, that is true, and we will have to view these experiments as one imperfect brush stroke in the overall painting. However, there have been some efforts to specifically isolate the effects of reward/palatability on obesity in rodents. In one study, investigators added a variety of flavor-enhancing substances to standard rodent pellets in an attempt to manipulate the diet's palatability without affecting nutritional factors, and called this regimen the "isocafeteria diet" (7). Here's what they found:
It was demonstrated that the daily presentation of a new choice of the palatable foods which composed the “isocafeteria diet” led also to a sustained increase in food intake and to overweight. Variety and high palatability are per se sufficient factors to overcome regulatory mechanisms.This effect was partially replicated by a second study that showed higher body fat gain in rats fed a nutritionally controlled diet high in fat and sugar with a variety of palatable flavorings added, compared to the same diet without flavors, although adding palatable flavors did not promote fat gain in the context of a low-fat, low-sugar diet (8). Dr. Anthony Sclafani showed that manipulating the palatability of a sweet solution (Polycose; similar to maltodextrin) using non-nutritive flavoring agents influenced fat gain in rats exposed to it (9). In baboons, adding preferred but non-nutritive flavors to standard monkey chow increased food intake and body weight (10). The investigators concluded:
These results indicate that flavored chows may be useful for producing a nonhuman primate behavioral model of obesity and for inducing animals to eat otherwise unpalatable diets.Many human studies have shown that people eat more food at a sitting if the food is higher palatability than if it is lower palatability (11). This is true even if palatability is manipulated using substances that have little or no impact on the nutritional quality of the food, including saccharin (sweet), monosodium glutamate (savory) and herbs/spices.
2. Decreasing the reward/palatability of the diet should cause fat loss in animals and humans that carry excess fat
One of the most striking weight loss studies I've seen was conducted in 1965 and involved feeding a bland liquid diet through a dispensing straw (12). Lean and obese volunteers were instructed to eat as much of the liquid food as they wanted, but they were permitted no other food. While lean volunteers ate a normal amount of calories and maintained weight, obese volunteers dramatically reduced their spontaneous calorie intake and lost fat rapidly, with one man losing 200 lbs in 255 days without hunger. This is exactly what one would expect if unpalatable/unrewarding food lowered the biologically "defended" level of fat mass. Interestingly, the diet was high in sugar but was otherwise very low in palatability/reward value. Similar findings have been reported by other investigators (13).
Monotonous, bland liquid diets remain one of the most effective fat loss tools in clinical practice. Similarly, diets that reduce major reward factors without deliberately calling for calorie restriction, such as low-fat and low-carbohydrate diets, all cause fat loss even though in some cases the diet changes that are implemented diametrically oppose one another (14, 15). The further reward is lowered, the more effective the diet is for appetite suppression and weight loss.
Rodent studies are consistent with those in humans. Returning rodents made obese using palatable high-fat diets or chocolate Ensure (or by gavage overfeeding) to ad libitum ordinary rodent chow makes them lose most of the excess fat, although some of it is often retained (16, 17, 18).
None of these studies are sufficiently controlled to isolate reward and/or palatability, therefore they can not by themselves prove the hypothesis, but they are nevertheless consistent with it.
3. Individual sensitivity to food reward should predict future fat gain
I'm aware of three studies that have investigated this question. In the first, researchers found that the reinforcing value of food relative to a non-food stimulus predicted fat gain over the next year in 7-10 year old children (19). In the second, the responsiveness of reward-related brain regions to imagining palatable vs. unpalatable foods (as assessed using fMRI) predicted body mass index (BMI) gains in adolescent girls, and this effect was modified by gene polymorphisms in dopamine receptor genes (20). The third study also used fMRI to demonstrate that greater activation in reward-related brain regions during exposure to appetizing food cues predicted greater BMI gains over time in adolescent girls (21).
Consistent with other findings, these studies indicate that heightened food reward sensitivity is a pre-existing state that predisposes to fat gain over time in susceptible people.
4. Brain circuitry that controls reward and hedonic processing should interact with circuits that influence food intake and body fatness
The hypothalamus, which is the main homeostatic regulator of body fatness, and reward/hedonic centers, share extensive and reciprocal anatomical connections. One major pathway runs between the nucleus accumbens (NAc), which is critical to reward and hedonic processing, and the lateral hypothalamus (LH), which is critical both for reward/hedonic processing and body fat homeostasis (22). The role of the LH in feeding and reward is illustrated by the fact that if you implant an electrode into a rat's LH and allow the rat to self-stimulate, it will do so compulsively as if it were administering a drug; LH stimulation also potently increases food intake (23). Lesioning the LH causes a decrease in food intake and fat mass (24).
The LH shares reciprocal connections with other hypothalamic regions important for body fat homeostasis such as the arcuate nucleus (ARC), and thus is likely a node for the relay and integration of information from homeostasis and reward/hedonic systems. The ARC is a major CNS receptor site for circulating signals of energy status, particularly leptin but also including insulin, ghrelin, amylin, glucose and fatty acids (25). It contains neuron populations (POMC/CART and NPY/AgRP) that are critical for the control of food intake and body fat mass. Manipulating the activity of these neurons profoundly influences food intake and fat mass (26, 27).
Reward, hedonic and homeostatic systems are in a tightly coupled CNS circuit, suggesting that they may reciprocally influence one another. Indeed, it is well established that homeostasis centers influence reward and palatability circuits, and this is evident in the fact that food tastes better when you're hungry than when you're full ("hunger is the best sauce"). The next section will demonstrate that reward and hedonic circuits influence homeostasis circuits as well.
5. Manipulation of motivational and hedonic circuits in the brain (e.g., by lesion, drugs or genetic manipulation) should impact food intake and body fatness
There is a tremendous amount of data related to this question, and I don't intend this to be anywhere near a comprehensive discussion of it. Manipulating reward or hedonic circuits can have a profound impact on food intake and body fatness. One example is that inhibiting NAc shell neurons using a glutamate receptor antagonist (DNQX) causes voracious feeding (28). The mechanism involves the stimulation of feeding centers in the LH and ARC, demonstrating the functional importance of connections from reward/hedonic centers to hypothalamic centers regulating food intake and body fatness (29, 30). Similarly, chemically lesioning the NAc shell increases weight gain in rats (31).
Opiate signaling in the CNS is a critical component of hedonic processing. Researchers have long recognized the ability of opiates to increase feeding, and the ability of blocking opiate signaling to decrease feeding (31, 32, 33). Opiate signaling in the striatum (particularly the NAc) increases the palatability of food, and this effect depends mostly on the mu opioid receptor subtype (34, 35). The mu opioid receptor blocker naloxone specifically decreases the palatability of food and the consumption of palatable food in rats and humans, but has little impact on the intake of less palatable food (36, 37, 38). Mice that lack the mu opioid receptor gene are lean and resistant to diet-induced obesity, suggesting that opioid-mediated food palatability may be a major part of the reason rodents become obese on refined high-fat diets (39). Activating opioid signaling in the striatum does not increase the amount a rodent will work for food, does not increase meal frequency, and does not reinforce behavior, thus striatal opioids regulate hedonic processing (liking) without affecting reward (wanting) (40).
It has been demonstrated many times in the scientific literature and in my own hands that rats will eat a substantial amount of palatable food even after they have already eaten less palatable food to fullness (41). This is the "dessert effect"-- you've just finished a large meal and you're stuffed, but you manage to "make room" for another 200 calories when the cake rolls around. This seems to be related to opioid signaling, as the activation of this system in the striatum increases both the liking of foods and meal size (42).
Dopamine is critically involved in reward functions, and it turns out to be a major regulator of food intake and body fatness as well. Injecting dopamine into the brain reduces food intake, and the primary site of action appears to be in the LH (43). Dopamine signaling in the striatum regulates the motivation to obtain food, but not the enjoyment or consumption of food that is readily accessible (44). Thus, in the striatum it impacts motivation/reward (wanting) but not palatability (liking).
Drugs that activate certain dopamine receptors can cause compulsive overeating of palatable foods and weight gain (45). However, the dopamine system is complex, involving multiple receptor subtypes that have different functions in different brain regions, and different drugs affect these subtypes differently. The dopamine D2 receptor in particular seems to be involved in food intake and body fatness. The D2 receptor activating drug bromocriptine causes fat loss in a variety of animal models of obesity, including humans in some studies (46). Obesity is associated with reduced D2 receptor density in the striatum, and there is evidence that reduced striatal D2 receptors can precede the development of obesity (47). There is also evidence in rodents that chronic exposure to highly palatable food can reduce D2 receptor density in the striatum (48), suggesting that exposure to hyperpalatable foods, in addition to genetics (discussed below), can lead to an obesity-promoting pattern of D2 receptor expression in the brain. This is part of a broader pattern of addiction-like changes that develop in the brains of animals chronically exposed to hyper-palatable/hyper-rewarding food.
If the effects of opioid and dopamine signaling on palatability and reward truly influence food intake and body fatness, then drugs that impact these systems should be able to cause fat loss in humans. Such a drug exists-- it's called Contrave, and it is a combination of naltrexone (an opioid receptor blocker that acts mostly at mu and kappa receptors) and bupropion (dopamine-norepinephrine re-uptake inhibitor). Contrave has shown effectiveness as a fat loss drug in humans, although side effects have prevented FDA approval (49). The proposed mechanism of this drug is via the indirect activation of POMC cells in the ARC, again suggesting that reward/hedonic circuits influence those that regulate body fatness.
One of the most interesting links between reward/hedonic circuits and hypothalamic circuits is via endocannabinoid signaling. It has been known for some time that marijuana increases the consumption of palatable foods and causes weight gain if smoked frequently enough (50). Today, we know that endocannabinoid signaling is intimately involved in hedonic and reward circuitry. Marijuana acts principally via the CB1 receptor (51). Mice lacking the CB1 receptor gene CNR1 are lean and resistant to diet-induced obesity (52). The CB1 receptor blocker Rimonabant is an effective fat loss drug in humans, and its effect appears to occur mostly if not exclusively in the brain (53, 54). Rimonabant is basically "reverse marijuana". Perhaps not surprisingly, it was pulled from the market by the FDA due to negative psychological side effects.
It is no coincidence that two of the most effective fat loss drugs developed in recent history, Contrave and Rimonabant, both target reward and hedonic centers in the brain.
6. Genetic differences that influence reward and/or hedonic circuits should correlate with differences in body fatness
Each person on this earth carries a unique combination of gene variants, and in some cases these influence physical characteristics and the response to environmental factors. Some of these variants are within dopamine receptor genes, and thus have the potential to influence food reward. Several studies have shown that polymorphisms of the D2 dopamine receptor associate with the risk of drug addiction, gambling addiction, and obesity (55, 56, 57, 58, 59, 60). Polymorphisms in other genes that influence dopamine signaling, including the D4 dopamine receptor, the dopamine re-uptake transporter, and catechol-O-methyltransferase, have also been found to associate with obesity risk (61, 62, 63, 64).
Polymorphisms in the CB1 cannabinoid receptor gene CNR1 also associate with body fatness (65, 66). This is consistent with the fact that the CB1 receptor knockout mouse is resistant to diet-induced obesity, and the CB1 blocking drug Rimonabant reduces body fatness in humans (67).
I have not come across any reports of opioid receptor gene polymorphism frequency in obesity.
These findings support the hypothesis that reward and hedonic circuits in the CNS influence body fatness, and that the pre-existing characteristics of these circuits can explain differences in the response to rewarding stimuli, and differences in body fatness, in the general population.
7. On a cultural level, obesity prevalence should track with changes in the reward/palatability value of prevailing diet patterns
The evidence in this category is the least well controlled, since many things have changed during the development of the "obesity epidemic". However, for the hypothesis to be convincing, at a minimum it has to be consistent with how diets in most cultures have changed as obesity has increased in recent years.
The obesity epidemic in the United States has paralleled a significant shift in food culture. Here are a few examples to illustrate the trends (68). In the last 50 years:
- Fast food spending increased from 2% to 18% of total food spending
- Soda consumption increased by 250%
- Consumption of fresh potatoes decreased by 50%
- Consumption of processed potatoes (mostly fries and chips) increased by 300%
- Between-meal snacks/drinks, particularly processed snacks and soda, have more than doubled (69)
Similar trends are apparent in all affluent nations as the food chain has become increasingly commercialized and gradually displaced traditional food culture, and invariably this has been accompanied by increased body fatness on a population level.
That's not to say that commercial food is responsible for all obesity. Obesity has been documented for thousands of years, if not longer, including in certain cultures that only had access to traditional ingredients (e.g., Polynesian islanders). However, in these cases (including Europe before industrialization), obesity was mostly restricted to affluent individuals who benefited from skilled chefs and choice ingredients, as well as a sedentary lifestyle.
Some people have brought up the examples of France and India as challenges to the food reward hypothesis, stating that both have a tradition of delicious food. That, of course, is true. However, it's important to remember that most people traditionally didn't eat foie gras and fatty spiced curries every day, and food that you eat in a restaurant or as a tourist doesn't necessarily represent peoples' day-to-day food choices. In France, which I can speak for because I spent a significant chunk of my life there, most meals are composed of relatively simple, fresh, home-cooked food. This is particularly true for the older generations. The food is not low in fat, or low in animal fat. It tastes good, but it isn't extravagant. Traditionally, people rarely ate at restaurants, which were expensive and considered a special treat. As the food system has industrialized, and commercial food has increasingly replaced home cooking, the prevalence of obesity has increased.
I'm not as familiar with traditional Indian food, and I don't want to pose as an expert here, but I'm quite confident that what we see in restaurants and cookbooks in the US is not what the average Indian person was eating 50 years ago, and for the most part still not today. The traditional Indian diet, before industrialization, differed by region but was generally very low in fat and relied heavily on plain starch foods such as rice, millet and wheat. This was supplemented with more exciting dishes based on legumes, vegetables, dairy and occasionally meat. But my understanding is that, for an ordinary meal, these were small portions compared to the plain starches that formed the backbone of the diet. I doubt the average Indian person 50 years ago was making complex dishes using liberal quantities of ghee and a dozen spices for dinner every night, or deep frying things. Spices and ghee aren't free and cooking like that takes time, skill and special equipment-- is it a priority if you don't have much disposable income?
As India has industrialized, and the diet has become more diverse, more fatty, more refined and more commercialized, obesity throughout the country has increased (particularly abdominal obesity), and India now has a serious diabetes and coronary heart disease problem on its hands.
If you made it this far, you're a hero! As you have seen, a number of mutually buttressing lines of evidence, including controlled diet studies, observational studies, neuroscience, pharmacology and genetics, support the hypothesis that food reward and palatability influence food intake and body fatness. It's not every day that you come across an idea in the diet-health literature that is consistently supported by so many independent lines of inquiry. This hypothesis isn't very sexy, which is one reason why it hasn't gained as much traction with the public as certain other ideas, but I'm not too concerned about that. I just want to understand what causes obesity, and this hypothesis has a lot of explanatory power.
The degree to which food reward and palatability have driven the modern obesity epidemic remains open to interpretation. However, in 2011, the evidence has accumulated to the point where it is clear that they play a role. I side with modern food reward researchers in thinking it is a major factor, although there are certainly others.