Although there is little evidence in humans of what foods may be addictive, animal models suggest that highly processed foods are associated with addictive-like eating. Rats with a propensity towards binge eating exhibit addictive-like behavior in response to highly processed foods, such as Oreo Double Stuf cookies or frosting, but not to their typical chow [28,29]. Rats maintained on a diet of highly processed foods, such as cheesecake, exhibit downregulation in the dopamine system that also occurs in response to drugs of abuse . Further, rats are motivated to seek out highly processed foods despite negative consequences (foot shock), which is another feature of an addiction . Therefore, at least in animal models, overconsumption of highly processed foods, but not standard rat chow, appears to produce some addictive-like characteristics. This reinforces the idea that not all foods are likely to be equally associated with addictive-like eating behaviors.
Animal research has also investigated whether food attributes typically added to highly processed foods, such as sugar and fat, are particularly implicated in “food addiction.” In animals, it appears that sugar may be most associated with addictive-like eating . Rats given intermittent access to sugar in their diet exhibit a number of behavioral indicators of addiction, such as binge consumption, tolerance, and cross-sensitization to other drugs of abuse . When the sugar is removed from the diet or when an opiate antagonist is administered, rats experience signs of opiate-like withdrawal, such anxiety, teeth chattering, and aggression [33–35]. Sugar bingeing has been shown to increase mu-opioid receptor binding  in a similar manner to drugs of abuse [37,38]. Bingeing on sucrose produces a repeated increase of dopamine, rather than the gradual decline over time, which is a hallmark of addictive substances [39,40]. Thus, behavioral and biological evidence in animal models suggest that sugar may be an addictive agent in highly palatable foods.
The prevalence of obesity in the United States continues to increase, with more than 85% of adults projected to be overweight or obese by 2030 . Health care costs associated with obesity currently comprise almost 10% of national health care expenditures  and are projected to increase to 15% over the next 15 years . There has been little success at preventing excessive weight gain or developing weight loss treatments that are effective long-term . Multiple causes contribute to the obesity epidemic, such as increased energy intake, increased availability and ease of access to foods, larger portion sizes, and decreased physical activity [4–6]. Although the causes of obesity are multifactorial, one potential contributing factor is the idea certain foods may be capable of triggering an addictive response in some individuals, which may lead to unintended overeating.
Gearhardt et al.  developed and validated the Yale Food Addiction Scale (YFAS), which uses DSM-IV criteria for substance dependence to quantify symptoms of addictive-like eating (see Table 1). “Food addiction” is characterized by symptoms such as loss of control over consumption, continued use despite negative consequences, and an inability to cut down despite the desire to do so . Addictive-like eating has been associated with increased impulsivity and emotional reactivity, which are similarly implicated in substance-use disorders . Thus, “food addiction” may share common behavioral attributes with other addictive disorders. Neuroimaging studies have also revealed biological similarities in patterns of reward-related dysfunction between “food addicts” and substance-dependent individuals. Individuals endorsing symptoms of “food addiction” exhibit increased activation in reward-related regions (e.g., striatum, medial orbitofrontal cortex) in response to food cues, consistent with other addictive disorders . Further, higher scores on the YFAS have been associated with a composite genetic index of dopamine signaling . This multilocus genetic profile has been related to dopamine signaling capacity, which may also be a risk factor for addictive disorders [12,13].
However, rats bingeing on sugar do not experience an increase in body weight . Thus, fat may also be an important food attribute for addictive-like eating, but through different mechanisms. Bingeing on fat-rich foods (e.g. shortening) is associated with an increase in body weight but may not result in opiate-like withdrawal symptoms . One explanation is that fat may alter effects on the opioid system or enhance the palatability of the food [38,39]. Interestingly, when rats binge on highly processed foods high in both sugar and fat, they experience changes in the dopamine system akin to drugs of abuse but do not exhibit signs of opiate-like withdrawal . This suggests that sugar and fat may both play important, yet distinct, roles in the addictive potential of highly processed foods.
Little is known about how these food characteristics might result in addictive-like eating in humans. Given the findings in animals, highly processed foods may be more likely to be consumed in an addictive manner. For drugs of abuse, processing may increase a substance’s addictive potential (e.g. processing grapes into wine) by elevating the dose, or concentration, of the addictive agent and expediting its rate of absorption into the bloodstream. Applying this logic to food attributes, it may follow that refined carbohydrates (e.g., sugar, white flour) and fat are important contributors to addictive-like eating. However, it is not just the presence of these nutrients, as they also appear in naturally occurring foods. Rather, the addictive potential of a food is likely to increase if the food is highly processed to increase the amount, or dose, of fat and/or refined carbohydrates and if the refined carbohydrates are absorbed into the bloodstream quickly (high GL). An essential next step in the consideration of “food addiction” is to determine which foods or food attributes pose the greatest risk in the development of addictive-like eating behaviors in humans.