Why we keep craving food even when we are full

Even when people feel full, images of food may continue to activate reward signals in the brain. A new neuroscience study suggests that learned responses to food cues could help explain why we sometimes eat despite not being hungry.

Study: Devaluation insensitivity of event related potentials associated with food cues. Image credit: Pixel-Shot/Shutterstock.com

The obesity epidemic is a major public health concern. Urgent research is focused on identifying modifiable risk factors. One such factor is eating when one is not hungry. This could indicate a failure of appetite regulatory mechanisms, driven in part by the ubiquitous presence of food cues in the modern world. A recent paper published in the journal Appetite examines this phenomenon by investigating how learned responses to food cues may contribute to eating in the absence of hunger.

How food cues may trigger eating without hunger

The researchers sought to explore the role played by food cues in overeating. Normally, such cues activate learned reward-related responses, resulting in eating activity. Conversely, the motivational value of food typically declines once satiety is reached, reducing the drive to eat. When this switch-off does not occur, it is termed devaluation insensitivity and may indicate that the individual has learned to eat without feeling hunger.

Eating is a response to multiple triggers: negative energy balance, feelings of hunger, the known palatability of a presented food, eating history, a low effort-to-reward ratio, and a rich mix of cultural factors.

Physiologically, ghrelin is a hormone released when blood sugar levels are low, triggering feelings of hunger. This motivates eating. Conversely, hormones such as leptin help signal satiety and can reduce neural responses to food stimuli. As satiety signals increase, the motivational value of food as a reward typically decreases. The neural response to food cues is accordingly reduced, called neural devaluation sensitivity.

However, this is not a failsafe mechanism, as seen by the obesity epidemic now raging. Epidemiological evidence suggests a major role for overeating in the current rise in obesity prevalence.

Food cues do prompt eating behavior. Children with high body weight overeat when exposed to such cues, and obese children recognize more food advertisements than non-obese children. Images of calorie-rich food trigger event-related potentials (ERPs) during brain tasks, even when they are incidental to the task. Food cues are conditioned stimuli associated with the food they represent. Once learned, the association elicits feeding behavior irrespective of satiety. This is termed external eating, as external cues trigger the action of eating.

The psychological theory of devaluation was used here to examine the cognitive processes that act between the food cue and the final behavioral outcome. Devaluation describes the loss of the previous value of a food after, for example, eating one’s fill of it. Its presence discriminates between goal-directed eating and habitual eating.

With goal-directed eating, the individual is motivated to eat because of hunger and stops eating once sated. In contrast, habitual eating is based on the reward experienced from eating in response to food cues when hungry, which reinforces the food cue-eating association. If powerful enough, this can disable the intrinsic motivational value of food and lead to automatic eating in response to food cues even when sated.

However, whether devaluation insensitivity reliably reflects habitual behavior in humans remains debated, and the concept has not been fully validated in human studies.

EEG experiment tests brain responses to food cues

The study used EEG to measure brain responses to food cues. Researchers recorded electrical signals from the scalp for up to 700 milliseconds after each cue to track reward-related activity while participants completed a two-step learning task in which food rewards were delivered with changing probabilities.

The study involved 90 university students aged 30 years or younger. Participants were screened to exclude those dieting, with eating disorders, neurological disorders, or body mass index outside a specified range. For each participant, two foods (one savory and one sweet) were selected from a pool of 11. Both foods were desirable to the participant, though not at the top of the scale. This was designed to encourage satiety without physically filling the participant up.

One of these was designated as the devalued food to be served midway through the task. Success in the task (in which images of the chosen foods were shown) was claimed to increase their chances of obtaining both foods, though only the devalued food would actually be served.

After the learning phases of the task, the participants were given a meal of the devalued food until they signaled they were sated. In the next phase, they repeated the task, but this time, rather than choose the devalued food, they could choose to win money instead. Finally, they were given the devalued food to assess their consumption, concluding the task.

Brain reward signals persist despite satiety and devaluation

After eating the food depicted in food cues to satiety, reward-associated ERPs elicited by those cues were expected to decrease. If not, this indicates insensitivity of the neural response to devaluation.

During the two learning phases of the task, valued and devalued foods were chosen 55 % and 65 % of the time (valued) versus 53 % and 65 % of the time (devalued), exceeding simple chance. After the meal, behavioral devaluation was observed. The devalued food was successfully chosen only 47 % of the time, compared to 56 % for the valued food.

The mean rating of the foods chosen by each participant also varied between pre- and post-meal ratings. The valued food remained highly rated, but the devalued food dropped significantly in rating after the meal.

The devaluation was only partial, however, since the sated participants did not consistently choose money rather than the now-devalued food. This might suggest behavioral devaluation insensitivity in some participants but not in others.

In sharp contrast to the significant behavioral and subjective devaluation of the chosen foods over the course of the trial, the ERPs showed no evidence of neural devaluation, even in reward-related EEG signals.

The authors suggest that this neural-behavioral dissociation may reflect habitual or model-free neural responses that persist even when behavior changes, although the study cannot conclusively demonstrate that a habit has been acquired.

This neural finding contradicts earlier studies, probably because of different methodologies. Importantly, the current study could not prove that an overeating habit had actually been acquired. Since no aversive outcome was provided, it is difficult to confirm that reward signals were indeed captured rather than signals of motivational salience rather than reward value.

Because EEG responses can reflect the general importance or attention-grabbing nature of a stimulus, rather than its reward value alone, interpreting these signals remains challenging. In addition, the experimental task used to separate goal-directed and habitual learning remains debated in the literature, which further limits strong conclusions about the mechanisms involved.

Food cues may override satiety through learned neural responses

Despite the limitations of the study, the absence of any neural devaluation in the EEG after presenting food cues, in the presence of behavioral and subjective devaluation effects, suggests that overeating may partly involve learned neural responses to food cues that persist even after satiety, with the regulation or inhibition of eating likely relying on later goal-directed control processes in the brain.

Download your PDF copy by clicking here.