Reward Systems Provide Insight into Obesity Epidemic

A coronal section showing amygdala responses to the predictive food odor (FO+) compared to the non-predictive food odor (FO-) and compared to the receipt of the caloric (drink) and non caloric (tasteless) solutions.

From Pavlov’s legendary experiment, we learned that when a dog hears a bell and is given a piece of steak simultaneously over and over, it eventually begins to salivate even if the bell is rung without the steak. Yet how can the results of this 100-yearold reward-learning experiment, shed light on the current obesity epidemic?

Dr. Dana Small, Assistant Professor of Psychiatry and associate fellow of the John B. Pierce Laboratory, probed the answer to this question by investigating the reward systems in humans and attempting to understand how these could impact human behavior, such as overeating.

In general terms, simple conditioning involves conditioning a neutral stimulus, one normally unrelated to a response, such that it eventually elicits a behavioral response, such as salivation. Functional MRI (fMRI) studies show that the amygdala, an area responsible for memory and emotion, is activated in simple conditioning.

However, the pertinent question for obesity research is whether the amygdala is activated due to affective conditioning (meaning that in Pavlov’s case, the bell itself develops a rewarding value) or due to a predictive signal (meaning that the prediction of steak is what is most rewarding).

Small tested this question by delivering food aromas that either predicted receipt of an associated drink (i.e. smell chocolate and receive chocolate milkshake) or receipt of a tasteless, non-caloric solution (i.e. smell strawberry and receive the tasteless solution).

This experimental design was possible because engineers at the Pierce Labs developed a custom-built olfactometer and gustometer, which enabled delivery of odors and liquids to subjects in the scanner, while keeping temperature, concentration, and pressure constant. Subjects learned the association between the odors and drinks but no affective conditioning occurred, likely because the odors were already well-learned, pleasant food cues.

The results of the experiments demonstrated that the amygdala is preferentially activated in response to predictive food odor, compared to the equally pleasant non-predictive food odor and compared to the receipt of the drink.

Other regions such as the insula and orbitofrontal cortex responded to aromas and caloric drinks. These findings are important for two reasons. First, they indicate that the amygdala encodes the predictive, and not the affective, signal. Second, they show that in humans, there are separable substrates encoding food sensations related to anticipation and consumption.

This latter finding has implications for understanding the obesity epidemic. For example, according to Small, “it may be the case that slight alterations in one signal but not another may lead to overeating. We know that this is likely the case for drug addiction and highly appetitive foods, such as chocolate, which lead to very similar brain responses to addictive drugs like cocaine. It is therefore not unreasonable to suppose similar processes underlie compulsive food and drug taking.”

Currently, Small is conducting tests to see how these separable reward signals differ as a function of body weight, dopaminergic genotypes, smoking status, and behavioral phenotypes.

She hopes that the findings will give us a better understanding of how individual differences in the neurophysiology of reward may contribute to the obesity epidemic, and possibly even present ways through which we can subdue this oversensitive reward system behaviorally.

Small, D. M., Veldhuizen, M. G., Felsted, J., Mak, E., & McGlone, F. (2008, March). Separable Substrates for Anticipatory and Consummatory Food Chemosensation. Neuron, 786-797.