What’s the first treat you pull out of your candy bag on Halloween? Probably your favorite guilty pleasure… but what if you could use pennies to train yourself to pick a candy you might not like the most but that might be healthier?
A new study finds that we may be able to train our brains to make healthier food choices. For the first time, researchers have shown that consistently rewarding people for choosing less desirable food makes people more likely to choose that food again in the future, even with no promise of reward. In other words, the training got people to pass up on the foods they wanted most.
“This kind of training to choose less-preferred items has not been used before,” says Tom Schonberg of the University of Texas, Austin, who with colleagues in the lab of Russell Poldrack, published their results in the Journal of Cognitive Neuroscience (JoCN). Schonberg’s team created personalized pairs of junk food items by asking individual participants to bid on items through an auction. The auction determined the value of each food item specific to each participant – for example, one person may prefer Doritos over M&Ms, while another may prefer the opposite.
The researchers then extensively trained participants to choose the lower value item. In multiple computer exercises, the participants were only rewarded – with 1 cent for each choice – when they chose the lower value item. At a later time, those trained chose to eat their less-desired foods, even when no monetary reward was offered.
In a replication of the results with another set of participants, the researchers scanned the participants in the fMRI to see how the training affected neural processes. As training progressed, the researchers found decreased recruitment of left dorsolateral pFC and bilateral parietal cortices – regions that previously have been associated with cognitive control. They also found increased brain connectivity for motor regions, suggesting that the training led to more habitual, automated responses.
CNS talked with Schonberg about this novel experiment and related work to better understand how we make food choices and what interventions can help change that.
Schonberg: Understanding human habitual behavior has always been one of my research goals – how long it takes for a certain behavior to become habitual and how we can change old habits. The work presented in the JoCN paper was conducted as part of a funded grant by the NIH that focuses on overcoming the persistence of habits to maintain behavioral change.
CNS: Can you describe briefly what we know about how the brain processes food choice?
Schonberg: We know that food choices are based on the value we place on different food options, much like for every other object. These values can come from many different factors of foods and are computed by the value-based decision making network in the brain. For example, brain areas that are active while choosing between healthy and non-healthy food items are similarly involved in decisions between short vs. long-term goals.
Studies by other labs have focused on measuring the brain response to processing food items. These studies involve the hedonic and metabolic responses to consumed food items. Hedonic feelings elicited by food contribute to the value assigned to it.
CNS: Were you surprised by any of the results?
Schonberg: The first time we ran this study we were surprised that training participants to choose items they like less (by reinforcing them with 1 cent per choice), later translated to more choices of these items when made for consumption and involved no monetary reinforcement.
This choice preference shift is achieved with less than one hour of training. Prolonged training over multiple days would potentially increase this preference shift.
CNS: How did you choose the junk food items determine their “value”?
Schonberg: We chose 60 snack food items from a database created by Antonio Rangel’s group at the California Institute of Technology because they have been validated in a number of previous studies. To determine the individual value of each snack item we asked participants to fast 4 hours prior to the experiment. Then we used an auction procedure, devised by economists in the 1960s and also used by Dr. Rangel’s group, in which participants receive $3 to express their individual willingness-to-pay for each item.
Participants know that at the end of the experiment there will be an auction and only if their bid on a randomly chosen item is higher than a randomly generated number can they buy this item – just like an auction in the real world. Taking into account that they are hungry, this procedure has been shown to produce reliable estimation of their “value” for these items.
CNS: What is the significance of the reduced recruitment of left dorsolateral pFC and bilateral parietal cortices in the trained individuals?
Schonberg: We believe our study has two main important findings: First, it shows that there is a reduction in this network with repeated choices of the reinforced items. This is consistent with the idea that practice reduces the need for top-down control over choices.
Second, we did not see other regions increasing their activity during training. However, we did find that there was a specific increase in functional connectivity between the left dorsolateral pFC and lower level motor and visual regions as training progressed – suggesting a transfer to more “habitual-like” responding.
CNS: How might these processes play out in real life? Can we do anything now to “train” ourselves for healthier eating?
Schonberg: Several efforts have been developed and aim to change food preferences – our approach has not been tried yet. I envision someone paying themselves a small amount after every successful choice. The key here is to try and maintain the same choices repeatedly over time until the value of the less preferred item changes.
In our study we did not use less preferred healthy items but there is no reason why it should not work the same way, as we prefer some healthy foods over others. More steps need to be taken before this approach can be suggested as a real-world intervention method. Science moves in baby steps….
CNS: What distinguishes this study from past studies in the area? Can you speak to how your study connects to a recent one by Antonio Rangel’s group in the Journal of Neuroscience?
Schonberg: We aim to induce behavioral change without any reliance on effortful self-control as most dieting plans do and fail. The Rangel group’s study presented a novel manipulation and very interesting results by using the emotion regulation technique to change preferences.
I think that at the end of the day, in order to achieve sustained behavioral change, we will need to combine different strategies. It will be interesting to think of a study that can address all these different mechanisms at the same time.
CNS: What is next for this work?
Schonberg: Our most current research focuses on even more automatic mechanisms for behavioral change. We are now using non-reinforced training to influence preferences.
The paper, “Influencing Food Choices by Training: Evidence for Modulation of Frontoparietal Control Signals,” by Tom Schonberg, Akram Bakkour, Ashleigh M. Hover, Jeanette A. Mumford, and Russell A. Poldrack, was published online in the Journal of Cognitive Neuroscience on October 11, 2013.
Media contact: Lisa M.P. Munoz, CNS Public Information Officer, email@example.com