Mapping fMRI responses to combinations of foods in the human brain

890 

Abstract Submission 

Hui-Kuan Chung¹, Philippe Tobler²

¹University of Zurich, Zurich, Zurich, ²University of Zurich, Zurich, Switzerland

Hui-Kuan Chung  
University of Zurich
Zurich, Zurich

Philippe Tobler  
University of Zurich
Zurich, Switzerland

Overweight is on the rise world-wide. One possibly contributing factor is that the pleasure derived from specific food combinations leads to more consumption than when the constituents are consumed separately. The classic pairing of beer and chips serves as a prime example. People often find it more satisfying to enjoy chips with beer than consuming them separately. These food combinations are known as complementary goods, where indulging in one often triggers the desire for the other. In contrast, substitute goods are interchangeable to some degree, like regular beer and low-calorie beer, and providing one typically reduces desire for the other. These phenomena underscore the importance of investigating the valuation process for combinations of goods, but we know still relatively little about the conditional valuation of the different relationships between foods. The present study investigates these relationships at both the behavioral and the neural level.

Thirty-eight healthy volunteers (22 women; BMI range = 18.79-26.47), participated in this study. Participants provided willingness to pay (WTP) bids for the same food items, sometimes in isolation (single condition) and sometimes in combination with other foods (paired condition) inside the MRI scanner. In the paired condition, participants were given a free food on top of the food item they were bidding for. This novel design allowed us to investigate how WTP differed between single and paired conditions (ΔWTP= WTPaired - WTPsingle) while controlling for the intrinsic value of the food item. We predicted that the valuation of a particular food item would change as a function of which other items it was paired with.
To identify the brain regions decoding different food relationships, we conducted a multi-voxel pattern analysis (MVPA) on the fMRI data. In the MVPA analysis. Specifically, the support vector machine classifier was trained to distinguish between high (i.e., complementary goods) versus low (i.e., substituting goods) ΔWTP. Additionally, the identified region was used as a seed region in a psychophysiological interaction (PPI) analysis, with high versus low ΔWTP serving as the psychological factor.

As predicted, ΔWTP increased when the combined foods were rated as more complementary but decreased when combined foods were rated as more substituting. These findings indicate that complementary and substituting relations between foods differentially affect valuation. Moreover, regression analyses revealed a positive relationship between increased combined valuation of an item and the self-reported frequency with which participants consumed the two items together.
Neurally, we observed that the occipital lobe, cerebellum, and hippocampus were involved in processing the value change related to combining foods (ΔWTP; whole-brain FWE <0.05). Furthermore, the hippocampus exhibited stronger functional connectivity with the putamen when foods were more complementary compared to when they were more substituting (whole-brain uncorrected <0.001).

Together, our behavioral and neural findings are compatible with the notion that memory contributes to the enhancement of value when goods are more complementary. By recognizing the role of memory in processing the valuation of combined foods, it paves the way for interventions aimed at promoting balanced and nutritious eating habits.

Decision Making ¹

Activation (eg. BOLD task-fMRI)

Classification and Predictive Modeling ²

BOLD fMRI

FUNCTIONAL MRI

Multivariate