Harnessing the power of disgust: a randomized trial to reduce high-calorie food appeal through implicit priming

Abstract

Background: In our increasingly obesogenic environment, in which high-calorie convenience foods are readily available, food choices can drastically affect weight and overall health. Learned food preferences, which are developed through repeated pairings with positively and negatively valenced stimuli, can contribute to obesity susceptibility if positive attitudes toward high-calorie foods are developed. Thus, the modification of automatic associations with food may be a viable strategy to promote healthier eating behaviors.

Objective: In this study, we investigated the ability of an implicit priming (IP) intervention to alter responses to visual food cues by using an evaluative conditioning approach. The main objective was to implicitly (i.e., below conscious perception) associate disgust with high-calorie foods with the aim of reducing liking of these foods.

Design: Participants were randomly assigned to active or control IP. In active IP (n = 22), high-calorie food images were implicitly primed with negatively valenced images, and low-calorie food images were implicitly primed with positively valenced images. In control IP (n = 20), all food images were primed with neutral images of fixation crosses. Food images were rated on the desire to eat immediately before and after IP.

Results: A significant main effect of calorie (high compared with low; P < 0.001) and a significant calorie-by-group (active compared with control) interaction (P = 0.025) were observed. Post hoc tests identified a significantly greater high-calorie rating decline after active IP than after control IP (P = 0.036). Furthermore, there was significantly greater change in high-calorie ratings than in low-calorie ratings in the active group (P = 0.001). Active IP effects extended to high-calorie foods not specifically included in the intervention, which suggested an effect generalization. Moreover, a greater change in high-calorie ratings than in low-calorie ratings persisted 3-5 d after active IP (P < 0.007), which suggested lasting effects.

Conclusion: This study provides initial evidence that IP can be used to alter high-calorie food preferences, which could promote healthier eating habits.

Trial registration: ClinicalTrials.gov NCT02347527.

Keywords: disgust; food preferences; implicit priming; nutrition; obesity.

© 2015 American Society for Nutrition.

Figures

FIGURE 1

Study flow diagram. IP, implicit priming.

FIGURE 2

Design of the implicit priming intervention in which a fixation cross on a black screen was shown for 500 ms and followed by an implicit prime presented for ∼20 ms. Following the implicit prime image, a food image was presented for 4 s. In the active condition, the implicit prime was either a positively valenced image paired with a subsequent low-calorie food image or a negatively valenced image paired with a subsequent high-calorie food image (as shown in figure). In the control condition, the implicit prime paired with both low- and high-calorie food images was a neutral image of 3 fixation crosses. Images shown are representative of those used in the study, although not actually from the International Affective Picture System database as per the International Affective Picture System use agreement. Images shown were modified from public domain images that are freely available under Creative Commons deed CC0.

FIGURE 3

Mean (±SEM) changes in high-calorie food ratings and low-calorie food ratings from pre-intervention to postintervention for active (n = 22) and control (n = 20) groups. A repeated-measures ANOVA identified a significant main effect of calorie with a greater change in high-calorie ratings compared with low-calorie ratings (P < 0.001), and a significant group (active compared with control) × calorie (high compared with low) interaction (P = 0.025). *Greater high-calorie change in the active group than in the control group, P = 0.036 (post hoc t test); **greater high-calorie change than low-calorie change within the active group, P = 0.001 (post hoc t test).

FIGURE 4

Mean (±SEM) changes in high-calorie food ratings and low-calorie food ratings separately for primed (i.e., included in the priming paradigm) and novel (i.e., not included in the priming paradigm) food images for active (n = 22) and control (n = 20) groups. A repeated-measures ANOVA identified a significant main effect of calorie with a greater change in high-calorie ratings than with low-calorie ratings (P = 0.001) and a significant group (active compared with control) × calorie (high compared with low) interaction (P = 0.017). *Greater high-calorie change than low-calorie change within the active group in primed images, P = 0.001 (post hoc t test); **greater high-calorie change than low-calorie change within the active group in novel images, P = 0.008 (post hoc t test).

FIGURE 5

Mean (±SEM) changes in high-calorie food ratings and low-calorie food ratings for the active group measured immediately after the intervention (Acute Effect; n = 22) and 3–5 d after the intervention (3-5 Day Retest; n = 20). A significant main effect of calorie was observed (P = 0.001), but effects of measurement timing (acute compared with retest) were not significant (P = 0.074), and there was no significant interaction between calories (high compared with low) and measurement timing (P = 0.225). *Significant difference between high-calorie acute change compared with low-calorie acute change, P = 0.001 (post hoc paired samples t test); **significant difference between high-calorie retest change compared with low-calorie retest change, **P = 0.007 (post hoc paired samples t test).