From attributes to value: Neural correlates of a front-of-package label on food decision-making - An fMRI study

Abstract

Diet-related chronic illnesses, such as obesity and diabetes, pose major global health challenges. To promote healthier choices, policymakers have introduced nudges like front-of-package labels such as the Nutri-Score, which simplify nutritional information. While previous research has examined the impact of front-of-package labels on product valuation via reward and cognitive control pathways, the neural mechanisms underlying attribute-specific changes in food perception (e.g., healthiness, tastiness) remain unclear. In this study, the impact of a Nutri-Score derived color-coded frame on the decision domains , perceived healthiness, perceived tastiness, and willingness-to-pay as a proxy for subjective value, was investigated. Forty healthy participants (28 females, age: M = 23.8 years, SD = 3.1 years) rated 63 food products in two conditions: without (control) and with (treatment) a color-coded frame while undergoing functional magnetic resonance imaging. The investigation focused on how the color-coded frame alters all three decision domains on a behavioral but also on the neural level. Additionally, the variation in neural and behavioral responses depending on the frame color was investigated. Overall, our results show that the color-coded frame significantly influenced WTP, healthiness, and tastiness ratings, with corresponding increases in neural activation in reward-related (ventromedial prefrontal cortex), cognitive control (dorsolateral prefrontal cortex), and homeostatic (thalamus) regions. Healthiness perception involved regions associated with value-based processing (ventromedial prefrontal cortex), as well as cognitive control (dorsolateral prefrontal cortex) when the color-coded frame was present, whereas tastiness perception engaged a network including the insula, brain regions related to reward valuation (ventromedial prefrontal cortex), and regions associated with cognitive control. These findings suggest that the color-coded frame, used here as a proxy for the Nutri-Score, influences food choices by modulating attribute-specific valuation and interacting with homeostatic mechanisms, rather than directly altering overall value computation. Understanding how external cues integrate with internal physiological signals to guide food choices may inform interventions aimed at improving dietary behavior and public health outcomes.

Fig 1. Experimental design.

A) The figure shows one control condition trial and one treatment condition trial. All control condition trials were presented first, followed by the treatment condition trials in the same order. Each trial started with a fixation dot that lasts between 1500 and 2500 milliseconds (ms), with the dot positioned at eye-level, as is the decision-domain-icon. Thereafter, a decision screen was displayed for up to 4000 ms. Participants were instructed to make a decision according to the respective criteria by choosing one of six options via key press within the time frame. Once a decision was made, the chosen option was highlighted with a thicker frame for 300 ms. If participants exceeded the four-second time range, a screen displaying the text “You were too slow.” appeared for 1000 ms before the next trial began. The rating task involved assessing 63 different products in a pseudo-randomized order. Each product was rated once in the no-frame condition and once in the frame condition, resulting in 189 trials for each condition. B) Each decision screen consisted of a product, a decision-domain-icon representing the ratings (WTP, healthiness, and tastiness) and the possible responses. Products were surrounded by a colored frame in green, yellow, or red, depending on their actual Nutri-Score. Due to copyright reasons products are edited in this figure and the original items are not shown.

Fig 2. Behavioral analysis of the rating task.

Violin plots presenting median and interquartile range for WTP (mean €), average healthiness ratings (-3(very unhealthy) to 3(very healthy)), and average tastiness ratings (-3(very untasty) to 3(very tasty)) in both the control condition and the treatment condition. Behavioral differences were analyzed using a paired t-test with Bonferroni correction. ‘***’ indicates p < .001, and ‘ns’ denotes not significant. No p-values between .05 and .001 were observed.

Fig 3. Conjunction analysis of WTP, healthiness, and tastiness; treatment > control: T-map is projected onto an MNI_152_2009 template, p -threshold of <.001 and FWE cluster correction at p < .05 (df = [1, 39]).

Abbreviations in the figure: dlPFC = dorsolateral prefrontal cortex (BA46), aPFC = anterior prefrontal gyrus (BA10), AG = angular gyrus (BA39), MTG = medial temporal gyrus (BA21).

Fig 4. Activation clusters derived from green treatment > green control condition for WTP (orange overlay), healthiness (green overlay), and tastiness (blue overlay): All t-maps are projected onto an MNI_152_2009 template, p < .001 and FWE cluster correction at p < .05 (df = [1, 39]) in the sequence first healthiness contrast, followed by tastiness contrast and then WTP contrast.

Abbreviations in the figure: ACC = anterior cingulate cortex (BA32), aPFC = anterior prefrontal cortex (BA10), dlPFC = dorsolateral prefrontal cortex (BA9), STG = superior temporal gyrus (BA22), vmPFC = ventromedial prefrontal cortex (BA11).

Fig 5. Activation clusters derived from yellow treatment > yellow control condition for WTP (orange overlay), healthiness (green overlay), and tastiness (blue overlay): All t-maps are projected onto an MNI_152_2009 template, p < .001 and FWE cluster correction at p < .05 (df = [1, 39]) in the sequence first healthiness contrast, followed by tastiness contrast and then WTP contrast.

Abbreviations in the figure: ACC = anterior cingulate cortex (BA32), aPFC = anterior prefrontal cortex (BA10), dlPFC = dorsolateral prefrontal cortex (BA46), IFG = inferior frontal gyrus (BA47), STG = superior temporal gyrus (BA22), vmPFC = ventromedial prefrontal cortex (BA11).

Fig 6. Activation clusters derived from red treatment > red control condition for WTP (orange overlay), healthiness (green overlay), and tastiness (blue overlay): All t-maps are projected onto an MNI_152_2009 template in the sequence first healthiness contrast, followed by tastiness contrast and then WTP contrast.

A p -threshold of <.001 and FWE cluster correction at p < .05 (df = [1, 39]) was used. Abbreviations in the figure: aPFC = anterior prefrontal cortex (BA10), dCa = dorsal caudate, dlPFC = dorsolateral prefrontal cortex (BA9), IFG = inferior frontal gyrus (BA47), STG = superiortemporal gyrus (BA22), vmPFC = ventromedial prefrontal cortex (BA11).