Developmental Changes in Food Perception and Preference

Abstract

Food choices are a key determinant of dietary intake, with brain regions, such as the mesolimbic and prefrontal cortex maturing at differential rates into adulthood. More needs to be understood about developmental changes in healthy and unhealthy food perceptions and preference. We investigated how food perceptions and preference vary as a function of age and how food attributes (taste and health) impact age-related changes. One hundred thirty-nine participants (8-23 years, 60 females) completed computerized tasks to rate high-calorie and low-calorie food cues for taste, health, and liking (preference), followed by 100 binary food choices based on each participant's ratings. Dietary self-control was considered successful when the healthier (vs. tastier) food was chosen. Self-control success ratio was the proportion of success trials over total number of choices. Beta-weights for health (β-health) and taste (β-taste) were calculated as each attribute's influence on food preference. Adiposity measurements included BMI z-score and waist-to-height ratio (WHtR). High-calorie foods were rated more tasty and less healthy with increasing age. Older participants liked high-calorie foods more (vs. younger participants), and β-taste was associated with age. Significant age-by-WHtR interactions were observed for health and taste ratings of high-calorie foods, β-taste, and marginally for preference of high-calorie foods. Stratifying by WHtR (high, low), we found age-related increases in taste and preference ratings of high-calorie foods in the high WHtR group alone. In contrast, age-related decreases in health ratings of high-calorie foods were significant in the low WHtR group alone. Age and β-taste were significantly associated in the high WHtR group and only marginally significant with low WHtR. Although participants rated low-calorie foods as less tasty and less healthy with increasing age, there was no association between age and preference for low-calorie foods. Participants made faster food choices with increasing age regardless of WHtR, with a significant age-by-WHtR interaction on reaction time (RT). There were no age-related effects in self-control success ratio and β-health. These results suggest that individual differences in age and central adiposity play an important role in preference for high-calorie foods, and a higher importance of food tastiness in food choice may contribute to greater preference for high-calorie foods with increasing age.

Keywords: adolescents; children; dietary self-control; eating behavior; food choice; pediatric obesity.

FIGURE 1

Food ratings. Participants rated 60 foods on taste, health and preference, using a 5-point scale with emoticons and words, by pushing buttons on the keyboard. Block and stimulus order were randomized across participants.

FIGURE 2

Food-choice mouse-tracking. Participants were asked to choose the food item they would like to eat. They completed several practice trials prior to starting the task. As shown in (A), participants clicked on the Start button, after which there was a blank screen of random duration (200–500 ms), and then the cursor would reappear in the bottom center of the screen and the food cues appeared on the screen, one at the left upper corner and one at the right upper corner of the screen. Participants moved the computer mouse to select a response and each mouse trajectory was recorded. Once the computer mouse entered the box containing the food cue, the trial ended, without the need to click. Trajectories were not visible to the participant. Trials were separated by a fixation cross of random duration (400–700 ms). One hundred binary choices were presented to each participant based on their individual ratings for taste and health from the food ratings task. Two representative mouse paths for Subject 1 are shown in (B) for trials on which the left-hand food item and the right-hand food item, respectively, were selected.

FIGURE 3

Scatterplots show ratings of 30 high-calorie and 30 low-calorie food cues by age, classified by WHtR status (categorized as high or low based on a median split for WHtR of 0.51). Data for participants with low WHtR are shown in blue circles and those with high WHtR are shown in orange circles. (A) Scatterplot shows the significant interaction of age-by-WHtR on taste ratings for high-calorie foods; age-related increases in taste ratings for high-calorie foods were significant in the high WHtR group but not in the low WHtR group. (B) Scatterplot shows the significant interaction of age-by-WHtR on health ratings; age-related decreases in health ratings for high-calorie foods were significant in the low WHtR group but not in the high WHtR group. (C) Scatterplot shows the marginally significant interaction of age-by-WHtR on preference ratings; age-related increases in preference for high-calorie foods were significant in the high WHtR group but not in the low WHtR group.

FIGURE 4

A linear regression model was created for each participant to measure how well the taste ratings predicted their preference for a specific food cue (i.e., β-weight). Scatterplot shows that older age was associated with a higher β-taste. Relationship between age and β-taste was significant in the high WHtR group, and marginally significant in the low WHtR group. Data for participants with low WHtR are shown in blue diamonds and those with high WHtR are shown in orange diamonds.