Taking a different perspective: mindset influences neural regions that represent value and choice

Abstract

Most choices are complex and can be considered from a number of different perspectives. For example, someone choosing a snack may have taste, health, cost or any number of factors at the forefront of their mind. Although previous research has examined neural systems related to value and choice, very little is known about how mindset influences these systems. In the current study, participants were primed with Health or Taste while they made decisions about snack foods. Some neural regions showed consistent associations with value and choice across Health or Taste mindsets. Regardless of mindset, medial orbitofrontal cortex (MOFC) tracked value in terms of taste, regions in left lateral prefrontal cortex (LPFC) tracked value in terms of health, and MOFC and dorsal anterior cingulate were associated with choice. However, activity in other neural regions was modulated by the mindset manipulation. When primed with Taste, rostral anterior cingulate tracked value in terms of taste whereas left amygdala and left putamen were associated with choice. When primed with Health, right LPFC and posterior MOFC tracked value in terms of health. The findings contribute to the neural research on decision-making by demonstrating that changing perspectives can modulate value- and choice-related neural activity.

Fig. 1

Decision-making task. Participants evaluate one food item, view a fixation cross, evaluate a second food item, view a fixation cross and then indicate their choice between the two items. Depending on the condition, participants were either primed to think about taste properties (Taste Mindset) or health properties (Health Mindset) while evaluating single items. Participants were instructed to always choose the item that they wanted, regardless of the Mindset condition. The two items were always from different categories: Healthy (e.g. almonds, granola) or Unhealthy (e.g. cupcake, cookie) in a balanced pseudorandom order. The same set of items was used in each condition but items were paired differently.

Fig. 2

Taste and health ratings for each food. Each point in the chart represents the average taste and health rating for a food. Green triangles represent Healthy category foods and blue squares represent Unhealthy category foods. Error bars represent standard error.

Fig. 3

Neural regions related to (A) Taste Value, (B) Health Value and (C) Choice that are insensitive to Mindset. (A) MOFC parametrically relates to Taste Value across Mindset conditions (Taste Mindset Taste Value and Health Mindset Taste Value conjunction, P < 0.005, 34 contiguous voxels). Charts on the right illustrate parametric relations with value in terms of taste by showing activation parameter estimates for foods rated low (rated two or lower), middle (rated three) and high (rated four or higher) on tastiness. (B) LPFC regions parametrically relate to value in terms of health across Mindset conditions (top: BA 8; bottom: BA 9; Taste Mindset Health Value and Health Mindset Health Value conjunction, P < 0.005, >34 contiguous voxels). Charts to the right of each brain image illustrate parametric relations to value in terms of health by showing activation parameter estimates for foods rated low (rated two or lower), middle (rated three) and high (rated four or higher) on healthiness. A ‘asterisks’ and solid line indicates that the parametric relationship is significant based on the criterion for the parametric modulation analysis (P < 0.005, >34 contiguous voxels). Ratings were categorized as low, middle or high to visualize the parametric relations, but no statistical tests were conducted based on the categorical estimates. (C) MOFC and DACC activity is greater for chosen than nonchosen foods across Taste and Health Mindset conditions. Maps show neural association with choice conjunction analysis (Taste Mindset Chosen vs Taste Mindset Non-chosen conjoined with Health Mindset Chosen vs Health Mindset Non-chosen, P < 0.005, >34 contiguous voxels). Bar charts to the right of each brain image show activation parameter estimates for chosen and nonchosen foods in each condition. ‘asterisks’ indicates a significant difference between parameter estimates for chosen and nonchosen foods. Error bars represent standard error.

Fig. 4

Taste compared to Health Mindset modulates the parametric relation between neural activity and Taste Value. An RACC cluster shows a significant parametric relation to value in terms of taste only in the Taste Mindset condition (Taste Mindset Taste Value > Health Mindset Taste Value contrast, P < 0.005, >34 contiguous voxels). Charts illustrate parametric relations with value in terms of taste by showing activation parameter estimates for foods rated low (rated two or lower), middle (rated three) and high (rated four or higher) on tastiness. A ‘Asterisks’ and solid line indicates that the parametric relationship is significantly greater than zero (P < 0.05), ‘n.s.’ and dotted line indicate that the parametric relationship is not significantly different from zero (P > 0.05). Ratings were categorized as low, middle or high to visualize the parametric relations, but no statistical tests were conducted based on the categorical estimates. Error bars represent standard error.

Fig. 5

Health compared to Taste Mindset modulates the parametric relation between neural activity and Health Value. In the left column, clusters show a significant parametric relation to value in terms of health only in the Health Mindset condition (Health Mindset Health Value > Taste Mindset Health Value contrast, P < 0.005, >34 contiguous voxels). In the right column, neural parametric relations to value in terms of health are illustrated by showing activation parameter estimates for foods rated low (rated two or lower), middle (rated three) and high (rated four or higher) on healthiness. ‘Asterisks’ and solid line indicates that the parametric relationship is significantly greater than zero (P < 0.05), ‘n.s.’ and dotted line indicate that the parametric relationship is not significantly different from zero (P > 0.05). Ratings were categorized as low, middle or high to visualize the parametric relations, but no statistical tests were conducted based on the categorical estimates. Error bars represent standard error.

Fig. 6

Taste compared to Health Mindset modulates brain regions associated with choice. Left amygdala and left putamen show an interaction of Choice and Mindset such that activity increases for chosen compared to nonchosen foods only in the Taste Mindset condition [(Taste Mindset Chosen vs Taste Mindset Non-chosen) vs (Health Mindset Chosen vs Health Mindset Non-chosen), P < 0.005, >34 contiguous voxels]. Bar charts to the right of each brain image show activation parameter estimates for chosen and nonchosen foods in each condition. ‘Asterisks’ indicates a significant difference between parameter estimates for chosen and nonchosen foods. Error bars represent standard error.