Neural substrates of reward magnitude, probability, and risk during a wheel of fortune decision-making task

Abstract

Economic decision-making involves the weighting of magnitude and probability of potential gains/losses. While previous work has examined the neural systems involved in decision-making, there is a need to understand how the parameters associated with decision-making (e.g., magnitude of expected reward, probability of expected reward and risk) modulate activation within these neural systems. In the current fMRI study, we modified the monetary wheel of fortune (WOF) task [Ernst, M., Nelson, E.E., McClure, E.B., Monk, C.S., Munson, S., Eshel, N., et al. (2004). Choice selection and reward anticipation: an fMRI study. Neuropsychologia 42(12), 1585-1597.] to examine in 25 healthy young adults the neural responses to selections of different reward magnitudes, probabilities, or risks. Selection of high, relative to low, reward magnitude increased activity in insula, amygdala, middle and posterior cingulate cortex, and basal ganglia. Selection of low-probability, as opposed to high-probability reward, increased activity in anterior cingulate cortex, as did selection of risky, relative to safe reward. In summary, decision-making that did not involve conflict, as in the magnitude contrast, recruited structures known to support the coding of reward values, and those that integrate motivational and perceptual information for behavioral responses. In contrast, decision-making under conflict, as in the probability and risk contrasts, engaged the dorsal anterior cingulate cortex whose role in conflict monitoring is well established. However, decision-making under conflict failed to activate the structures that track reward values per se. Thus, the presence of conflict in decision-making seemed to significantly alter the pattern of neural responses to simple rewards. In addition, this paradigm further clarifies the functional specialization of the cingulate cortex in processes of decision-making.

Figure 1

The wheel of fortune task. A. The four task stimuli comprise two types of wheel. In one type, the wheels present the options of a high-probability/high-magnitude reward (high-prob./high-mag.) vs. a low-probability/low-magnitude reward (low-prob./low mag.). In the other type, the two options are a high-probability/low-magnitude (high-prob./low mag.) reward vs. a low-probability/high-magnitude reward (low-prob./high mag.). B. Example of a single task trial. Analyses focused on brain activation during the 4 s of the selection period. C. The three main contrasts are made explicit in this figure: The magnitude contrast compares the selection of the high-probability/high-magnitude option vs. the high-probability/low-magnitude option. The probability contrast compares the selection of the high-probability/high-magnitude option vs. the low-probability/high-magnitude option. The risk contrast compares the selection of the high-probability/low-magnitude option vs. the low-probability/high-magnitude option.

Figure 1

The wheel of fortune task. A. The four task stimuli comprise two types of wheel. In one type, the wheels present the options of a high-probability/high-magnitude reward (high-prob./high-mag.) vs. a low-probability/low-magnitude reward (low-prob./low mag.). In the other type, the two options are a high-probability/low-magnitude (high-prob./low mag.) reward vs. a low-probability/high-magnitude reward (low-prob./high mag.). B. Example of a single task trial. Analyses focused on brain activation during the 4 s of the selection period. C. The three main contrasts are made explicit in this figure: The magnitude contrast compares the selection of the high-probability/high-magnitude option vs. the high-probability/low-magnitude option. The probability contrast compares the selection of the high-probability/high-magnitude option vs. the low-probability/high-magnitude option. The risk contrast compares the selection of the high-probability/low-magnitude option vs. the low-probability/high-magnitude option.

Figure 2

Areas activated by the magnitude contrast and percent signal changes from baseline within these regions: High-magnitude vs. low-magnitude reward selections with probability held constant. Activations are shown for: (A) medial OFC (7, 54, 16); (B) right insula (38, −7, −5); and (C) left amygdala (−23, −6, −13) (see Table 2). The associated percent signal changes from baseline for each of these regions are shown immediately beneath for A and B and on the right for C.

Figure 3

Regions of significant activation within the cingulate gyrus and percent signal changes from baseline within these regions. Activations are shown for: (A) Magnitude contrast: regions of mid-cingulate gyrus and posterior cingulate which showed significantly greater BOLD responses when making selections associated with higher magnitude rather than lower magnitude rewards (−1, −13, 44 and −5, −52, 25 respectively) (Table 2); (B) Probability contrast: cingulate gyrus which showed significantly greater BOLD responses when making lower probability as opposed to higher probability selections (1, 21, 45) (Table 3); (C) Risk contrast: cingulate gyrus which showed significantly greater BOLD responses when making ‘risky’ as opposed to ‘safe’ selections (−4, 22, 44) (Table 3). The associated percent signal changes from baseline for each of these regions are shown immediately to the right or beneath the brain slices.

Figure 4

(A) Region of middle frontal gyrus activation (−34, 29, 42) showing greater BOLD responses for low-probability as opposed to high-probability selections (Table 3); (B) Region of intraparietal lobule (−55, −46, 40) that showed greater BOLD responses for ‘risky’ as opposed to ‘safe’ selections (Table 3). The associated percent signal changes from baseline for each of these regions are shown immediately beneath.