Frames, biases, and rational decision-making in the human brain

Abstract

Human choices are remarkably susceptible to the manner in which options are presented. This so-called "framing effect" represents a striking violation of standard economic accounts of human rationality, although its underlying neurobiology is not understood. We found that the framing effect was specifically associated with amygdala activity, suggesting a key role for an emotional system in mediating decision biases. Moreover, across individuals, orbital and medial prefrontal cortex activity predicted a reduced susceptibility to the framing effect. This finding highlights the importance of incorporating emotional processes within models of human choice and suggests how the brain may modulate the effect of these biasing influences to approximate rationality.

Fig. 1

The financial decision-making task. At the beginning of each trial, participants were shown a message indicating the starting amount of money that they would receive (e.g., ‘You receive £50’) (duration 2 s). Subjects were instructed that they would not be able to retain the whole of this initial amount, but would next have to choose between a sure option and a gamble option (4 s). The sure option was presented in the Gain frame trials (A) as an amount of money retained from the starting amount (e.g., keep £20 of the £50) and in the Loss frame trials (B) as an amount of money lost from the starting amount (e.g., lose £30 of the £50). The gamble option was represented as a pie chart depicting the probability of winning (green) or losing (red) all of the starting money. The expected outcomes of the gamble and sure options were equivalent. Gain frame trials were intermixed pseudo-randomly with Loss frame trials. No feedback concerning trial outcomes was given during the experiment.

Fig. 2

Behavioral results. (A) Percentages of trials in which subjects chose the gamble option in the Gain frame and the Loss frame. Subjects showed a significant increase in the percentage of trials in which the gamble option was chosen in the Loss frame with respect to the Gain frame [61.6%> 42.9% (P < 0.001, t₁₉ = 8.06)]. The dashed line represents riskneutral behavior (choosing the gamble option in 50% of trials). Error bars denote SEM. (B) Each bar represents, for each individual subject, the percentage difference between how often subjects chose the gamble option in the Loss frame as compared to the Gain frame. A hypothetical value of zero represents a complete indifference to the framing manipulation (i.e., fully ‘rational’ behavior). All participants, to varying degrees, showed an effect of the framing manipulation.

Fig. 3

fMRI results (A) Interaction contrast [(G _sure + L_gamble) - (G_gamble + L_sure)] : brain activations reflecting subjects’ behavioral tendency to choose the sure option in the Gain frame and the gamble option in the Loss frame (i.e., in accordance with the frame effect). Bilateral amygdala (Amyg) activation [MNI space coordinates (x, y, z)]: left hemisphere, −14, 2, −24 (peak Z score = 3.97); right hemisphere, 12, 2, −20 (Z score = 3.82). (C) Reverse interaction contrast [(G_gamble + L_sure) − (G_sure + L _gamble)]:brain activations reflecting the decision to choose counter to subjects’ general behavioral tendency. Anterior cingulate cortex (ACC) activation: 2, 24, 44 (Z score = 3.65); −2, 8, 56 (Z score = 3.78). Effects in (A) and (C) were significant at P < 0.001; for display purposes they are shown at P < 0.005. (B and D) Plots of percentage signal change for peaks in right amygdala (12, 2, −20) (B) and ACC (2, 24, 44) (D). Error bars denote SEM.

Fig. 4

Rationality across subjects: fMRI correlational analysis. Regions showing a significant correlation between rationality index [between-subjects measure of susceptibility to the framing manipulation; see (14)] and the interaction contrast image [(G _sure+ L_gamble) - (G_gamble + L_sure)] are highlighted. (A) Orbital and medial prefrontal cortex (OMPFC) [MNI space coordinates (x, y, z)]: VMPFC (left panel), −4, 34, −8 (Z score = 4.56); OMPFC and R-OFC circled in right panel [R-OFC: 24, 30, −10 (Z score = 5.77)]. Effects were significant at P < 0.001; for display purposes they are shown at P < 0.005. (B) Plot of the correlation of parameter estimates for R-OFC with the rationality index for each subject (r = 0.8, P < 0.001).