Figure 1.
On each trial two options were presented for two seconds. This presentation was followed by the appearance of a yellow cross, which signaled a maximum of 1.5 s for indicating the preferred option by pressing one of two buttons on a computer mouse. Thereafter, a feedback screen indicating the subject's choice was presented for 0.5 s plus the difference between 1.5 s and the reaction time (RT) to make sure that the total time of choice plus feedback was 2 s. This was followed immediately with the next trial.
Table 1.
Effect of state on risk behavior: Representative agent in the non-parametric approach.
Table 2.
Effect of state on risk behavior: Representative agent in EUT approach.
Table 3.
Effect of state on relative value: Representative agent in EUT approach.
Figure 2.
Average effect of state on risk behavior: Within-subject.
A) The average (across subjects) of the proportion to choose the lottery option out of the total number of choices made in both states for all reward types. B) The average (across subjects) of the fitted risk parameters (α) in both states for all reward types. Data represents the mean ± s.e.m. * p<0.05; ** p<0.002; *** p<0.0001.
Figure 3.
Correlation of risk behavior between states: Non-parametric.
(A) Correlations between the proportions to choose the lottery option in the satiated state and in the deprived state across all subjects for all reward types. Each point represents the proportions in both states for a single subject. (B) Correlations between the proportions to choose the lottery option in the satiated state and the difference in proportions (deprived - satiated) across all subjects for all reward types. The black line represents the unity line. The red line represents the least square fit. The green circle highlights an example subject that increased her proportion to choose the lottery option when in the deprived state. The red circle highlights an example subject that decreased her proportion to choose the lottery option when in the deprived state. Details regarding the regression in A can be found in Table 1.
Figure 4.
Correlation of risk behavior between the to states: EUT.
(A) Correlations between the fitted risk parameters in the satiated state and in the deprived state across all subjects for all reward types. Each point represents the fitted risk parameters in both states for a single subject. (B) Correlations between the fitted risk parameters in the satiated state and the difference in the risk parameters (deprived - satiated) across all subjects for all reward types. The risk parameters for all reward types in each subject were jointly estimated. The black line represents the unity line. The red line represents the least square fit. The green circle highlights an example subject that became less risk averse when in the deprived state. The red circle highlights an example subject that became more risk averse when in the deprived state. Details regarding the regression in A can be found in Table 2.
Table 4.
Figure 5.
A) Standard deviation (STD) in the non-parametric approach. B) Standard deviation in the EUT approach. The average standard errors of the distributions across subjects of the proportion to choose the lottery option (A) and the fitted risk parameters (B) for all reward types in both states. A lower average standard error demonstrates convergence towards a similar value and lower variance in the deprived condition. For comparing the averages and conducting a significance test, we bootstrapped the standard errors across subjects for all reward types in each state. Note that the data represents mean ± STD (and not s.e.m.). *** p<0.0001.