Fig. 2

Experimental task (following Choi et al.¹³). a A Trial: Subjects were presented a visualization of a budget line with 50/50 lotteries between two accounts, labeled X and Y. Each point on the budget line represents a different lottery between the X and Y accounts. Subjects used a trackball to choose their preferred lottery (a bundle of the X and Y accounts) out of all possible lotteries along the line. For example, as depicted in Fig. 2a, the bundle (11,72) corresponds to a lottery with a 50% chance to win 11 tokens (account X), and a 50% chance to win 72 tokens (account Y), where 1 token equals 5 NIS (~$1.5). We varied and randomized the budget lines (slopes and endowments) across trials and subjects. At the end of the experiment, one trial was randomly selected as well as one of the accounts. The subject received the tokens she had allocated to the selected account in the selected trial. b Behavioral example: Given this budget line, the subject could choose A (10,35) with a 50% chance of winning 10 tokens (account X) and 50% chance of winning 35 tokens (account Y); or similarly, B (45,10). An extremely risk-seeking subject would choose C, the lottery with the maximal expected payoff, yielding a 50% chance of winning 60 tokens and a 50% chance of winning nothing. By contrast, an extremely risk-averse subject would choose D, the intersection with the 45-degree line. This bundle is a degenerate lottery, which allocates the same number of tokens for both X and Y accounts. c Timeline: Inside the scanner, subjects had a maximum of 12 s time window to make their choice, followed by a 9 s variable ITI. If subjects made a choice before the end of the 12 s time window, the remaining time was added to the ITI. There were 27 trials in each block, 4 blocks, for a total of 108 trials. Subjects completed a pre-scan questionnaire (see Supplementary Note 7 for an English version) and a practice block with a trackball outside the scanner to make sure the instructions and procedures were clear