The following repository includes a four-armed bandit task performed online with punishment and reward conditions manipulated between blocks. **Participants.** 178 prolific workers (age mean = 26.1, range 18 to 51; 101 males, 76 females, 1 other) completed an online experiment in return for monetary compensation. All participants reported normal or corrected vision, no current or past psychiatric or neurological diagnosis. The study protocol was approved by the Research Ethics Council of Tel-Aviv University and all participants signed informed consent before participating in the study. **Procedure.** Participants were invited to take part in the study using the Prolific online platform. Participants completed a Reinforcement Learning task of four cards and two reward conditions. On each trial of the task, two cards of the four were offered by the computer, and participants were asked to pick one. Each card could lead to a reward on an independent drifting probability across trials (Figure 1). The difference between conditions was in whether participants won extra points or avoided the loss of points. The task was preceded by instructions, a short practice session, and a quiz about the instructions. After completing the quiz, participants were told that they are given “100 game-coins to start with”. At the end of the experiment, participants were paid a fixed amount (£2.5) plus a bonus (of £1 or £1.5) based on their performance. **Reinforcement Learning Task.** Here, participants were asked to play a card game. Each trial was preceded by a 1sec fixation point presented in the middle of the screen. Then, two cards (abstract fractal images, see SI) were randomly assigned to one of the two sides of the screen (i.e, left and right). Participants then choose a card freely using a right/left corresponding response-key press (‘s’ or ‘k’ keys in a QWERTY keyboard; until response with 6sec deadline). After a choice has been made, the chosen card remained on the screen for 500 ms, while the other choice option disappeared. Then an outcome (1, 0, or -1 game-coin) appeared in the middle of the screen for 1 sec. The task included four blocks of 50 trials, two of each reward condition. Every block started with a screen that introduced the subject with four new cards to be used in the block and two possible outcomes. Positive blocks, in which the subject could either gain one game-coin or no-coin, were indicated by a green frame. Negative blocks, in which the outcome could either be a loss of a game-coin or no-coin, were indicated by a red frame. **Self-report data** for exploratory reasons all participants completed the OCI-R, and a partial sample also completed BDI, STAI , and SPQ Data was collected during 2021 (part of Inbal Alon Master's thesis) A component with the code converting raw json is also included (not visible for the public) - email us for access if needed.