Fig. 1.

Schematic task timeline. Two participants simultaneously played in a PGG. Each participant was first asked to choose how much they wanted to contribute to the public good. Participants were first presented with an instruction screen with the sentence ‘How do you want to allocate your MU this round?’ during 3 s. Then the payoff matrix appeared with the choice options of the participant depicted as rows and the choice options of the partner depicted in columns. They could navigate between rows to make their choice using two buttons of an MR-compatible response box placed in the subject’s right hand and validated their choice at any time using a third button. This choice period was self-paced, thus introducing some natural variability in trial time course. Their choice was shown during 2 s. Then, a second instruction screen displaying ‘How do you think the other will allocate his or her MU in this round?’ was presented during 3 s. The payoff matrix appeared and they could choose the expected contribution of the other by navigating between columns of the matrix (self-paced). This choice was also shown during 2 s. A screen displayed ‘Please wait for the other to respond’ during 500 ms followed by a black screen displayed until the other participant had completed their choice, with a minimum of 6 s. The feedback screen, displayed during 16 s, then showed both participants’ contributions to the public account as well as the participant’s payoff.