Fig. 2: Task paradigms for Exps. 3 and 4.

Exp. 3 investigated effort-based decision-making. The design is similar to that of Exp. 2. Initial calibration involved setting the hand-held dynamometer on the basis of each participant’s MVC. The participants then familiarized themselves with various effort levels, squeezing the handle to match the effort indicated by a yellow line; a higher line denoted more effort. They practised this twice per effort level. The core task involved deciding whether the reward (apples) on offer was worth the effort assigned to it. The participants were informed that ten of their choices would be randomly selected at the experiment’s end to physically execute for obtaining apples. Exp. 4 investigated effort-based decision-making under uncertainty. ITI, intertrial interval. a, Post-training (see below), the participants underwent 200 trials of accept/reject decisions, balancing rewards (credits) against effort levels. Decisions were made considering the presence or absence of uncertainty, particularly in estimating the hidden circle’s location on the basis of dot configurations (as detailed in Methods). The absence of uncertainty was signified by directly showing the purple circle’s location. b,c, After the experiment, 24 trials were randomly chosen from the accepted decisions. Here, the participants exerted physical effort (b) to earn the chance to place a blue disk (c) where they thought the hidden circle was. In 12 trials, the purple circle’s location was shown, and in the other 12, it was not. Performance accuracy determined the final credits earned. Training occurred in three stages: (1) an interactive tutorial on Circle Quest (as in Exps. 1 and 2) introduced the scoring function and assessed localization accuracy (Fig. 8a); (2) the participants rated their confidence in locating the hidden circle on a scale of 0 to 100, with dot configurations reflecting those in the decision-making phase and additional catch trials to broaden the uncertainty range (subjective uncertainty estimates in Fig. 8b); and (3) effort calibration and familiarization mirrored Exp. 3. Hand icons from BioRender.com and macrovector on Freepik.