The role of the striatum in effort-based decision-making in the absence of reward

Abstract

Decision-making involves weighing costs against benefits, for instance, in terms of the effort it takes to obtain a reward of a given magnitude. This evaluation process has been linked to the dorsal anterior cingulate cortex (dACC) and the striatum, with activation in these brain structures reflecting the discounting effect of effort on reward. Here, we investigate how cognitive effort influences neural choice processes in the absence of an extrinsic reward. Using functional magnetic resonance imaging in humans, we used an effort-based decision-making task in which participants were required to choose between two options for a subsequent flanker task that differed in the amount of cognitive effort. Cognitive effort was manipulated by varying the proportion of incongruent trials associated with each choice option. Choice-locked activation in the striatum was higher when participants chose voluntarily for the more effortful alternative but displayed the opposite trend on forced-choice trials. The dACC revealed a similar, yet only trend-level significant, activation pattern. Our results imply that activation levels in the striatum reflect a cost-benefit analysis, in which a balance is made between effort discounting and the intrinsic motivation to choose a cognitively challenging task. Moreover, our findings indicate that it matters whether this challenge is voluntarily chosen or externally imposed. As such, the present findings contrast with classical findings on effort discounting that found reductions in striatum activation for higher effort by finding enhancements of the same neural circuits when a cognitively challenging task is voluntarily selected and does not entail the danger of losing reward.

Figure 1.

Sequence of events in case of a correct response on both the choice cue and flanker stimulus. A trial started with the presentation of a choice cue, imposed (left) or voluntary (right), and two gray squares denoting the two possible locations. Participants had to respond to the choice cue in a 1500 ms interval. From the presentation of the choice cue, a variable time interval (1500–6500 ms) started, after which the flanker stimulus was presented on one of the two locations. The flanker stimulus remained on the screen for 300 ms. The screen turned black after the response on the flanker stimulus or when the response deadline (1000 ms) was exceeded. From the presentation of the flanker stimulus, a variable time interval (1500–6500 ms) started until the beginning of the next trial.

Figure 2.

Activation map of the whole-brain voluntary > imposed choice contrast (left) and high-conflict location > low-conflict location choice contrast (right).

Figure 3.

A, Activation map of the whole-brain interaction contrast (choice type × conflict location) during the moment of choice. B, Mean β estimates for each conflict location condition (high-conflict vs low-conflict), separately for imposed choice trials and voluntary choice trials in the head of caudate nucleus. Error bars represent ±1 SEM.

Figure 4.

A, C, Anatomical ROIs for the VS and the dACC. B, D, Mean β estimates for each conflict location condition (high-conflict vs low-conflict), separately for imposed choice trials and voluntary choice trials in the VS and the dACC, respectively. Error bars represent ±1 SEM.