Conditioning task switching behavior

- Reward studies often focus on the reinforcement of simple tasks.
- This study disproportionally rewarded task repetitions versus alternations.
- People rewarded more for task alternations show more voluntary task switching.
- These findings suggest that abstract control representations can be conditioned.

Most reward studies focus on the reinforcement of simple tasks or stimulus-response rules. However, recent theories (re)emphasized that cognitive control representations should adhere to the same reinforcement learning principles as do more basic stimulus and response representations. This study focused on the act of switching between different tasks, and investigated the effects of disproportionally rewarding task alternations or repetitions in a cued task switching paradigm on subsequent voluntary task switching behavior (i.e., when participants could choose which task to perform). The results show that subjects who were more rewarded for task alternations (relative to those more rewarded for repetitions) showed more task switching behavior. Moreover, this increased task switching behavior also came with a cost, with participants more rewarded for task repetitions showing a better task focus (i.e., smaller task-rule congruency effects). These results demonstrate that reward can reinforce more abstract control representations, beyond low-level stimulus or response representations.