Model-free decision making resists improved instructions and is enhanced by stimulus-response associations

Abstract

Human behaviour is driven by two types of processes running in parallel: goal-directed and habitual, each supported by different computational-learning mechanisms, model-free and model-based respectively. In model-free strategies, stimulus-response associations are strengthened when actions are followed by a reward and weakened otherwise. In model-based learning, previous to selecting an action, the current values of the different possible actions are computed based on a detailed model of the environment. Previous research with the two-stage task suggests that participants' behavior usually shows a mixture of both strategies. But, interestingly, a recent study by da Silva and Hare (2020) found that participants deploy a purely model-based behavior when they are given detailed instructions about the structure of the task. In the present study, we reproduce this essential experiment using a larger sample size (N=59). However, our results do not suggest a sole model-based behaviour, but rather a hybrid one. Furthermore, an additional experiment shows that slight changes in the task, like a consistent stimulus-response mapping, can encourage reliance on model-free strategies, even if participants are presented with improved instructions. This suggests that the model-free marker, as measured by the two-stage task, is related to S-R learning.