Enhancing metacognitive reinforcement learning using reward structures and feedback

How do we learn to think better, and what can we do to promote such metacognitive learning? Here, we propose that cognitive growth proceeds through metacognitive reinforcement learning. We apply this theory to model how people learn how far to plan ahead and test its predictions about the speed of metacognitive learning in two experiments. In the first experiment, we find that our model can discern a reward structure that promotes metacognitive reinforcement learning from one that hinders it. In the second experiment, we show that our model can be used to design a feedback mechanism that enhances metacognitive reinforcement learning in an environment that hinders learning. Our results suggest that modeling metacognitive learning is a promising step towards promoting cognitive growth.