Investigation of the relationship between memory and metamemory using the game of Concentration

Abstract

Metamemory is one of the metacognitive functions and involves monitoring and controlling one's memory. In this study, we analyze the relationship between memory and metamemory using an agent-based model focused on a memory-based game. The task used is a solo version of the simplified classic card matching game "Concentration," in which the agent has three memories for cards: type, position, and the correspondence between type and position. Over time, the intensity of these memories decreases, resulting in forgetting, and noise in the memories results in memory misunderstandings. In the experiment, individuals with metamemory abilities improve their scores by adopting a strategy of not using memories that fall below a certain strength threshold for card selection. The results of varying the noise intensity showed that the improvement in scores due to metamemory was greater at higher noise intensities. Therefore, the effect of metacognition is more pronounced when the noise is stronger, suggesting a complementary relationship between memory and metamemory. This finding supports some of the results of previous studies of metamemory evolution using neural networks.