Previous research has linked both rapid eyemovement (REM) sleep and Stage 2 sleep to procedural memory consolidation. The present study sought to clarify the relationship between sleep stages and procedural memory consolidation by examining the effect of initial skill level in this relationship in young adults. In-home sleep recordings were performed on participants before and after learning the pursuit rotor task. We divided the participants into low- and high-skill groups based on their initial performance of the pursuit rotor task. In high-skill participants, there was a significant increase in Stage 2 spindle density after learning, and there was a significant correlation between the spindle density that occurred after learning and pursuit rotor performance at retest 1 week later. In contrast, there was a significant correlation between changes in REM density and performance on the pursuit rotor task during retest 1 week later in low-skill participants, although the actual increase in REM density failed to reach significance in this group. The results of the present study suggest the presence of a double dissociation in the sleep-related processes that are involved in procedural memory consolidation in low- and high-skill individuals. These results indicate that the changes in sleep microarchitecture that take place after learning depend on the initial skill level of the individual and therefore provide validation for the model proposed by Smith et al. [Smith, C. T., Aubrey, J. B., & Peters, K. R. Different roles for REM and Stage 2 sleep in motor learning. Psychologica Belgica, 44, 79–102, 2004]. Accordingly, skill level is an important variable that needs to be considered in future research on sleep and memory consolidation.