Introduction Relearned fine-motor skills, like typing on a mirrored keyboard, are supposed to require suppression of over-practiced motor skills, like typing on a regular keyboard. Interestingly, performance on the habitual skill often worsens after practicing such an unusual skill. The aim of our study was to investigate whether sleep modifies this interfering effect. Materials and methods 25 males (25.44±4.56years) had to practice touch typing of words with 5 letters length as rapidly and accurately as possible on a regular keyboard and on a mirrored keyboard. The training period for the regular keyboard consisted of four 3min blocks. This regular typing period was followed by three times four 3min blocks on the mirrored keyboard. Testing on the regular and mirrored keyboard (each typing condition was tested during two 3min blocks) occurred after 8h of diurnal wakefulness (wake group, n=11) or nocturnal sleep (sleep group, n=14). The sleep group spent two nights with polysomnography (baseline and experimental night) in the sleep laboratory. Fine-motor performance was measured by the number of correctly typed letters per 30s. Sleep was scored visually according to AASM criteria and sleep spindles were detected automatically (The Siesta Group, Vienna, Austria). Results A 2×2×2 ANOVA for repeated measures with the within-subject factors TIME (pre vs. post sleep/wakefulness) and CONDITION (regular vs. mirrored typing) and the between subject factor GROUP (sleep vs. wake) revealed a significant interaction between TIME×CONDITION×GROUP (F1,23=9.959, p=0.004). Subjects in the sleep group showed a significant decrease in regular typing speed after nocturnal sleep whereas mirrored typing did not change. On the other hand, for subjects in the wake group we found a significant deterioration in mirrored typing but no change in regular typing. Furthermore, we could demonstrate a significant correlation (r14=0.644, p=0.013) between fast (13–15Hz) sleep spindle number during sleep stage N2 and overnight gains in mirrored typing. Conclusion Our results indicate an increased retroactive interference during regular keyboard typing after sleep which may occur because of a more effective consolidation of the mirrored keyboard typing skill during sleep in comparison to wakefulness. Additionally, we provide evidence that fast sleep spindle number during N2 promotes unlearning of an overlearned automated motor skill and facilitates learning of a replacement skill. Acknowledgement This study was funded by the Austrian Science Fund (P25000).