Here is an bit of work Zhang et al. on consolidation of motor memory that more clearly confirms what I know from my own experience of trying to learn a new piano piece. If I watch a video of myself playing a passage where I have difficulty with the notes, I remember the notes better than if I actually play them several times - the actual movement appears to get in the way of forming a motor memory of it. (The same effects can happen with mentally visualizing the movements, a trick known to many athletes and performers).
Practicing a motor task can induce neuroplastic changes in the human primary motor cortex (M1) that are subsequently consolidated, leading to a stable memory trace. Currently, little is known whether early consolidation, tested several minutes after skill acquisition, can be improved by behavioral interventions. Here we test whether movement observation, known to evoke similar neural responses in M1 as movement execution, can benefit the early consolidation of new motor memories. We show that observing the same type of movement as that previously practiced (congruent movement stimuli) substantially improves performance on a retention test 30 min after training compared with observing either an incongruent movement type or control stimuli not showing biological motion. Differences in retention following observation of congruent, incongruent, and control stimuli were not found when observed 24 h after initial training and neural evidence further confirmed that, unlike motor practice, movement observation alone did not induce plastic changes in the motor cortex. This time-specific effect is critical to conclude that movement observation of congruent stimuli interacts with training-induced neuroplasticity and enhances early consolidation of motor memories. Our findings are not only of theoretical relevance for memory research, but also have great potential for application in clinical settings when neuroplasticity needs to be maximized.