My mother forced me to learn piano. I initially wasn't enthusiastic, and put in the minimum amount of practice to get away without a scolding from my teacher. By the time I was a teenager, though, and developed a more complex emotional life, I started to love playing. Beethoven's sonatas are a great way to release teenage angst.
One of the downsides of graduate school (and undergrad for that matter) was that I left my parents' piano, and for the past seven (!) years, the only time I can play is when I go home. Each year, given the time constraints, I choose a handful of pieces to practice; this holiday they were by Brahms and Chopin. The first day or two of practice were pretty dissonant as I regained muscle memory. For some reason, my hands kept playing sevenths instead of octaves.
By the third day I had regained some semblance of motor control, but still found myself progressing slowly through the pieces. This is when I remember the most important technique of practicing piano: break it down and repeat the hard parts. If I'm slipping on a specific measure, or even a few notes, I need to isolate that section and repeat 5-10 times until I can play it almost blindfolded. * Then I incorporate it into the surrounding measures, and finally into the whole piece. **
Once I remembered how to practice, things went more smoothly. I still was not nearly as proficient as I once was, but there was a vague semblance of musicality to my playing.
So what does this have to do with science?
I have been interviewing on and off for post-docs for the last eighteen (!) months (this is another story). For my last interviews in October/November, I was practicing the talk I had given so many times before. The hardest parts of my talk are the intro slide, and some data heavy slides that require a little explanation. Every time I would practice my talk, I would stumble in the same few places.
It was then that I realized I needed to practice my talk like I practiced piano. I pulled out the problem slides, and took clear notes on the point of each slide, and the easiest way to walk through the figures. Then once I had mastered the slide, I put it in the context of the previous and next slide. And finally I practiced my full talk once again, including the formerly problem slides, and it all went mellifluously.
* I need to write a whole 'nother post about the motor skills involved in playing piano. To be able to play well, you need to basically memorize the piece in terms of motor coordination, because you can't possibly read the notes fast enough to command your hand (or at least I can't sight-read that fast). But you still need to sheet music to remind you where you are, and to cue the motor commands. Similarly, while you undoubtedly memorize how to shift your hand an octave, I'm convinced there is some visual feedback on hand positioning out of the corner of your eye.
** The newer Rock Bands have a practice feature, which I appreciate for the effort, but is sorely lacking in utility. The portions of each song to practice are just too long, when usually it's just one set of notes that's tricky. It was so much more satisfying to be able to look at the sheet music and repeat it ad nauseum without waiting for reloads.