I sometimes think reading scientific studies about creative activities like art and music are just hilarious, the way they a) state the obvious, or b) try to make a fairly ethereal concept into a hard science thing.
Take this study: Dimensions of Music Improvisation, by Michele Biasutti and Luigi Frezza, published in Creativity Research Journal, 21(2&3) 2009. In their abstract, the duo say they found “significant Pearson correlations between flow and anticipation, between flow and musical practice, between anticipation and basic skills, between repertoire and emotive communication, between repertoire and feedback, between musical practice and basic skills.”
In essence what they’re saying is the more skilled you are, the better you can improv music. Well, duh. They also mention that certain instruments are easier to translate over to other instruments, which again is a no-brainer.